Systems and methods for personalized identification of host-directed therapies targeting immune response to infection

Abstract

Disclosed herein are methods for subtyping patients with a respiratory infection (e.g., community acquired pneuomonia or severe acute respiratory infection) to identify likely responders to corticosteroid therapy. The methods involve analyzing values of a plurality of biomarkers, examples of which include a first biomarker selected from procalcitonin, IL-6, and TREM-1, and a second biomarker selected from Pentraxin-3, MIP-3α, IL-1ra, IL-8, and NGAL. The methods enable personalized treatment decisions for patients with respiratory infections, potentially improving patient outcomes by targeting corticosteroid therapy to those most likely to benefit while avoiding harm to non-responders.

Description

Attorney Docket No: PNS-001WOSYSTEMS AND METHODS FOR PERSONALIZED IDENTIFICATION OF HOST- DIRECTED THERAPIES TARGETING IMMUNE RESPONSE TO INFECTIONCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 689,273 filed August 30, 2024 and U.S. Provisional Patent Application No. 63 / 716,101 filed November 4, 2024, the entire disclosure of which is hereby incorporated by reference in its entirety for all purposes.BACKGROUND

[0002] Respiratory infections include severe acute respiratory infection (SARI) and pneumonia. Pneumonia is an acute infection of alveoli in the lungs, the air sacs where oxygen and carbon dioxide are exchanged. When that infection is acquired in the community, as opposed to in the hospital, it is community-acquired pneumonia (CAP). Various pathogens, including Sars-CoV-2, can cause CAP and a variety of risk factors impact susceptibility, including age, chronic conditions, and smoking. Non-COVID CAP is a leading cause of hospitalization and death in the United States, where an estimated 6 million cases and 1.5 million hospitalizations occur each year. Local tissue damage caused by the pathogen prompts local and systemic inflammatory responses, which becomes dysregulated in over half of patients hospitalized with CAP and can result in long and expensive hospital admissions, long-term morbidity, or death. The current understanding of CAP emphasizes its impact on multiple systems and potential for both acute and long-term illness and death. An estimated 10-15% of patients die within 30 days after hospitalization for CAP, and 30% of all patients hospitalized with CAP and 50% of patients admitted to the intensive care unit (ICU) with CAP die within 1 year.

[0003] There is an unmet clinical need for host-directed therapeutics for respiratory infections. The central tenet to CAP therapy remains source control through antimicrobial therapy and supportive care (e.g., intravenous fluids and vasopressors to support blood pressure and resultant organ perfusion, and non-invasive and invasive mechanical ventilation to support oxygenation and respiration). These therapies are limited because they do not treat the dysregulated inflammatory host response that drives much of the morbidity and mortality. Additional therapies such as immunoglobulin therapy and anti-inflammatory strategies have been studied, but their use remains controversial.Attorney Docket No: PNS-001WOSUMMARY

[0004] In one aspect provided is a method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise: a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

[0005] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, each additional biomarker different from the first biomarker and the second biomarker and selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase- associated lipocalin (NGAL). In an embodiment, the plurality of biomarkers further comprise two, three, four, five, or six additional markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL- 8, and neutrophil gelatinase-associated lipocalin (NGAL).

[0006] In an embodiment, the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8. In an embodiment, the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a. In an embodiment, the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein. In an embodiment, the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, sequential organ failure assessment (SOFA) Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.Attorney Docket No: PNS-001WO

[0007] In an embodiment, the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon- a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.

[0008] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more core inflammatory response biomarkers. In an embodiment, the one or more core inflammatory response biomarkers are different from the first biomarker and the second biomarker and selected from IL-ip, TNF-a, IL-6, C-Reactive Protein, Pentraxin-3, IL-lra, IL- 10, IL-8, NGAL, TREM-1, G-CSF, MIP-3a, PCT, Granzyme B, and TRAIL. In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more immune cell activation biomarkers or chemokines. In an embodiment, the one or more immune cell activation features or chemokines are selected from Neutrophils, WBC, Lymphocytes, Monocytes, Platelets, MCP-1, MIP-la, MIP-1 (CCL4), IP-10, PD-L1, GM-CSF, FLT3 Ligand, IL-15, IL-7, Interferon-a, and Interferon-y. In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more endothelia and coagulation markers.

[0009] In an embodiment, the one or more endothelia and coagulation markers are selected from Thrombomodulin, Tissue Factor, VCAM-1, E-Selectin, Angiopoietin-2, Angiopoietin- 1, VEGF, TGF-a, and Leptin. In an embodiment, the plurality of biomarkers further compriseAttomey Docket No: PNS-001WO one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more metabolic and organ function features. In an embodiment, the one or more metabolic and organ function features are selected from Lactate, Blood Urea Nitrogen, Creatinine, Bilirubin, AST, ALT, ALP, Albumin, Calcium, Potassium, Sodium, Chloride, Total CO2, Hemoglobin, RBC, RDW. In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more clinical parameters or interventions. In an embodiment, the one or more clinical parameters or interventions are selected from Systolic BP, Diastolic BP, Heart Rate, Respiratory Rate, Temperature, Pulse Oximetry, SOFA Kidney, SOFA Cardiovascular, SOFA Respiratory, SOFA Coagulation, SOFA Liver, SIRS, Shock, Vasopressors, and Ventilator. In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more demographics or comorbidities. In an embodiment, the one or more demographics or comorbidities are selected from Age, Gender, Diabetes, Chronic Kidney Disease, Chronic Liver Disease, Active Cancer, CHF, Emphysema, Atelectasis, Rheumatoid Arthritis, Lupus, Immunodeficiency, Acute Bronchitis, Asphyxiation, Nicotine Use, and Glucocorticoid Deficiency.

[0010] In an embodiment, the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone. In an embodiment, the corticosteroid is administered intravenously, intramuscularly, or orally.

[0011] In an embodiment, values of at least a subset of the plurality of biomarkers are obtained by performing an immunoassay. In an embodiment, the immunoassay is a quantitative lateral flow assay. In an embodiment, performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies. In an embodiment, values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient. In one aspect provided is a method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise: a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-Attorney Docket No: PNS-OOIWO associated lipocalin (NGAL); and generating a prediction of a subtype for the patient with the respiratory infection based on the values of the plurality of biomarkers.

[0012] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, each additional biomarker different from the first biomarker and the second biomarker and selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase- associated lipocalin (NGAL). In an embodiment, the plurality of biomarkers further comprise two, three, four, five, or six additional markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL- 8, and neutrophil gelatinase-associated lipocalin (NGAL). In an embodiment, the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8. In an embodiment, the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a. In an embodiment, the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein. In an embodiment, the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-1 (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium. In an embodiment, the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen,Attorney Docket No: PNS-001WO nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.

[0013] In an embodiment, the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid. In an embodiment, the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

[0014] In an embodiment, the method further comprises: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid. In an embodiment, obtaining or having obtained values for at least a subset of the plurality of biomarkers comprises performing an immunoassay. In an embodiment, the immunoassay is a quantitative lateral flow assay. In an embodiment, performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies. In an embodiment, values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient. In one aspect provided is a non-transitory computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform the method as defined in any of the embodiments described herein. In one aspect provided is a method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise two or more of procalcitonin, IL-6, triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

[0015] In an embodiment, the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8. In an embodiment, the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a. In an embodiment, the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein. In an embodiment, the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, SOFA Kidney,Attorney Docket No: PNS-OOIWOMIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium. In an embodiment, the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM- CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer. In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more core inflammatory response biomarkers. In an embodiment, the one or more core inflammatory response biomarkers are different from the first biomarker and the second biomarker and selected from IL-ip, TNF-a, IL-6, C-Reactive Protein, Pentraxin-3, IL-lra, IL- 10, IL-8, NGAL, TREM-1, G-CSF, MIP-3a, PCT, Granzyme B, and TRAIL. In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more immune cell activation biomarkers or chemokines. In an embodiment, the one or more immune cell activation features or chemokines are selected from Neutrophils, WBC, Lymphocytes, Monocytes, Platelets, MCP-1, MIP-la, MIP-1 (CCL4), IP-10, PD-L1, GM-CSF, FLT3 Ligand, IL-15, IL-7, Interferon-a, and Interferon-y. In an embodiment, the plurality of biomarkers furtherAttorney Docket No: PNS-001WO comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more endothelia and coagulation markers. In an embodiment, the one or more endothelia and coagulation markers are selected from Thrombomodulin, Tissue Factor, VCAM-1, E-Selectin, Angiopoietin-2, Angiopoietin-1, VEGF, TGF-a, and Leptin.

[0016] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more metabolic and organ function features. In an embodiment, the one or more metabolic and organ function features are selected from Lactate, Blood Urea Nitrogen, Creatinine, Bilirubin, AST, ALT, ALP, Albumin, Calcium, Potassium, Sodium, Chloride, Total CO2, Hemoglobin, RBC, RDW.

[0017] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more clinical parameters or interventions. In an embodiment, the one or more clinical parameters or interventions are selected from Systolic BP, Diastolic BP, Heart Rate, Respiratory Rate, Temperature, Pulse Oximetry, SOLA Kidney, SOLA Cardiovascular, SOLA Respiratory, SOLA Coagulation, SOLA Liver, SIRS, Shock, Vasopressors, and Ventilator.

[0018] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more demographics or comorbidities. In an embodiment, the one or more demographics or comorbidities are selected from Age, Gender, Diabetes, Chronic Kidney Disease, Chronic Liver Disease, Active Cancer, CHE, Emphysema, Atelectasis, Rheumatoid Arthritis, Lupus, Immunodeficiency, Acute Bronchitis, Asphyxiation, Nicotine Use, and Glucocorticoid Deficiency.

[0019] In an embodiment, the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone. In an embodiment, the corticosteroid is administered intravenously, intramuscularly, or orally.

[0020] In an embodiment, values of at least a subset of the plurality of biomarkers are obtained by performing an immunoassay. In an embodiment, the immunoassay is a quantitative lateral flow assay. In an embodiment, performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies. In an embodiment, values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.Attorney Docket No: PNS-001WO

[0021] In one aspect provided is a method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise two or more of procalcitonin, IL-6, triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL); and generating a prediction of a subtype for the patient with the respiratory infection based on the values of the plurality of biomarkers.

[0022] In an embodiment, the plurality of biomarkers further comprise one or more additional biomarkers, each additional biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL- 8, and neutrophil gelatinase-associated lipocalin (NGAL). In an embodiment, the plurality of biomarkers further comprise two, three, four, five, or six additional markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL). In an embodiment, the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin- 3, and IL- 8. In an embodiment, the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL-10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a. In an embodiment, the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein. In an embodiment, the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.

[0023] In an embodiment, the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon- a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes,Attorney Docket No: PNS-001WOGlucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use. In an embodiment, the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Eymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Fiver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Fiver Disease, Diabetes, Chloride, and Active Cancer.

[0024] In an embodiment, the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid. In an embodiment, the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

[0025] In an embodiment, the method further comprises: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid. In an embodiment, obtaining or having obtained values for at least a subset of the plurality of biomarkers comprises performing an immunoassay. In an embodiment, the immunoassay is a quantitative lateral flow assay. In an embodiment, performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies. In an embodiment, values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient. In one aspect provided is a non-transitory computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform the method as defined in any of the embodiments described herein.Additionally disclosed herein is a method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise two or more of PCT, IE-6, Pentraxin-3, MIP-3a, IF-lra, IE-8, NGAE, TREM-1, Angiopoietin-2, IE- ip, VCAM-1, Eactate, Thrombomodulin, G-CSF, IL-10, Granzyme B, TNF-a, C-Reactive Protein, TGF-a, Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1,Attorney Docket No: PNS-OOIWOBilirubin, Albumin, SOFA Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la, RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, RBC, Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use. In one aspect, the plurality of biomarkers comprise two or more of IL-6, IL-8, TNF-a, MIP-la, IL-ip, and IL- Ira. In one aspect, the plurality of biomarkers comprise two or more of IL-2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, and Troponin. In one aspect, the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid. In one aspect, the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone. In one aspect, methods disclosed herein further comprise: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid.

[0026] Additionally disclosed herein is a method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise two or more of PCT, IL-6, Pentraxin-3, MIP-3a, IL-lra, IL-8, NGAL, TREM-1, Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL-10, Granzyme B, TNF-a, C-Reactive Protein, TGF-a, Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E- Selectin, IL-15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la, RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, RBC, Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use; and generating a prediction of a subtype for the patient with the respiratoryAttorney Docket No: PNS-001WO infection based on the values of the plurality of biomarkers. In one aspect, the plurality of biomarkers comprise two or more of IL-6, IL-8, TNF-a, MIP-la, IL-ip, and IL- Ira. In one aspect, the plurality of biomarkers comprise two or more of IL-2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, and Troponin. In one aspect, the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid. In one aspect, the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone. In one aspect, methods disclosed herein further comprise: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid. In one aspect, obtaining or having obtained values for at least a subset of the plurality of biomarkers comprises performing an immunoassay. In one aspect, the immunoassay is a quantitative lateral flow assay. In one aspect, performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies. In one aspect, values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.BRIEF DESCRIPTION OF THE DRAWINGS

[0027] These and other features, aspects, and advantages of the present invention will become better understood with regard to the following description and accompanying drawings.

[0028] Figure (FIG.) 1A depicts an overall system environment for subtyping a subject, in accordance with an embodiment.

[0029] FIG. IB is an example block diagram of the activity prediction system, in accordance with an embodiment.

[0030] FIG. 2 depicts an example set of training data, in accordance with an embodiment.

[0031] FIG. 3 shows a flow process for generating a prediction of a subtype for a patient, in accordance with an embodiment.

[0032] FIG. 4 illustrates an example computer for implementing the entities shown in FIGS. 1A-1B and 2-3.

[0033] FIG. 5 shows an example flow diagram for predicting a patient subtype, which involves a protein biomarker reader and an Al biomarker algorithm.

[0034] FIG. 6 shows exemplary markers for predicting a patient subtype.Attorney Docket No: PNS-001WO

[0035] FIG. 7 shows an example detailed flow diagram for subtyping a patient as either a responder (CAP SR Positive) or a non-responder (CAP SR Negative) to a steroid.

[0036] FIG. 8 shows example clustering of responders and non-responders to steroid (left) and the average treatment effect of responders and non-responders (right).

[0037] FIG. 9 shows the performance of the example Steroids ImmunoScore, which substantially outperformed all 3 severity scores (e.g., PSI, CURB-65, and SOFA).

[0038] FIG. 10 shows the correlation between steroid responders and upregulation of proinflammatory cytokines.

[0039] FIGs. 11A-11C shows the differential expression of certain inflammatory markers (e.g., cytokines and chemokines, G-CSF and GM-CSF, and Granzyme B) in responders and non-responders.

[0040] FIG. 12A shows the thresholding for categorizing patients above and below the threshold cutoff. The maximal value of cortiCAP with an upper bound to the 95% confidence interval of the AlPW-estimated conditional average treatment effect (CATE) below 0 was -0.007, corresponding to CATE point estimate of -12.5%.

[0041] FIG. 12B shows the average treatment effect of responders and non-responders.

[0042] FIGs. 13A-13B show standard curve and regression fits for four protein biomarkers (angiopoietin-2, IL-8, IP- 10, and Pentraxin-3).

[0043] FIGs. 14A-14B show the results of the validation, which demonstrated a strong association between concentrations measured on the Protein Biomarker Reader and Luminex xMAPs.DETAILED DESCRIPTIONI. Definitions

[0044] Terms used in the claims and specification are defined as set forth below unless otherwise specified.

[0045] The terms “subject” and “patient” are used interchangeably and encompasses a cell, tissue, or organism, human or non-human, whether in vivo, ex vivo, or in vitro, male or female.

[0046] The term “mammal” encompasses both humans and non-humans and includes but is not limited to humans, non-human primates, canines, felines, murines, bovines, equines, and porcines.Attorney Docket No: PNS-001WO

[0047] The term “sample” can include a single cell or multiple cells or fragments of cells or an aliquot of body fluid, such as a blood sample, taken from a subject, by means including venipuncture, excretion, ejaculation, massage, biopsy, needle aspirate, lavage sample, scraping, surgical incision, or intervention or other means known in the art. Examples of an aliquot of body fluid include amniotic fluid, aqueous humor, bile, lymph, breast milk, interstitial fluid, blood, blood plasma, cerumen (earwax), Cowper’s fluid (pre-ejaculatory fluid), chyle, chyme, female ejaculate, menses, mucus, saliva, urine, vomit, tears, vaginal lubrication, sweat, serum, semen, sebum, pus, pleural fluid, cerebrospinal fluid, synovial fluid, intracellular fluid, and vitreous humour.

[0048] The terms “marker,” “markers,” “biomarker,” and “biomarkers” encompass any of demographic information (e.g., age or gender), clinical laboratory data, vital measurements, respiratory information, comorbidities, clinical scores, and a biological marker. Example biological markers include, without limitation, lipids, lipoproteins, proteins, cytokines, chemokines, growth factors, peptides, nucleic acids, genes, and oligonucleotides, together with their related complexes, metabolites, mutations, variants, polymorphisms, modifications, fragments, subunits, degradation products, elements, and other analytes or sample-derived measures. A marker can also include mutated proteins, mutated nucleic acids, variations in copy numbers, and / or transcript variants, in circumstances in which such mutations, variations in copy number and / or transcript variants are useful for generating a predictive model, or are useful in predictive models developed using related markers (e.g., non-mutated versions of the proteins or nucleic acids, alternative transcripts, etc.).

[0049] The term "antibody" is used in the broadest sense and specifically covers monoclonal antibodies (including full length monoclonal antibodies), polyclonal antibodies, multispecific antibodies (e.g., bispecific antibodies), and antibody fragments that are antigen-binding so long as they exhibit the desired biological activity, e.g., an antibody or an antigen-binding fragment thereof.

[0050] "Antibody fragment," and all grammatical variants thereof, as used herein are defined as a portion of an intact antibody comprising the antigen binding site or variable region of the intact antibody, wherein the portion is free of the constant heavy chain domains (i.e. CH2, CH3, and CH4, depending on antibody isotype) of the Fc region of the intact antibody. Examples of antibody fragments include Fab, Fab', Fab'-SH, F(ab')2, and Fv fragments; diabodies; any antibody fragment that is a polypeptide having a primary structureAttorney Docket No: PNS-001WO consisting of one uninterrupted sequence of contiguous amino acid residues (referred to herein as a "single-chain antibody fragment" or "single chain polypeptide").

[0051] The term “biomarker panel” refers to a set biomarkers that are informative for predicting likely responders or non-responders to a steroid (e.g., corticosteroid). For example, values of biomarkers in the biomarker panel can be informative for predicting likely responders or non-responders to a steroid (e.g., corticosteroid). In various embodiments, a biomarker panel can include 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, or 50 biomarkers.

[0052] The term “obtaining a dataset associated with a sample” encompasses obtaining a set of data determined from at least one sample. Obtaining a dataset encompasses obtaining a sample and processing the sample to experimentally determine the data. The phrase also encompasses receiving a set of data, e.g., from a third party that has processed the sample to experimentally determine the dataset. Additionally, the phrase encompasses mining data from at least one database or at least one publication or a combination of databases and publications. A dataset can be obtained by one of skill in the art via a variety of known ways including stored on a storage memory.

[0053] It must be noted that, as used in the specification, the singular forms “a,” “an” and “the” include plural referents unless the context clearly dictates otherwise.II. System Environment Overview

[0054] FIG. 1A depicts an overview of a system environment 100 for predicting a subtype for a subject, in accordance with an embodiment. The system environment 100 provides context in order to introduce a marker quantification assay 120, clinical data 125, and a prediction system 130 for generating a subtype prediction 140.

[0055] In various embodiments, the subject 110 undergoes clinical testing to generate clinical data 125. For example, clinical data 125 can include any of demographic information of the subject 110, clinical laboratory data of the subject 110, vitals of the subject 110, respiratory information of the subject 110, comorbidities of the subject 110, and / or clinical scores of the subject 110. Example demographic information, clinical laboratory data, vitals, respiratory information, comorbidities, and clinical scores are disclosed herein (e.g., in at least Table 6). In various embodiments, clinical data 125 can be obtained from electronic health record (EHR) of the subject.Attorney Docket No: PNS-001WO

[0056] In various embodiments, a test sample is obtained from the subject 110. The sample can be obtained by the subject or by a third party, e.g., a medical professional. Examples of medical professionals include physicians, emergency medical technicians, nurses, first responders, psychologists, phlebotomists, medical physics personnel, nurse practitioners, surgeons, dentists, and any other obvious medical professional as would be known to one skilled in the art.

[0057] The test sample can, in various embodiments, be tested to determine values of one or more markers e.g., by performing the marker quantification assay 120 and / or can be tested to determine clinical data 125. The marker quantification assay 120 determines quantitative values of one or more biomarkers from the test sample. In various embodiments, the one or more biomarkers are protein biomarkers. Example protein biomarkers are disclosed herein (e.g., in at least Tables 1 and 6). The marker quantification assay 120 may be an immunoassay, and more specifically, a multi-plex immunoassay, examples of which are described in further detail below. The expression levels of various biomarkers can be obtained in a single run using a single test sample obtained from the subject 110. The quantified expression values of the biomarkers are provided to the prediction system 130.

[0058] Generally, the prediction system 130 includes one or more computers, embodied as a computer system 400 as discussed below with respect to FIG. 4. Therefore, in various embodiments, the steps described in reference to the prediction system 130 are performed in silico. The prediction system 130 analyzes at least the received biomarker values from the marker quantification assay 120 to generate a subtype prediction for the subject 110. In various embodiments, the prediction system 130 analyzes both the received biomarker values from the marker quantification assay 120 and the clinical data 125 to generate a subtype prediction for the subtype 110. The subtype prediction can be used to categorize the subject 110 in one of a plurality of categories. In various embodiments, the subtype prediction can be used to categorize the subject 110 into one of a responder category or non-responder category.

[0059] In various embodiments, the marker quantification assay 120 and the prediction system 130 can be employed by different parties. For example, a first party performs the marker quantification assay 120 which then provides the results to a second party which implements the prediction system 130. For example, the first party may be a clinical laboratory that obtains test samples from subjects 110 and performs the assay 120 on the testAttorney Docket No: PNS-001WO samples. The second party receives the expression values of biomarkers resulting from the performed assay 120 and analyzes the expression values using the prediction system 130.

[0060] In various embodiments, the clinical data 125 can be obtained by a party that differs from the party that operates the prediction system 130. For example, a first party obtains the clinical data 125 which then provides the clinical data to a different party which implements the prediction system 130. For example, the first party may be a physician’s office or a hospital that obtains clinical data 125 from subjects 110. The second party receives the clinical data 125 and analyzes at least the clinical data 125 using the prediction system 130.

[0061] Reference is now made to FIG. IB which depicts a block diagram illustrating the computer logic components of the prediction system 130, in accordance with an embodiment. Specifically, the prediction system 130 may include a model training module 150, a model deployment module 160, and a training data store 170.

[0062] Each of the components of the prediction system 130 is hereafter described in reference to two phases: 1) a training phase and 2) a deployment phase. More specifically, the training phase refers to the building and training of one or more predictive models based on training data that includes values of biomarkers (e.g., quantitative values of markers such as protein biomarkers and / or values of clinical data) obtained from individuals. Therefore, the predictive models are trained to generate a subtype prediction for a subject based on biomarker values (e.g., quantitative values of markers such as protein biomarkers and / or values of clinical data). During the deployment phase, a predictive model is applied to biomarker values from a test sample obtained from a subject of interest and clinical data 125 of the subject of interest (e.g., quantitative values of markers such as protein biomarkers and / or values of clinical data) to generate a subtype prediction for the subject of interest.

[0063] In some embodiments, the components of the prediction system 130 are applied during one of the training phase and the deployment phase. For example, the model training module 150 and training data store 170 (indicated by the dotted lines in FIG. IB) are applied during the training phase whereas the model deployment module 160 is applied during the deployment phase. In various embodiments, the training phase and the deployment phase can be performed to enable continuously trained models. For example, the model training module 150 can train a model that the model deployment module 160 can subsequently deploy. The same model can undergo additional training by the model training module 150 (e.g., continuously trained using, for example, new training data that is obtained). Therefore,Attorney Docket No: PNS-001WO as the model is continuously trained, it can exhibit improved prediction capacity when analyzing samples during deployment.

[0064] In various embodiments, the components of the prediction system 130 can be performed by different parties depending on whether the components are applied during the training phase or the deployment phase. In such scenarios, the training and deployment of the predictive model are performed by different parties. For example, the model training module 150 and training data store 170 applied during the training phase can be employed by a first party (e.g., to train a predictive model) and the model deployment module 160 applied during the deployment phase can be performed by a second party (e.g., to deploy the predictive model).III. Predictive modelIII.A. Trainins a Predictive model

[0065] During the training phase, the model training module 150 trains one or more predictive models using training data comprising values of biomarkers. Referring to FIG. IB, the training data may be stored in the training data store 170. In various embodiments, the prediction system 130 obtains training data comprising values of biomarkers from a third party. The third party may have processed test samples to determine the biomarker values and / or may have obtained biomarker values in the form of clinical data e.g., from electronic health records. In various embodiments, the training data further includes reference ground truths that indicate an outcome, such as a responder or non-responder. As an example, the training data includes reference ground truths that identify a responder or a non-responder e.g., to a treatment, such as a steroid treatment.

[0066] Reference is made to FIG. 2, which depicts an example set of training data 250, in accordance with an embodiment. As shown in FIG. 2, the training data 250 includes data corresponding to multiple individuals (e.g., column 1 depicting individual 1, 2, 3, 4...). For each individual, the training data 250 includes values (e.g., Al, Bl, A2, B2, etc.) for different markers obtained from the corresponding individual. In some embodiments, the values are determined by the marker quantification assay 120 shown in FIG. 1A. In some embodiments, the values are of clinical data 125 shown in FIG. 1A and can be obtained from electronic health records. Although FIG. 2 depicts 4 individuals and 2 different markers (marker A and marker B), the training data 250 may include tens, hundreds, or thousands of individuals as well as tens, hundreds, or thousands of markers.Attorney Docket No: PNS-001WO

[0067] As shown in FIG. 2, a first training example (e.g., first row) of the training data refers to individual 1 and corresponding value of marker A (e.g., Al) and the value of marker B (e.g., Bl). Similarly, the second training example (e.g., second row) of the training data refers to individual 2 and corresponding value of marker A (e.g., A2) and the value of marker B (e.g., B2). Individuals 3 and 4 have corresponding marker values as shown in FIG. 2.

[0068] As shown in FIG. 2, the training data 250 further includes a reference ground truth (“Indication” column) that identifies whether the corresponding individual was a responder or non-responder e.g., to a treatment such as a steroid treatment. For example, referring to the first training example (e.g., first row), a “Responder” indication can reflect that Individual 1 responded to a treatment e.g., a steroid treatment. Similarly, a “non-responder” indication (e.g., individual 3 or individual 4) reflects that the corresponding individual did not respond to a treatment e.g., a steroid treatment.

[0069] In various embodiments, the predictive model is any one of a regression model (e.g., linear regression, logistic regression, or polynomial regression), decision tree, random forest, support vector machine, Naive Bayes model, k-means cluster, or neural network (e.g., feedforward networks, convolutional neural networks (CNN), deep neural networks (DNN), autoencoder neural networks, generative adversarial networks, or recurrent networks (e.g., long short-term memory networks (LSTM), bi-directional recurrent networks, deep bidirectional recurrent networks), or any combination thereof. In particular embodiments, the predictive model is a random forest model. In particular embodiments, the predictive model is a logistic regression model.

[0070] The predictive model can be trained using a machine learning implemented method, such as any one of a linear regression algorithm, logistic regression algorithm, decision tree algorithm, support vector machine classification, Naive Bayes classification, K-Nearest Neighbor classification, random forest algorithm, deep learning algorithm, gradient boosting algorithm, and dimensionality reduction techniques such as manifold learning, principal component analysis, factor analysis, autoencoder regularization, and independent component analysis, or combinations thereof. In various embodiments, the predictive model is trained using supervised learning algorithms, unsupervised learning algorithms, semi- supervised learning algorithms (e.g., partial supervision), weak supervision, transfer, multi-task learning, or any combination thereof.

[0071] In particular embodiments, the predictive model is trained using unsupervised learning algorithms. Example unsupervised learning approaches include clustering,Attorney Docket No: PNS-001WO endotyping, dimensional reduction (e.g., principle component analysis (PCA), embeddings and manifold learning), and causal inference procedures. Clustering algorithms can identify patterns in the underlying data and can reveal distinct phenotypes among subjects with respiratory infections. Thus, subjects in different clusters can respond differently to steroid therapy. PCA is used for dimensionality reduction, which enables the simplification of complex datasets into largely uncorrelated variables that capture the information in the complex datasets. This reduces the number of variables while preserving the needed information in the datasets. Manifold learning is another unsupervised dimensional reduction technique that learns the high-dimensional structure of the data without the use of predetermined classifications. Manifold learning is a particular non-linear dimension reduction technique. Causal inference procedures focuses on inferring causal relationships without direct supervision and is particularly useful in scenarios where labeled data is scarce or not available.

[0072] In various embodiments, the predictive model has one or more parameters, such as hyperparameters or model parameters. Hyperparameters are generally established prior to training. Examples of hyperparameters include the learning rate, depth or leaves of a decision tree, number of hidden layers in a deep neural network, number of clusters in a k- means cluster, penalty in a regression model, and a regularization parameter associated with a cost function. Model parameters are generally adjusted during training. Examples of model parameters include weights associated with nodes in layers of neural network, support vectors in a support vector machine, and coefficients in a regression model.

[0073] The model training module 150 trains one or more predictive models, each predictive model receiving, as input, one or more biomarkers. In various embodiments, the model training module 150 constructs a predictive model that receives, as input, expression values of two biomarkers. In various embodiments, the model training module 150 constructs a predictive model that receives, as input, expression values of three biomarkers. In various embodiments, the model training module 150 constructs a predictive model that receives, as input, expression values of four biomarkers. In some embodiments, the model training module 150 constructs a predictive model for more than four biomarkers. For example, a predictive model receives, as input, expression values of 8 biomarkers (e.g., 8 biomarkers categorized as Tier 1 in Table 2 or any of their corresponding substitute biomarkers in Table 10). As another example, the predictive model receives, as input, expression values of 17 biomarkers (e.g., 17 biomarkers categorized as Tier 1 or Tier 2 in Table 2 or any of theirAttorney Docket No: PNS-001WO corresponding substitute biomarkers in Table 10). As another example, the predictive model receives, as input, expression values of 21 biomarkers (e.g., 21 biomarkers categorized as Tier 1, Tier 2, or Tier 3 in Table 2 or any of their corresponding substitute biomarkers in Table 10).

[0074] A predictive model is iteratively trained using, as input, the values of the biomarkers for each individual. For example, referring again to FIG. 2, an iteration involves providing a training example (e.g., a row of the training data) that includes the value of biomarkers (e.g., “Al” and “Bl”) for a particular individual (e.g., individual 1). Each predictive model is trained on reference ground truth data that includes the indication (e.g., the responder or nonresponder outcome). In various embodiments, over training iterations, a predictive model is trained (e.g., the parameters are tuned) to minimize a prediction error between a subtype prediction outputted by the predictive model and the ground truth data. In various embodiments, the prediction error is calculated based on a loss function, examples of which include a LI regularization (Lasso Regression) loss function, a L2 regularization (Ridge Regression) loss function, or a combination of LI and L2 regularization (ElasticNet).III.B. Deplo ins a Predictive model

[0075] During the deployment phase, the model deployment module 160 (as shown in EIG. IB) obtains values for a plurality of biomarkers and generates a prediction of a subtype by analyzing the values of the plurality of biomarkers. In some embodiments, the subject has been previously diagnosed with a disease, such as a respiratory infection (e.g., SARI or pneumonia) or sepsis. Here, the deployment of the predictive model enables in silica prediction of a patient subtype based on at least the values of a plurality of biomarkers derived from the subject.

[0076] Generally, methods involve deploying a predictive model that analyzes values of biomarkers of a biomarker panel, such as a biomarker panel disclosed herein. Bor example, the biomarker panel may be a multivariate biomarker panel that includes 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, or 50 biomarkers. In various embodiments, the values of the plurality of biomarkers are provided as input to the predictive model. The predictive model analyzes the values of the plurality of biomarkers and generates a subtype prediction. In particular embodiments, the plurality of biomarkers include at least one or more protein biomarkers and one or more clinical laboratory data.Attorney Docket No: PNS-001WO

[0077] In various embodiments, the subtype prediction is a predicted score. The predicted score can be useful for determining whether the subject is a likely responder or a nonresponder e.g., to a steroid treatment. As used herein, a “responder” is an individual who benefits as a result of the treatment (e.g., improved patient outcome and / or amelioration of disease symptoms). As used herein, a “non-responder” is an individual who does not benefit from a treatment (e.g., worst or no change to patient outcome and / or worsening or no change to disease symptoms). For example, a non-responder includes an individual who is harmed by the treatment e.g., steroid treatment.

[0078] In various embodiments, the predicted score outputted by the prediction model is compared to one or more reference scores to determine a subtype. Reference scores refer to previously determined scores, further described below as “responder scores” or “non- responder scores,” that correspond to patients that responded or patients that did not respond to a treatment e.g., a steroid treatment. For example, the one or more scores may be “responder scores” corresponding to patients with disease (e.g., a respiratory infection such as SARI or pneumonia, or sepsis) that responded to a treatment e.g., a steroid treatment. The responder score can be a statistical combination of scores (e.g., an average or a median) corresponding to patients with disease (e.g., a respiratory infection such as SARI or pneumonia, or sepsis) that responded to a treatment e.g., a steroid treatment. As another example, the one or more scores may be “non-responder scores” corresponding to patients with disease (e.g., a respiratory infection such as SARI or pneumonia, or sepsis) that did not respond to a treatment e.g., a steroid treatment. The non-responder score can be a statistical combination of scores (e.g., an average or a median) corresponding to patients with disease (e.g., a respiratory infection such as SARI or pneumonia, or sepsis) that did not respond to a treatment e.g., a steroid treatment.

[0079] In various embodiments, the predicted score outputted by the prediction model can be compared to one or more reference scores and therefore, the subject is categorized into a subtype based on the comparison. For example, the predicted score outputted by the prediction model can be compared to a responder score and / or to a non-responder score to categorize the subject into a subtype.

[0080] In one embodiment, the predicted score outputted by the prediction model can be compared to the responder score. In various embodiments, the subject can be classified as a likely non-responder if the predicted score of the subject is significantly different (e.g., p- value < 0.05) in comparison to the responder score. In one embodiment, the predicted scoreAttorney Docket No: PNS-001WO outputted by the prediction model can be compared to the non-responder score. The subject can be classified as a likely responder if the predicted score of the subject is significantly different (e.g., p-value <0.05) in comparison to the non-responder score.

[0081] In some embodiments, the predicted score outputted by the prediction model is compared to both the responder score and the non-responder score. For example, the subject can be classified as a likely responder if the predicted score of the subject is significantly different (e.g., p-value < 0.05) in comparison to the non-responder score and not significantly different (e.g., p-value >0.05) in comparison to the responder score. As another example, the subject can be classified as a likely non-responder if the predicted score of the subject is significantly different (e.g., p-value < 0.05) in comparison to the responder score and not significantly different (e.g., p-value >0.05) in comparison to the non-responder score.

[0082] In some embodiments, the predicted score outputted by the prediction model is compared to a threshold value (e.g., a cutoff). The threshold value may be a pre-determined value that distinguishes subjects that are likely responders and other subjects that are likely non-responders to a treatment e.g., a steroid treatment. In various embodiments, the threshold value is determined by analyzing a patient cohort of responders and non-responders (e.g., to a treatment such as a steroid treatment). To identify the threshold value, various different candidate threshold values are explored, where each candidate threshold value separates a first plurality of subjects and a second plurality of subjects. For example, the average treatment effect can be determined for each candidate threshold value. The threshold value corresponding to the largest proportion of patients with a statistically significant treatment benefit can be identified as the pre-determined threshold value for further use.

[0083] In various embodiments, depending on the determined subtype of the subject, the subject can undergo treatment. In other words, the assessment can guide the treatment of the subject. For example, if the subject is classified as a likely responder to a treatment e.g., a steroid, the subject can be administered the treatment. In some embodiments, if the subject is classified as a likely non-responder to the treatment e.g., a steroid, the treatment can be withheld from the subject.IV. Biomarker Panel

[0084] Disclosed herein are methods for predicting a patient subtype by implementing a biomarker panel including one or more biomarkers. For example, methods involve implementing a biomarker panel to categorize a patient as a likely responder or a likely nonAttorney Docket No: PNS-001WO responder to a therapeutic, e.g., a steroid therapeutic. As used herein, the terms “marker,” “markers,” “biomarker,” and “biomarkers” encompass any of demographic information (e.g., age or gender), clinical laboratory data, vital measurements, respiratory information, comorbidities, clinical scores, and a biological marker. Example biological markers include, without limitation, lipids, lipoproteins, proteins, cytokines, chemokines, growth factors, peptides, nucleic acids, genes, and oligonucleotides, together with their related complexes, metabolites, mutations, variants, polymorphisms, modifications, fragments, subunits, degradation products, elements, and other analytes or sample-derived measures. In particular embodiments, a biomarker includes a protein biomarker. Example demographic information (e.g., age or gender), clinical laboratory data, vital measurements, respiratory information, comorbidities, clinical scores, and protein markers are disclosed herein (e.g., see Tables 1, 2, and 6).

[0085] In various embodiments, clinical scores disclosed herein include sequential organ failure assessment (SOFA) scores, such as SOFA Respiratory, SOFA Coagulation, SOFA Cardiovascular, SOFA Kidneys, and SOFA Fiver. Further details of SOFA scores are described in Eambden S, et al., The SOFA score-development, utility and challenges of accurate assessment in clinical trials. Crit Care. 2019 Nov 27;23(1):374, which is incorporated by reference in its entirety.

[0086] In various embodiments, the biomarker panel involves one biomarker (e.g., a univariate biomarker panel). In various embodiments, the univariate biomarker panel includes any one of the biomarkers shown in Tables 1, 2, 5, or 6. In various embodiments, the univariate biomarker panel includes any one of the biomarkers categorized as “Tier 1” in Table 5. In various embodiments, the univariate biomarker panel includes any one of the biomarkers categorized as “Tier 2” in Table 5. In various embodiments, the univariate biomarker panel includes any one of the biomarkers categorized as “Tier 3” in Table 5. In various embodiments, the univariate biomarker panel includes any one of the biomarkers categorized as “Tier 4” in Table 5.

[0087] In some embodiments, methods for subtyping patients with a respiratory infection involve implementing a multivariate biomarker panel. In such embodiments, the multivariate biomarker panel includes two or more biomarkers. In various embodiments, the multivariate biomarker panel includes two biomarkers. In various embodiments, the multivariate biomarker panel includes 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22,Attorney Docket No: PNS-001WO23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, or 50 biomarkers.

[0088] In various embodiments, the biomarker panel includes two or more of the biomarkers shown in Tables 1, 2, 5, or 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1” in Table 5. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1” or “Tier 2” in Table 5. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1”, “Tier 2”, or “Tier 3” in Table 5. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1”, “Tier 2”, “Tier 3” , or “Tier 4” in Table 5.

[0089] In various embodiments, the biomarker panel includes two or more of the biomarkers shown in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 2” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1” or “Tier 2” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 3” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1” or “Tier 3” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1”, “Tier 2”, or “Tier 3” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1” or “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 2” and “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 3” and “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1, “Tier 2”, and “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1”, “Tier 3”, and “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 2”, “Tier 3”, and “Tier 4” in Table 6. In various embodiments, the biomarker panel includes two or more of the biomarkers categorized as “Tier 1”, “Tier 2”, “Tier 3” , or “Tier 4” in Table 6.Attorney Docket No: PNS-001WO

[0090] In various embodiments, the biomarker panel includes two or more biomarkers selected from procalcitonin (PCT), IL-6, Pentraxin-3, MIP-3a, IL- Ira, IL-8, NGAL, and TREM-1. In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL). In various embodiments, the biomarker panel includes at least a first biomarker of procalcitonin or IL-6, and at least a second biomarker of Pentraxin-3 or IL-8. In various embodiments, the biomarker panel includes any one of the combinations of 1) PCT and Pentraxin-3, 2) PCT and IL-8, 3) IL-6 and Pentraxin-3, or 4) IL- 6 and IL- 8.

[0091] In various embodiments, the biomarker panel includes a plurality of biomarkers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL). In various embodiments, the biomarker panel includes two, three, four, five, or six markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase- associated lipocalin (NGAL).

[0092] In various embodiments, the biomarker panel includes each of procalcitonin, IL-6, Pentraxin-3, and IL-8. In various embodiments, the biomarker panel includes procalcitonin, IL-6, Pentraxin-3, and IL-8, and further includes one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a. In various embodiments, the biomarker panel includes procalcitonin, IL- 6, Pentraxin-3, and IL-8, and further includes Angiopoietin-2, Lactate, and C-Reactive Protein.

[0093] In various embodiments, the biomarker panel includes procalcitonin, IL-6, Pentraxin- 3, IL-8, Angiopoietin-2, Lactate, and C-Reactive Protein and further includes one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, sequential organ failure assessment (SOFA) Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la. In various embodiments, the biomarker panel includes procalcitonin, IL-6, Pentraxin-3, IL-8, Angiopoietin-2, Lactate, and C-Reactive Protein and further includes two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen,Attorney Docket No: PNS-001WOBilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.

[0094] In various embodiments, the biomarker panel includes procalcitonin, IL-6, Pentraxin- 3, IL-8, Angiopoietin-2, Lactate, C-Reactive Protein, Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium, and further includes one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use. In various embodiments, the biomarker panel includes procalcitonin, IL-6, Pentraxin-3, IL-8, Angiopoietin-2, Lactate, C-Reactive Protein, Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium, and further includes two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer. In various embodiments, the biomarker panel includes procalcitonin, IL-6, Pentraxin-3, IL-8, Angiopoietin-2, Lactate, C-Reactive Protein, Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium, and further includes each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.Attorney Docket No: PNS-001WO

[0095] In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL), and further includes one or more core inflammatory response biomarkers (as identified in Table 2). In various embodiments, the one or more core inflammatory response biomarkers are selected from IL-ip, TNF-a, IL-6, C-Reactive Protein, Pentraxin-3, IL- Ira, IL- 10, IL-8, NGAL, TREM-1, G-CSF, MIP-3a, PCT, Granzyme B, and TRAIL.

[0096] In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL), and further includes one or more immune cell activation biomarkers or chemokines (as identified in Table 2). In various embodiments, the one or more immune cell activation features or chemokines are selected from Neutrophils, WBC, Lymphocytes, Monocytes, Platelets, MCP-1, MIP-la, MIP-1 (CCL4), IP-10, PD-L1, GM-CSF, FLT3 Ligand, IL- 15, IL-7, Interferon-a, and Interferon- y. In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL- 6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase- associated lipocalin (NGAL), and further includes one or more endothelia and coagulation markers. In various embodiments, the one or more endothelia and coagulation markers are selected from Thrombomodulin, Tissue Factor, VCAM-1, E-Selectin, Angiopoietin-2, Angiopoietin-1, VEGF, TGF-a, and Leptin.

[0097] In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL), and further includes one or more metabolic and organ function features. In various embodiments, the one or more metabolic and organ function features are selected from Lactate, Blood Urea Nitrogen, Creatinine, Bilirubin, AST, ALT, ALP, Albumin, Calcium, Potassium, Sodium, Chloride, Total CO2, Hemoglobin, RBC, RDW.

[0098] In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1),Attorney Docket No: PNS-001WO and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL), and further includes one or more clinical parameters or interventions. In various embodiments, the one or more clinical parameters or interventions are selected from Systolic BP, Diastolic BP, Heart Rate, Respiratory Rate, Temperature, Pulse Oximetry, SOFA Kidney, SOFA Cardiovascular, SOFA Respiratory, SOFA Coagulation, SOFA Liver, SIRS, Shock, Vasopressors, and Ventilator.

[0099] In various embodiments, the biomarker panel includes a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL), and further includes one or more demographics or comorbidities. In various embodiments, the one or more demographics or comorbidities are selected from Age, Gender, Diabetes, Chronic Kidney Disease, Chronic Liver Disease, Active Cancer, CHF, Emphysema, Atelectasis, Rheumatoid Arthritis, Lupus, Immunodeficiency, Acute Bronchitis, Asphyxiation, Nicotine Use, and Glucocorticoid Deficiency.V. Assays

[0100] As shown in FIG. 1A, the system environment 100 involves implementing a marker quantification assay 120 for evaluating levels of one or more biomarkers (e.g., protein biomarkers). Examples of an assay (e.g., marker quantification assay 120) for one or more markers include DNA assays, microarrays, polymerase chain reaction (PCR), RT-PCR, Southern blots, Northern blots, antibody-binding assays, enzyme-linked immunosorbent assays (ELISAs), flow cytometry, protein assays, Western blots, nephelometry, turbidimetry, chromatography, mass spectrometry, immunoassays, including, by way of example, but not limitation, RIA, immunofluorescence, immunochemiluminescence, immunoelectrochemiluminescence, immunoprecipitation, or competitive immunoassays. In particular embodiments, the marker quantification assay 120 is an immunoassay, such as a lateral flow immunoassay. The information from the assay can be sent to a computer system. The information can be quantitative and / or qualitative, such as observing patterns or fluorescence, which can be translated into a quantitative measure by a user or automatically by a reader or computer system.Attorney Docket No: PNS-001WO

[0101] Various immunoassays designed to quantitate markers can be used in screening including multiplex assays. Measuring the concentration of a target marker in a sample or fraction thereof can be accomplished by a variety of specific assays. For example, a conventional sandwich type assay can be used in an array, ELISA, RIA, etc. format. Other immunoassays include Ouchterlony plates that provide a simple determination of antibody binding and lateral flow immunoassays. Additionally, Western blots can be performed on protein gels or protein spots on filters, using a detection system specific for the markers as desired, conveniently using a labeling method.

[0102] Protein based analysis, using an antibody that specifically binds to a polypeptide (e.g. marker), can be used to quantify the marker level in a test sample obtained from a subject. In various embodiments, an antibody that binds to a marker can be a monoclonal antibody. In various embodiments, an antibody that binds to a marker can be a polyclonal antibody. For multiplex analysis of markers, arrays containing one or more marker affinity reagents, e.g. antibodies can be generated. Such an array can be constructed comprising antibodies against markers. Detection can utilize one or a panel of marker affinity reagents, e.g. a panel or cocktail of affinity reagents specific for one, two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, twenty seven, twenty eight, twenty nine, thirty, thirty one, thirty two, thirty three, thirty four, thirty five, thirty six, thirty seven, thirty eight, thirty nine, forty, forty one, forty two, forty three, forty four, forty five, forty six, forty seven, forty eight, forty nine, or fifty or more markers.

[0103] In various embodiments, the multiplex assay involves the use of oligonucleotide labeled antibody probes that bind to target biomarkers and allow for subsequent quantification of biomarkers. One example of a multiplex assay that involves oligonucleotide labeled antibody probes is the Proximity Extension Assay (PEA) technology (Olink Proteomics). Briefly, a pair of oligonucleotide labeled antibodies bind to a biomarker, wherein the two oligonucleotide sequences are complementary to one another. Thus, only when both antibodies bind to the target biomarker will the oligonucleotide sequences hybridize with one another. Mismatched oligonucleotide sequences (which occurs due to non-specific binding of antibodies or cross -reactivity of antibodies) will not hybridize and therefore, will not result in a readout. Hybridized oligonucleotide sequences undergo nucleic acid extension and amplification, followed by quantification using microfluidic qPCR.Attorney Docket No: PNS-001WO

[0104] In various embodiments, the multiplex assay involves the use of bead conjugated antibodies (e.g., capture antibodies) that enable the binding and detection of biomarkers. One example of a multiplex assay involving bead conjugated antibodies is Luminex’s xMAP® Technology. Here, bead conjugated antibodies are added to the sample along with biotinylated detection antibodies. Both antibodies are specific to the biomarkers of interest and therefore, form an antibody-antigen sandwich. Streptavidin is further added, which binds to the biotinylated detection antibodies and enables detection of the complex. The Luminex 200™ or FlexMap® analyzer are employed to identify and quantify the amount of the biomarker in the sample. In various embodiments, the multiplex assay represents an improvement over Luminex’s xMAP® technology, such as the Multi- Analyte Profile (MAP) technology by Myriad Rules Based Medicine (RBM), Inc.

[0105] In various embodiments, prior to implementation of a marker quantification assay 120 (e.g., an immunoassay), a sample obtained from a subject can be processed. In various embodiments, processing the sample enables the implementation of the marker quantification assay 120 to more accurately evaluate expression levels of one or more biomarkers in the sample.

[0106] In various embodiments, the sample from a subject can be processed to extract biomarkers from the sample. In one embodiment, the sample can undergo phase separation to separate the biomarkers from other portions of the sample. For example, the sample can undergo centrifugation (e.g., pelleting or density gradient centrifugation) to separate larger and / or more dense entities in the sample (e.g., cells and other macromolecules) from the biomarkers. Other examples include filtration (e.g., ultrafiltration) to phase separate the biomarkers from other portions of the sample.

[0107] In various embodiments, the sample from a subject can be processed to produce a sub-sample with a fraction of biomarkers that were in the sample. In various embodiments, producing a fraction of biomarkers can involve performing a protein fractionation procedure. One example of protein fractionation procedures include chromatography (e.g., gel filtration, ion exchange, hydrophobic chromatography, or affinity chromatography). In particular embodiments, the protein fractionation procedure involves affinity purification or immunoprecipitation where biomarkers are bound by specific antibodies. Such antibodies can be immobilized on a support, such as a magnetic particle or nanoparticle or a plate.

[0108] In various embodiments, the sample from the subject is processed to extract biomarkers from the sample and further processed to produce a sub-sample with a fraction ofAttorney Docket No: PNS-001WO extracted biomarkers. Altogether, this enables a purified sub-sample of biomarkers that are of particular interest. Thus, implementing an assay (e.g., an immunoassay) for evaluating expression levels of the biomarkers of particular interest can be more accurate and of higher quality. In various embodiments, the biomarkers of particular can be biomarkers of a biomarker panel, embodiments of which are described herein (e.g., a biomarker panel including one or more biomarkers of Table 6).VI. Therapeutic Agents and Compositions for Therapeutic Agents

[0109] Methods disclosed herein involve categorizing patients into different subtypes e.g., responder or non-responder to therapeutic agent, such as a steroid treatment. In various embodiments, a therapeutic agent disclosed herein is a steroid, such as a corticosteroid. Example corticosteroids include any one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone. In particular embodiments, the corticosteroid is hydrocortisone. In particular embodiments, the corticosteroid is methylprednisolone. In particular embodiments, the corticosteroid is cortisone. In particular embodiments, the corticosteroid is prednisone. In particular embodiments, the corticosteroid is prednisolone. In particular embodiments, the corticosteroid is dexamethasone.

[0110] In various embodiments, the therapeutic agent (e.g., the steroid such as a corticosteroid) is formulated as a pharmaceutical composition for administering to a subject (e.g., a subject determined to be a likely responder using the methods disclosed herein). A pharmaceutical composition administered to a subject includes an active agent such as a steroid e.g., corticosteroid described above. The steroid e.g., corticosteroid is present in a therapeutically effective amount, i.e., an amount sufficient when administered to treat a disease (e.g., respiratory infection such as SARI or pneumonia). The compositions can also include various other agents to enhance delivery and efficacy, e.g. to enhance delivery and stability of the active ingredients. Thus, for example, the compositions can also include, depending on the formulation desired, pharmaceutically-acceptable, non-toxic carriers or diluents, which are defined as vehicles commonly used to formulate pharmaceutical compositions for animal or human administration. The diluent is selected so as not to affect the biological activity of the combination. Examples of such diluents are distilled water, buffered water, physiological saline, PBS, Ringer’s solution, dextrose solution, and Hank’s solution. In addition, the pharmaceutical composition or formulation can include otherAttorney Docket No: PNS-001WO carriers, adjuvants, or non-toxic, nontherapeutic, nonimmunogenic stabilizers, excipients and the like. The compositions can also include additional substances to approximate physiological conditions, such as pH adjusting and buffering agents, toxicity adjusting agents, wetting agents and detergents. The composition can also include any of a variety of stabilizing agents, such as an antioxidant.

[0111] The therapeutic agent (e.g., steroid such as a corticosteroid) and / or pharmaceutical composition including the therapeutic agent described herein can be administered in a variety of different ways. Examples include administering a therapeutic agent (e.g., steroid such as a corticosteroid) and / or a composition containing a pharmaceutically acceptable carrier via oral, intranasal, rectal, topical, intraperitoneal, intravenous, intramuscular, subcutaneous, subdermal, transdermal, intrathecal, or intracranial method. As used herein, an oral administration can encompass administration through a feeding tube. In particular embodiments, the corticosteroid is administered intravenously, intramuscularly, or orally.VII. Respiratory Infections

[0112] Disclosed herein are methods for subtyping patients with a respiratory infection. Examples of respiratory infection include severe acute respiratory infection (SARI) and pneumonia, such as community-acquired pneumonia (CAP). Generally, pneumonia is an infection that inflames the air sacs in one or both lungs. The air sacs may fill with fluid or pus (purulent material), causing cough with phlegm or pus, fever, chills, and difficulty breathing. A variety of organisms, including bacteria, viruses and fungi, can cause pneumonia. When that infection is acquired in the community, as opposed to in the hospital, it is community- acquired pneumonia (CAP). SARI refers to any infection of the respiratory system that may interfere with normal breathing. SARI can affect the upper respiratory system (ranging from the sinuses to the vocal cords) or the lower respiratory system (ranging from the vocal cords and to the lungs), or both.

[0113] In various embodiments, a SARI can be defined by one or more of the following 4 criteria: a) Onset in the last 10 days, b)Results in hospitalization, c) at least one symptom or sign of respiratory illness (defined as purulent sputum, new or worsened cough, new or worsened dyspnea, tachypnea with respiratory rate >22 breaths per minute, rales, or bronchial breath sounds), and d) at least one symptom or sign of acute infection (defined as a temperature >38°C or <36°C, feverishness, chills, altered mental status, a white blood cellAttorney Docket No: PNS-001WO count of >12,000 / mm3, <4,000 / mm3, or >10% immature neutrophils, or imaging findings consistent with acute respiratory infection).VIII. Computer Implementation

[0114] The disclosed methods, including the methods of subtyping patients, are, in some embodiments, performed on one or more computers. For example, the building and deployment of a predictive model and database storage can be implemented in hardware or software, or a combination of both. In one embodiment of the invention, a machine-readable storage medium is provided, the medium comprising a data storage material encoded with machine readable data which, when using a machine programmed with instructions for using said data, is capable of displaying any of the datasets and execution and results of a predictive model of this invention. Such data can be used for a variety of purposes, such as patient monitoring, treatment considerations, and the like. The invention can be implemented in computer programs executing on programmable computers, comprising a processor, a data storage system (including volatile and non-volatile memory and / or storage elements), a graphics adapter, a pointing device, a network adapter, at least one input device, and at least one output device. A display is coupled to the graphics adapter. Program code is applied to input data to perform the functions described above and generate output information. The output information is applied to one or more output devices, in known fashion. The computer can be, for example, a personal computer, microcomputer, or workstation of conventional design.

[0115] Each program can be implemented in a high level procedural or object oriented programming language to communicate with a computer system. However, the programs can be implemented in assembly or machine language, if desired. In any case, the language can be a compiled or interpreted language. Each such computer program is preferably stored on a storage media or device (e.g., ROM or magnetic diskette) readable by a general or special purpose programmable computer, for configuring and operating the computer when the storage media or device is read by the computer to perform the procedures described herein. The system can also be considered to be implemented as a computer-readable storage medium, configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner to perform the functions described herein.Attorney Docket No: PNS-OOIWO

[0116] The signature patterns and databases thereof can be provided in a variety of media to facilitate their use. “Media” refers to a manufacture that contains the signature pattern information of the present invention. The databases of the present invention can be recorded on computer readable media, e.g. any medium that can be read and accessed directly by a computer. Such media include, but are not limited to: magnetic storage media, such as floppy discs, hard disc storage medium, and magnetic tape; optical storage media such as CD-ROM; electrical storage media such as RAM and ROM; and hybrids of these categories such as magnetic / optical storage media. One of skill in the art can readily appreciate how any of the presently known computer readable mediums can be used to create a manufacture comprising a recording of the present database information. "Recorded" refers to a process for storing information on computer readable medium, using any such methods as known in the art. Any convenient data storage structure can be chosen, based on the means used to access the stored information. A variety of data processor programs and formats can be used for storage, e.g. word processing text file, database format, etc.

[0117] In some embodiments, the disclosed methods, including the methods of subtyping patients, are performed on one or more computers in a distributed computing system environment (e.g., in a cloud computing environment). In this description, “cloud computing” is defined as a model for enabling on-demand network access to a shared set of configurable computing resources. Cloud computing can be employed to offer on-demand access to the shared set of configurable computing resources. The shared set of configurable computing resources can be rapidly provisioned via virtualization and released with low management effort or service provider interaction, and then scaled accordingly. A cloudcomputing model can be composed of various characteristics such as, for example, on- demand self-service, broad network access, resource pooling, rapid elasticity, measured service, and so forth. A cloud-computing model can also expose various service models, such as, for example, Software as a Service (“SaaS”), Platform as a Service (“PaaS”), and Infrastructure as a Service (“laaS”). A cloud-computing model can also be deployed using different deployment models such as private cloud, community cloud, public cloud, hybrid cloud, and so forth. In this description and in the claims, a “cloud-computing environment” is an environment in which cloud computing is employed.Attorney Docket No: PNS-001WOVIII.A. Example Computer

[0118] FIG. 4 illustrates an example computer 400 for implementing the entities shown in FIGs. 1A-1B and 2-3. The computer 400 includes at least one processor 402 coupled to a chipset 404. The chipset 404 includes a memory controller hub 420 and an input / output (VO) controller hub 422. A memory 406 and a graphics adapter 412 are coupled to the memory controller hub 420, and a display 418 is coupled to the graphics adapter 412. A storage device 408, an input device 414, and network adapter 416 are coupled to the I / O controller hub 422. Other embodiments of the computer 400 have different architectures.

[0119] The storage device 408 is a non-transitory computer-readable storage medium such as a hard drive, compact disk read-only memory (CD-ROM), DVD, or a solid-state memory device. The memory 406 holds instructions and data used by the processor 402. The input interface 414 is a touch-screen interface, a mouse, track ball, or other type of pointing device, a keyboard, or some combination thereof, and is used to input data into the computer 400. In some embodiments, the computer 400 may be configured to receive input (e.g., commands) from the input interface 414 via gestures from the user. The graphics adapter 412 displays images and other information on the display 418. The network adapter 416 couples the computer 400 to one or more computer networks.

[0120] The computer 400 is adapted to execute computer program modules for providing functionality described herein. As used herein, the term “module” refers to computer program logic used to provide the specified functionality. Thus, a module can be implemented in hardware, firmware, and / or software. In one embodiment, program modules are stored on the storage device 408, loaded into the memory 406, and executed by the processor 402.

[0121] The types of computers 400 used by the entities of FIGs. 1A-1B or 2-3 can vary depending upon the embodiment and the processing power required by the entity. For example, the prediction system 130 can run in a single computer 400 or multiple computers 400 communicating with each other through a network such as in a server farm. The computers 400 can lack some of the components described above, such as graphics adapters 412, and displays 418.IX. Kit Implementation

[0122] Also disclosed herein are kits for performing patient subtyping (e.g., of patients with respiratory infection e.g., pneumonia or severe acute respiratory infection (SARI)) in aAttorney Docket No: PNS-001WO subject. Such kits can include reagents for detecting levels of one or more biomarkers and instructions for determining a patient subtype based on at least the detected levels of the one or more biomarkers.

[0123] The detection reagents can be provided as part of a kit. Thus, the invention further provides kits for detecting the presence of a panel of biomarkers of interest in a biological test sample. A kit can comprise a set of reagents for generating a dataset via at least one protein detection assay (e.g., immunoassay) that analyzes the test sample from the subject. In various embodiments, the set of reagents enable detection of one or more biomarkers from Table 1. In particular embodiments, the set of reagents enable detection of biomarkers (e.g., protein biomarkers) categorized as Tier 1, Tier 2, Tier 3, or Tier 4 biomarkers in Table 3 or Table 6. In particular embodiments, the set of reagents enable detection of biomarkers (e.g., protein biomarkers) categorized as Tier 1, Tier 2, or Tier 3 biomarkers in Table 6. In certain aspects, the reagents include one or more antibodies that bind to one or more of the markers. The antibodies may be monoclonal antibodies or polyclonal antibodies. In some aspects, the reagents can include reagents for performing ELISA including buffers and detection agents. In some aspects, the reagents can include reagents for performing an immunoassay (e.g., such as a lateral flow immunoassay) including buffers and detection agents.

[0124] A kit can include instructions for use of a set of reagents. For example, a kit can include instructions for performing at least one biomarker detection assay such as an immunoassay, a protein-binding assay, an antibody-based assay, an antigen-binding proteinbased assay, a protein-based array, an enzyme-linked immunosorbent assay (ELISA), flow cytometry, a protein array, a blot, a Western blot, nephelometry, turbidimetry, chromatography, mass spectrometry, enzymatic activity, proximity extension assay, and an immunoassay selected from RIA, immunofluorescence, immunochemiluminescence, immunoelectrochemiluminescence, immunoelectrophoretic, a competitive immunoassay, and immunoprecipitation.

[0125] In various embodiments, the kits include instructions for practicing the methods disclosed herein (e.g., methods for building or deploying a predictive model to predict a patient subtype). These instructions can be present in the subject kits in a variety of forms, one or more of which can be present in the kit. One form in which these instructions can be present is as printed information on a suitable medium or substrate, e.g., a piece or pieces of paper on which the information is printed, in the packaging of the kit, in a package insert, etc. Yet another means would be a computer readable medium, e.g., diskette, CD, hard-drive,Attorney Docket No: PNS-001WO network data storage, etc., on which the information has been recorded. Yet another means that can be present is a website address which can be used via the internet to access the information at a removed site. Any convenient means can be present in the kits.X. Systems

[0126] Further disclosed herein are systems for analyzing levels of biomarkers for predicting patient subtypes. In various embodiments, such a system can include a set of reagents for detecting expression levels of biomarkers (e.g., protein biomarkers) in the biomarker panel, an apparatus configured to receive a mixture of the set of reagents and a test sample obtained from a subject to measure the levels of the biomarkers, and a computer system communicatively coupled to the apparatus to obtain the measured levels and to implement the predictive model to generate the patient subtype prediction.

[0127] The set of reagents enable the detection of levels of the biomarkers in the biomarker panel. In various embodiments, the set of reagents involve reagents used to perform an assay, such as an assay or immunoassay as described above. For example, the reagents include one or more antibodies that bind to one or more of the biomarkers (e.g., protein biomarkers). The antibodies may be monoclonal antibodies or polyclonal antibodies. As another example, the reagents can include reagents for performing ELISA including buffers and detection agents. In some aspects, the reagents can include reagents for performing an immunoassay (e.g., such as a lateral flow immunoassay) including buffers and detection agents.

[0128] The apparatus is configured to detect levels of biomarkers in a mixture of a reagent and test sample. For example, the apparatus can determine levels of biomarkers through an immunologic assay or assay for nucleic acid detection. The mixture of the reagent and test sample may be presented to the apparatus through various conduits, examples of which include wells of a well plate (e.g., 96 well plate), a vial, a tube, and integrated fluidic circuits. As such, the apparatus may have an opening (e.g., a slot, a cavity, an opening, a sliding tray) that can receive the container including the reagent test sample mixture and perform a reading to generate values of biomarkers. Examples of an apparatus include a plate reader (e.g., a luminescent plate reader, absorbance plate reader, fluorescence plate reader), a spectrometer, and a spectrophotometer.

[0129] The computer system, such as example computer 400 described in FIG. 4, communicates with the apparatus to receive the quantitative values of biomarkers. TheAttorney Docket No: PNS-001WO computer system implements a predictive model to analyze the values of the biomarkers to predict a patient subtype.EXAMPLES

[0130] Below are examples of specific embodiments for carrying out the present invention. The examples are offered for illustrative purposes only and are not intended to limit the scope of the present invention in any way. Efforts have been made to ensure accuracy with respect to numbers used (e.g., amounts, temperatures, etc.), but some experimental error and deviation should be allowed for.Example 1: Steroids ImmunoScore Satisfies Criteria for Breakthrough Device Designation

[0131] The Steroids ImmunoScore is a rapid turnaround, artificial intelligence / machine learning (AI / ML) -enabled companion diagnostic device intended to guide corticosteroid therapy for patients with non-COVID community-acquired pneumonia (CAP). It utilizes an Al biomarker algorithm that inputs labs, vitals, demographic information, novel protein biomarkers, and other information to classify a patient as “CAP steroid responsive positive” (CAP SR+) or “CAP steroid-responsive negative” (CAP SR-). CAP SR+ indicates a patient signature expected to benefit from corticosteroid therapy. CAP SR- indicates a patient signature expected to be harmed by or not benefit from corticosteroid therapy.

[0132] The Steroids ImmunoScore meets both criteria for Breakthrough Device Designation. Criterion 1: The device provides for more effective treatment or diagnosis of life-threatening or irreversibly debilitating human disease or conditions. The Steroids ImmunoScore meets criterion 1, because it provides for more effective treatment of non- COVID CAP, a life-threatening disease. It does this by discriminating patients who will benefit (CAP SR+) from those who will not benefit or will be harmed (CAP SR-). This determination serves as a guide for the safe and effective use of corticosteroid therapy in patients with non-COVID CAP. As evidence of this, in an independent hold-out verification cohort, CAP SR+ patients had an estimated 9.3% absolute risk reduction in 30-day mortality when treated with corticosteroids, while CAP SR- patients had an estimated 9.9% absolute risk increase in mortality. CAP SR+ patients had biological profiles consistent with elevated levels of proinflammatory cytokines, suggesting an empiric biological rationale for whyAttorney Docket No: PNS-001WO corticosteroids — which are known to dampen inflammatory response — may benefit these patients.

[0133] The Steroids ImmunoScore incorporates several innovative technologies, including causal inference AI / ML to guide treatment and rapid turnaround immunoassays for protein biomarkers. It constitutes novel technology that has the potential to lead to a clinical improvement in the treatment of non-COVID CAP, a life-threatening disease.

[0134] There are no approved companion diagnostics or immunomodulators for non- COVID CAP.

[0135] The Steroids ImmunoScore offers significant advantages over the standard of care, because it equips providers with a diagnostic test that accurately stratifies patients for the safe and effective use of corticosteroids. In a verification analysis, it substantially outperformed existing methods for identifying patients who may benefit from corticosteroids, including the pneumonia severity index (PSI), CURB-65, and the sequential organ failure assessment (SOFA) score. Unlike the Steroids ImmunoScore, none of these alternatives identified patients likely to be harmed, many of whom received corticosteroids under the standard of care. There are no approved companion diagnostics or immunomodulators for non-COVID CAP.

[0136] The Steroids ImmunoScore empowers providers to target corticosteroid therapy based on each patient’s unique biological signature. A verification analysis of an external, hold-out cohort showed that the Steroids ImmunoScore accurately guides this decision. In addition, this study indicated that the Steroids ImmunoScore would avoid potential serious harm to patients in the CAP SR- subgroup that occurs under the current standard of care.

[0137] The Steroids ImmunoScore meets all criteria for breakthrough device designation. It is a novel technology that provides for better treatment of a life-threatening disease with significant unmet need. It has the potential to deliver life-saving corticosteroid therapy to patients who will benefit, and simultaneously prevent harm currently being done in the standard of care to patients who do not. The expectation of clinical and technical success is high. If successful, the Steroids ImmunoScore would become the first ever companion diagnostic and immunomodulator approved by FDA for patients with CAP.Attorney Docket No: PNS-001WOExample 2: Steroids ImmunoScore

[0138] The Steroids ImmunoScore is a companion diagnostic that provided rapid turnaround measurements of an Al biomarker to guide treatment. As shown in FIG. 5, the Steroids ImmunoScore has 2 main components: the Al Biomarker Algorithm and the Protein Biomarker Reader. The Al Biomarker algorithm analyzed a collection of objectively measured inputs, including clinical lab results, protein biomarker concentrations, demographic information, vital measurements, and other information, and indicated beneficial response to steroids; it included secure, HIPAA-compliant cloud-based supporting software that received clinical data and protein measurements and provided a web-based, EMR integrable result interface, as well as a network interface for integration. The Protein Biomarker Reader is a multiplexed rapid turnaround instrument and immunoassay cassettes for quantitative lateral flow assay measurement of protein biomarker concentrations (Angiopoietin-2, IL-6, IL-8, IP- 10, and Pentraxin-3) in patient whole blood or plasma.

[0139] Al Biomarker Algorithm: The Steroids ImmunoScore classified a patient as CAP steroid-responsive positive (CAP SR+) or negative (CAP SR-). This binary classification was a dichotomization of a continuous Al Biomarker, cortiCAP, at a cutoff set during algorithm derivation. The algorithm classified patients with a cortiCAP value below the cutoff as CAP SR+ and above the cutoff as CAP SR-.

[0140] CortiCAP is a single number that represented the estimated mortality benefit or harm that a particular patient with CAP will experience when treated with corticosteroids. It is a combination of many inputs (clinical labs, vitals, protein biomarkers) and correlated with responsiveness to corticosteroid therapy to identify patients likely to benefit. Patients in the study were classified as receiving corticosteroid therapy if they received the equivalent of at least 100 mg hydrocortisone of a systemic corticosteroid per day for at least 1 day beginning within 24 hours after they were suspected of serious infection. The EHR parameters included vitals, demographic information, routine lab measurements, and 2 protein biomarkers: procalcitonin and C-reactive protein. The 5 additional protein biomarkers are Angiopoietin-2, IL-6, IL-8, IP- 10, and Pentraxin-3, which are assayed using the Protein Biomarker Reader.

[0141] CortiCAP was estimated in a derivation cohort of 724 patients with non-COVID CAP in 2 steps. In the first step, supervised causal inference AI / ML algorithms made individualized predictions of the treatment effect of corticosteroids using a comprehensive set of 43 protein biomarkers and 46 parameters obtained from hospital EHRs. The protein biomarkers are shown in EIG. 6.Attorney Docket No: PNS-001WO

[0142] Treatment effect was defined as the absolute risk difference in 30-day mortality. The algorithm training procedure also used data from over 9,000 patients to estimate the probability of 30-day mortality and propensity scores. The patient- specific predictions of the treatment effect were regressed on a reduced, parsimonious set of covariates that can be measured in a clinical setting using standard assessments, laboratory machines, and the Protein Biomarker Reader. These covariates included 46 parameters available from hospital EHRs, including 2 protein biomarkers — procalcitonin and C-reactive protein — and 5 additional protein biomarkers — Angiopoietin-2, IL-6, IL-8, IP- 10, and Pentraxin-3 — which the Protein Biomarker Reader can measure. The full set of inputs to the Steroids ImmunoScore is shown in Table 6.

[0143] The derivation analysis identified a cutoff to dichotomize the continuous cortiCAP marker to define 2 classes of signatures: CAP SR+, where treatment with corticosteroids is expected to reduce 30-day mortality, and CAP SR-, where no benefit or harm is expected. This procedure assessed each possible cutoff to select an optimal cutoff based on the largest proportion of patients with a statistically significant treatment benefit. The goal was to balance the magnitude of the estimated treatment effect and the relative size of the subpopulation classified as steroid responsive. The algorithm classifies patients with a cortiCAP value below this cutoff as CAP SR+, and those above as CAP SR-.

[0144] The Al Biomarker Algorithm is supported by cloud-based software that receives protein measurements from the Protein Biomarker Reader and streaming clinical data to execute the Al Biomarker Algorithm in real time and return the resulting diagnosis. It was developed in accordance with PDA guidances for software as a medical device (SaMD). The software provides interfaces for order initiation and delivery of data to the Al Biomarker Algorithm via secure, HL7 FHIR-compliant APIs. Results are available programmatically via a FHIR API or in an HTML web form for embedding in platforms like hospital EHRs. The software also provides authentication and security associated with all commands. Due to the streaming nature of the clinical data, the device was deployed with integration engines that interface with EHRs. This is analogous to the approach taken by most laboratory devices in clinical labs and followed an architectural paradigm identical to that used in the Sepsis ImmunoScore, an AI / ML SaMD diagnostic for sepsis that was granted marketing authorization by FDA in 2024 (DEN230036).

[0145] The Steroids ImmunoScore is operated like a typical laboratory test: it is ordered, input data is collected, a one-time result is generated, and the result is returned. When aAttorney Docket No: PNS-001WO healthcare professional seeks to diagnose a patient as CAP SR+ or CAP SR-, they can create an order for the device to initiate the test. (The Steroids ImmunoScore can be configured so that order initiation occurs within the EHR or so that orders may be initiated externally, like many standard clinical laboratory machines.) The Steroids ImmunoScore processed the input data and produced a one-time result within a specific time interval from the time at which the order was placed. Producing this diagnosis rapidly is critical for delivering care in CAP, because it is a dynamic condition that can change rapidly. The key difference from traditional laboratory instruments lies in the output generation process, which utilized AI / ML to produce the device's output. The result can then be made available directly in the hospital EHR or via a standalone web interface.Protein Biomarker Reader (reader): The Protein Biomarker Reader includes a multichannel fluorimeter and a single-use, disposable immunoassay cassette. The multi-channel fluorimeter quantifies the concentrations of biomarker proteins in whole blood or plasma by analyzing the results of a cassette, a multiplex of quantitative lateral flow assays (qLFA).

[0146] The disposable cassette was based upon a low-cost nitrocellulose membrane solidphase vertical immunoassay. The cassette contained up to 6 independent channels that each housed one 2.3 mm wide test strip. Each test strip was printed with up to 3 independent immunoassays and a control immunoassay. All immunoassay reagents were dried into the test strips within the cartridge. A whole blood or plasma sample was applied onto a self- contained, single use cassette that was inserted into the reader, and red blood cells were filtered out on-board the cassette. The biomarker of interest was measured in the plasma fraction of the sample that transferred onto the solid phase nitrocellulose membrane.

[0147] The reader is capable of \measuring up to 18 biomarker proteins, producing lab quality data in as few as 30 minutes, and transmitting the results to the Steroids ImmunoScore software via a network connection. Rapid turnaround measurements of the biomarker inputs of the Steroids ImmunoScore are critical for guiding treatment of CAP, because it can be a highly dynamic disease marked by rapid patient deterioration. The reader measures Angiopoietin-2, IL-6, IL-8, IP- 10, and Pentraxin-3 and inputs the concentrations of these markers into the Steroids ImmunoScore Al Biomarker Algorithm.

[0148] Clinical workflow: FIG. 7 depicts a typical workflow for the Steroids ImmunoScore. The Steroids ImmunoScore functions as a diagnostic test, allowing healthcare providers to order and view the test results for a particular patient, similar to a laboratory test.Attorney Docket No: PNS-001WOWhen the Steroids ImmunoScore is ordered by a clinician, its software awaits current measurements for its clinical inputs. The most common deployment of order initiation and data transmission uses an integration engine interfacing with the hospital EHR. Part of that integration can include prompting the care team to place orders for any clinical inputs to the algorithm.

[0149] The software also waits for clinical staff to draw the patient’s blood and measure protein biomarker concentrations using the Protein Biomarker Reader, which measures the protein biomarkers and transmits the results to the software. Once valid measurements are available for all device inputs, the software triggers execution of the Al Biomarker Algorithm. If one or more measurements are not yet available, the system will wait until it receives all necessary values. The output is a diagnosis of either CAP SR+, indicating the patient would likely benefit from corticosteroid therapy, or CAP SR-. This result can be made available to clinicians within an EHR, in the Steroids ImmunoScore Patient View, or via a standalone web interface.

[0150] Results and user interface: The Steroids ImmunoScore can provide a simple user display, the Steroids ImmunoScore Patient View, which is intended to be viewed by clinicians.

[0151] The Patient View can be rendered via a web interface or directly in the EHR. The display includes the following elements:• Binary classification of the patient as CAP SR+ or CAP SR-;• All input parameters, measurements, and the time at which they were generated, including protein concentrations from the Protein Biomarker Reader, labs and, vitals extracted from the EHR;• A ranking and visual depiction of the contribution of each input parameter to the classification as quantified by Shapley values, which are individualized feature importance scores that use a game-theoretic approach to quantifying the contribution of each input feature.Example 3: Derivation of Steroids ImmunoScore

[0152] Derivation Cohort: The Steroids ImmunoScore Al Biomarker Algorithm was derived on 724 patients with non-COVID CAP recruited into the NOSIS Atlas, a large biobank linkedAttorney Docket No: PNS-001WO to clinical data to overcome disease heterogeneity in CAP, sepsis, and other infection related syndromes.

[0153] Patients receiving at least 100 mg daily of a systemic corticosteroid beginning close to blood culture order constituted the treatment group. Patients receiving no systemic steroids during their hospital stay constituted the control group. Patients receiving a lower dose of corticosteroids or receiving them later in their hospital stay were excluded. A sensitivity analysis showed that this exclusion had minimal impact. Patients with CO VID- 19, asthma, or chronic obstructive pulmonary disease (COPD) were excluded, because they were so likely to receive corticosteroids that an observational analysis of this cohort could not reliably estimate treatment effects for them.

[0154] Causal ML Algorithm: Causal AI / ML algorithms estimated cortiCAP, an Al biomarker encoding the effect of corticosteroid therapy on 30-day mortality. Treatment effect was defined as the expected absolute risk difference. This is the average absolute difference in 30-day mortality when corticosteroids were given versus not given. CortiCAP was estimated in 2 steps and then dichotomized to provide binary classification of patients in a third step. In the first step, causal forest estimated individual treatment effects, which is an individualized estimate of the impact of steroids on mortality based on each patient’ s complete biological profile, including 43 protein biomarkers and 46 clinical parameters. Missing parameters were imputed for model derivation. To improve the accuracy of this causal forest model, supervised learning regression models were trained on larger sets of patients for propensity scores (n=9,605) and an outcome model (n=8,199; used by causal forest for orthogonalization. In the second step, the out-of-bag estimates of individual treatment effects were regressed on a reduced, parsimonious set of features to produce the continuous cortiCAP Al biomarker. The motivation for a parsimonious feature set was to restrict to available clinical features and novel biomarkers that can be measured on the Protein Biomarker Reader. In the third step, a doubly robust causal inference procedure identified an optimal cortiCAP cutoff for classifying patients as CAP SR+ or CAP SR-. This binary classifier generates the output of the Steroids ImmunoScore.

[0155] An independent verification analysis of 369 patients with non-COVID CAP found a strong treatment benefit for the CAP SR+ group and a strong harm signal for the CAP SR- group. This indicates that the Steroids ImmunoScore algorithm accurately identifies patients as benefiting from or being harmed by corticosteroids. The patients in the verification cohort were not used in deriving the algorithm. The Steroids ImmunoScore classified each patient inAttorney Docket No: PNS-001WO the verification cohort as CAP SR+ or CAP SR-. The average treatment effects of corticosteroids were estimated using Bayesian logistic regression for (1) the overall verification cohort, (2) the CAP SR+ subgroup, and (3) the CAP SR- subgroup. The regression model included covariates to adjust for confounding that could be caused by bias in which patients were treated with corticosteroids. The Steroids ImmunoScore classified 46.3% of the verification cohort as CAP SR+.Table A: Effects of Corticosteroids in the Verification Cohortverification cohort was an absolute risk increase of + 1% in 30-day mortality (Table A). This means that if the entire cohort were given corticosteroids the prevalence of 30-day mortality would be an estimated 1% higher than if corticosteroids were withheld from the entire cohort. This effect was not statistically significant. Finding no significant treatment effect of corticosteroids when the entire non-COVID CAP population is grouped together is consistent with the literature.

[0157] Substantial Treatment Benefit for CAP SR+: The estimated average treatment effect for the CAP SR+ group was an absolute risk reduction of 9.3% in 30-day mortality (Table A). The posterior probability that corticosteroids have a beneficial effect in the CAP SR+ group was 94%.

[0158] Substantial Treatment Harm for CAP SR-: The estimated average treatment effect for the CAP SR- group was an absolute risk increase of 9.9% in 30-day mortality (Table A). The posterior probability that corticosteroids have a harmful effect in the CAP SR+ group was 95%.

[0159] These results are depicted in FIG. 8 (right) along with the locations of CAP SR+ and CAP SR- patients on a UMAP of over 9,000 patients. This shows that many CAP SR+ patients (triangles identified as “Positive”) localize in the upper region characterized by elevated inflammatory markers, while CAP SR- patients (circles identified as “Negative”)Attorney Docket No: PNS-001WO are less concentrated in this zone (FIG. 8; left). The results of this verification analysis confirmed that the Steroids ImmunoScore accurately classified patients who will on average benefit from corticosteroids as CAP SR+ and patients who on average will be harmed as CAP SR-.Example 4: Steroids ImmunoScore outperforms severity-based approaches

[0160] The verification analysis also confirmed that using the Steroids ImmunoScore to guide corticosteroid therapy was superior to approaches that used clinical measures of CAP severity. These severity measures are alternatives that could be used to guide corticosteroids for non-COVID CAP. This comparison demonstrated that the Steroids ImmunoScore provided for much better treatment than these alternatives. The recent CAPE COD trial and updated guidelines that recommend corticosteroids for severe CAP motivated this comparison.

[0161] The analysis compared the Steroids ImmunoScore against 3 validated measures of CAP severity: pneumonia severity index (PSI), CURB-65, and SOFA. (The confusion component of CURB-65 was not included, because it could not be reliably determined from patient EHRs.) The same regression approach used for the Steroids ImmunoScore estimated the effects of corticosteroid treatment in less severe and more severe patients. This analysis was repeated for each of the 3 severity measures.

[0162] The Steroids ImmunoScore substantially outperformed all 3 severity scores (FIG. 9). The estimated average treatment effects for subgroups defined using the severity scores were not significant (i.e., 80% posterior credible intervals include 0.) The point estimates were negative for the more severe PSI (100+) and SOFA (2+) groups, which suggested a beneficial effect. The direction of these effects was consistent with CAPE COD and recent guidelines. However, the effects were not significant in this analysis. The less severe SOFA group (0-1) was the evidence that any of the severity measures may be able to identify patients who are harmed by corticosteroids. This effect was small and not significant. There is no indication CURB-65 is useful for guiding corticosteroids in this population. Using CAP severity to guide corticosteroid therapy is much worse than using the Steroids ImmunoScore. This further demonstrated that the Steroids ImmunoScore provided more effective treatment for non-COVID CAP than the standard of care.Attorney Docket No: PNS-001WO

[0163] The clinical and biological signatures of patients in the CAP SR+ and CAP SR- groups support the hypothesis that corticosteroids may benefit CAP SR+ populations because they reduce inflammation. Patients that are CAP SR+ tended to have elevated levels of inflammatory marker expression, including cytokines and chemokines, G-CSF and GM-CSF, and Granzyme B (FIGs. 11A-11C). CAP SR+ patients (blue triangles) were also more likely to be located within the upper, hyperinflammatory region of the UMAP (FIG. 10, left) as compared to CAP SR- patients (red circles). Coloring the UMAP with a composite of multiple cytokines and chemokines, including IL-ip, IL- Ira, IL-2, IL-4, IL-6, IL-8, and IFN- y, depicts the degree of inflammation for each patient (FIG. 10, right). The CAP SR+ group also had elevated neutrophils and neutrophil activity as determined by neutrophil numbers, elevated neutrophil to lymphocyte ratio (NLR), and neutrophil gelatinase-associated lipocalin (NGAL) concentrations.

[0164] One of the primary mechanisms of corticosteroid activity is its anti-inflammatory properties. Corticosteroids are active on many types of immune cells and repress transcription of genes associated with cytokines and chemokines, reducing cell adhesion molecules and molecules involved in the initiation and maintenance of the inflammatory response. They also promote anti-inflammatory and immunosuppressive signals, such as lipocortin and annexin AL The elevated inflammatory signals, increased neutrophil numbers and activation and increased adhesion molecules in CAP SR+ patients suggest that corticosteroid treatment may benefit these patients by reducing the elevated immune cell activity that can contribute to increased mortality when patient host response becomes dysregulated.

[0165] The CAP SR- population, in contrast, has a low inflammatory signal. This makes the anti-inflammatory properties of corticosteroids less likely to benefit these patients. There is also evidence that corticosteroid treatment in some patients may not be as safe as previously considered, and short-term use of low-dose corticosteroids to treat CAP can cause adverse events that may include secondary infections, hyperglycemia, gastrointestinal bleeding, hypernatremia, neuropsychiatric disorders, and muscle weakness.

[0166] Corticosteroids have powerful anti-inflammatory effects. CAP SR+ patients tended to have elevated levels of inflammatory markers, which can be indicative of a dysregulated host response. This suggests that corticosteroids may benefit the CAP SR+ population by reducing inflammation.

[0167] In conclusion, Non-COVID CAP is a life-threatening disease that afflicts millions of Americans each year. This Example detailed an independent verification study of 369Attorney Docket No: PNS-001WO patients, finding strong evidence that the Steroids ImmunoScore will achieve clinical success in providing more effective treatment for non-COVID CAP:

[0168] 1. In the independent verification cohort, CAP SR+ patients exhibited an estimated absolute risk reduction of 9.3% and CAP SR- exhibited an absolute risk increase of 9.9% for 30-day mortality.

[0169] 2. Standard measures of severity (PSI, CURB-65, and SOFA) were much inferior to the Steroids ImmunoScore in distinguishing patients who benefit from those who do not receive benefit (or are harmed).

[0170] 3. The signatures of various features in the CAP SR+ and CAP SR- groups, including both labs and novel biomarkers, suggest a plausible biological rationale for why corticosteroids could benefit and not benefit these groups respectively. This was illustrated with both univariate and multivariate methods. Quite simply these results strongly suggest that the Steroids ImmunoScore may be able to substantially improve treatment and reduce the risk of mortality for the millions of patients hospitalized with non-COVID CAP each year.Example 4.1 Methods

[0171] This was a retrospective, observational, multi-center study assessing for differential response to corticosteroid therapy in patients with non-COVID CAP recruited into the NOSIS dataset and biobank. Patients suspected of bacteremia, inferred from the order of a blood culture, and CAP, inferred from ICD codes, were eligible for inclusion in the study. Patients were enrolled into NOSIS at participating sites by identifying a list of eligible patients, and research or hospital personnel evaluating for the presence of remnant samples from a lithium heparin with gel separator (light green top) tube in the clinical lab. Patients with a sample that was available and collected were included, and those with no remnant sample from a lithium heparin with gel separator (light green top) tube available or for whom no sample was collected were excluded. Patients with diagnosis codes indicating COVID- 19, COPD, or asthma were also excluded from this investigation.

[0172] Patients with CAP without CO VID- 19, COPD, or asthma whose biobanked plasma had been assayed for protein biomarkers prior to March 14, 2024, served as the derivation cohort (n=724). Subsequent to the analysis of the derivation cohort, biobanked plasma samples were assayed for 369 additional patients to serve as a verification cohort. Larger sets of patients from the NOSIS dataset were used for estimating propensity scores (n= 9,605) and estimating the marginal risk of 30-day mortality (n=8,199), which were importantAttorney Docket No: PNS-001WO components in the treatment effect analyses. Verification cohort patients were not used in the training of propensity score and mortality models.

[0173] All cause 30-day mortality was the primary outcome, and the treatment effect of interest was the mean absolute risk difference in 30-day mortality between treated and not treated groups.

[0174] Patients were classified as eligible for the treatment group if they (1) received their first dose of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, or dexamethasone via an intravenous, intramuscular, or oral route no earlier than 1 hour prior to and no later than 24 hours after baseline, and (2) if they received the equivalent of at least 100 mg hydrocortisone within 24 hours of that initial dose. Baseline was time of first blood culture order, which is the time at which patients were recruited into the study. Patients were eligible for the control group if they received no systemic steroids during their hospital stay. Patients who received less than 100 mg hydrocortisone equivalents within 24 hours of their initial dose or whose initial dose of steroids was later in their hospital stay were excluded. The impact of this exclusion was assessed in a sensitivity analysis.

[0175] Clinical Data and Specimen Collection: Patient encounter data were gathered directly through an offline EHR extraction and a transfer of deidentified data. Data elements were abstracted from the EHR and included demographic information, coded ICD-9 or ICD- 10 diagnoses, medications, vital sign measurements, clinical laboratory test results (e.g., chemistry laboratory testing results, lactic acid), infection-related laboratory measurements (e.g., complete blood cell count with differential, C-reactive protein, procalcitonin), secondary outcomes metrics, microbiology results, and relevant orders (e.g., antibiotic administration).

[0176] Comorbidities were based on ICD-9 or ICD-10-CM codes. All institutions provided deidentified data deliveries using a secure file transfer protocol. In addition to patient data extracted from EHRs, remnant plasma samples leftover from blood drawn during routine clinical care each comprising four 175 pL aliquots were collected, frozen at -80°C, and shipped to a central laboratory. Plasma samples were collected from lithium heparin with gel separator (light green top) tubes. Each sample comprises four 175 pL aliquots of plasma frozen at -80°C from when samples were available up to 3 days prior to enrollment and prospectively during hospitalization.

[0177] Plasma Protein Biomarker Assays: One plasma discard sample per patient was assayed for 43 protein biomarkers (FIG. 6) for a subsample of patients in the NOSISAttorney Docket No: PNS-001WO biobank. Operating sequentially for each site based on the date of patient enrollment, we selected the sample closest to baseline. Patients without a biobanked sample collected within 3 hours of baseline were excluded. Protein concentrations were measured using multiplexed Luminex xMAP assays developed by R&D Systems. Laboratory quality control procedures adhered to the FDA

[0178] Data Preprocessing: Parameters extracted from the patient EHR and the protein biomarker data were temporally aligned at patient baseline. For vitals, the most extreme value measured within a 12 hour window centered at baseline was used. For other features, the value closest to baseline measured no earlier than 24 hours before and no later than 6 hours after baseline was used. Focus was on EHR features that were objective, standardized measurements, including physical characteristics, lab measurements, vital measurements, biomarker measurements, and demographic information. Patients for whom more than 30% of their clinical features or more than 30% of their protein biomarker features were missing were excluded from the analysis. Each feature was normalized using ordered quantile normalization, which is robust to protein data, using the bestNormalize R package. Missing values were imputed using the bag imputation procedure implemented in the caret R package.

[0179] Causal machine learning tools were used to combine 43 protein biomarkers and 46 clinical features and learn a treatment assignment rule that classifies patients into two groups, one that appeared to benefit from corticosteroid therapy (CAP SR+) and the other that did not (CAP SR-). The treatment assignment rule was verified in the verification cohort. Treatment effect was assessed as the expected absolute risk difference, i.e., the average difference in 30- day mortality when corticosteroids were given versus not given. Formally, let Y(l) denote the outcome (30-day mortality) under treatment, and Y(0) denote the outcome under no treatment. The conditional average treatment effect (CATE) was defined as the expected risk difference conditional on CAP SR diagnosis, i.e.,CATE(CAP SR+) = E[F(1) - F(0) I CAP SR+], CATE(CAP SR-) = E[F(1) - F(0) I CAP SR-]. where E[] is an expectation. The remainder of this section details the derivation and validation of the treatment assignment rule.

[0180] Step 1: Estimate Al Biomarker: Machine learning methods were used to distill high-dimensional patient data into a 1 -dimensional Al biomarker, cortiCAP, a single number that represents the estimated mortality benefit or harm that a particular patient with CAP will experience when treated with corticosteroids. Causal forest, a tree-based ensemble adaptedAttorney Docket No: PNS-001WO from random forest for estimating conditional average treatment effects, was used for supervised, nonparametric estimation of heterogeneous treatment effects of steroids in the derivation cohort. Causal forest estimates the average treatment effect conditional on the entirety of the data available for each patient, which we denote as the patient’s individual treatment effect.

[0181] An important step in the causal forest estimation of treatment effects was orthogonalization using estimates of the propensity score e(x) and the marginal outcome m(x), where e(x) estimates the conditional probability of treatment, E[WIX=r], and m( ) estimates the conditional probability of the outcome, E[yiX=x], where W G {0,1 } denotes treatment with steroids, and Y £{0,1 } denotes 30-day mortality. To optimize this orthogonalization step, we trained propensity score and outcome models on larger sets of patients (n=9,605 for propensity scores; n=8,199 for the outcome model) from the NOSIS data set, including patients without CAP.

[0182] Leveraging more data enables better estimates of the propensity scores and marginal outcome, thereby increasing the precision of the causal forest model. Using these mortality and propensity score models, individual treatment effects were estimated for the 724 patients of the derivation cohort using an honest causal forest of 10,000 trees. The out-of-bag estimates of the individual treatment effect are individualized estimates of the effect of corticosteroid therapy on the risk of 30-day mortality for each patient. The out-of- bag estimates were regressed on a parsimonious subset of features that are more readily clinically deployable to derive the cortiCAP Al biomarker.

[0183] Step 2: Set Cutoff for Patient Classification: A cutoff was selected to dichotomize the continuous cortiCAP Al biomarker to derive a classifier that classifies patients as likely to benefit from corticosteroids (CAP SR+) or not (CAP SR-). A doubly robust, augmented inverse propensity weighting (AIPW) procedure was used to estimate treatment benefit as a function of cortiCAP, similar to the rank average treatment effect approach. For each possible cutoff (i.e., each value of cortiCAP in the derivation cohort) this estimates the CATE for patients below that threshold. The cutoff corresponding to the largest proportion of patients with a statistically significant treatment benefit was identified. To be conservative — in part due to the fact that this value often preceded a jump in the point estimate of the conditional average treatment effect — the cutoff was selected that classified 5% fewer patients in the derivation cohort as CAP SR+.Attorney Docket No: PNS-001WO

[0184] Step 3: Evaluate Classification Schema in Verification Cohort: Patients in the verification cohort were classified as CAP SR+ or CAP SR- by computing the cortiCAP Al biomarker and applying the cutoff set in step 2. The overall average treatment effect in the entire verification cohort as well as the conditional average treatment effects of the CAP SR+ and CAP SR- subgroups were estimated using G-computation with covariate-adjusted Bayesian logistic regression via Markov chain Monte Carlo (MCMC). This technique, which is substantially different from the causal machine learning tools employed in steps 1 and 2, was selected for 3 reasons: (1) the interpretation of its results are straightforward, (2) covariate-adjusted techniques tend to be more robust than inverse probability weighted approaches, which are sensitive to small propensity scores, and (3) causal machine learning tools are sometimes prone to overfitting and spurious findings. This represents a conservative approach to verification using a pre- specified, less flexible, parametric model rather than risking spurious validation by a causal machine learning algorithm that may overfit in both the verification as well as derivation.

[0185] Six covariates were included in the logistic regression model to adjust for treatment assignment bias. A principal component analysis (PCA) was conducted on 8,199 patients to distill 76 clinical and biomarker features at baseline to 5 principal components that summarized patient clinical and biological state. The sixth feature was a mortality score, the predicted probability of 30-day mortality from a random forest supervised learning algorithm trained on the same 8,199 patients. No verification cohort patients were used in the PCA or mortality score training. A UMAP derived on 9,655 patients was used to visualize results.Example 4.2 Results

[0186] Features, outcomes, auxiliary scores and priority rules were compared for the derivation and verification cohorts (Table B). Treatment and control groups were generally similar between cohorts. On average, lactate, mechanical ventilation within 24 hours, vasopressors within 24 hours, and SOFA cardiovascular were higher in the verification cohort treatment group than in that of the derivation cohort, while hydrocortisone equivalent dosage, respiratory rate, and SOFA respiratory were higher in the derivation cohort treatment group.Table B: Characteristics of derivation and verification cohortsAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-001WO

[0187] The maximal value of cortiCAP with an upper bound to the 95% confidence interval of the AlPW-estimated CATE below 0 was -0.007, corresponding to CATE point estimate of -12.5% (FIG. 12A). At this value of cortiCAP, 40.7% of the derivation cohort would have been classified as steroids responsive. The cutoff for CAP SR classification was set to be -0.012, the value of the priority rule that classified 5% less of derivation cohort (35.7%) into the steroids recommended group. Applying this cutoff to the verification cohort classified 46.3% (171 out of 369) patients as CAP SR+ and 53.7% (198 out of 369) as CAP SR-.

[0188] The posterior distribution for the interaction between steroid administration and CAP SR+ was highly negative with a 95% credible interval that did not overlap 0 (log odds ratio: -1.81, 95% credible interval: -3.35 — 0.27; Table C). (A negative effect corresponds to a reduction in expected mortality and therefore a treatment benefit.) Regression coefficient credible intervals for the main effect of steroid administration, the mortality score, and the 5thprinciple component also did not overlap 0. The posterior mean of the average treatment effect was +1%, i.e., a +1% mean risk difference (FIG. 12B). The posterior mean for the conditional average treatment effect of the SR CAP+ group was -9.3% and the posterior probability of benefit was 0.94 (i.e., the posterior probability that the CATE was less than 0 was 0.94; FIG. 12B). For the SR CAP- group, the posterior mean was +9.9% and the posterior probability of harm was 0.95 (i.e., the posterior probability that the CATE was greater than 0 was 0.95; FIG. 12B).Table C: Posterior distributions of the logistic regression model parameters

[0189] The clinical and biological characteristics of the verification cohort were evaluated to derive insights into potential signatures that led to differences inAttorney Docket No: PNS-001WO responsiveness to corticosteroids (Table D). The CAP SR+ population consisted of 46.3% of the patients (171 / 369). Compared to the CAP SR- group, CAP SR+ group had a higher mean age (68.5 years vs. 64.8 years), a higher percentage of males (64.3% vs. 53.0%), and a higher prevalence of heart failure, chronic kidney disease, and chronic liver disease. In addition, the CAP SR+ population was observed to have a higher 30-day mortality rate (20.5% vs. 10.6%) and was more likely to have critical care interventions, such as vasopressors or mechanical ventilation, within 24 hours of blood culture order.Table D: Clinical characteristics of the CAP SR+ and CAP SR- populations

[0190] Patients classified as CAP SR+ were also more likely to have elevated markers of inflammation than patients in CAP SR-. The CAP SR+ patients (blue triangles) were more likely to be located within the upper, hyperinflammatory region of the UMAP (FIG.10, left) as compared to CAP SR- patients (red circles). A coloring scheme was also applied to this UMAP to approximate the degree of inflammation for each patient, using a composite score consisting of multiple cytokines and chemokines, including IL-ip, IL- Ira, IL-2, IL-4, IL- 6, IL-8, and IFN-y (FIG. 10, right). Patients in the CAP SR+ group tended to have higher concentrations of multiple inflammatory biomarkers, including IL- ip, TNF-a, IL-6, MCP-1,Attorney Docket No: PNS-001WO and IL- 8, compared to patients in the CAP SR- group (FIGs. 11A-11C). Furthermore, this group also had elevated neutrophils and neutrophil activity as determined by neutrophil numbers, elevated neutrophil to lymphocyte ratio (NLR), and neutrophil gelatinase- associated lipocalin (NGAL) concentrations. These patients also had elevated cell adhesion molecules, including VCAM-1 and soluble E-selectin. Finally, coagulation factors, such as thrombomodulin and tissue factor, were also elevated in CAP SR+ versus CAP SR-.

[0191] These results provided strong evidence that the Al Biomarker Algorithm of the Steroids ImmunoScore can be used to prospectively identify subgroups of patients with non- COVID CAP who respond differentially to corticosteroid therapy. The posterior mean of the average treatment effect in the verification cohort was close to 0, estimating the average change in mortality risk as a 1% increase in the probability of mortality. This was consistent with previous clinical trials, which have failed to provide convincing evidence for a treatment benefit across hospitalized or severe CAP populations. Posterior distributions for the CATEs of patients in the verification cohort classified as CAP SR+ or CAP SR- were far from zero. The posterior means differed by nearly 20% (+9.9% vs. -9.3%), and their posterior distributions indicate a high probability of a beneficial treatment effect on average in the CAP SR+ group (0.94) and a high probability of a harmful treatment effect on average in the CAP SR- group (0.95).

[0192] The prior distributions in the Bayesian verification analysis were not informative and made no use of any data from the derivation cohort. As a sensitivity analysis, treatment effects in the verification cohort were also estimated using frequentist methods, fitting the logistic regression via maximum likelihood and quantifying uncertainty in the ATE and CATE estimates using 1,000 bootstrap replicates. As expected, the results were very similar to the Bayesian model, with point estimates of -9.8% for the CAP SR+ group (one-sided p- value: 0.054) and +10.9% for the CAP SR- group (one-sided p-value: 0.050).

[0193] The clinical and biological characteristics of patients in the SR CAP+ and SR CAP- groups suggest a biological rationale for responsiveness to corticosteroids in the CAP SR+ group. One of the primary mechanisms of corticosteroid activity is its anti-inflammatory properties as they repress transcription of genes associated with cytokines and chemokines, reducing cell adhesion molecules and molecules involved in the initiation and maintenance of the inflammatory response. Corticosteroids are active on many types of immune cells, including innate cells, such as macrophages, eosinophils, mast cells, and dendritic cells, as well as on adaptive cells, such as lymphocytes. In addition to reducing pro-inflammatoryAttorney Docket No: PNS-001WO signals, corticosteroids also promote anti-inflammatory and immunosuppressive signals, such as lipocortin and annexin Al. The data generated here for patients that are CAP SR+ indicate they tend to have biological profiles consistent with high levels of inflammatory marker expression, including cytokines and chemokines, G-CSF and GM-CSF, and Granzyme B, all of which are indicative of elevated innate immune cell activity (FIGs. 11A-11C).

[0194] It has also been reported that corticosteroids can reverse the increased vascular permeability observed in septic patients. Data generated in this analysis demonstrated that patients in the CAP SR+ population tended to have elevated adhesion molecules VCAM-1 and E-selectin. While it is unknown whether corticosteroid treatment has a direct impact on these adhesion molecules or whether this is an indirect effect due to general immunosuppression, it is plausible that this patient population may benefit from treatment that improves vascular permeability and homeostatic function.

[0195] Use of the cortiCAP biomarker signature to select CAP SR+ patients demonstrated potential for a significant improvement in determination of treatment response when compared to the overall non-COVID CAP population. CAP SR+ patients tend to have elevated inflammatory signals, increased neutrophil numbers and activation, increased adhesion molecules, changes in coagulation markers, and dysregulation in kidney function. Thus, corticosteroid treatment may reduce the elevated immune cell activity as well as have the potential to reduce vascular permeability, coagulation dysfunction, and acute kidney injury that collectively contribute to elevated mortality in this patient population.

[0196] In addition, this biomarker analysis has identified a population of patients that are steroid unresponsive, CAP SR-, who may be harmed by the treatment. These patients in general have a low inflammatory signal, low cytokines and chemokines, and tend not to have an elevation in kidney markers that would suggest kidney dysfunction. There is also evidence that corticosteroid treatment in some patients may not be as safe as previously considered, and short-term use of low-dose corticosteroids to treat CAP can cause adverse events that may include secondary infections, hyperglycemia, gastrointestinal bleeding, hypernatremia, neuropsychiatric disorders, and muscle weakness.

[0197] These analyses are subject to the usual caveats of an observational analysis of causal effects, as there can be no guarantee that the impact of selection bias in treatment assignment has been fully mitigated. A further limitation of the analyses presented here is that the exclusion of patients who received steroids later in their hospital stay may introduce bias into the estimates of the CATE. In particular, this procedure may not be a consistent estimator ofAttorney Docket No: PNS-001WO the CATE for a group of patients that can be identified prospectively. To mitigate concerns of bias from excluding these patients, a sensitivity analysis was conducted in which these patients were not excluded and were included in the control group and the verification analysis repeated. The results were very similar, with an overall posterior mean average treatment effect of -0.4%. The posterior means of the conditional average treatment effect was -10.1% in the SR CAP+ and +8.0% for the SR CAP- group.Example 4.3 NOSIS Atlas

[0198] The NOSIS Atlas is a large biobank linked to clinical data built by Prenosis to overcome disease heterogeneity in CAP, sepsis, and other infection-related syndromes. Recruitment into the Atlas is ongoing, and as of October 2024 included 112,000 biospecimens from 28,000 patients recruited from a network of 11 US hospitals. Patients are eligible for recruitment if they have a blood culture ordered in the hospital or a confirmed viral test. The NOSIS Atlas was designed to overcome disease heterogeneity by combining clinical data with novel protein biomarker data at a large scale. The two primary sources of data are (1) biological (protein) data generated from remnant plasma collected during routine clinical care, and (2) clinical and clinical laboratory data extracted from hospital EHRs before, during, and after the relevant healthcare episode.

[0199] Biological Data: Augmenting clinical data with novel biomarkers provides insight into patient biology. Patient plasma is assayed for 43 proteins, including measures of endothelial cell signaling, inflammation, immune function and response, renal dysfunction, energy metabolism, and coagulation, among others, using multiplexed Luminex xMAP immunoassays.

[0200] Clinical Data: A broad set of real- world data from hospital EHRs was abstracted, the data, including demographic data, comorbidities, vitals, lab results, electronically documented interventions, ICD-10 codes, outcomes, and other parameters. Using well- established quality guidelines, data engineers curated these clinical data and imported them to a standardized database that minimizes variation between sites. Physician adjudicators augmented the EHR data with retrospective expert adjudication as needed.Attorney Docket No: PNS-001WOExample 4.4 Protein Biomarker Measurements

[0201] Protein biomarker measurements made by the Protein Biomarker Reader and initial versions of the disposable cassettes were compared to gold standard measurements made on Luminex instruments in the central lab. Performance on 4 proteins (Angiopoietin-2, IL- 8, IP- 10, and Pentraxin-3) was assessed using a calibration panel and in 2 blinded patient sample comparisons. The biomarker reader generated strong standard curves on an R&D Systems calibration panel, demonstrating strong spike recovery in plasma samples with an R2 > 0.94 for each marker (FIGs. 13A-13B). In a first blinded sample comparison, clinical samples from the NOSIS biobank were assayed on both the Luminex platform and the Protein Biomarker Reader for all 4 markers. Both platforms use control material using the R&D systems recombinant antigen source. The correlation between the concentration interpolated using the Protein Biomarker Reader standard curve with the concentration interpolated using the Luminex system was used to determine the performance of the rapid turnaround assays. Strong correlation values indicate a good agreement between both platforms for the concentration of target protein in each sample (FIGs. 13A-13B). A second, larger blinded patient comparison study demonstrated a strong association between concentrations measured on the Protein Biomarker Reader and Luminex xMAPs. Subsequent to this study, immunoassay chemistry was refined, and was assessed in a third blinded comparison study (FIGs. 14A-14B).Example 5: Combinations of Biomarkers Successfully Predict Likely Responders to Steroid

[0202] Various models were constructed with varying input features (varying combinations of input biomarkers) to evaluate ability to predict likely responders to steroid treatment. .87?+ (steroid responsive positive) was the binary label assigned by the Steroids ImmunoScore algorithm. Derivation cohort refers to patients the algorithm was trained on. Validation cohort refers to patients the algorithm was validated / verified on.

[0203] The total number of features evaluated included the 53 features used in the Steroids ImmunoScore algorithm plus the 34 more biomarkers used in deriving it but that are not included as input features in the algorithm. Total number of features = 87 (all 87 features identified in Table 5).Attorney Docket No: PNS-001WOExample 5.1 - Tiered Biomarkers based on Univariate Performance

[0204] First, the 87 features were grouped into tiers based on their univariate performance. Specifically, the AUC was computed for each feature for discriminating SR+ from SR- in the Steroids ImmunoScore validation cohort. Features were grouped into 4 tiers (e.g., tiers 1-4) based on their univariate AUC. In total, a total of 8 biomarkers were included in tier 1 for exhibiting an AUC between 0.75-1. A total of 11 biomarkers were included in tier 2 for exhibiting an AUC between 0.7-0.74. A total of 25 biomarkers were included in tier 3 for exhibiting an AUC between 0.65-0.69. A total of 43 biomarkers were included in tier 4 for exhibiting an AUC of less than 0.65. The identities and quantities of biomarkers in each of the 4 tiers are shown in Tables 3 and 4.

[0205] Next, combinations of the biomarkers were evaluated for the classification performance (e.g., performance for identifying likely responders to steroid treatment). In the Steroids ImmunoScore derivation cohort, logistic regression models were fit to predict SR+. Specifically, the biomarker combinations included1. All pairwise combinations of features (3,741 total combinations)2. All 3-way combinations of features (105,995 total combinations)3. All subsets of tier 1 features (255 total combinations)4. All combinations of exactly 1 feature from each of the 4 tiers (94,600 total combinations)

[0206] The top performing combinations of biomarkers are shown in Tables 7-9. Specifically, Table 7 shows the top performing bivariate (pairwise) combinations of features (e.g., all combinations achieving AUC > 0.7). A total of 2,994 combinations achieved AUC > 0.6. Table 8 shows the top trivariate (3-way) combinations of features (e.g., all combinations achieving AUC > 0.85). A total of 97,842 combinations achieved AUC > 0.6. Table 9 shows the top performing combinations of tier 1 features. Table 10 shows the top performing combinations of exactly 1 biomarker from each of tiers 1, 2, 3, and 4. A total of 94,601 combinations achieved AUC > 0.74.Example 5.2 - Tiered Biomarkers based on Pathway

[0207] The 87 features (identified in Table 5) were grouped into tiers according to their pathway associations. Each biomarker and corresponding pathway category is shown in Table 2. Additionally, Table 2 shows the tiering of the pathways (e.g., Core InflammatoryAttorney Docket No: PNS-001WOResponse is Tier 1). The biomarkers were grouped into pathways heuristically based upon literature review with the input of clinical biomarker experts.

[0208] Combinations of the biomarkers in Tier 1 (Core Inflammatory Response) were evaluated for predicting likely responders to steroid treatment. Table 11 shows the top performing combinations of tier 1, core inflammatory response biomarkers. Altogether, 32,765 different combinations of 2+ core inflammatory response biomarkers achieved AUC values of greater than 0.7.Example 6: Additional Combinations of Biomarkers Successfully Predict Likely Responders to Steroid

[0209] A summary measure of inflammation was defined by conducting a principal component analysis of normalized concentrations of 6 protein biomarkers known for their roles in inflammatory pathways: IL-6, IL-8, TNF-a, MIP-la, IE-ip, and IL-lra on 8,199 patients. The first principal component of these biomarkers was used as a summary measure of inflammation and, using the derivation cohort of 724 patients, dichotomized it at a cutoff set as described in Example 4.1 “Step 2: Set Cutoff for Patient Classification”. Bayesian logistic regression was applied with G-computation to the Steroids ImmunoScore verification cohort. There was a 14.8% reduction in absolute mortality in patients classified as SR+ using this procedure and a 6.5% increase in mortality in patients classified as SR-.

[0210] Additional combination of biomarkers involving one or more of IL-2, IL-4, Cystatin- C, D-Dimer, GDF15, SlOOb, Protein C, and Troponin were evaluated for predicting likely responders to steroid treatment. Table 12 shows the biomarker combinations involving one or more of IL-2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, and Troponin.Attorney Docket No: PNS-OOIWOTABLESTable 1 : Protein BiomarkersAttorney Docket No: PNS-OOIWOTable 2: Biomarker Pathway CategorizationsAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOTable 3. Feature tiersTable 4. Feature tier classificationAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOTable 6. Exemplary markers of a panelCategoryFeatureTierAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOTable 9: Tier 1 CombinationsAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOAttorney Docket No: PNS-OOIWOTable 10: Additional Biomarker CombinationsAttorney Docket No: PNS-OOIWO

Claims

1. Attorney Docket No: PNS-OOIWOCLAIMS1. A method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise: a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

2. A method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise: Procalcitonin, Pentraxin-3, Age, Sex, Monocytes, Neutrophils, Platelets, White Blood Cells, Lymphocytes, Red Cell Distribution Width, Hemoglobin, Red Blood Cells, Albumin, Bilirubin, Blood Urea Nitrogen, Creatinine, Potassium, Sodium, Chloride, Total CO2, Calcium, Alanine Aminotransferase, Alkaline Phosphatase, Aspartate Aminotransferase, Lactate, , C-Reactive Protein, Temperature, Heart Rate, Respiratory Rate, Systolic Blood Pressure, Diastolic Blood Pressure, Pulse Oximetry, Vasopressors, Shock, Mechanical Ventilation, Diabetes (Charlson Comorbidity Index), Chronic Liver Disease, Chronic Kidney Disease, Congestive Heart Failure, Active Cancer, Angiopoietin-2, IL-8, IP- 10, IL-6, SIRS, SOFA Respiratory, SOFA Coagulation, SOFA Cardiovascular, SOFA Kidneys, SOFA Liver.

3. The method of claim 1, wherein the plurality of biomarkers further comprise one or more additional biomarkers, each additional biomarker different from the first biomarker and theAttorney Docket No: PNS-OOIWO second biomarker and selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

4. The method of claim 3, wherein the plurality of biomarkers further comprise two, three, four, five, or six additional markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

5. The method of any one of claims 1 and 3-4, wherein the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8.

6. The method of any one of claims 1 and 3-5, wherein the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G- CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a.

7. The method of any one of claims 1 and 3-4, wherein the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein.

8. The method of any one of claims 1 and 3-7, wherein the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL- 15, PD-L1, Bilirubin, Albumin, sequential organ failure assessment (SOFA) Kidney, MIP- ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la.

9. The method of any one of claims 1 and 3-8, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.

10. The method of any one of claims 1 and 3-8, wherein the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use.Attorney Docket No: PNS-OOIWO11. The method of any one of claims 1 and 3-8, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.

12. The method of claim 1, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more core inflammatory response biomarkers.

13. The method of claim 12, wherein the one or more core inflammatory response biomarkers are different from the first biomarker and the second biomarker and selected from IL-ip, TNF-a, IL-6, C-Reactive Protein, Pentraxin-3, IL- Ira, IL- 10, IL-8, NGAL, TREM-1, G- CSF, MIP-3a, PCT, Granzyme B, and TRAIL.

14. The method of any one of claims 1 or 12-13, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more immune cell activation biomarkers or chemokines.

15. The method of claim 14, wherein the one or more immune cell activation features or chemokines are selected from Neutrophils, WBC, Lymphocytes, Monocytes, Platelets, MCP-1, MIP-la, MIP-ip (CCL4), IP-10, PD-L1, GM-CSF, FLT3 Ligand, IL-15, IL-7, Interferon- a, and Interferon-y.

16. The method of any one of claims 1 or 12-15, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more endothelia and coagulation markers.

17. The method of claim 14, wherein the one or more endothelia and coagulation markers are selected from Thrombomodulin, Tissue Factor, VCAM-1, E-Selectin, Angiopoietin-2, Angiopoietin-1, VEGF, TGF-a, and Leptin.

18. The method of any one of claims 1 or 12-17, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more metabolic and organ function features.Attorney Docket No: PNS-OOIWO19. The method of claim 18, wherein the one or more metabolic and organ function features are selected from Lactate, Blood Urea Nitrogen, Creatinine, Bilirubin, AST, ALT, ALP, Albumin, Calcium, Potassium, Sodium, Chloride, Total CO2, Hemoglobin, RBC, RDW.

20. The method of any one of claims 1 or 12-19, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more clinical parameters or interventions.

21. The method of claim 20, wherein the one or more clinical parameters or interventions are selected from Systolic BP, Diastolic BP, Heart Rate, Respiratory Rate, Temperature, Pulse Oximetry, SOFA Kidney, SOFA Cardiovascular, SOFA Respiratory, SOFA Coagulation, SOFA Liver, SIRS, Shock, Vasopressors, and Ventilator.

22. The method of any one of claims 1 or 12-21, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more demographics or comorbidities.

23. The method of claim 22, wherein the one or more demographics or comorbidities are selected from Age, Gender, Diabetes, Chronic Kidney Disease, Chronic Liver Disease, Active Cancer, CHF, Emphysema, Atelectasis, Rheumatoid Arthritis, Lupus, Immunodeficiency, Acute Bronchitis, Asphyxiation, Nicotine Use, and Glucocorticoid Deficiency.

24. The method of any one of claims 1-23, wherein the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

25. The method of any one of claims 1-24, wherein the corticosteroid is administered intravenously, intramuscularly, or orally.

26. The method of any one of claims 1-25, wherein values of at least a subset of the plurality of biomarkers are obtained by performing an immunoassay.

27. The method of claim 26, wherein the immunoassay is a quantitative lateral flow assay.

28. The method of claim 26 or 27, wherein performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies.

29. The method of any one of claims 1 and 3-25, wherein values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.

30. A method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise:Attorney Docket No: PNS-OOIWO a first biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), and a second biomarker selected from Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL); and generating a prediction of a subtype for the patient with the respiratory infection based on the values of the plurality of biomarkers.

31. A method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise: Age, Sex, Monocytes, Neutrophils, Platelets, White Blood Cells, Lymphocytes, Red Cell Distribution Width, Hemoglobin, Red Blood Cells, Albumin, Bilirubin, Blood Urea Nitrogen, Creatinine, Potassium, Sodium, Chloride, Total CO2, Calcium, Alanine Aminotransferase, Alkaline Phosphatase, Aspartate Aminotransferase, Lactate, Procalcitonin, C-Reactive Protein, Temperature, Heart Rate, Respiratory Rate, Systolic Blood Pressure, Diastolic Blood Pressure, Pulse Oximetry, Vasopressors, Shock, Mechanical Ventilation, Diabetes (Charlson Comorbidity Index), Chronic Liver Disease, Chronic Kidney Disease, Congestive Heart Failure, Active Cancer, Angiopoietin-2, Pentraxin-3, IL-8, IP- 10, IL-6, SIRS, SOFA Respiratory, SOFA Coagulation, SOFA Cardiovascular, SOFA Kidneys, SOFA Liver; and generating a prediction of a subtype for the patient with the respiratory infection based on the values of the plurality of biomarkers.

32. The method of claim 30, wherein the plurality of biomarkers further comprise one or more additional biomarkers, each additional biomarker different from the first biomarker and the second biomarker and selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

33. The method of claim 32, wherein the plurality of biomarkers further comprise two, three, four, five, or six additional markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).Attorney Docket No: PNS-001WO34. The method of any one of claims 30 and 32-33, wherein the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8.

35. The method of any one of claims 30 and 32-34, wherein the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G- CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a.

36. The method of any one of claims 30 and 32-35, wherein the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein.

37. The method of any one of claims 30 and 32-35, wherein the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL- 15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la.

38. The method of any one of claims 30 and 32-35, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.

39. The method of any one of claims 30 and 32-35, wherein the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use.

40. The method of any one of claims 30 and 32-39, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver,Attorney Docket No: PNS-001WOTemperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.

41. The method of any one of claims 30 and 32-40, wherein the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid.

42. The method of claim 41, wherein the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

43. The method of any one of claims 30 and 32-42, further comprising: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid.

44. The method of any one of claims 30 and 32-43, wherein obtaining or having obtained values for at least a subset of the plurality of biomarkers comprises performing an immunoassay.

45. The method of claim 44, wherein the immunoassay is a quantitative lateral flow assay46. The method of claim 44 or 45, wherein performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies.

47. The method of any one of claims 30 and 32-46, wherein values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.

48. A non-transitory computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform the method of any one of claims 30-43.

49. A method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise two or more of procalcitonin, IL-6, triggering receptor expressed on myeloid cells 1 (TREM- 1), Pentraxin-3, MIP-3a, IL- Ira, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

50. The method of claim 49, wherein the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8.Attorney Docket No: PNS-OOIWO51. The method of claim 49 or 50, wherein the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a.

52. The method of any one of claims 49-51, wherein the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein.

53. The method of any one of claims 49-51, wherein the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL- 15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la.

54. The method of any one of claims 49-53, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.

55. The method of any one of claims 49-53, wherein the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use.

56. The method of any one of claims 49-53, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twenty three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.Attorney Docket No: PNS-OOIWO57. The method of claim 49, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more core inflammatory response biomarkers.

58. The method of claim 57, wherein the one or more core inflammatory response biomarkers are different from the first biomarker and the second biomarker and selected from IL-ip, TNF-a, IL-6, C-Reactive Protein, Pentraxin-3, IL- Ira, IL- 10, IL-8, NGAL, TREM-1, G- CSF, MIP-3a, PCT, Granzyme B, and TRAIL.

59. The method of any one of claims 49 or 57-58, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more immune cell activation biomarkers or chemokines.

60. The method of claim 59, wherein the one or more immune cell activation features or chemokines are selected from Neutrophils, WBC, Lymphocytes, Monocytes, Platelets, MCP-1, MIP-la, MIP-ip (CCL4), IP-10, PD-L1, GM-CSF, FLT3 Ligand, IL-15, IL-7, Interferon- a, and Interferon-y.

61. The method of any one of claims 49 or 57-60, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more endothelia and coagulation markers.

62. The method of claim 61, wherein the one or more endothelia and coagulation markers are selected from Thrombomodulin, Tissue Factor, VCAM-1, E-Selectin, Angiopoietin-2, Angiopoietin-1, VEGF, TGF-a, and Leptin.

63. The method of any one of claims 49 or 57-62, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more metabolic and organ function features.

64. The method of claim 63, wherein the one or more metabolic and organ function features are selected from Lactate, Blood Urea Nitrogen, Creatinine, Bilirubin, AST, ALT, ALP, Albumin, Calcium, Potassium, Sodium, Chloride, Total CO2, Hemoglobin, RBC, RDW.

65. The method of any one of claims 49 or 57-64, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more clinical parameters or interventions.

66. The method of claim 65, wherein the one or more clinical parameters or interventions are selected from Systolic BP, Diastolic BP, Heart Rate, Respiratory Rate, Temperature, Pulse Oximetry, SOFA Kidney, SOFA Cardiovascular, SOFA Respiratory, SOFA Coagulation, SOFA Liver, SIRS, Shock, Vasopressors, and Ventilator.Attorney Docket No: PNS-OOIWO67. The method of any one of claims 49 or 57-66, wherein the plurality of biomarkers further comprise one or more additional biomarkers, wherein the one or more additional biomarkers comprise one or more demographics or comorbidities.

68. The method of claim 67, wherein the one or more demographics or comorbidities are selected from Age, Gender, Diabetes, Chronic Kidney Disease, Chronic Liver Disease, Active Cancer, CHF, Emphysema, Atelectasis, Rheumatoid Arthritis, Lupus, Immunodeficiency, Acute Bronchitis, Asphyxiation, Nicotine Use, and Glucocorticoid Deficiency.

69. The method of any one of claims 49-68, wherein the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

70. The method of any one of claims 49-69, wherein the corticosteroid is administered intravenously, intramuscularly, or orally.

71. The method of any one of claims 49-70, wherein values of at least a subset of the plurality of biomarkers are obtained by performing an immunoassay.

72. The method of claim 71, wherein the immunoassay is a quantitative lateral flow assay73. The method of claim 71 or 72, wherein performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies.

74. The method of any one of claims 49-73, wherein values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.

75. A method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise two or more of procalcitonin, IL-6, triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase- associated lipocalin (NGAL); and generating a prediction of a subtype for the patient with the respiratory infection based on the values of the plurality of biomarkers.

76. The method of claim 75, wherein the plurality of biomarkers further comprise one or more additional biomarkers, each additional biomarker selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase- associated lipocalin (NGAL).Attorney Docket No: PNS-OOIWO77. The method of claim 76, wherein the plurality of biomarkers further comprise two, three, four, five, or six additional markers selected from procalcitonin, IL-6, and triggering receptor expressed on myeloid cells 1 (TREM-1), Pentraxin-3, MIP-3a, IL-lra, IL-8, and neutrophil gelatinase-associated lipocalin (NGAL).

78. The method of any one of claims 75-77, wherein the plurality of biomarkers comprise each of procalcitonin, IL-6, Pentraxin-3, and IL-8.

79. The method of any one of claims 75-78, wherein the plurality of biomarkers further comprise one or more of Angiopoietin-2, IL-ip, VCAM-1, Lactate, Thrombomodulin, G- CSF, IL- 10, Granzyme B, TNF-a, C-Reactive Protein, and TGF-a.

80. The method of any one of claims 75-79, wherein the plurality of biomarkers further comprise Angiopoietin-2, Lactate, and C-Reactive Protein.

81. The method of any one of claims 75-79, wherein the plurality of biomarkers further comprise one or more of Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL- 15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-ip (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-la.

82. The method of any one of claims 75-81, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, or each of Creatinine, Blood Urea Nitrogen, Bilirubin, Albumin, SOFA Kidney, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, SOFA Cardiovascular, and Calcium.

83. The method of any one of claims 75-81, wherein the plurality of biomarkers further comprise one or more of red cell distribution width (RDW), alanine transaminase (ALT), Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, congestive heart failure (CHF), Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, red blood cells (RBC), Age, Ventilator, Interferon-a, Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon-y, Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use.

84. The method of any one of claims 75-83, wherein the plurality of biomarkers further comprise two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen, sixteen, seventeen, eighteen, nineteen, twenty, twenty one, twenty two, twentyAttorney Docket No: PNS-OOIWO three, twenty four, twenty five, twenty six, or each of RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Gender, RBC, Age, Ventilator, Potassium, Hemoglobin, Sodium, Chronic Liver Disease, Diabetes, Chloride, and Active Cancer.

85. The method of any one of claims 75-83, wherein the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid.

86. The method of claim 85, wherein the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

87. The method of any one of claims 75-86, further comprising: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid.

88. The method of any one of claims 75-87, wherein obtaining or having obtained values for at least a subset of the plurality of biomarkers comprises performing an immunoassay.

89. The method of claim 88, wherein the immunoassay is a quantitative lateral flow assay90. The method of claim 88 or 89, wherein performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies.

91. The method of any one of claims 75-90, wherein values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.

92. A non-transitory computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform the method of any one of claims 75-90.

93. The method of any one of claims 1-91, wherein the respiratory infection is a severe acute respiratory infection (SARI) or pneumonia.

94. The method of claim 93, wherein the pneumonia is community-acquired pneumonia (CAP).

95. A method for treating a patient with a respiratory infection, the method comprising: administering to the patient with the respiratory infection a therapeutically effective amount of a corticosteroid; wherein the patient was classified in a patient subtype identified as responders to the corticosteroid based on values of a plurality of biomarkers determined from the patient, wherein the plurality of biomarkers comprise two or more of PCT, Pentraxin-3, IL-6, MIP-3 a , IL- Ira, IL-8, NGAL, TREM-1, Angiopoietin-2,Attorney Docket No: PNS-001WOIL-1 , VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10, Granzyme B, TNF- a , C-Reactive Protein, IL-2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, Troponin, TGF- a , Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-1 0 (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-1 a , RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, RBC, Age, Ventilator, Interferon- a , Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon- y , Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use.

96. The method of claim 95, wherein the plurality of biomarkers comprise two or more of IL- 6, IL-8, TNF-a, MIP-la, IL-1 , and IL- Ira.

97. The method of claim 95, wherein the plurality of biomarkers comprise two or more of IL- 2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, and Troponin.

98. The method of any one of claims 95-97, wherein the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid.

99. The method of claim 98, wherein the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

100. The method of any one of claims 95-99, further comprising: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid.

101. A method for subtyping patients with a respiratory infection, the method comprising: obtaining or having obtained values for a plurality of biomarkers for a patient with the respiratory infection, wherein the plurality of biomarkers comprise two or more of PCT, Pentraxin-3, IL-6, MIP-3 a , IL- Ira, IL-8, NGAL, TREM-1, Angiopoietin-2, IL-1 P , VCAM-1, Lactate, Thrombomodulin, G-CSF, IL- 10,Attorney Docket No: PNS-001WOGranzyme B, TNF- a , C-Reactive Protein, IL-2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, Troponin, TGF- a , Creatinine, MCP-1, Blood Urea Nitrogen, VEGF, E-Selectin, IL-15, PD-L1, Bilirubin, Albumin, SOFA Kidney, MIP-1 (CCL4), FLT3 Ligand, Systolic BP, Neutrophils, AST, Respiratory Rate, Diastolic BP, SIRS, ALP, WBC, Shock, Tissue Factor, SOFA Cardiovascular, Calcium, MIP-1 a , RDW, ALT, Pulse Oximetry, Vasopressors, Heart Rate, GM-CSF, CHF, Total CO2, Platelets, Monocytes, SOFA Respiratory, Lymphocytes, SOFA Coagulation, Interleukin-7, IP- 10, Chronic Kidney Disease, SOFA Liver, Temperature, Male, RBC, Age, Ventilator, Interferon- a , Potassium, Leptin, Hemoglobin, Sodium, Chronic Liver Disease, Interferon- y , Diabetes, Glucocorticoid Deficiency, Angiopoietin-1, Asphyxiation, Chloride, Active Cancer, Acute Bronchitis, Rheumatoid Arthritis, Immunodeficiency, TRAIL, Lupus, Emphysema, Atelectasis, and Nicotine Use; and generating a prediction of a subtype for the patient with the respiratory infection based on the values of the plurality of biomarkers.

102. The method of claim 101, wherein the plurality of biomarkers comprise two or more of IL-6, IL-8, TNF-a, MIP-la, IL-1 , and IL- Ira.

103. The method of claim 101, wherein the plurality of biomarkers comprise two or more of IL-2, IL-4, Cystatin-C, D-Dimer, GDF15, SlOOb, Protein C, and Troponin.

104. The method of any one of claims 101-103, wherein the predicted subtype of the patient is one of a responder to a corticosteroid or a non-responder to a corticosteroid.

105. The method of claim 104, wherein the corticosteroid is one of hydrocortisone, methylprednisolone, cortisone, prednisone, prednisolone, dexamethasone, triamcinolone, betamethasone, and fludrocortisone.

106. The method of any one of claims 101-105, further comprising: administering the corticosteroid to the patient based on the predicted subtype of the patient being a responder to the corticosteroid.

107. The method of any one of claims 101-106, wherein obtaining or having obtained values for at least a subset of the plurality of biomarkers comprises performing an immunoassay.

108. The method of claim 107, wherein the immunoassay is a quantitative lateral flow assayAttorney Docket No: PNS-001WO109. The method of claim 107 or 108, wherein performing the immunoassay comprises contacting a test sample with a plurality of reagents comprising antibodies.

110. The method of any one of claims 101-109, wherein values of at least a subset of the plurality of biomarkers are obtained from an electronic health record (EHR) of the patient.

Images