Raman spectroscopy system, liquid blood sample holder, and method of interrogating a liquid blood sample

Abstract

The Raman spectroscopy system generally has a sample holder receiver receiving a sample holder in a measurement position, the sample holder having a body having a depression and a metallic foil covering the depression, the metallic foil forming a concave surface receiving a liquid blood sample, the concave surface exposed to an optical path when the sample holder is in the measurement position; a Raman spectroscopy measurement unit performing a Raman spectroscopy measurement on the liquid blood sample via the optical path, said Raman spectroscopy measurement including interrogating the liquid blood sample with a Raman excitation beam propagating through the liquid blood sample prior to detection, and generating a Raman spectrum resulting from said propagation; and a controller communicatively coupled to the Raman spectroscopy measurement unit, the controller configured for accessing said Raman spectrum; comparing said Raman spectrum to reference data; and generating a signal based on said comparison.

Description

RAMAN SPECTROSCOPY SYSTEM, LIQUID BLOOD SAMPLE HOLDER, AND METHOD OF INTERROGATINGA LIQUID BLOOD SAMPLE FIELD

[0001] The improvements presented in this disclosure relate to the screening fordisease(s) such as lung cancer, and most particularly relate to systems and methods for the screening of disease(s) from a liquid blood sample using Raman spectroscopy.BACKGROUND

[0002] The screening for disease(s) is of importance in today’s society. Taking the lung cancer for example, this disease is the second most common type of non-skin cancer in the United States. Lung cancer is also the leading cause of cancer death in men and in women. Existing screening techniques for lung cancer typically rely on costly imaging and invasive biopsies. With only 5.8% of the 14.2 million eligible Americans undergoing disease screening, there thus exists a significant gap which underscores the urgent need for more accessible disease screening solutions. Moreover, current screening techniques for lung cancer involve ionizing radiation in the form of X-rays which is associated with high levels of false positives, putting patients under unnecessary stress. Although existing screening techniques are satisfactory to a certain degree, there remains room for improvement.SUMMARY

[0003] The recurrence rates of some diseases vary significantly depending on the type and stage of the disease(s) at the time of diagnosis. Taking the lung cancer for example again, the most common type of lung cancer is Non-Small Cell Lung Cancer (NSCLS), which makes up about 85% of cases. Recurrence rates for NSCLC generally depend on the cancer stage at diagnosis: Stage I: 5-19%, Stage II: 11-27%, Stage III: 24-40% recurrence rate. Therefore, an ideal screening solution would not only detect the early presence of a disease, but would also identify a stage thereof (hereinafter referred to as “staging”). Indeed, in the case of lung cancer, detecting the disease at an early stage, and identifying its stage, are particularly important practically as it can enable rapid intervention, appropriate treatment selection, and sound prognosis to be communicated to the patient.

[0004] There is thus a need in the industry for a low-cost, non-invasive disease screening device which would be invaluable for the early detection and / or staging of disease(s). The present disclosure suggests a system and method for screening (including detecting and / or staging) of one or more disease(s) simultaneously from a liquid blood sample using Raman spectroscopy. The system and method described herein was demonstrated to be rapid, compatible with common blood extraction techniques such as finger pricks, low-cost, reagent-free, user-friendly, and integrated into a portable equipment of relatively small footprint. Furthermore, the Raman spectroscopy system does not use ionizing radiation and can be fit into current clinical workflows, which are highly sought after characteristics of an optimal clinical device.

[0005] In accordance with a first aspect of the present disclosure, there is provided a Raman spectroscopy system for interrogating a liquid blood sample, the Raman spectroscopy system comprising: a frame; a sample holder receiver mounted to the frame, the sample holder receiver configured for receiving a sample holder in a measurement position, the sample holder having a body having a depression and a metallic foil covering the depression, the metallic foil forming a concave surface receiving the liquid blood sample, the concave surface exposed to an optical path when the sample holder is in the measurement position; a Raman spectroscopy measurement unit mounted to the frame and configured for performing a Raman spectroscopy measurement on the liquid blood sample via the optical path when the sample holder is in the measurement position, said Raman spectroscopy measurement including interrogating the liquid blood sample with a Raman excitation beam propagating once through the liquid blood sample, reflecting on the metallic foil, and propagating once again through the liquid blood sample prior to detection, and generating a Raman spectrum resulting from said propagation; and a controller communicatively coupled to the Raman spectroscopy measurement unit, the controller configured for accessing said Raman spectrum; comparing said Raman spectrum to reference data; and generating a signal based on said comparison.

[0006] Further in accordance with the first aspect of the present disclosure, the metallic foil can for example be an aluminum foil.

[0007] Still further in accordance with the first aspect of the present disclosure, the metallic foil can for example be adhered to the body of the sample holder.

[0008] Still further in accordance with the first aspect of the present disclosure, the depression can for example have a dimension ranging between about 0.5 cm and 1 cm.

[0009] Still further in accordance with the first aspect of the present disclosure, the liquid blood sample can for example have a volume ranging between about 15 and 200 pL, preferably between 25 and 100 pL, and most preferably between 40 and 60 pL.

[0010] Still further in accordance with the first aspect of the present disclosure, the sample holder receiver can for example have a sample securing mechanism movable to a receiving position, in which the sample holder is received away from the measurement position, and to a close position, in which the sample holder is in the measurement position.

[0011] Still further in accordance with the first aspect of the present disclosure, the measurement position can for example be inside the sample holder receiver.

[0012] Still further in accordance with the first aspect of the present disclosure, the depression can for example be a hole recessed in the body of the sample holder.

[0013] Still further in accordance with the first aspect of the present disclosure, the hole can for example be a through hole extending through the body.

[0014] Still further in accordance with the first aspect of the present disclosure, the body of the sample holder can for example include polymer material.

[0015] Still further in accordance with the first aspect of the present disclosure, the sample holder can for example have an elongated prismatic shape having a length ranging between about 2 and 10 cm, a width ranging between about 1 and 4 cm, and a thickness ranging between about 0.2 and 1 cm.

[0016] In accordance with a second aspect of the present disclosure, there is provided a liquid blood sample holder for a Raman spectroscopy measurement, the liquid blood sample holder comprising: a body having a depression; and a metallic foil covering the depression ofthe body, the metallic foil forming a concave surface receiving a liquid blood sample having a volume below 200 pL, the liquid blood sample holding on to the metallic foil due to surface tension; wherein, during the Raman spectroscopy measurement, a Raman excitation beam propagates once through the liquid blood sample, reflects on the metallic foil, and propagating once again through the liquid blood sample prior to detection.

[0017] Further in accordance with the second aspect of the present disclosure, the metallic foil can for example be an aluminum foil.

[0018] Still further in accordance with the second aspect of the present disclosure, the metallic foil can for example be adhered to the body of the sample holder.

[0019] Still further in accordance with the second aspect of the present disclosure, the depression can for example have a dimension ranging between about 0.5 cm and 1 cm.

[0020] Still further in accordance with the second aspect of the present disclosure, the depression can for example be a hole recessed in the body of the sample holder.

[0021] Still further in accordance with the second aspect of the present disclosure, the hole can for example be a through hole extending through the body.

[0022] Still further in accordance with the second aspect of the present disclosure, the body of the sample holder can for example include a polymer material.

[0023] In accordance with a third aspect of the present disclosure, there is provided a method for interrogating a liquid blood sample, the method comprising: receiving a liquid blood sample on a sample holder, the sample holder having a body having a depression and a metallic foil covering the depression of the body, the metallic foil forming a concave surface receiving the liquid blood sample; moving the sample holder into a measurement position relative to a Raman spectroscopy measurement unit, in which the concave surface is exposed to an optical path; using the Raman spectroscopy measurement unit, performing a Raman spectroscopy measurement on the liquid blood sample via the optical path when the sample holder is in the measurement position, said Raman spectroscopy measurement including interrogating the liquid blood sample with a Raman excitation beam propagating once throughthe liquid blood sample, reflecting on the metallic foil, and propagating once again through the liquid blood sample prior to detection, and generating a Raman spectrum resulting from said propagation; and using a controller communicatively coupled to the Raman spectroscopy measurement unit, accessing said Raman spectrum, comparing said Raman spectrum to reference data, and generating a signal based on said comparison.

[0024] Further in accordance with the third aspect of the present disclosure, the liquid blood sample can for example have a volume below 200 pL.

[0025] Still further in accordance with the third aspect of the present disclosure, said receiving the liquid blood sample can for example include a step of extracting one or more drops of liquid blood from a patient and a step of depositing the one or more drops of liquid blood onto the concave surface of the sample holder.

[0026] Still further in accordance with the third aspect of the present disclosure, the step of extracting the one or more drops of liquid blood from the patient and the step of performing the Raman spectroscopy measurement can for example be performed within 30 minutes from one another, preferably within 10 minutes from one another, and most preferably within 5 minutes from one another.

[0027] Still further in accordance with the third aspect of the present disclosure, said method can for example be performed under 2 minutes, preferably under 1 minutes, and most preferably under 30 seconds.

[0028] Still further in accordance with the third aspect of the present disclosure, the signal can for example be indicative of whether the liquid blood sample is associated to a healthy status or to an unhealthy status.

[0029] Still further in accordance with the third aspect of the present disclosure, the signal can for example be indicative of one or more diseases identified in the liquid blood sample.

[0030] Still further in accordance with the third aspect of the present disclosure, said signal can for example be indicative of a stage of at least one of the one or more diseases identified in the liquid blood sample.

[0031] All technical implementation details and advantages described with respect to a particular aspect of the present invention are self-evidently mutatis mutandis applicable for all other aspects of the present invention.

[0032] Many further features and combinations thereof concerning the present improvementswill appearto those skilled in the art following a reading of the instant disclosure.DESCRIPTION OF THE FIGURES

[0033] In the figures,

[0034] Fig. 1 is an oblique view of an example of a Raman spectroscopy system, showing a sample holder receiving unit, a Raman spectroscopy device, and a controller, in accordance with one or more embodiments;

[0035] Fig. 2 is a top plan view of an example of a liquid blood sample holder, shown with a depression, in accordance with one or more embodiments;

[0036] Fig. 3 is an oblique view of the liquid blood sample holder of Fig. 2, shown with a metallic foil covering the depression, in accordance with one or more embodiments;

[0037] Fig. 3A is a sectional view of the liquid blood sample holder of Fig. 3, taken along section 3A-3A of Fig. 3A, in accordance with one or more embodiments;

[0038] Fig. 4 is a graph showing Raman spectra associated with different types of metallic foil covering the depression of the sample holder, in accordance with one or more embodiments;

[0039] Fig. 5 includes Raman spectra measured on different liquid blood samples using the Raman spectroscopy system of Fig. 1 , showing slight differences in spectral contents for different healthy and / or unhealthy groups of liquid blood samples, in accordance with one or more embodiments;

[0040] Fig. 6 is a flow chart of a method for screening one or more diseases in a liquid blood sample using a Raman spectroscopy system, in accordance with one or more embodiments;

[0041] Fig. 7 is a schematic view of an example of a computing device of a controller of the Raman spectroscopy system of Fig. 1 , in accordance with one or more embodiments.

[0042] Fig. 8 is an exemplary workflow for lung cancer screening incorporating the use of a Raman spectroscopy system, in accordance with one or more embodiments;

[0043] Fig. 9 is a graph showing Raman spectra obtained by testing liquid blood plasma and liquid blood serum using the Raman spectroscopy system of Fig. 8, in accordance with one or more embodiments;

[0044] Fig. 10 is a graph showing Raman spectra obtained by testing different volumes of a same liquid blood sample, in accordance with one or more embodiments;

[0045] Fig. 11A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples known to exhibit benign tumors and a second group of liquid blood samples known to exhibit Stage l-ll lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0046] Fig. 11 B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of benign tumors and Stage l-ll lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0047] Fig. 11C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0048] Fig. 11 D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0049] Fig. 12A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples known to exhibit benign tumors and a second group of liquid blood samples known to exhibit Stage lll-IV lung cancer,with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0050] Fig. 12B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of benign tumors and Stage lll-IV lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0051] Fig. 12C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0052] Fig. 12D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0053] Fig. 13A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples known to exhibit benign tumors and a second group of liquid blood samples known to exhibit Stage l-IV lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0054] Fig. 13B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of benign tumors and Stage l-IV lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0055] Fig. 13C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0056] Fig. 13D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0057] Fig. 14A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples known to exhibit Stage l-ll lung cancer and a second group of liquid blood samples known to exhibit Stage lll-IV lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0058] Fig. 14B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of Stage l-ll lung cancer and Stage lll-IV lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0059] Fig. 14C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0060] Fig. 14D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0061] Fig. 15A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples associated to a healthy control group and a second group of liquid blood samples known to exhibit Stage l-ll lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0062] Fig. 15B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of healthy control group and Stage l-ll lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0063] Fig. 15C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0064] Fig. 15D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0065] Fig. 16A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples associated to a healthy control group and a second group of liquid blood samples known to exhibit Stage lll-IV lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0066] Fig. 16B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of healthy control group and Stage lll-IV lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0067] Fig. 16C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0068] Fig. 16D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0069] Fig. 17A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples associated to a healthy control group and a second group of liquid blood samples known to exhibit Stage l-IV lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0070] Fig. 17B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of healthy control group and Stage l-IV lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0071] Fig. 17C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0072] Fig. 17D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0073] Fig. 18A is a graph showing Raman spectra measured using the Raman spectroscopy system of Fig. 8 on a first group of liquid blood samples associated to both a healthy control group and benign tumors and a second group of liquid blood samples known to exhibit Stage l-IV lung cancer, with black dashed lines showing peaks used in the machine learning models, in accordance with one or more embodiments;

[0074] Fig. 18B is a confusion matrix showing efficiency of Raman spectroscopy-based prediction of the combined healthy control and benign tumor groups and Stage l-IV lung cancer from the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0075] Fig. 18C is a graph showing receiver operating characteristic (ROC) curves for machine learning models discriminating between the spectral fingerprints of the liquid blood samples of the two groups, in accordance with one or more embodiments;

[0076] Fig. 18D is a graph showing a box-and-whisker plot showing the relative intensities of different Raman bands between the liquid blood samples of the two groups, with * representing a p-value of <0.05, ** representing a p-value of < 0.01, and *** representing a p-value of < 0.001, in accordance with one or more embodiments;

[0077] Fig. 19A is an oblique view of an example of a Raman spectroscopy system, showing a sample holder receiver unit, in accordance with one or more embodiments;

[0078] Fig. 19B is an enlarged view of the sample holder receiver unit moved in the open position, in accordance with one or more embodiments; and

[0079] Fig. 19C is an oblique view of an excitation source and a detector of the Raman spectroscopy system of Fig. 19A, in accordance with one or more embodiments.DETAILED DESCRIPTION

[0080] Fig. 1 shows an example of a Raman spectroscopy system 100 for interrogating a liquid blood sample, in accordance with an embodiment. As depicted, the Raman spectroscopy system 100 has a frame 102, a sample holder receiver 104, a Raman spectroscopy measurement unit 106 and a controller 108.

[0081] In this example, the frame 102 is provided in the form of an enclosure which encloses the sample holder receiver 104, the Raman spectroscopy measurement 106 and the controller 108. More specifically, the enclosure can be provided in the form of an openable case having a sturdy construction, including cushioning for example, for securing its components during transportation. However, in some other embodiments, the frame can be provided in the form of an open frame, such as a tabletop, and the like.

[0082] As depicted, the sample holder receiver 104 is mounted to the frame 102. In this example, the sample holder receiver 104 is mounted inside the case. The sample holder receiver 104 is configured for receiving a sample holder in a measurement position. As best shown in Figs. 2 and 3, the sample holder 120 has a body 122 having a depression 124 and a metallic foil 126 covering the depression 124. The metallic foil 126 thus forms a concave surface 128 for receiving the liquid blood sample 130, as best illustrated in Fig. 3A. The concave surface 128 preferably has a continuous clean shape with no discontinuities such as cracks, for instance. In this example, the depression 124 is provided in the form of a through hole extending through a thickness of the body 122. However, in some other embodiments, the depression can be a partial hole or even a concave-shaped recess in the body. In any case, when the sample holder 120 is in the measurement position, the concave surface 128, and thereby the liquid blood sample 130, are exposed an optical path 132 of the Raman spectroscopy measurement unit.

[0083] Depending on the embodiment, the metallic foil can be provided in the form of a metallic tape adhered to the body of the sample holder. Examples of such metallic foil can include, but are not limited to, aluminum foil, silverfoil, gold foil, to name a few examples. Stillreferring to Fig. 3A, the depression 124 generally has a dimension d ranging between about 0.5 cm and 1 cm. The concave surface 126 is sized and shaped to receive at most 200 pL of liquid blood. Indeed, depending on the embodiment, the liquid blood sample has a volume ranging between about 15 and 200 pL, preferably between 25 and 100 pL, and most preferably between 40 and 60 pL. It is understood that since the volume of the liquid blood sample is relatively low, the liquid blood sample will hold on to the concave surface thanks to surface tension. Accordingly, as soon as the liquid blood sample is deposited on the concave surface, the sample holder can be moved around without causing any spillage and loss of the liquid blood sample.

[0084] In the illustrated embodiment, the body 122 of the sample holder 120 includes polymer material. Indeed, in this embodiment, the body 122 has been three-dimensionally printed into shape using a 3D printer and the like. The body 122 can include any other suitable material including, but not limited to, metal, glass, ceramic, composite materials, and the like. It was found convenient to size and shape the body 122 into an elongated prismatic shape. For instance, the body 122 can have a length L ranging between about 2 and 10 cm, a width W ranging between about 1 and 4 cm, and a thickness T ranging between about 0.2 and 1 cm.

[0085] Referring back to Fig. 1 , the Raman spectroscopy measurement unit 106 is also mounted to the frame 102, and inside the case in this example. The Raman spectroscopy measurement unit 106 is configured for performing a Raman spectroscopy measurement on the liquid blood sample 130 via the optical path 132 when the sample holder is in the measurement position. As explained in greater detail below, the Raman spectroscopy measurement typically includes a step of interrogating the liquid blood sample with a Raman excitation beam. The Raman excitation beam propagates once through the liquid blood sample, reflects on the metallic foil, and propagates once again through the liquid blood sample prior to detection. As such, the concave surface of the sample holder acts as a reflective concave surface focusing the Raman excitation beam towards the liquid blood sample in some embodiments.

[0086] It was found that the use of the metallic foil in conjunction with the body of the sample holder is advantageous in many aspects. For instance, the metallic foil (and the liquid bloodsample deposited thereon) can be easily removed between consecutive measurements. As metallic foil such as aluminum foil or tape is relatively cheap, the costs associated to the aluminum foil for each screening can be very low. Moreover, the disposability of the metallic foil can allow the body to be re-used in different measurements, again reducing the overall cost per screening. Moreover, the reflectivity of the metallic foil, and its concave shape, which allows not only one but two passes of the Raman excitation signal through the liquid blood sample, can increase the amount of Raman emission, which can in turn open the door for more accurate measurement using so little volume of liquid blood. Furthermore, a number of different metallic foil were tested and all obtained similar Raman spectra for an identical blood sample, such as shown in Fig. 4. For all of these reasons, it was found particularly valuable to use the sample holder described herein in the Raman spectroscopy system.

[0087] The controller 108 is communicatively coupled to the Raman spectroscopy measurement unit. The communicative coupling can be wired, wireless, or a combination of both depending on the example. During use of the Raman spectroscopy system, the controller is configured for accessing the Raman spectrum, comparing the Raman spectrum to reference data, and for generating a signal based on the comparing. For instance, the controller can be provided in the form of a laptop computer which can access the Raman spectrum via an accessible memory, from an internal computer and / or from a remote computer. The Raman spectrum can be generated in many ways. The Raman spectrum can be presented in the form of data representing an array or matrix of numbers or values, preferably showing a measurand indicative of a Raman emission intensity (e.g., such as an intensity value, a number of photon counts, SNR) as a function of a measurand indicative of spectral content or Raman shift. For instance, the Raman spectrum can include information showing intensity as a function of Raman shift. Such information may be referred to as a raw Raman spectrum. In some other embodiments, the generated Raman spectrum is processed to remove noise or some spectral information which is deemed to be irrelevant. Whether the generated Raman spectrum is raw or processed, the Raman spectrum can be stored on an accessible memory, communicated to an internal computer and / or shared to a remote computer via a wired or wireless communication protocol.

[0088] In some embodiments, the reference data includes a reference Raman spectrum associated to a medical condition. In these embodiments, the comparing step includes comparing the Raman spectrum to the reference Raman spectrum, and determining whether the medical condition is present in the liquid blood sample based on the comparison. The comparing step can involve a full comparison between the measured Raman spectrum and the reference Raman spectrum. However, in some other embodiments, the comparison can only be partial. In these latter embodiments, the comparing step includes the comparison of Raman emission content present within a given spectral region of the Raman spectrum to reference Raman emission content present within the given spectral region of the reference Raman spectrum. For instance, if it has been determined that Raman emission content present within a given spectral region is indicative of a particular medical condition, the comparing step can be limited to that given spectral region. In some embodiments, the given spectral region be a fingerprint region extending between about 600 cm-1and about 1400 cm-1, depending on the embodiment. The spectral region can be expressed in terms of wavenumber units (e.g., in cm-1), expressed in terms of a wavelength (e.g., in nm) and / or expressed in terms of a frequency (e.g., GHZ). However, it was found convenient to express the spectral region in terms of wavenumber units as this notation is independent of nanometers and excitation wavelength, for instance. This strategy can help enhance computational processing time. In this way, the computer may be able to determine whether the medical condition is present in the liquid blood sample based on that partial comparison alone. There can be more than one spectral region of interest for any given medical condition in any given Raman spectra. In some specific embodiments, the medical condition to be assessed is a disease positive such as a cancer positive. In this specific embodiment, this medical condition can be expressed by a reference Raman emission content having a signal-to-noise (SNR) ratio of at least about 3, preferably at least 4, and most preferably at least 5. In some embodiments, relative or normalized intensity is either above or below a corresponding intensity threshold. In some embodiments, the Raman emission content to be assessed is a ratio of a peak intensity value of the measured Raman spectrum to a peak intensity value of the reference Raman spectrum. Any type of difference or set of differences between the measured Raman emission spectra and the reference Raman emission spectra can be identified, and then be associated with a given medical condition, depending on the embodiment. In this embodiment, comparing the Raman emission content of the measuredRaman spectrum to the reference Raman emission content can help determine whether the liquid blood sample cancer positive. Other examples of assessable medical conditions can include, but are not limited to, cancer positive or negative, smoker or non-smoker, cancerous or healthy, respiratory disease(s), diabetes, heart disease(s), dental disease(s), sexually transmitted infection(s), viral hepatitis, vitamin deficiencies, mineral deficiencies and the like. Such reference data can stem from reference measurements made onto real or synthetic reference liquid blood samples presenting a given medical condition. Preferably, the conditions under which the liquid blood sample of interest and the reference liquid blood samples are measured are similar, and the spectral regions inside which Raman emission content is to be compared are similar as well, thereby ensuring that any difference uncovered at the comparing step stem from a difference in the presence of the medical condition, for instance.

[0089] Fig. 5 shows different Raman spectra obtained from liquid blood samples associated to benign tumors / healthy control groups superposed to Raman spectra obtained from liquid blood samples associated to lung cancer of different stages. As can be appreciated, slight differences in Raman spectra can be observed and measured to determine whether a liquid blood sample is healthy or unhealthy. Identification of one or more diseases can stem from the Raman spectra, in addition to the identification of a stage of the disease(s).

[0090] Fig. 6 shows an example of a method 600 for interrogating a liquid blood sample, in accordance with an embodiment. The method 600 can be performed with any of the Raman spectroscopy systems described herein.

[0091] At step 602, a liquid blood sample is received on a sample holder. The sample holder has a body which has a depression and a metallic foil which covers the depression of the body.

[0092] At step 604, the sample holder is moved into a measurement position relative to a Raman spectroscopy measurement unit. In this position, the concave surface of the sample holder is exposed to an optical path.

[0093] At step 606, a Raman spectroscopy measurement is performed on the liquid blood sample via the optical path when the sample holder is in the measurement position. The Raman spectroscopy measurement can be performed by the Raman spectroscopy measurement unit. In this step, the liquid blood sample is interrogated with a Raman excitation beam which is propagated through the liquid blood sample. More specifically, the Raman excitation beam is propagated once through the liquid blood sample, reflected on the metallic foil, and then propagated once again through the liquid blood sample prior to detection. After the propagation of the Raman excitation beam and the detection of the resulting signals, the step 606 includes a step of generating at least one Raman spectrum.

[0094] At step 608, using a controller communicatively coupled to the Raman spectroscopy measurement unit, the Raman spectrum is accessed, then compared to reference data and a signal based on the comparison is generated. In some embodiments, the generated signal is indicative of whether the liquid blood sample is associated to a healthy status or to an unhealthy status. For instance, the signal can be indicative of one or more diseases identified in the liquid blood sample. In some staging-enabled embodiment, the signal can also be indicative of a stage of at least one of the disease(s) identified in the liquid blood sample. It is understood that although the lung cancer is frequently used as an example in this disclosure, the method described herein can be used to screen for a number of different diseases including, but not limited to, any type of cancers (e.g., lung cancer, breast cancer, pancreatic cancer, blood cancer, colorectal cancer, prostate cancer), respiratory diseases such as asthma, and the like. Indeed, uni-cancer screening, multi-cancer or non-cancer disease screening can be enabled by the present method.

[0095] It in intended that the metallic foil can be formed into a concave surface in the depression for receiving the liquid blood sample. In some embodiments, the depression recessed into the body to form a concave shape. In these embodiments, when covered by the metallic foil, the concave shape of the depression forms the concave surface. In other embodiments, the depression is a partial hole or a through hole inside which the metallic foil recedes to form the concave surface. In these latter embodiments, the metallic foil can be obtained by gently pressing the metallic foil within the depression. This pressing step should be interrupted before tearing the metallic foil, in which case a new metallic foil may have to beapplied onto the depression of the body. While the metallic foil can be disposable between consecutive iterations of the method 600, the body of the sample holder can be re-used several times.

[0096] As shown at optional step 610, the step 602 of receiving the liquid blood sample can include extracting one or more drops of liquid blood from a patient and depositing the drop(s) of liquid blood onto the concave surface of the sample holder. An example of such a liquid blood extraction technique includes, but is not limited to, finger pricking. In these embodiments, the liquid blood sample can have a volume below 200 pL. For example, the volume of the liquid blood sample can range between about 15 and 200 pL, preferably between 25 and 100 pL, and most preferably between 40 and 60 pL. It is understood that since the volume of the liquid blood sample is relatively low, the liquid blood sample will hold on to the concave surface thanks to surface tension. Accordingly, as soon as the liquid blood sample is deposited on the concave surface, the sample holder can be moved around without causing any spill and loss of the liquid blood sample.

[0097] It is understood that as the blood sample is provided in the liquid form, therefore no drying step is necessary. Accordingly, the method 600 can be performed quite rapidly. For instance, only 30 minutes, preferably 10 minutes and most preferably 5 minutes may lapse between the step of extracting the drop(s) of liquid blood from the patient and the step of performing the Raman spectroscopy measurement detailed below. In some preferred embodiments, the method 600 can be performed under 2 minutes, preferably under 1 minutes, and most preferably under 30 seconds.

[0098] Referring now to Fig. 7, the controller can be provided as a combination of hardware and software components. The hardware components can be implemented in the form of a computing device 700, an example of which is described with reference to Fig. 7. The computing device 700 can have a processor 702, a memory 704, and I / O interface 706. Instructions 708 for operating the Raman spectroscopy system can be stored on the memory 904 and accessible by the processor 702.

[0099] The processor 702 can be, for example, a general-purpose microprocessor or microcontroller, a digital signal processing (DSP) processor, an integrated circuit, a field-programmable gate array (FPGA), a reconfigurable processor, a programmable read-only memory (PROM), a programmable logic controller (PLC), or any combination thereof.

[0100] The memory 704 can include a suitable combination of any type of computer-readable memory that is located either internally or externally such as, for example, randomaccess memory (RAM), read-only memory (ROM), compact disc read-only memory (CDROM), electro-optical memory, magneto-optical memory, erasable programmable readonly memory (EPROM), and electrically-erasable programmable read-only memory (EEPROM), Ferroelectric RAM (FRAM) or the like.

[0101] Each I / O interface 706 enables the computing device 700 to interconnect with one or more input devices, such as keyboard(s), mouse(s), sensor(s), or with one or more output devices such as monitor screen(s), accessible memory system(s), and external network(s).

[0102] Each I / O interface 706 enables the controller to communicate with other components, to exchange data with other components, to access and connect to network resources, to server applications, and perform other computing applications by connecting to a network (or multiple networks) capable of carrying data including the Internet, Ethernet, plain old telephone service (POTS) line, public switch telephone network (PSTN), integrated services digital network (ISDN), digital subscriber line (DSL), coaxial cable, fibre optics, satellite, mobile, wireless (e.g. Wi-Fi, WiMAX), SS7 signaling network, fixed line, local area network, wide area network, and others, including any combination of these.

[0103] The computing device 700 and any software application that can be ran by the computing device 700 are meant to be examples only. Other suitable embodiments of the controller can also be provided, as it will be apparent to the skilled reader.

[0104] Example - Cost-effective, rapid, plasma-based lung cancer screening and staging via Raman spectroscopy in a national biobank

[0105] A rapid screening test that also allows staging would thus be invaluable. The Ramanbased rapid lung cancer screening test described here addresses this challenge with a portable, minimally invasive, low-cost and user-friendly solution. Designed to integrate seamlessly into existing cancer screening protocols, the screening device acts as an efficientintermediary between initial imaging and invasive procedures. Achieving up to 87% accuracy, this innovative technique is validated by clinical studies. The method and system proposed herein are reagent-free and low-cost by circumventing the use of metal nanoparticles or nano surfaces, which may require refrigeration or pipetting, which adds a few steps to the overall testing procedure. Moreover, the method and system proposed herein involves testing on liquid blood as opposed to dried blood samples, which requires long drying times. Indeed, the typical 10 uL drop of biofluid can take up to 45 minutes to satisfactorily dry. Accordingly, the method and system described herein can perform tests more rapidly than conventional testing techniques relying on dried blood samples. Additionally, the method and system described herein does not require the use of a microscope objective, which often times have a small field-of-view and focal plane, thereby causing manipulation errors which can lead to false positives, false negatives and low-quality signal when operated by unexperienced users. Furthermore, the method and system described herein can rely on a relatively small amount of liquid blood, e.g., between 30 and 100 uL, as opposed to conventional liquid blood testing techniques which require at least 200 mL, which is incompatible with finger prick blood extraction. In these high-volume cases, the sample needs to be centrifuged and pipetted, which means such testing procedures are not user-friendly and rapid.

[0106] Current lung cancer diagnosis relies on the Lung Computed Tomography Screening Reporting and Data System (Lung-RADS). Low dose computed tomography (LCDT) uses X-rays to scan the patients’ lungs and has an extremely high false-positive rate of 96.4%. Chest radiography is another method for lung cancer diagnosis, also relying on X-rays and also has a false positive rate of 94.5%. Both techniques expose the patient to ionising radiation which can damage DNA and increase likelihood of cancer. These false positives may be associated with benign lung nodules i.e. lung nodules which are harmless long term. Lung-RADS introduces more frequent LCDT (3-6 months) screening in the period after the initial detection of lung nodules. This allows clinicians to assess whether nodules are benign or malignant (changed since the previous scan). A Lung-RADS score of 3 or 4A means that the patient falls into an indeterminate category, and this occurs in 11.7% of patients. These patients are put under stress whilst they await further testing, which includes invasive procedures and further exposure to ionising radiation. Furthermore, the false positive result remains around 12.8%. A recent study has reported that a biochemical screening test that costs less than $250 permeasurement would prove to be cost-effective. The method and system disclosed herein can meet this criteria as the system could be rented to hospitals and the sample analysis uses an aluminum well of cost <$1 , and a highly trained clinician is not necessary to operate the device.

[0107] A blood-based method of cancer screening and staging would reduce the requirement for biopsies, and reduces the patient’s exposure to potentially harmful ionising radiation. The analysis of blood can be rapid and integrated into a point-of-care disease screening system. A point-of-care device would require small quantities of blood compatible with finger pricks (<100 ptL), rapid analysis time (<20 min), low cost (<$250 per test) and a portable system. Furthermore, reagents, contrast agents or nanoparticles often require specialised storage conditions, so removing the necessity for these would also allow such a test to be used in resource-poor settings. One possible solution for point-of-care, blood-based analysis is Raman spectroscopy.

[0108] Raman spectroscopy uses lasers to detect molecular changes, including those associated with diseases. In Raman spectroscopy (RS), laser light is shone at a sample, and inelastically scattered light results from the interaction of light with molecules in the sample. The Raman scattered light can be detected using a Raman spectrometer. This light contains information about the concentration and identity of molecules in a sample and can be used as a “molecular barcode”. The Raman spectrum of a molecule consists of a series of peaks where the height of the peaks is related to the concentration of the molecule and the position of the peaks on the x-axis is related to the identity of a molecule. As the molecular composition of biofluids change with disease state of an individual, Raman spectroscopy can be used to gain molecular information from biofluids and hence may be used to build machine learning (ML) models to detect disease. Raman spectroscopy can be used to create diagnostic tests without the use of costly reagents. Reagents may also require refrigeration or specialized storage, whereas Raman spectroscopic techniques can bypass this step.

[0109] Fig. 8 shows an example of a Raman spectroscopy system for the screening of disease(s), in accordance with an embodiment. As depicted, a liquid blood sample is first drawn from a patient. The liquid blood sample can be drawn using a conventional finger prick technique. The amount of volume required for the Raman spectroscopy system ranges between about 15 and 200 pL, preferably between 25 and 100 pL, and most preferablybetween 40 and 60 pL. The liquid blood sample is thus deposited onto a disposable sample holder which is insertable into a slot of the Raman spectroscopy system. Then, the liquid blood sample is processed using Raman spectroscopy, the results of which can be processed using traditional or trained modules to determine the presence of one or more disease(s), and / or stage, if any. The use of the Raman spectroscopy system can be embedded in known clinical workflows where it can be used instead or, or in addition to, conventional LDCT screening tests.

[0110] In this example, 381 blood plasma samples were analyzed from the Quebec Respiratory Health Network Biobank. These blood plasma samples were taken from healthy controls (n=96) and patients with Stage 1-2 lung cancer (n=98), Stage 3-4 cancer (n=50), benign tumors (n=37) and other lung conditions (n=100). 51% of samples were from women and 49% from men. Samples were obtained from non-smokers (33%), smokers (25%) and ex-smokers (42%). Lung cancer types include epidermal carcinoma, adenocarcinoma and neuroendocrine tumors. Blood plasma had been obtained from volunteers as follows. Plasma was collected in 4ml K2-EDTA BD Vacutainers with mauve stoppers. Blood was centrifuged immediately or conserved at 4 °C for a maximum of 3 hours before centrifugation. Centrifugation was carried out at 3200g at room temperature for 15 min, and the upper layer (plasma) was taken and pipetted into 200 pl aliquots, before freezing at -80 °C. Table 2 shows the demographics of lung cancer biobank.

[0111] Table 1 - Demographics of lung cancer biobank from the Quebec Respiratory Health Network (RRSQ) Biobank, in which samples were matched where possible to reduce confounding variables. However, all Stage lll / IV lung cancer patients had been smokers at some point in their life.Control Non-tumoral Benign Stage l / ll Stage 111 / IIV Age, average (std. dev.) 61 (5) 62 (5) 62 (6) 64 (5) 66 (5) SexMale 43 (47) 42 (43) 18 (50) 48 (49) 21 (46) Female 49 (53) 53 (54) 18 (50) 47 (48) 24 (52) N / A - 3 (3) - 2 (2) 1 (2) Histology, n (%)Squamous cell carcinoma — — — 26 (27) 15 (33) Adenocarcinoma — — — 62 (64) 26 (57) Neuroendrocine tumor — — — 6 (6) 3 (7) Sarcomatoid carcinoma — — — — 1 (2) Other — — — — 1 (1) Tabagism, n (%)Non-smoker 41 (45) 65 (66) 6 (17) 8 (8)Smoker 3 (3) 27 (28) 4 (11) 47 (48) 8 (17) Ex-smoker 47 (52) 3 (3) 25 (69) 40 (41) 37 (80) Passive smoker — — 1 (3) — — N / A - 3 (3) - 2 (2) 1 (2) Ethnicity, n (%)Caucasian 91 (100) 95 (97) 36 (100) 94 (97) 45 (98) Indigenous — — — 1 (1) —

[0112] N / A“3 (3)“ 2 (2)1 (2)

[0113] Samples were analyzed with the custom, portable, rapid single-point Raman spectrometer in the form of a suitcase such as the one shown in Fig. 8. The system uses a 785 nm laser source (Model I0785MM1500M4S, Innovative Photonic Solutions, USA) with an output of 1.5 Wand a spectral bandwidth < 2 nm. The other main constituent of the system is a spectrometer (HT model, EmVision, USA) composed of a diffraction grating and a charge-coupled device (CCD) camera (Newton 920, Oxford Instruments, USA) resulting in a spectral resolution < 8.7 cm1.

[0114] Before each measurement the system CCD sensor was cooled to -80°C. Calibration of the x-axis (Raman shift) was determined from a spectrum acquired using acetaminophen powder prior to each measurement. The system response was determined using the fluorescence spectrum of a standard reference material (SRM 2214, National Institute of Standards and Technology, NIST, USA).

[0115] For each blood plasma spectrum, a 50-60 ul sample was analyzed after being pipetted into an aluminum well as described above. Spectra were obtained using a laser power of 861 mW, with 3 measurements per drop. Automatic exposure control was implemented, with an average exposure time of 1.2 + / - 0.3 s, a minimum exposure time of 380 ms and amaximum exposure time of 3000 ms. Fifty accumulations were obtained per sample, but only the first 20 accumulations were required and used in the analysis.

[0116] A number of data processing steps were required to extract a Raman spectrum out of a raw unprocessed acquisition. In the following order, the steps are 1) background (dark noise) subtraction, obtained with the laser turned off, 2) removal of cosmic ray events, 3) x-axis calibration by corresponding Raman peaks from Tylenol measurements to their known position in cm1, 4) correction by the instrument response function (y-axis calibration) obtained from a calibration material (NIST 785 nm Raman standard), 5) truncation of the pixels below 450 cm-1 , 6) averaging of successive accumulations performed on the same sample, 7) baseline removal using the BubbleFill algorithm with adjusted minimal bubble width, 8) further constrain the final Raman spectrum between 600 and 1500 cm-1 , to remove the water peak present at higher wavenumbers, and 9) standard normal variate (SNV) normalization. All spectra are individually interpolated between 600 and 1500 cm1in steps of 1 cm1.

[0117] A metric known as the quality factor (QF) was calculated on the final SNV-normalized spectra to separate low quality Raman spectra from high quality spectra. The QF ranges from 0 to 1. Lower QFs are often linked with higher stochastic noise and lower Raman photon counts, and the spectra often have poorly defined Raman peaks. For blood plasma, it was determined that all spectra had sufficient QF to be used in machine learning models (QF>0.5), so no spectra were rejected on the basis of QF. The average QF is 0.70 ± 0.01.

[0118] To ensure classification results were derived from Raman bands rather than noise, we reduced each Raman spectrum to a set of 20 bands. Gaussians were fitted to any peak with an SNV-normalized intensity of higher than 0.35 to extract peak height, width and position (with 0.2 cm1tolerance to position). Only the peaks present in at least half of samples were considered, to reduce risk of artifacts. There correspond to the 20 Raman bands. For each peak, we kept all the spectral bins corresponding to the full width at half maximum, such that the total number of features (spectral bins) is thus reduced from 900 to 359.

[0119] Each classification model was trained following a 5-fold cross validation to assess the model performance. To do so, the dataset is split into 5 subsets, 4 of these subsets (training set) are used to train a model which is then applied on the remaining set (validationset) to evaluate the performance. This is repeated until each subset was used once as a validation set.

[0120] Independently for each fold, a feature selection algorithm would be applied on the reduced feature set described above, followed by the classification algorithm. The feature selection algorithm is a support vector machine (SVM) with a linear kernel and L1 regularization. The classification algorithm is also a SVM with linear kernel, but with L2 regularization.

[0121] Specific parameters of these two algorithms needed to be adjusted, which was done via a grid search running over multiple combinations. The regularization parameter C for the feature selection, was varied between 0.05 and 0.5 and the maximum number of features between 10 and 30. The regularization parameter C for the classification algorithm ranged between 0.01 and 1. For each combination, the performance was assessed via the 5-fold cross validation described above, where predictions on the validation sets are compared with the known label associated with each sample. The prediction probabilities on the validation set were used to compute a receiver operating characteristic (ROC) curve. The ROC curve gives the area under curve (AUC) but also enables to extract a threshold that optimizes specificity and sensitivity which is used to turn prediction probabilities into binary predictions to make a confusion matrix (CM). From the confusion matrix, we have the number of false / true positives and false / true negatives. The combination of AUC and CM values was used to select the bet set of parameters and thus the best classification model, which we report results on.

[0122] The models described in Table 2 were tested with the correspond number of samples for each class

[0123] Table 2 - Accuracies, sensitivities, specificities an AUC (area under curve) for each machine learning model.M. Negative group Positive group Accuracy Sensitivity Specificity AUC description description [%] [%] [%] Model 1 Benign Cancer stage l / ll 82 87 67 0.80 Model 2 Benign Cancer stage lll / l\Z 86 90 81 0.91 Model 3 Benign Cancer stage I— IV 76 78 69 0.79 Model 4 Cancer stage l / ll Cancer stage lll / IV 84 84 84 0.91 Model 5 Control Cancer stage l / ll 81 84 78 0.86 Model 6 Control Cancer stage lll / IV 87 83 88 0.93 Model 7 Control Cancer stage I— IV 81 79 85 0.88 Model 8 Control / benign Cancer stage I— IV 75 73 78 0.82

[0124] RS and Al can discriminate between blood plasma samples from patients with benign tumors and lung cancer.

[0125] Machine learning (ML) models built using Raman spectroscopy (RS) measurements from liquid blood plasma samples obtained in <2 minutes were able to discriminate between blood plasma samples from patients with benign tumors and Stages l-IV lung cancer (Figs.11, 12 and 13). Peaks associated with 750, 756, 888, 900 and 1337 cm-1 increased in cancer samples whilst peaks at 1127 and 1207 cm-1 decreased (Fig. 14). Model 3, designed to discriminate between benign nodules and lung cancer regardless of stage, did so with 76% accuracy, 78% sensitivity and 69% specificity. Model 1 classified benign nodules and early stage (l-ll) cancers with a higher accuracy of 82%, a sensitivity of 87% and a specificity of 67%. Model 2 classified benign nodules and late stage (lll-IV) cancers with an accuracy of 86%, a sensitivity of 90% and a specificity 81%. Model 2 likely outperforms Model 1 as later stage cancer is often associated with spread to lymph nodes, metastasis, or growth affecting other organs or lung structures. This in turn can bring about more major changes in metabolism than for earlier stage cancers. Furthermore, later stage cancers are associated with an increase in circulating tumor cells (CTCs) which can impact the Raman spectrum of blood plasma. Model 3 is likely less accurate than Models 1 and 2 as the ML model must take into account the differences between the spectra between early stage and late stage cancers. The increase in accuracy with increase in information about pathology (e.g. cancer stage, tumor type) has already been described in our Ember et al. 2024 paper concerning brain cancer.

[0126] The Raman-based device combined with machine learning (ML) currently achieves 75%-87% accuracy, depending on the classification (see Table 2). The combination of RS analysis of liquid blood plasma in less than two minutes and machine learning allows classification of patients by whether they have benign nodules or lung cancer, which could be a critical and cost-effective method to use adjacent to the existing clinical workflow. It is also able to classify lung cancer samples based on whether they are from patients with Stage l-ll or Stage lll-IV cancers with 84% accuracy. This could be invaluable for streamlining patient triaging, prognosis and treatment plans. Finally, the device is able to discriminate between healthy control samples and those with cancer, which could be integrated into a powerful population-wide lung cancer screening tool or tool for assessment of cancer recurrence. This is achieved with a device costing less than $100k, using a few tens of pL of plasma, in liquid form and without the addition of reagents or nanoparticles. Our Raman-based device addresses the challenge of low uptake of lung cancer screening with a non-invasive and low-cost solution. The device is also portable and user-friendly.

[0127] The next step towards point-of-care lung cancer device is integration of fingerpick analysis. This is not necessary for success of the device as venous blood draws are commonplace in a hospital setting. However, replacing venous blood draw with finger prick analysis would reduce the time needed from trained clinicians. The device will also be tested on a large scale in blood plasma samples fresh from patients who are undergoing lung cancer diagnosis.

[0128] As discussed above, the Raman spectroscopy system described herein can be a rapid, user-friendly test for multiple types of lung cancer and lung cancer stage from very small quantities of liquid-form blood products e.g. plasma or serum. This may be serum or plasma, as both yield very similar Raman spectra such as shown in Fig. 9. More specifically, Fig. 9 shows Raman spectra of liquid blood plasma and liquid blood serum. The liquid blood samples include plasma heparin (blue line), plasma obtained using K3-EDTA (yellow line), serum with SST (red line) and serum obtained without any chemical (green line). This figure shows that any of these liquid blood samples can work, as there isn’t any significant differences between the measured Raman spectra.

[0129] Fig. 10 shows exemplary Raman spectra obtained by screening different volume (less than 200 pL) of liquid blood sample such as blood plasma or serum using the disposable sample holder. This graph shows that almost no loss in signal was observed for the different volumes (15 pL shown by blue line), 30 pL shown by red line, and 50 pL shown by green line. Larger volumes may also be used by these are not compatible with point-of-care use. A laser light is shone at a blood product and the light Raman scattered back from the sample contains information about structure, bonding, and concentration of molecules in the sample. The Raman scattered light contains a molecular fingerprint of the blood product. Machine learning (ML) models differentiate between molecular fingerprints with 79-93% accuracy depending on lung cancer stage, the results of which are shown in Figs. Figs. 11-18.

[0130] Figs 19A-C show an example of a custom, portable, low-cost Raman spectroscopy system 1900 which was used to analyze blood plasma from 381 blood plasma samples from the Quebec Respiratory Health Network Biobank. These blood plasma samples were taken from healthy controls and patients with Stage 1-2 lung cancer, Stage 3-4 cancer, benign tumors, and other lung conditions. The custom Raman spectrometer was equipped with a 785 nm excitation wavelength and analyzes a 15-200 pL liquid drop of blood plasma within 2 minutes. Machine learning models were built from Raman spectra obtained from blood plasma from volunteers from the biobank. Accuracy, sensitivity, and specificity were assessed.

[0131] As shown in this example, the Raman spectroscopy system 1900 has a sample holder receiver 1904 provided in an ejectable format. More specifically, the sample holder receiver 1904 has a sample securing mechanism 1911 movable to a receiving position O, in which the sample holder 1920 is received away from the measurement position, and to a close position C, in which the sample holder 1920 is in the measurement position. In this way, the sample holder 1920 can be secured within the sample securing mechanism 1911 and then moved into the measurement position which is hidden inside a casing of the sample holder receiver 1904, where it is exposed to the optical path of the Raman spectroscopy measurement unit. By doing so, ambient light is prevented from affecting the Raman spectroscopy measurement. Fig. 19C more specifically shows the optical components of the Raman spectroscopy measurement unit 1906 which may be hidden inside the case of the Raman spectroscopy system 1900.

[0132] In this embodiment, the Raman spectroscopy measurement unit 1906 has a Raman excitation source 1950 and a Raman emission detector 1952 which are both communicatively coupled to a computer 1908. As depicted, the coupling is wired in this embodiment. However, it is envisaged that the coupling can be wireless in some other embodiments. As shown, the Raman excitation source 1950 is in this case provided in the form of a fibered laser source operating at about 785 nm. The Raman spectroscopy measurement unit 1906 includes a Raman interrogation path having a fiber cable, a collimating lens, a band-pass filter, a dichroic mirror and a focusing lens providing a Raman excitation beam having a required beam dimension. A Raman emission path is also provided. The Raman emission path can have a focusing lens, a dichroic mirror, a high-pass filter, an injection lens and a fiber bundle collecting as much of the Raman emission as possible and communicating it to the Raman emission detector. In this embodiment, the Raman emission detector 1952 is provided in the form of a spectrometer. An example of such the fiber bundle is described in PCT Patent Application Publication No. WO 2019 / 051 ,602 A1 , the contents of which are hereby incorporated by reference. The use of the fiber bundle has been found to be convenient as it can enhance the collected amount of Raman emission content. However, in some other embodiments the fiber bundle can only be optional as a standard Raman emission collecting fiber may be used. It is intended that the Raman spectroscopy unit described herein is meant to be exemplary only. Indeed, it is noted that other examples of Raman spectroscopy units can be used in some other embodiments. The optical components shown in this specific example are meant to be exemplary only, as in some other embodiments other optical components can equivalently be used. For instance, in this example some optical components are shared between the Raman interrogation path and the Raman emission path. However, in some other embodiments, the optical components of the Raman interrogation and emission paths and are independent from one another.

[0133] The system and method presented herein were found to be advantageous on many different levels. For instance, a key step in the screening method proposed herein is that the blood plasma is in liquid form, and not dried form, as in the majority of other existing Raman spectral studies. It is difficult to obtain Raman spectra from liquid samples due to the low concentration of Raman-active molecules compared to dried samples. Moreover, the sample is analyzed using a single-point system rather than a microscope (using a microscopeobjective), rendering it an extremely rapid experimental protocol. The majority of other existing Raman spectral studies use Raman microscopes as the Raman signal may be superior in dried samples. As evidenced in Figs. 11-18, the system and method presented herein can discriminate between stages of lung cancer in addition to simple presence / absence of lung cancer. Other existing Raman studies focus on NSCLC alone, or do not report whether findings are applicable to multiple types of lung cancer. Moreover, the sample is relatively small (less than or equal to 200 p). This is compatible with extraction of plasma or serum from finger pricks, making the system and method described herein user-friendly, minimally-invasive, and rapid. It is also noted that the sample is reagent-free and does not use nanoparticles or nano surfaces, reducing the cost and potential toxicity of the analysis.

[0134] Indeed, Raman-techniques involving nanoparticles or nano surfaces, such as surface-enhanced Raman spectroscopy (SERS), typically rely on the addition of metal nanoparticles to the sample. This is costly, difficult to reproduce and adds an extra pipetting step. The extra step increases the time until diagnosis and reduces how user-friendly the platform is. The system and method described herein does not need the sample to be dried before testing is performed. Indeed, the extra time required for drying the sample is omitted with the system and method presented herein. The system and method described herein use liquid form blood plasma samples, reducing total analysis time and fitting into the clinical workflow more easily, allowing the device to be considered “point-of-care.” Moreover, the system and method disclosed herein does not use microscope objective and instead focuses light onto the entire sample. This can a) integrate information from the entire sample, reducing the likelihood that bubbles or particles can have an impact on the reading, b) reduce user error due to lack of focussing, as the focal plane of microscopes if very small, and / or c) reduce the cost of the device as Raman microscope objectives may be costly. The system and method described herein can thus be sold at lower price points than competitor devices. By performing testing on volumes of equal to or more than 200 pL, the system and method described herein are incompatible with finger prick blood extraction. As a result, liquid blood samples can originate from venous blood draws or finger pricks (small volumes). This can minimise the invasiveness of the process and reduces the need for trained phlebotomists for blood extraction. The system and method described herein can also detect the stage of a disease. As demonstrated in Figs. 11-18, the system and method described herein is able to detect thestage of lung cancer More specifically, model 4 (Fig. 15) is able to discriminate between Stage l-ll cancer and Stage lll-IV cancer with 84% accuracy, 84% sensitivity and 84% specificity. This can give clinicians the ability to give patients an indication of their prognosis, and could allow triaging of patients, allowing later stage patients to be biopsied and treated earlier.

[0135] As can be understood, the examples described above and illustrated are intended to be exemplary only. The scope is indicated by the appended claims.

Claims

WHAT IS CLAIMED IS:1 . A Raman spectroscopy system for interrogating a liquid blood sample, the Raman spectroscopy system comprising:a frame;a sample holder receiver mounted to the frame, the sample holder receiver configured for receiving a sample holder in a measurement position, the sample holder having a body having a depression and a metallic foil covering the depression, the metallic foil forming a concave surface receiving the liquid blood sample, the concave surface exposed to an optical path when the sample holder is in the measurement position; a Raman spectroscopy measurement unit mounted to the frame and configured for performing a Raman spectroscopy measurement on the liquid blood sample via the optical path when the sample holder is in the measurement position, said Raman spectroscopy measurement including interrogating the liquid blood sample with a Raman excitation beam propagating once through the liquid blood sample, reflecting on the metallic foil, and propagating once again through the liquid blood sample prior to detection, and generating a Raman spectrum resulting from said propagation; and a controller communicatively coupled to the Raman spectroscopy measurement unit, the controller configured for accessing said Raman spectrum; comparing said Raman spectrum to reference data; and generating a signal based on said comparison.

2. The Raman spectroscopy system of claim 1 wherein the metallic foil is an aluminum foil.

3. The Raman spectroscopy system of claim 1 or 2 wherein the metallic foil is adhered to the body of the sample holder.

4. The Raman spectroscopy system of any one of claims 1 to 3 wherein the depression has a dimension ranging between about 0.5 cm and 1 cm.

5. The Raman spectroscopy system of any one of claims 1 to 4 wherein the liquid blood sample has a volume ranging between about 15 and 200 pL, preferably between 25 and 100 pL, and most preferably between 40 and 60 pL.

6. The Raman spectroscopy system of any one of claims 1 to 5 wherein the sample holder receiver has a sample securing mechanism movable to a receiving position, in which the sample holder is received away from the measurement position, and to a close position, in which the sample holder is in the measurement position.

7. The Raman spectroscopy system of any one of claims 1 to 6 wherein the measurement position is inside the sample holder receiver.

8. The Raman spectroscopy system of any one of claims 1 to 7 wherein the depression is a hole recessed in the body of the sample holder.

9. The Raman spectroscopy system of claim 8 wherein the hole is a through hole extending through the body.

10. The Raman spectroscopy system of any one of claims 1 to 9 wherein the body of the sample holder includes polymer material.

11. The Raman spectroscopy system of any one of claims 1 to 10 wherein the sample holder has an elongated prismatic shape having a length ranging between about 2 and 10 cm, a width ranging between about 1 and 4 cm, and a thickness ranging between about 0.2 and 1 cm.

12. A liquid blood sample holder for a Raman spectroscopy measurement, the liquid blood sample holder comprising:a body having a depression; anda metallic foil covering the depression of the body, the metallic foil forming a concave surface receiving a liquid blood sample having a volume below 200 pL, the liquid blood sample holding on to the metallic foil due to surface tension;wherein, during the Raman spectroscopy measurement, a Raman excitation beam propagates once through the liquid blood sample, reflects on the metallic foil, and propagating once again through the liquid blood sample prior to detection.

13. The liquid blood sample holder of claim 12 wherein the metallic foil is an aluminum foil.

14. The liquid blood sample holder of claim 12 or 13 wherein the metallic foil is adhered to the body of the sample holder.

15. The liquid blood sample holder of any one of claims 12 to 14 wherein the depression has a dimension ranging between about 0.5 cm and 1 cm.

16. The liquid blood sample holder of any one of claims 12 to 15 wherein the depression is a hole recessed in the body of the sample holder.

17. The liquid blood sample holder of claim 16 wherein the hole is a through hole extending through the body.

18. The liquid blood sample holder of any one of claims 12 to 17 wherein the body of the sample holder includes a polymer material.

19. A method for interrogating a liquid blood sample, the method comprising:receiving a liquid blood sample on a sample holder, the sample holder having a body having a depression and a metallic foil covering the depression of the body, the metallic foil forming a concave surface receiving the liquid blood sample;moving the sample holder into a measurement position relative to a Raman spectroscopy measurement unit, in which the concave surface is exposed to an optical path;using the Raman spectroscopy measurement unit, performing a Raman spectroscopy measurement on the liquid blood sample via the optical path when the sample holder is in the measurement position, said Raman spectroscopy measurement including interrogating the liquid blood sample with a Raman excitation beam propagating once through the liquid blood sample, reflecting on the metallic foil, and propagating once again through the liquid blood sample prior to detection, and generating a Raman spectrum resulting from said propagation; andusing a controller communicatively coupled to the Raman spectroscopy measurement unit, accessing said Raman spectrum, comparing said Raman spectrum to reference data, and generating a signal based on said comparison.

20. The method of claim 19 wherein the liquid blood sample has a volume below 200 pL.

21. The method of claim 19 or 20 wherein said receiving the liquid blood sample includes a step of extracting one or more drops of liquid blood from a patient and a step of depositing the one or more drops of liquid blood onto the concave surface of the sample holder.

22. The method of claim 21 wherein the step of extracting the one or more drops of liquid blood from the patient and the step of performing the Raman spectroscopy measurement are performed within 30 minutes from one another, preferably within 10 minutes from one another, and most preferably within 5 minutes from one another.

23. The method of any one of claims 19 to 22 wherein said method is performed under 2 minutes, preferably under 1 minutes, and most preferably under 30 seconds.

24. The method of any one of claims 19 to 23 wherein the signal is indicative of whether the liquid blood sample is associated to a healthy status or to an unhealthy status.

25. The method of any one of claims 19 to 24 wherein the signal is indicative of one or more diseases identified in the liquid blood sample.

26. The method of claim 25 wherein said signal is indicative of a stage of at least one of the one or more diseases identified in the liquid blood sample.

Images