Disease progression prediction

Abstract

Embodiments are presented for predicting the likelihood of a patient to progression to the late stage of a particular medical condition (e.g., known medical disease). These embodiments provide a predictive model and a clustering analysis method. The prediction model and the clustering analysis method can be executed in parallel or in sequence. Alternatively, only one of the predictive model and the clustering analysis method may be performed, for example, to improve processing speed and reduce the use of hardware resources.

Description

Background Art

[0001] The likelihood of progression to an advanced stage of the disease may vary for different patients. While a first patient may take several years to progress from stage one cancer to stage four cancer, a second patient may experience this progression in just a few months, and a third patient may never see cancer progress beyond stage two. Depending on the likelihood and extent to which the disease may progress in a particular patient, different medical procedures or treatments may be prescribed for the patient. In any case, early diagnosis of the disease can generally improve the chances of recovery and reduce the risk of fatality caused by the progression of the disease. Summary of the invention

[0002] Embodiments of the present disclosure include computer-implemented methods and systems for analyzing and predicting the progress of medical conditions (e.g., specific diseases) in patients. Analysis and prediction are based on the clinical characteristics of patients measured or observed at multiple moments during disease progression. Therefore, the present embodiment can predict the possibility of the patient's progression to the late stage of the disease based on the patient's current or past clinical characteristics (also referred to herein as "medical characteristics"). The patient's clinical characteristics can include one or more of the following information: the patient's demographic data (e.g., age, sex, place of birth, and living area), medical history (e.g., past or present diagnosis, medical prescriptions, and medical procedures), biomarker information (e.g., tumor size), body mass index, and lifestyle (e.g., smoking and drinking habits).

[0003] In some embodiments of the present embodiment, the prediction model predicts the possibility of the patient's progression to the late stage of the disease based on the clinical characteristics of the patient measured at one or more moments. The model is trained by using training data including a set of clinical characteristics of training patients. The clinical characteristics of each patient are measured at multiple moments during the patient's disease progression process. Once the model is trained, the model can predict the possibility of the patient's (e.g., test patient) disease progression to, for example, a predetermined late stage.

[0004] In some embodiments, cluster analysis is performed to find patterns in medical features and thus identify medical features associated with disease progression. The results of the cluster analysis can be used to find clinical features associated with disease progression and the proportion of patients associated with such features. Clinical features can be used in predictive models to improve the accuracy of predictions.

[0005] The present disclosure also provides one or more non-transitory computer-readable storage media, which are coupled to one or more processors and have instructions stored thereon, which when executed by the one or more processors cause the one or more processors to perform operations according to an embodiment of the method provided herein.

[0006] The present disclosure further provides a system for implementing the method provided herein. The system includes one or more processors, and a computer-readable storage medium, the computer-readable storage medium is coupled to the one or more processors and has instructions stored thereon, which, when executed by the one or more processors, cause the one or more processors to perform operations according to the embodiments of the method provided herein.

[0007] The method according to the present disclosure may include any combination of aspects and features described herein. That is, the method according to the present disclosure is not limited to the combination of aspects and features specifically described herein, but also includes any combination of aspects and features provided.

[0008] Among other advantages, the present embodiment provides the following benefits. Since the present embodiment trains the prediction model based on the patient's progression process, the present embodiment provides a more accurate estimate of the disease progression of a particular patient compared to the conventional methods using a single or separate treatment and evaluation results (such as survival rate and recurrence rate). Further, the prediction provided by the present system can include time, development stage, or any other parameter when predicting disease progression. For example, the system can indicate when and with how high accuracy a particular patient will progress from one stage of cancer to another stage (e.g., from a phase I cancer progresses to a phase II cancer, from a phase I cancer progresses to a phase III cancer, from a phase I cancer progresses to a phase IV cancer, from a phase II cancer progresses to a phase III cancer, from a phase II cancer progresses to a phase IV cancer, etc.).

[0009] In addition, the present embodiment uses the clinical characteristics of the patient when making predictions. Compared with the conventional methods that only consider treatment characteristics (such as survival rate and recurrence rate), the present embodiment provides personalized predictions for patients. These personalized predictions provide an estimate of the possibility or process of the progression to the late stage of the disease for a particular patient based on the medical characteristics specific to the particular patient. The late stage of the disease can be predefined; it can be a specific stage of the disease, which can be defined based on the degree of spread of the disease in the patient's body, based on the size of the tumor in the patient's body, based on the medical treatment that the patient has as an option for treating the disease, based on the medical treatment or drugs that the patient has taken, etc. The embodiment can predict the different stages of disease progression based on the corresponding predefined parameters of each stage.

[0010] Embodiments may use the predictive model to estimate the likelihood of disease progression. Embodiments may also use cluster-based analysis to complement predictive model building by confirming or rejecting features already identified by the predictive model. It may be possible to confirm whether there is overlap in findings between the two approaches, detect batch effects, and check whether there is any complementary value addition. Thus, the two models can be used to confirm, reject, or modify each other's results.

[0011] The details of one or more embodiments of the present disclosure are set forth in the accompanying drawings and the description below. Other features and advantages of the present disclosure will be apparent from the description and drawings, and from the claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 An example system is depicted that can be used to perform embodiments of the present disclosure.

[0013] Figure 2A and Figure 2B are example processes that may be performed by the system according to an implementation of the present disclosure.

[0014] Figure 3 is an example process that may be performed by a clustering module according to an embodiment of the present disclosure.

[0015] Figure 4 Schematic diagrams showing example computing devices and mobile computing devices that can perform the methods described in this disclosure.

[0016] Like reference numbers in the various drawings indicate like elements. DETAILED DESCRIPTION

[0017] The present disclosure provides methods and systems for predicting the likelihood of a medical condition (e.g., a disease) in a patient progressing to one or more advanced stages. The disease can be any known disease, such as any type of cancer or tumor. In an example, the disease is cutaneous squamous cell carcinoma (CSCC), which is the second most common type of skin cancer and currently causes approximately 7,000 deaths per year in the United States due to the advanced stages of the cancer.

[0018] This embodiment can reduce diagnostic delays, thereby reducing the chance of patients progressing to late stages of the disease, and improving the treatment rate and survival rate of patients. For example, a common diagnostic and treatment procedure for CSCC is that a dermatologist initially diagnoses the patient with cancer, and depending on the dermatologist's diagnosis, the patient is referred to an oncologist. This process is both time-consuming and subjective-that is, it is affected by the dermatologist's observations. This embodiment reduces (or even eliminates) the impact of these two problems. The procedures disclosed herein can be used as early as the initial diagnosis of the patient's disease, and are not affected by a particular doctor's misdiagnosis or delayed diagnosis.

[0019] In some embodiments, cluster analysis is used to find patterns in the data and give details of the features that are important in those patterns. It can be an unsupervised machine learning method, where the machine does not know whether the patient has an advanced form of the disease. Any feature obtained from the cluster analysis can be the dominant pattern in the data.

[0020] On the other hand, the predictive model can specifically distinguish between advanced forms of the disease and non-advanced forms of the disease. In some embodiments, the features identified by the predictive model are compared with the features identified by the cluster analysis. If the features identified in the predictive model overlap with the features identified by the cluster analysis, the model can be recognized as capturing the dominant pattern in the data. If there is a mismatch between the features, the predictive model can be modified to reduce (e.g., minimize or eliminate) the mismatch.

[0021] Figure 1 An example system 100 is depicted that can be used to perform embodiments of the present disclosure. The system 100 receives input data 128 and provides output 130. During a training phase, the input data 128 is a set of clinical characteristics (i.e., medical features) of training patients 102a to 102n, and the output data 130 can be a set of identified predictive features or feature clusters. During an interference phase, the input data can be the medical characteristics (i.e., features) of a particular patient, and the output 130 is a prediction of the progression of the particular patient in a disease, e.g., an estimate of the likelihood that the patient will progress to a predetermined advanced stage of the disease, or a prediction of the course of the patient's progression to an advanced stage of the disease.

[0022] The system 100 includes a data preprocessor module 124, a predictive model module 110, and a clustering module 116. The system 100 may also include a storage device 126 and an evaluator module 122. The system 100 is capable of communicating with an external device (e.g., external device 132) to send or receive data, such as receiving input data, storing output data, and / or transmitting output for presentation to a user.

[0023] The data preprocessor 124 performs initial operations on the input data, for example, to clean, classify, identify specific features of interest, etc. The data preprocessor 124 may have one or more submodules, for example, a group generator 104, a time point mapper 106, a feature identifier 108, etc. Some of these submodules may only operate at specific stages of the operation of the system 100. For example, the group generator 104 may only operate during the training stage. The data preprocessor 124 transmits the preprocessed data to one or both of the prediction model module 110 and the clustering module 116.

[0024] When the system 100 is in the training phase, each of the prediction model module 110 and the clustering module 116 can use the input data to identify corresponding predictive features that can later be used to predict the likelihood of a particular patient progressing to an advanced stage of the disease. The identified predictive features can be stored in a data storage device that can be part of the system 100 (e.g., the storage device 126) or an external device that communicates with the system 100 (e.g., the external device 132). When the system is in the interference phase, each of the prediction model module 110 and the clustering module 116 predicts the corresponding likelihood of a particular patient progressing to an advanced stage of the disease based on information about the corresponding features identified by the prediction model module 110 or the clustering module 116 in the training phase.

[0025] In the training phase, the system 100 receives training data as medical features (eg, F1, F2, ..., Fn) of a plurality of patients (eg, patients 102a to 102n). The medical features of each patient include information indicating whether the patient has progressed to an advanced stage of a particular disease.

[0026] The data preprocessor 124 divides the training data into a plurality of patient groups. For example, the group generator 104 divides the patients 102a to 102n into an advanced group and a non-advanced group. The advanced group includes patients whose diseases have progressed to a predetermined advanced stage. The non-advanced group includes patients whose diseases have not yet progressed to a predetermined advanced stage.

[0027] Each medical feature (F1, F2, ..., Fn) of a patient is a clinical characteristic of the patient that is measured at multiple moments during the course of the patient's disease progression. Figure 1 In the example training data shown, each of the medical features F1 to Fn of the patient 102a is measured at three times t0, t1, and t2 during the disease progression of the patient 102a. However, the number of times for measuring, recording, or using medical features is not limited.

[0028] Since each patient may have a different disease progression process, the moment at which the measurement occurs may also differ from the first patient to the second patient. In order to compare relevant medical features to each other, the time point mapper 106 of the data preprocessor module 124 identifies the absolute moments t0, t1, and t2 of interest and converts these absolute moments relative to the time when each process begins and ends into relative time points T0, T1, T2 for the process.

[0029] For example, while the first patient's process may take five years to progress from stage one cancer to stage four cancer, the second patient's process may take two years to progress from stage one cancer to stage four cancer. The number and timing of feature measurements and diagnoses made by the first patient during the first process may be different than the number and timing of feature measurements made by the second patient during the second process. Therefore, the time point mapper 106 converts the absolute measurement times (t0, t1, t2) to relative time points (T0, T1, T2) so that the system 100 can compare the features measured at the relative time points relative to each process.

[0030] Each of these relative time points is a predefined moment during the patient's disease progression process. The patient's disease progression process may begin at a specified time period before or after the initial diagnosis of the patient's disease. For example, T0 may indicate the initial point of the process, which may be one year prior to the initial diagnosis.

[0031] The process can end at a predefined end time point. For patients 102a to 102n that progress to a particular late stage of the disease, the end time point can be set to the time when the patient is diagnosed as having an advanced stage. For patients that have not yet progressed to an advanced stage, the end time point can be defined, for example, relative to the median of the end time points of patients that have progressed to an advanced stage. In the example presented above, T2 can represent the corresponding end time point of the process for each training patient. T1 can be a predetermined value between T0 and T2, for example, six months after the initial diagnosis.

[0032] Alternatively or additionally, the mapping from absolute moments to relative time points may be completed before the training data enters the system 100. For example, each of the moments t0, t1, and t2 of the patient 102a may actually be a relative time point relative to the disease progression of the patient 102a.

[0033] Once relative time points are obtained, the time point mapper 102 selects a predetermined number (e.g., three) of time points or a specific set of time points so that the system will focus on the features measured at those time points. For example, the specific set can be T0 as the starting point of the process, T1 as six months after the initial diagnosis of the disease, and T2 as the end point of the process. Therefore, if the medical features of patient 102j (a patient among training data patients 102a to 102n) have been measured at multiple moments mapped to time points T0, T1, T2, and T3, the system will skip the features measured at T3. Depending on the desired processing speed and accuracy, the time point mapper 106 can select more or less time points. Selecting a smaller number of time points will improve processing speed, and selecting more number will improve the accuracy of identifying appropriate prediction features and estimating the possibility of disease development. Alternatively or additionally, the operator can select or can participate in selecting one or more time points in these time points.

[0034] The time point mapper 106 can perform mapping and selection in any order. It can map first and then select, select first and then map, or map and select in parallel. For example, the time point mapper 106 can only target specific moments and only select specific moments (e.g., the beginning of a process, six months from the first diagnosis, and the end of a process), and only map those specific moments to corresponding time points.

[0035] The data preprocessor 124 may also select one or more features from a plurality of features included in the training data. For example, the feature selector 108 may select features F1 to Fj from features F1 to Fn measured for the patient 102a. The data preprocessor 124 may select the same features (e.g., features F1 to Fj) at the same time point (e.g., time points T0, T1, T2) for all training patients 102a to 102n.

[0036] The data preprocessor 124 provides the generated groups (e.g., late groups and non-late groups), time points (e.g., time points T0, T1, T2), and selected features (e.g., features F1 to Fj) to each of the prediction model module 110 and the clustering module 116.

[0037] Continuing with the training phase, the prediction model module 110 trains the prediction model by using the preprocessed training data. The training of the prediction model includes identifying prediction features. The prediction features can be features that distinguish patients in the advanced group from those in the non-advanced group at various time points.

[0038] To determine whether a particular feature among the selected features F1 to Fj is a predictive feature, the model-based (MB) feature identifier 112 may determine a first number of patients having the particular feature in the non-advanced group and a second number of patients having the particular feature in the advanced group at each of the selected time points T0, T1, and T2.

[0039] MB feature identifier 112 obtains the delta value at each time point by calculating the difference between the first quantity and the second quantity at each time point. MB feature identifier 112 determines whether the feature is identified at a specific time (e.g., T0 and T1) or when moving from a first time point (e.g., T0) to a second time point (e.g., T2) later than the first time point.

[0040] In some embodiments, the MB feature identifier 112 uses an iterative partitioning procedure to partition patients (e.g., into two groups) based on features that best distinguish patients who progress to an advanced form of the disease from patients who do not progress to an advanced form. The partitioning can be based on corresponding feature values ​​of one or more features. For example, the MB feature identifier 112 can use the following features to partition patients, i.e., using antiproliferative drugs without taking medication. This procedure is repeated until different groups of advanced patients and non-advanced patients are formed, or until it is no longer possible to distinguish patients based on the feature values ​​of a particular feature. The MB feature identifier 112 can repeatedly apply this procedure to all features to obtain various combinations of features to distinguish advanced patients from non-advanced patients. The MB feature identifier 112 can also rank features based on their ability to distinguish advanced patients from non-advanced patients.

[0041] In some embodiments, module 112 checks whether a particular feature is associated with a target disease in which other features are present, and makes a statistical determination of the importance of the particular feature. The importance is determined based on the frequency of occurrence of the particular feature among patients with the target disease, or based on the value or range of values ​​of the particular feature among patients with a particular disease.

[0042] In some embodiments, a delta value provides additional information about the relative importance of a particular feature between multiple time points (e.g., T0 to T2). In some embodiments, if both a predictive model and the change in delta value among multiple time points confirm that a particular feature is an indicator of progression of a target disease, the importance level of a particular feature relative to a target disease can be confirmed.

[0043] As a first example, if 1000 patients from the advanced cohort and 500 patients from the non-advanced cohort use drug X at T0, the delta value for the feature "use drug X" for T0 will be 500. If 1500 patients from the advanced cohort and 800 patients from the non-advanced cohort use drug X at T2, the delta value for the feature at T2 will be 700. Assuming that T2 is later than T0, the delta value for the feature increases by 40% from T0 to T2 (i.e., (700-500) / 500). Now, if the threshold is 20%, the MB feature identifier 112 identifies the feature "use drug X" as a predictive feature because the 40% increase exceeds the threshold of 20%. However, as a second example, if only 950 patients from the non-advanced cohort used drug X at T2, "use of drug X" would not be identified as a predictive feature because the Δ value only increased by 10% (i.e., (550-500) / 500) from T0 to T2.

[0044] In some embodiments, the MB feature identifier 112 determines one or more corresponding predictive powers for each predictive feature. In some embodiments, the predictive power of a feature can be calculated as the difference between the values ​​of the feature measured at different time points (e.g., T2 and T0, or T1 and T0). In some embodiments, the predictive power can be calculated as the difference normalized by a reference value (e.g., the value of the feature measured at T0). For example, in the first example discussed in the previous paragraph, the predictive power of the feature "use of drug X" would be 40% for predicting a medical feature at T2 based on the features measured at T0. The predictive model module 110 can later use the predictive power of these features in the interference stage when using each feature to predict disease progression in a new patient.

[0045] In some embodiments, the MB feature identifier 112 identifies the predictive feature by applying a tree-based approach to the corresponding values ​​of the medical features received from the data preprocessor 124 to distinguish patients in the advanced cohort from patients in the non-advanced cohort at various time points, for example, according to the procedure explained earlier. The MB feature identifier 112 makes the distinction in an iterative manner based on the corresponding values ​​of the medical features.

[0046] The prediction model module 110 may store the information of the prediction feature in a data storage device (e.g., storage device 126), or transmit the prediction feature for presentation on another device (e.g., external device 132). The information of the prediction feature may include a corresponding identifier associated with the prediction feature, a prediction capability, a measurement time point, etc.

[0047] In the interference phase, the system 100 receives medical features of a particular patient (e.g., a new patient) and predicts the likelihood of the disease progressing to an advanced stage in the patient. The data preprocessor 124 can optionally preprocess the medical features of the particular patient, for example to select specific time points and features that the prediction model 100 uses to identify predictive features. In some embodiments, the preprocessed data includes only predictive features for which the data preprocessor 124 can obtain information from the storage device 126 or from the prediction model module 110. In some embodiments, the preprocessed data includes only a subset of the time points that the prediction model 100 uses to identify predictive features.

[0048] In some embodiments, the system 100 does not use one or more submodules of the data preprocessor 124 during the interference phase. For example, the predictive model module 110 does not use the feature selector 108, but instead works based on the medical features of the specific patient received by the system 100 and skips any medical features that are not associated with any of the predictive features.

[0049] In the intervention phase, the model-based (MB) estimator 114 of the prediction model module 110 can use the prediction model trained in the training phase to predict the model-based likelihood of disease progression for a particular patient. The model-based likelihood can be predicted as an estimate of the chance of progression to a predetermined late stage of the disease for a particular patient.

[0050] The input data to the MB estimator 114 is the preprocessed data from the data preprocessor 124, and the predictive features identified by the MB feature identifier 112. The output data of the MB estimator 114 is the model-based likelihood. The MB estimator 114 generates the output by, for example, retrieving information of the predictive features that the MB feature identifier 112 has identified in the training phase from the storage device 126. The MB estimator 114 then uses the predictive features on the trained predictive model and applies the model to the medical features of a particular patient to estimate the model-based likelihood.

[0051] In some embodiments, the MB estimator 114 uses each of the predictive features according to the predictive power associated with the predictive features. For example, if the MB feature identifier 112 has determined that the predictive power of a first predictive feature is 20% and the predictive power of a second predictive feature is 40% for a particular disease, the MB estimator 114 considers the second predictive feature to be a more decisive factor than the first predictive feature in estimating the likelihood of progression to an advanced stage of the disease.

[0052] MB estimator 114 may provide the estimated likelihood to an external device, such as for presentation or storage.

[0053] The clustering module 116 includes a cluster-based (CB) feature identifier 118 that clusters patients 102a to 102n based on a first feature subset (e.g., features F2, F4, ..., F10) measured at a specific time point (e.g., T0 or T2). In contrast to the training phase in the prediction model, the clustering module 116 does not use information about the patient's disease progression. For example, the occurrence rate of advanced cSCC or non-advanced cSCC can be hidden from the clustering algorithm. The generated clusters represent corresponding patient subsets that share a predetermined similarity in the first feature subset. In some examples, the features are measured on a binary basis, and the similarity means the presence or absence of the feature in the patient; for example, the presence or absence of a drug in the patient. In some examples, the features are measured on a continuous basis, and the similarity means the closeness of the feature on two different patients. For example. As another example, the age of the patient at the T0 time point, for example, one feature indicates a patient aged 30 to 40 at T0, and another feature indicates a patient aged 40 to 45 at T0.

[0054] The CB feature identifier 118 identifies one or more cluster subsets from the generated clusters, each cluster subset having a greater than a threshold number of patients whose disease has progressed to a predetermined late stage (e.g., 50 patients, or 60% of the patients in the cluster). For each cluster in the identified cluster subsets, the CB feature identifier 118 identifies a corresponding second feature subset (also referred to as a "dominant" feature subset) that is shared among the patients in the cluster. In some embodiments, if more than a second threshold number of patients in the cluster (e.g., more than 70%) have a feature, the feature may be in the dominant feature subset. In some embodiments, if more than a second threshold number of patients with an advanced stage of the disease in the cluster have a feature, the feature may be in the dominant feature subset.

[0055] The CB feature identifier 118 outputs a second feature subset (ie, a dominant feature subset). The output may be transmitted to the evaluator module 122, stored in the storage device 126, or transmitted to an external device (eg, 132) for presentation or further processing.

[0056] In some embodiments, clustering module 116 uses the output of CB feature identifier 118 to determine the importance of each feature in the dominant feature subset relative to the advanced form of the disease. For example, clustering module 116 can overlay disease-related information (e.g., progression of the disease) for each patient in the cluster, calculate a delta value for each feature within each cluster, and determine the relative importance of the feature or the correlation between the feature and progression to the advanced form of the disease based on the change in delta between multiple time points (e.g., T0 to T2). In practice, clustering module 116 determines the correlation between the features observed or measured in patients with advanced disease. In some embodiments, the features identified as the most important features are selected by prediction model module 110 (e.g., by evaluator module 122) to verify or confirm the predictive features identified by MB feature identifier 112.

[0057] In some embodiments, the evaluator module 122 compares the dominant features output from the CM feature identifier 118 with the predicted features output from the MB feature identifier 112 to identify feature mismatches between the dominant features and the predicted features. In the event that a mismatch is found, for example, the predicted features are found to be not included in the dominant features or the dominant features are not included in the predicted features, the evaluator module can make changes to reduce or minimize the mismatch.

[0058] In some embodiments, the changes include editing the predictive features by, for example, eliminating additional predictive features that do not match among the dominant features, adding dominant features that do not match among the predictive features, or making changes to values ​​corresponding to the predictive features (e.g., age ranges) to match values ​​corresponding to the corresponding dominant features. The evaluator module 122 may retrieve the predictive features from the MB feature identifier 112 or from the storage device 126. The evaluator module 122 may transmit the modified predictive features to the storage device 126 to replace the previous predictive features. The evaluator module 122 may transmit the modified or approved predictive features to the MB estimator 114 for use in the likelihood estimation process.

[0059] In some embodiments, the changes (made by the evaluator module 122 after discovering a mismatch between the dominant feature and the predictive feature) include modifying a parameter of the predictive model so that the model provides a predictive feature that is also a dominant feature identified by 118. In some embodiments, the parameters of the predictive model are modified so that the predictive feature identified by the modified model matches the dominant feature.

[0060] In some embodiments, the evaluator module 122 evaluates the predictive feature based on all dominant features identified by the CB feature identifier 118. In some embodiments, the evaluator module 122 evaluates the predictive feature based on dominant features in one or more specific subsets of the dominant features identified by the CB feature identifier. For example, a specific subset may include the patient's age at T0, use of specific medications, and Mohs surgery. When evaluating the predictive feature age, the evaluator module 122 considers whether other dominant features included in the subset (i.e., use of specific medications and Mohs surgery) are also among the predictive features.

[0061] In some embodiments, the evaluator module 122 considers the dominant features included in each subset as relevant features, and also adds the correlations between the corresponding predictive features. In the above example, once the evaluator module 122 adds Mohs surgery and the use of specific medications to the predictive features, the evaluator module 122 can also assign the correlations between the three features of Mohs surgery, the use of specific medications, and age based on the corresponding dominant feature subsets.

[0062] The CB feature identifier 118 can identify one or more subsets of dominant features from a particular cluster. For example, assuming the second threshold number is 70%, if more than 70% of the patients in the cluster have all three features of male, HIV positive, and actinic keratosis, the CB feature identifier 118 can create a subset that includes these three dominant features. Moreover, if more than 70% of the patients in the same cluster have features of being older than 50 years old, using specific medications (e.g., topical antiproliferative immunomodulators), undergoing specific medical procedures (e.g., incisional biopsy of the ear, lip, or eyelid), and belonging to a specific race, the CB feature identifier 118 includes these four features in another dominant feature subset formed by the same specific prediction cluster.

[0063] Each dominant feature subset can be associated with a specific time point associated with the time when the corresponding feature in the subset was measured or observed. For example, a first subset can include features measured at a moment associated with time point T0: HIV positive and actinic keratosis, while a second subset can include the same features measured at a later moment associated with time point T1: HIV positive and actinic keratosis.

[0064] The CB feature identifier 118 can associate each leading feature subset with a corresponding likelihood of disease progression to an advanced stage. The CB feature identifier 118 associates the likelihood with the predictive feature subset by comparing how many patients in the group share the same value (or share values ​​within a particular range) for the feature for the group of patients that progressed to an advanced form of the disease and the group of patients that did not progress to an advanced form of the disease.

[0065] The CB feature identifier 118 may output information about the dominant feature subset for storage, for example, in the storage device 126 or an external data storage device, or for transmission for presentation to a user (e.g., via the external device 132). As described above, the CB feature identifier 118 may transmit the dominant feature subset to the evaluator module 122, which may verify the accuracy of the predictive features identified by the MB feature identifier 112. The evaluator module may

[0066] In some embodiments, each of the prediction model module 110 and the clustering module 116 transmits its output to an external device 132, for example, for presentation to a user or for further processing. The prediction model module 110 may transmit the calculated likelihood or the identified predictive features to an external device. The clustering module 116 may transmit a subset of the dominant features identified from the created clusters to an external device.

[0067] Although the above description provides a method for predicting the likelihood of a patient progressing to a predetermined late stage of a disease, a similar method can be taken to predict a patient's timeline of disease progression. For example, the patient's disease progression at time points T1 and T2 can be estimated based on data measured before moments t1 and t2 associated with T1 and T2 (e.g., at moment t0 (associated with time point T0)); and the patient's disease progression at time point T2 can be estimated based on data measured before moment t2 associated with time point T2 (e.g., at moments t0 and t1 (associated with time points T0 and T1)). The more time points selected by the time point mapper 106, the greater number of moments that can be used as a reference for the timeline when predicting the user's likelihood of disease progression.

[0068] The system 100 can recommend a drug or medical procedure for the patient based on one or more of the predicted probabilities. For example, based on the disease progression of a particular patient and the clinical characteristics of the patient, the system can recommend one or more drugs or medical procedures that were found to be more effective or commonly used in training patients 102a to 102n who have not progressed to an advanced stage of the disease.

[0069] Figure 2A and Figure 2B Depicted are example processes 200 and 220 that can be performed by a computing system (e.g., system 100) according to embodiments of the present disclosure. Process 200 can be performed to provide a set of predictive features for training a prediction model. Process 220 can be performed to predict the likelihood that a patient has or develops a particular disease or an advanced stage of a particular disease based on the predictive features.

[0070] The system receives input data (202) including a set of medical features associated with a group of patients, for example as training data. The system may receive the set of medical features by communicating with an external device. Each feature in the set of medical features is measured at multiple moments during the progression of a disease in a corresponding patient in the group of patients (e.g., 102a to 102n).

[0071] The system preprocesses the input data (204). The preprocessing includes mapping moments to corresponding time points and forming patient groups. The mapping includes associating each moment with a corresponding time point that represents the relativity of the moment to the patient's overall disease progression. The groups may include a late stage group and a non-late stage group. The late stage group may include patients whose disease has progressed to a predetermined late stage. The non-late stage group may include patients whose disease has not yet progressed to a predetermined late stage.

[0072] The system identifies one or more predictive features from the medical features (206), such as by applying a predictive model to the set of medical features. The predictive features distinguish the advanced cohort from the non-advanced cohort at various time points. To determine whether a particular feature in the set of medical features is a predictive feature for a particular disease, the system calculates the difference between the number of patients with the particular feature in the advanced (or first) cohort and the number of patients with the particular feature in the non-advanced (or second) cohort (208). The system then determines whether the difference in these numbers increases or decreases when moving from the first time point to the second time point. If the difference increases or decreases by more than a threshold value (210), the system identifies the particular feature as a predictive feature (212).

[0073] The module transmits the predicted features (214), such as to store the predicted features in a storage device or to present the predicted features on another device (e.g., Figure 1 On device 132).

[0074] exist Figure 2B In the process 220 depicted in FIG. 1 , the system receives medical characteristics of a patient ( 222 ), e.g., as test data. The system obtains information predicting the characteristics ( 224 ). The system then applies a predictive model to the medical characteristics of the patient to predict the likelihood that the patient will develop or have a particular disease or progress to an advanced form of the disease ( 226 ). The system transmits the likelihood ( 228 ), e.g., to be stored or presented on another device.

[0075] Figure 3is an example process that can be performed by a clustering module (e.g., 116) according to an embodiment of the present disclosure. The module receives a set of medical features (302), for example, as training data. The module can receive the set of medical features by communicating with an external device. Each feature in the set of medical features is measured at multiple moments (e.g., t0, t1, t2) during the disease progression of a corresponding patient in a group of patients (e.g., 102a to 102n).

[0076] The module preprocesses the input data (303). The preprocessing may include mapping moments to corresponding time points. The mapping includes associating each moment with a corresponding time point (e.g., T0, T1, T2) that represents the relativity of the moment to the patient's overall disease progression.

[0077] The module clusters patients in the group of patients, for example based on a first subset of features (304). The first subset of features may be measured or observed at a particular moment in time. Each cluster represents a corresponding subset of patients that share a predetermined similarity in the first subset of features.

[0078] The module identifies clusters of patients having a greater than a first threshold number of disease progression to a particular advanced stage (306). The module identifies one or more second feature subsets that are common among the patients in the identified clusters (308). The module sets the identified second feature subsets as dominant features associated with progression to an advanced stage (i.e., a predetermined advanced form) of the disease (310). The module may transmit the dominant features (312), e.g., for storage or presentation.

[0079] Figure 4 An example of a computing device 400 and an example of a mobile computing device that can be used to implement the techniques described herein are shown. For example, Figure 1 The system 100 in the embodiment of the present invention may be in the form of a computing device 400, a mobile computing device 450, or a combination thereof. The computing device 400 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The mobile computing device is intended to represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, and other similar computing devices. The components shown here, their connections and relationships, and their functions are exemplary only and are not intended to limit the embodiments of the present invention described and / or claimed in this document.

[0080] The computing device 400 includes a processor 402, a memory 404, a storage device 406, a high-speed interface 408 connected to the memory 404 and a plurality of high-speed expansion ports 410, and a low-speed interface 412 connected to a low-speed expansion port 414 and the storage device 406. Each of the processor 402, the memory 404, the storage device 406, the high-speed interface 408, the high-speed expansion port 410, and the low-speed interface 412 is interconnected using various buses and can be installed on a common motherboard or in other appropriate ways. The processor 402 can process instructions for execution within the computing device 400, including instructions stored in the memory 404 or on the storage device 406, to display graphical information of a GUI on an external input / output device (such as a display 416 coupled to the high-speed interface 408). In other embodiments, multiple processors and / or multiple buses, as well as multiple memories and multiple types of memories, can be used as needed. In addition, multiple computing devices can be connected, wherein each device provides part of the necessary operations (for example, as a server group, a group of blade servers, or a multi-processor system).

[0081] Memory 404 stores information within computing device 400. In some implementations, memory 404 is one or more volatile memory units. In some implementations, memory 404 is one or more non-volatile memory units. Memory 404 may also be other forms of computer-readable media, such as magnetic disks or optical disks.

[0082] Storage device 406 can provide mass storage for computing device 400. In some embodiments, storage device 406 can be or include a computer-readable medium, such as a floppy disk device, a hard disk device, an optical disk device, or a magnetic tape device, a flash memory or other similar solid-state memory device, or an array of devices, including devices in a storage area network or other configuration. A computer program product can be tangibly embodied in an information carrier. A computer program product can also include instructions that, when executed, perform one or more methods, such as the above-described methods. A computer program product can also be tangibly embodied in a computer or machine-readable medium, such as memory 404, storage device 406, or memory on processor 402.

[0083] The high-speed interface 408 manages bandwidth-intensive operations of the computing device 400, while the low-speed interface 412 manages less bandwidth-intensive operations. This allocation of functions is exemplary only. In some embodiments, the high-speed interface 408 is coupled to the memory 404, the display 416 (e.g., through a graphics processor or accelerator), and to a high-speed expansion port 410 that can accept various expansion cards (not shown). In an embodiment, the low-speed interface 412 is coupled to the storage device 406 and the low-speed expansion port 414. The low-speed expansion port 414 (which may include various communication ports (e.g., USB, Bluetooth, Ethernet, wireless Ethernet)) can be coupled to one or more input / output devices, such as a keyboard, a pointing device, a scanner, or a network device (such as a switch or router), for example, through a network adapter.

[0084] Computing device 400 can be implemented in a variety of different forms, as shown. For example, it can be implemented as a standard server 420, or multiple times in a group of such servers. In addition, it can also be implemented in a personal computer (such as laptop computer 422). It can also be implemented as part of a rack server system 424. Alternatively, the components in computing device 400 can be combined with other components in a mobile device (not shown) (such as mobile computing device 450). Each such device can contain one or more of computing device 400 and mobile computing device 450, and the entire system can be composed of multiple computing devices that communicate with each other.

[0085] The mobile computing device 450 includes a processor 452, a memory 464, an input / output device (such as a display 454), a communication interface 466, and a transceiver 468, among other components. The mobile computing device 450 may also be equipped with a storage device, such as a micro drive or other device, to provide additional storage. Each of the processor 452, the memory 464, the display 454, the communication interface 466, and the transceiver 468 are interconnected using various buses, and several of the components may be mounted on a common motherboard or in other suitable ways.

[0086] Processor 452 can execute instructions within mobile computing device 450, including instructions stored in memory 464. Processor 452 can be implemented as a chipset including separate multiple analog and digital processors. Processor 452 can provide, for example, coordination of other components of mobile computing device 450, such as control of a patient interface, applications running on mobile computing device 450, and wireless communications of mobile computing device 450.

[0087] The processor 452 can communicate with the patient through a control interface 458 and a display interface 456 coupled to a display 454. The display 454 can be, for example, a TFT (thin film transistor liquid crystal display) display or an OLED (organic light emitting diode) display, or other suitable display technology. The display interface 456 may include appropriate circuitry for driving the display 454 to present graphical information and other information to the patient. The control interface 458 can receive commands from the patient and convert them for submission to the processor 452. In addition, an external interface 462 can provide communication with the processor 452 to enable near-area communication of the mobile computing device 450 with other devices. The external interface 462 can provide, for example, wired communication in some embodiments, or wireless communication in other embodiments, and multiple interfaces can also be used.

[0088] The memory 464 stores information within the mobile computing device 450. The memory 464 may be implemented as one or more computer-readable media, one or more volatile memory units, or one or more non-volatile memory units. An expansion memory 474 may also be provided and connected to the mobile computing device 450 via an expansion interface 472, which may include, for example, a SIMM (single in-line memory module) card interface. The expansion memory 474 may provide additional storage space for the mobile computing device 450, and may also store applications or other information for the mobile computing device 450. Specifically, the expansion memory 474 may include instructions for executing or supplementing the above-mentioned processes, and may also include security information. Thus, for example, the expansion memory 474 may be provided as a security module for the mobile computing device 450, and may be programmed with instructions that allow the secure use of the mobile computing device 450. In addition, secure applications and other information may also be provided via a SIMM card, such as placing identification information on the SIMM card in an unbreakable manner.

[0089] The memory may include, for example, flash memory and / or NVRAM memory (non-volatile random access memory), as discussed below. In some embodiments, a computer program product is tangibly embodied in an information carrier. The computer program product includes instructions that, when executed, perform one or more methods, such as the methods described above. The computer program product may be a computer or machine readable medium, such as memory 464, expansion memory 474, or memory on processor 452. In some embodiments, the computer program product may be received, for example, via transceiver 468 or external interface 462 in the form of a propagated signal.

[0090] The mobile computing device 450 can communicate wirelessly via a communication interface 466, which may include digital signal processing circuitry when necessary. The communication interface 466 can provide communications in various modes or protocols, such as GSM voice calls (Global System for Mobile Communications), SMS (Short Message Service), EMS (Enhanced Short Message Service), or MMS messages (Multimedia Short Message Service), CDMA (Code Division Multiple Access), TDMA (Time Division Multiple Access), PDC (Personal Digital Cellular), WCDMA (Wideband Code Division Multiple Access), CDMA2000, or GPRS (General Packet Radio Service), etc. Such communications can be performed, for example, using radio frequencies via a transceiver 468. In addition, short-range communications can also be performed, such as using Bluetooth, WiFi, or other such transceivers (not shown). In addition, a GPS (Global Positioning System) receiver module 470 can provide additional wireless data related to navigation and location to the mobile computing device 450, which can be used appropriately by applications running on the mobile computing device 450.

[0091] Mobile computing device 450 may also communicate audio using audio codec 460, which may receive spoken information from the patient and convert it into usable digital information. Audio codec 460 may also generate audible sounds for the patient, such as through a speaker (e.g., in a receiver of mobile computing device 450). Such sounds may include sounds from voice phone calls, may include recorded sounds (e.g., voice messages, music files, etc.), and may also include sounds generated by applications running on mobile computing device 450.

[0092] Mobile computing device 450 can be implemented in many different forms, as shown. For example, it can be implemented as a cellular phone 480. It can also be implemented as part of a smart phone 482, a personal digital assistant, or other similar mobile devices.

[0093] Various implementations of the systems and techniques described herein can be implemented in digital electronic circuitry, integrated circuitry, specially designed ASICs (application specific integrated circuits), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system that includes at least one programmable processor that can be special purpose or general purpose, coupled to receive data and instructions from a storage system, at least one input device, and at least one output device, and to transmit data and instructions to the storage system, at least one input device, and at least one output device.

[0094] These computer programs (also referred to as programs, software, software applications, or code) include machine instructions for a programmable processor and may be implemented in high-level procedural and / or object-oriented programming languages ​​and / or assembly / machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and / or device (e.g., magnetic disks, optical disks, memories, programmable logic devices (PLDs)) for providing machine instructions and / or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0095] To provide interaction with a patient, the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube), or LCD (liquid crystal display) monitor) for displaying information to the patient, and a keyboard and pointing device (e.g., a mouse or trackball), wherein the patient can provide input to the computer via the keyboard and pointing device. Other types of devices can also be used to provide interaction with the patient; for example, the feedback provided to the patient can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the patient can be received in any form, including sound, voice, or tactile input.

[0096] The systems and techniques described herein can be implemented in a computing system that includes a back-end component (e.g., as a data server) or includes a middleware component (e.g., an application server), or includes a front-end component (e.g., a client computer with a graphical patient interface or a web browser through which a patient can interact with an implementation of the systems and techniques described herein), or any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any digital data communication form or medium (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0097] A computing system may include clients and servers. Clients and servers are generally remote from each other and generally interact through a communication network. The relationship of client and server arises from computer programs running on the respective computers and having a client-server relationship to each other.

[0098] In some embodiments, a method is performed by a system having one or more computers. The method includes: receiving a set of medical features, each feature in the set of medical features is measured at multiple moments during the progression of a disease in a corresponding patient in a set of patients; mapping each moment to a corresponding time point representing the relativity of the moment to the overall disease progression of the patient; forming an advanced group and a non-advanced group from the set of patients, the advanced group including patients whose disease progresses to a predetermined advanced form, and the non-advanced group including patients whose disease does not progress to the predetermined advanced form; applying a predictive model to the set of medical features to identify one or more predictive features that distinguish the advanced group from the non-advanced group at various time points; and storing information of the one or more predictive features in a data storage device.

[0099] The method may further include: receiving data about medical characteristics of a particular patient from a client device; retrieving information about the one or more predictive characteristics from the storage device; applying the predictive model to the data to predict the likelihood of the particular patient progressing to an advanced stage of the disease based on the one or more predictive characteristics; and transmitting the likelihood to the client device for presentation. The method may also include: applying the predictive model to the data to predict a timeline for the progression of the disease in the particular patient; and transmitting the timeline to the client device for presentation. The method may include: recommending a medication or medical procedure for the patient based on the likelihood.

[0100] The disease progression process for a patient may start from a predetermined time period prior to the initial diagnosis of the patient's disease. The disease progression process for a first patient in the advanced stage cohort may end at the time the patient is diagnosed as having an advanced stage of the disease. The disease progression process for a second patient in the non-advanced stage cohort may end at a time calculated based on the median duration of the process for the patients in the advanced stage cohort.

[0101] The at least one medical characteristic may include clinical characteristics of the patients, including one or more of the following: the patient's age, gender, geographic origin, medical diagnosis, prescription information, medical procedures, biomarker information, body mass index, smoking and drinking habits, and laboratory test results.

[0102] The method may further include: clustering patients in the group of patients based on a first feature subset measured at a particular time point, each cluster representing a corresponding subset of patients sharing a predetermined similarity in the first feature subset; identifying a cluster of patients having a greater than a first threshold number of patients that have progressed to the predetermined advanced form; identifying a second feature subset shared between more than a second threshold number of patients in the identified cluster; and associating the second feature subset with the predetermined advanced form of the disease. The method may also include: comparing the second feature subset with the one or more predictive features to identify feature mismatches between the second feature subset and the predictive features, and in response: modifying the predictive model to reduce the feature mismatches. The predictive model may be modified so that the one or more predictive features match the second feature subset.

[0103] The predictive model may identify the predictive features by applying a tree-based approach to corresponding values ​​of the medical features in the set of medical features so as to iteratively distinguish patients in the advanced stage group from patients in the non-advanced stage group at various time points based on the corresponding values ​​of the medical features.

[0104] The predictive model may identify a specific feature as a predictive feature in the following manner: determining a first number of patients having the specific feature in the non-late stage group and a second number of patients having the specific feature in the late stage group at each time point among the multiple time points; for each time point, calculating a corresponding Δ value between the first number and the second number at the time point; and determining that the first Δ value calculated for the first time point is greater than the second Δ value calculated for the second time point by more than a specific threshold, and in response: identifying the specific feature as a predictive feature.

[0105] A number of embodiments of the present disclosure have been described. However, it should be understood that various modifications may be made without departing from the spirit and scope of the present disclosure. Therefore, other embodiments are within the scope of the claims.

Claims

1. A method comprising: Receiving, by a system having one or more computers, a set of medical features, each feature in the set of medical features measured at a plurality of moments throughout the progression of a disease in a corresponding patient in a group of patients; mapping each moment to a corresponding time point representing the relativity of that moment to the overall disease progression of that patient; forming an advanced cohort and a non-advanced cohort from the group of patients, the advanced cohort comprising patients whose disease progressed to a predetermined advanced form and the non-advanced cohort comprising patients whose disease did not progress to the predetermined advanced form; applying a predictive model to the set of medical features to identify one or more predictive features that distinguish the advanced cohort from the non-advanced cohort at various time points; as well as Information about the one or more predicted features is stored in a data storage device.

2. The method of claim 1, further comprising: receiving data regarding a medical characteristic of a particular patient from a client device; Retrieving information of the one or more prediction features from the storage device; applying the predictive model to the data to predict the likelihood of the particular patient progressing to an advanced stage of the disease based on information about the one or more predictive features; as well as The likelihood is transmitted to the client device for presentation.

3. The method of claim 2, further comprising: applying the predictive model to the data to predict a timeline for progression of the disease in the particular patient; as well as The timeline is transmitted to the client device for presentation.

4. The method of any one of claims 1 to 3, further comprising recommending a medication or medical procedure for the patient based on the likelihood.

5. The method according to any one of claims 1 to 4, wherein: The disease progression process for a patient begins at a predetermined time period prior to the initial diagnosis of the patient's disease.

6. The method according to any one of claims 1 to 5, wherein: the course of disease progression for the first patient in the advanced stage cohort ends at the time the patient is diagnosed as having an advanced stage of the disease, and Therein, the disease progression course of the second patient in the non-advanced group ends at a time calculated based on the median duration of the course of the patients in the advanced group.

7. The method according to any one of claims 1 to 6, wherein: The at least one medical characteristic includes clinical characteristics of the patients, including one or more of the following: the patient's age, gender, geographic origin, medical diagnosis, prescription information, medical procedures, biomarker information, body mass index, smoking and drinking habits, and laboratory test results.

8. The method according to any one of claims 1 to 7, further comprising: clustering the patients in the group of patients based on a first subset of features measured at a particular time point, each cluster representing a corresponding subset of patients sharing a predetermined similarity in the first subset of features; identifying a cluster of patients having greater than a first threshold number of disease progressions to the predetermined advanced form; identifying a second subset of features that are common among more than a second threshold number of patients in the identified cluster; as well as The second subset of features is associated with being a predetermined advanced form of the disease.

9. The method of claim 8, further comprising comparing the second feature subset to the one or more predictive features to identify a feature mismatch between the second feature subset and the predictive features, and in response, The prediction model was modified to reduce these feature mismatches.

10. The method of claim 9, wherein: The predictive model is modified so that the one or more predictive features match the second subset of features.

11. The method according to any one of claims 1 to 10, wherein: The predictive model identifies the predictive features by applying a tree-based approach to corresponding values ​​of the medical features in the set of medical features to iteratively distinguish patients in the advanced stage group from patients in the non-advanced stage group at various time points based on the corresponding values ​​of the medical features.

12. The method according to any one of claims 1 to 10, wherein: The predictive model identifies specific features as predictive features in the following ways: determining, at each time point in the plurality of time points, a first number of patients in the non-advanced group having the particular characteristic and a second number of patients in the advanced group having the particular characteristic; For each time point, calculating a corresponding Δ value between the first quantity and the second quantity at the time point; as well as determining that the first delta value calculated for the first time point is greater than the second delta value calculated for the second time point by more than a certain threshold, and in response, This particular feature is identified as a predictive feature.

13. A system comprising: One or more computers, and one or more storage devices storing operable instructions which, when executed by the one or more computers, cause the one or more computers to perform the method of any one of claims 1 to 12.

14. A computer storage medium encoded with a computer program comprising operative instructions which, when executed by a data processing apparatus, cause the data processing apparatus to perform the method of any one of claims 1 to 12.

Images