Method and device for predicting lifetime of CRLM patient, medium and program product

Abstract

The invention provides a method, a system, equipment, a medium and a program product for predicting the lifetime of a CRLM patient, further discloses a method, a system, equipment, a medium and a program product for predicting the RAS gene mutation risk of the CRLM patient, and relates to the field of intelligent medical treatment. The vein phase image and the arterial phase image are respectively input into a classifier to calculate DL scores of different periods of a subject, the DL scores of different periods are integrated, and the function of predicting the lifetime is realized based on an integration result. Meanwhile, the RAS gene state is evaluated based on the classifier for calculating the DL score, the survival rate and the lifetime of the subject are predicted based on the RAS gene state, targeted drugs suitable for being treated are recognized, and patients most likely to benefit from cetuximab or bevacizumab treatment are recognized.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical treatment, and more specifically, to a method, device, medium and program product for predicting the survival period of CRLM patients. Background Art

[0002] Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer-related death worldwide. Previous studies have shown that approximately 25-30% of patients diagnosed with CRC develop liver metastasis (LM) during the course of the disease, and LM significantly affects CRC prognosis.

[0003] The standard treatment for unresectable colorectal liver metastases (CRLM) usually involves chemotherapy. Different patients need to choose appropriate drug treatment. For individual patients, finding appropriate early diagnosis and treatment and prognostic evaluation markers is very important to improve the survival of CRLM patients. It is even more important to predict the survival of CRLM patients after treatment through reasonable methods. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a method, device, medium and program product for predicting the survival period of CRLM patients; the method of the present invention calculates the DL scores of the subjects at different stages by inputting the venous phase image and the arterial phase image into a classifier respectively, integrates the DL scores of the different stages, and realizes the function of predicting the survival period based on the integrated result.

[0005] The first aspect of the present application discloses a method for predicting the survival period of CRLM patients, the method comprising:

[0006] S101: Acquire arterial phase and venous phase images of the subject including the target area;

[0007] S102: inputting the arterial phase image into a classifier to calculate the arterial phase DL score, inputting the venous phase image into a classifier to calculate the venous phase DL score, and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score;

[0008] S103: Predicting the survival period of the subject based on the total DL score.

[0009] In some embodiments, the method further comprises: obtaining a CEA value and / or a DCR value of the subject; predicting the survival period of the subject based on any one or more of the DL score, CEA value, and DCR value;

[0010] Optionally, when predicting the survival of a subject based on the DL score, CEA value and DCR value, the survival of the subject is predicted by calculating the total score of the DL score, CEA value and DCR value.

[0011] In some embodiments, when the total score is calculated based on the DL score, CEA value, and DCR value, the calculation method of the total score includes: Y=2.995×10-5×CEA+(-1.581×DCR)+(-6.483×10-1×DL), where Y is the total score.

[0012] In some embodiments, between S101 and S102, the method further includes: segmenting an arterial phase ROI region from the arterial phase image, and segmenting a venous phase ROI region from the venous phase image; calculating an arterial phase DL score based on the arterial phase ROI region, and calculating a venous phase DL score based on the venous phase ROI region; and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score.

[0013] The second aspect of the present application discloses a method for predicting the risk of RAS gene mutation in CRLM patients, the method comprising:

[0014] S201: Acquire CECT images of a subject; the CECT images include venous phase images and arterial phase images;

[0015] S202: Inputting the CECT image into a classifier to obtain a predicted RAS gene mutation risk.

[0016] The method for constructing the classifier comprises:

[0017] S2021: Obtain training set samples including images of the arterial phase (AP) and the venous phase (VP) and RAS gene status labels corresponding to the samples; the RAS gene status labels include mutation and wild type;

[0018] S2022: inputting the arterial phase image into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining an arterial phase classifier based on the arterial phase image; inputting the venous phase image into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a venous phase classifier based on the venous phase image;

[0019] S2023: Using an integrated classifier to combine the arterial phase classifier and the venous phase classifier to obtain the classifier.

[0020] In some embodiments, the method further comprises: predicting the survival rate of the subject based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, obtaining a result that the subject's survival rate is high; if the RAS gene mutation risk is high risk, obtaining a result that the subject's survival rate is low;

[0021] Optionally, the survival period of the subject is predicted based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, a result indicating that the subject has a long survival period is obtained; if the RAS gene mutation risk is high risk, a result indicating that the subject has a short survival period is obtained;

[0022] Optionally, the survival rate or survival period includes any one or more of the following years: 1 year, 3 years, 5 years.

[0023] In some embodiments, the method further comprises: identifying the type of targeted drug suitable for the subject based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, the targeted drug suitable for the subject is cetuximab; if the RAS gene mutation risk is high risk, the targeted drug suitable for the subject is bevacizumab;

[0024] Optionally, between S2021 and S2022, the construction method also includes: segmenting an arterial phase ROI area from the arterial phase image, and segmenting a venous phase ROI area from the venous phase image; inputting the arterial phase ROI area into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining an arterial phase classifier based on the arterial phase image; inputting the venous phase ROI area into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a venous phase classifier based on the venous phase image.

[0025] The third aspect of the present application discloses a computer device, which includes: a memory and a processor; the memory is used to store a computer program; and the processor executes the computer program to implement the steps of the above method.

[0026] A fourth aspect of the present application discloses a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0027] A fifth aspect of the present application discloses a computer program product, including a computer program, which implements the steps of the above method when executed by a processor.

[0028] This application has the following beneficial effects:

[0029] 1. This application innovatively discloses a method for predicting the survival of CRLM patients. This method calculates and integrates the DL scores of subjects at different stages based on venous phase images and arterial phase images. The DL score is used as an independent prognostic factor for patients with colorectal liver metastasis to predict the survival of the subjects. The calculation of the DL score is based on a prediction classifier (DL-AP+VP) constructed based on venous phase and arterial phase images focusing on the liver metastasis area. Furthermore, this application found three independent predictive factors in the construction of the prediction classifier. Based on any one or more of the three factors, the prognosis of patients with colorectal liver metastasis can be predicted, and a specific calculation formula is obtained when the prognosis prediction effect is achieved based on the three.

[0030] 2. This application innovatively uses CECT images of LM in CRLM patients to develop a DL model to predict RAS mutation status, realizing a non-invasive, accurate and repeatable method to evaluate RAS mutation status and improve the accuracy of targeted treatment decisions; it is based on the accurate and rapid quantification of RAS mutation status to predict the survival rate and survival period of the subjects, identify targeted drugs suitable for treatment, and identify patients most likely to benefit from cetuximab or bevacizumab treatment.

[0031] This study aimed to address this gap by developing and validating a DL radiomics (DLR) model based on LM contrast-enhanced computed tomography (CECT) images to assess RAS gene status. Our results are expected to facilitate precise, noninvasive treatment adjustments in the management of CRLM. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 is a schematic diagram of a method flow chart provided by the first aspect of an embodiment of the present invention;

[0034] Figure 2 is a schematic diagram of a system for predicting the survival period of CRLM patients provided by an embodiment of the present invention;

[0035] Figure 3 is a schematic diagram of a computer device provided by an embodiment of the present invention;

[0036] Figure 4 is a schematic diagram of the architecture of an exemplary computing device provided by an embodiment of the present invention;

[0037] Figure 5is a schematic diagram of a storage medium provided by an embodiment of the present invention;

[0038] Figure 6 is a schematic diagram of a method flow chart provided by the second aspect of an embodiment of the present invention;

[0039] Figure 7 is a schematic diagram of a system for predicting the risk of RAS gene mutation in CRLM patients provided by an embodiment of the present invention;

[0040] Figure 8 is a patient inclusion flow chart provided by an embodiment of the present invention;

[0041] Fig. 9 is a research flow chart provided by an embodiment of the present invention; Fig. 9 A is data preprocessing, Fig. 9 B is the DL model building process, Fig. 9 C is risk stratification, and data preprocessing mainly relies on manual selection of ROI and image preprocessing. Model construction consists of two different deep learning model structures (arterial phase and venous phase) and model integration methods.

[0042] Fig.10 It is a ROC curve result diagram of 5 models in the training cohort, internal test cohort and external test cohort provided by the embodiment of the present invention; Fig.10 A is the ROC curve of the training cohort. Fig.10 B is the ROC curve of the internal test cohort. Fig.10 C is the ROC curve plot of the external testing cohort.

[0043] Fig.11 It is a nomogram for predicting survival rate using three independent predictive factors (CEA, DCR, DL score) provided in an embodiment of the present invention; Fig.11 A is the nomogram for prediction by three independent predictors. Fig.11 BD shows the TC, ITC, and ETC calibration curves to evaluate the performance of this DL nomogram. Fig.11 EG shows the evaluation of the model's survival advantage by calculating the net benefit at different risk threshold probabilities.

[0044] Fig.12 is a schematic diagram of comparing survival rates between two RAS mutation states in a training cohort, an internal test cohort, and an external test cohort provided by an embodiment of the present invention; Fig.12 A shows that the cumulative 1-, 3-, and 5-year overall survival rates of the low-risk RAS mutation status group were significantly higher than those of the high-risk group; Fig.12 B shows that the cumulative 1-, 3-, and 5-year OS rates in the low-risk RAS mutation status group of ITC were significantly higher than those in the RAS mutation status group; Fig.12C shows that the cumulative 1-, 3-, and 5-year OS rates in the ETC low-risk RAS mutation status group were significantly higher than those in the RAS mutation status group;

[0045] Fig.13 The PFS data of patients with different RAS mutation risk levels treated with cetuximab or bevacizumab provided in the embodiments of the present invention; Fig.13 A shows that among patients with low-risk RAS mutations, patients who used cetuximab had a longer PFS than those who did not use cetuximab. Fig.13 B shows that in the RAS high-risk mutation group, the median PFS in the bevacizumab group was significantly longer than that in the non-bevacizumab group. DETAILED DESCRIPTION

[0046] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0047] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0049] Figure 1 The present invention provides a method for predicting the survival of CRLM patients. Specifically, the method comprises the following steps:

[0050] S101: Acquire arterial phase and venous phase images of the subject including the target area;

[0051] In some embodiments, the term "subject" or "test subject" or "test sample" used herein refers to any animal (e.g., mammal), including but not limited to humans, non-human primates, rodents, etc., which will be the recipient of a specific treatment. Generally, the terms "subject" and "patient" are used interchangeably herein when referring to human subjects. Preferably, the subject is a human.

[0052] S102: inputting the arterial phase image into a classifier to calculate the arterial phase DL score, inputting the venous phase image into a classifier to calculate the venous phase DL score, and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score;

[0053] In some embodiments, the overall DL score is obtained based on the constructed classifier (DL-AP+VP).

[0054] S103: predicting the survival period of the subject based on the total DL score;

[0055] In some embodiments, the method further comprises: obtaining a CEA value and / or a DCR value of the subject; predicting the survival period of the subject based on any one or more of the DL score, CEA value, and DCR value;

[0056] Optionally, when predicting the survival of a subject based on the DL score, CEA value and DCR value, the survival of the subject is predicted by calculating the total score of the DL score, CEA value and DCR value.

[0057] In some embodiments, when the total score is calculated based on the DL score, CEA value, and DCR value, the total score is calculated by: Y=2.995×10-5×CEA+(-1.581×DCR)+(-6.483×10-1×DL), where Y is the total score, CEA is carcinoembryonic antigen, and DCR is disease control rate.

[0058] In some embodiments, between S101 and S102, the method further includes: segmenting an arterial phase ROI region from the arterial phase image, and segmenting a venous phase ROI region from the venous phase image; calculating an arterial phase DL score based on the arterial phase ROI region, and calculating a venous phase DL score based on the venous phase ROI region; and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score.

[0059] The second aspect of the present application discloses a method for predicting the risk of RAS gene mutation in CRLM patients, the method comprising:

[0060] S201: Acquire CECT images of a subject; the CECT images include venous phase images and arterial phase images;

[0061] S202: Inputting the CECT image into a classifier to obtain a predicted RAS gene mutation risk.

[0062] The method for constructing the classifier comprises:

[0063] S2021: Obtain training set samples including images of the arterial phase (AP) and the venous phase (VP) and RAS gene status labels corresponding to the samples; the RAS gene status labels include mutation and wild type;

[0064] In some embodiments, between S2021 and S2022, the construction method further includes: segmenting an arterial phase ROI area from the arterial phase image, and segmenting a venous phase ROI area from the venous phase image; inputting the arterial phase ROI area into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining an arterial phase classifier based on the arterial phase image; inputting the venous phase ROI area into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a venous phase classifier based on the venous phase image.

[0065] S2022: inputting the arterial phase image into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining an arterial phase classifier based on the arterial phase image; inputting the venous phase image into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a venous phase classifier based on the venous phase image;

[0066] S2023: Using an integrated classifier to combine the arterial phase classifier and the venous phase classifier to obtain the classifier.

[0067] In some embodiments, the voting separator - voting classifier, also known as an integrated classifier, is a technique in machine learning that combines multiple classification models to improve prediction capabilities. Specifically, this method involves the following steps: 1) Individual learner training: First, multiple different classification models are trained independently. These models can be of the same type (such as several logistic regression models) or different types (such as decision trees, support vector machines, random forests, etc.); 2) Prediction aggregation: For each sample to be predicted, all individual learners will generate a prediction result; 3) Voting decision: Majority Voting: Select the most common category as the final prediction result; Weighted Voting: Give different weights according to the performance of each learner (such as accuracy), and determine the prediction category after weighted summation; Threshold voting: Set a threshold, and only when the prediction of a certain category exceeds the threshold, it will be determined as the final category.

[0068] In some embodiments, a total of 185 eligible CRC patients were identified, with a mean age of 64.3±11.2 years, of which 132 were male. They were randomly divided into a training cohort (TC, n=88), an internal test cohort (ITC, n=39), and an external test cohort (ETC, n=58) in a 7:3 ratio. There was no significant difference in the proportion of RAS wild-type between the different cohorts (TC: 51.1%, ITC: 61.5%, ETC: 62.1%; P=0.278). In TC, RAS wild-type patients showed shallower primary tumor infiltration and larger LM diameter compared with RAS mutant patients (P=0.044 and 0.026, respectively). 35 RAS wild-type patients received cetuximab treatment, and 28 patients with RAS mutation status received bevacizumab treatment.

[0069] DLR feature extraction for predicting RAS mutations

[0070] A total of 1,126 radiomic features and 2,589 DL features were extracted from each ROI in the CECT stage. Fourteen significant radiomic features associated with RAS mutation status were selected, and the XGBoost model showed superior performance in developing R scores compared to other classifiers. In addition, DL signatures from AP and VP ROIs were input into FCNN to indicate the probability of RAS mutation status. The research process of the DLR model for predicting RAS mutation status is shown in Figure 2. Fig. 9 shown.

[0071] Performance comparison of RAS mutation prediction models

[0072] This study involved the establishment of five models: radiomics model, AP-DL model, VP-DL model, AP+VP-DL model, and DL-R model. The performance of all models was then examined in three cohorts ( Fig.10 The AP+VP-DL model based on the two-phase DL features had the best discrimination (AUC, 0.98) compared with the radiomics model (AUC, 0.90), AP-DL model (AUC, 0.93), and VP-DL model (AUC, 0.87). The DL-R model (AUC, 0.97) was used to predict the RAS mutation status in ETC.

[0073] The area under the curve (AUC), sensitivity (SEN), specificity (SPE), positive predictive value (PPV), and negative predictive value (NPV) of the five models are shown in Table 1. In terms of quantitative indicators, such as AUC and PPV, the AP+VP-DL model is generally superior to other prediction models. Multivariate analysis further confirmed that CEA, DCR, and AP+VP-DL scores were significantly associated with OS (Table 2). Therefore, we selected the DL model as the risk stratification of RAS mutation status before treatment.

[0074] The method for constructing the DL-AP model includes: obtaining the arterial phase CECT images of the training set samples and the RAS gene status labels corresponding to the samples; the RAS gene status labels include high risk and low risk; segmenting the arterial phase ROI region from the arterial phase images; inputting the arterial phase ROI region into the machine learning model for training to obtain the predicted classification results, comparing the predicted classification results with the classification labels, optimizing the model according to the comparison results, and obtaining the DL-AP model based on the arterial phase images;

[0075] The method for constructing a DL-VP model includes: obtaining a venous phase CECT image of a training set sample and a RAS gene status label corresponding to the sample; the RAS gene status label includes high risk and low risk; segmenting a venous phase ROI region from the venous phase image; inputting the venous phase ROI region into a machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a DL-VP model based on the venous phase image;

[0076] The method for constructing a radiomics model includes: obtaining CECT images of training set samples; segmenting a ROI region from the CECT image, and extracting radiomics features based on the ROI region; inputting the radiomics features into a machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a radiomics model based on radiomics features; radiomics is the application of computer image processing to convert image data of a region of interest (ROI) into high-dimensional feature data that can be mined.

[0077] The method for constructing the DL-R model includes: combining the radiomics model and the DL-AP+VP model to obtain the DL-R model.

[0078] DL nomogram for predicting survival in CRLM

[0079] According to multivariate Cox regression analysis, CEA, DCR (disease control rate), and AP+VP-DL score (a composite score of the arterial and venous phases that may be derived from imaging data) were significantly associated with OS in patients with CRLM (Table 2). This study identified three independent predictors of OS in patients with CRLM: CEA (HR: 1.185; 95% CI: 1.028-3.645; P = 0.004), DCR (HR: 0.177; 95% CI: 0.039-0.807; P = 0.025), and DL score (high-risk RAS mutation status vs. low-risk RAS mutation status; HR: 5.252; 95% CI: 2.174-7.549; P < 0.001). These predictors were used to develop the DL nomogram, as shown in Table 2. Fig.11 A. Independently identified risk factors were incorporated into the nomogram using the following formula: Y = 2.995 × 10-5 × CEA + (-1.581 × DCR) + (-6.483 × 10-1 × DL) score. Fig.11 The performance of this DL nomogram was evaluated by the calibration curves of TC, ITC, and ETC shown in BD. These calibration curves showed strong agreement between the predicted 5-year OS probability and the actual 5-year OS probability in patients with CRLM, indicating that the DL nomogram accurately estimated the survival outcomes.

[0080] Table 1. Comparison of the performance of all models for RAS status prediction

[0081]

[0082] Numbers are 95% confidence intervals. Bold values ​​represent optimal AUC values.

[0083] AUC, area under; ACC, accuracy; NPV, negative predictive value; PPV, positive predictive value;

[0084] TC, training cohort; ITC, internal test cohort; ETC, external test cohort. DL-AP, deep learning in the arterial phase; DL-VP, deep learning in the venous phase; DL: deep learning in both the arterial and venous phases; DL-R, deep learning radiomics.

[0085] Table 2 Factors associated with OS in univariate and multivariate Cox regression analysis

[0086]

[0087]

[0088] Abbreviations: Ci, confidence interval; kps, Karnofsky performance status; DCR, disease control rate; CEA, carcinoembryonic antigen; WBC, white blood cell; LY, lymphocyte; LDH, lactate dehydrogenase; OS, overall survival.

[0089] A decision curve analysis (DCA) was performed to evaluate the survival advantage of the model by calculating the net benefit at different risk threshold probabilities. The net benefit was determined by subtracting the rate of false positives from true positives, taking into account the proportional hazards of necessary treatment relative to the negative effects of unnecessary treatment. The DCA showed that the nomogram provided a significant advantage ( Fig.11 The consistency index (CI) of the Cox model (Table 3) showed that the DL model performed better than the DL-C model in ETC (CI: 0.5543; 95% CI: 0.4546-0.6540; P = 0.0016).

[0090] Table 3 Cox model consistency index (CI)

[0091]

[0092] Deep learning model for identifying survival benefits

[0093] The cumulative 1-year, 3-year, and 5-year overall survival rates of the low-risk RAS mutation status group were 99%, 36%, and 13%, respectively, which were significantly higher than those of the high-risk group (100%, 29%, and 5%). The high-risk RAS mutation status group in TC (log-rank 153 test, P = 0.025) ( Fig.12A). Similarly, the cumulative 1-year, 3-year, and 5-year OS rates of the low-risk RAS mutation status group were 91%, 43%, and 13%, respectively, which were significantly higher than those of the high-risk RAS mutation status group predicted in ITC (log-rank test, P = 0.0021) ( Fig.12 B). In addition, the cumulative 1-year, 3-year, and 5-year OS rates of the low-risk RAS mutation status group were 96%, 35%, and 15%, respectively, which were significantly higher than those of the high-risk RAS mutation status group predicted in ETC (log-rank test, P = 0.03) ( Fig.12 C).

[0094] Prediction of response to targeted therapy

[0095] exist Fig.13 In the study, PFS data of patients with different RAS mutation risk levels treated with cetuximab or bevacizumab were presented. Among patients with low-risk RAS mutations, patients who received cetuximab had a longer PFS than those who did not receive cetuximab ((19 vs 9 months, 95% CI: 13-, 8-12; P < 0.001; Fig.13 A)), in the RAS high-risk mutation group, the median PFS of the bevacizumab group was 11 months ((95%CI: 8-13 months)), which was significantly longer than the 6 months ((95%CI, 5-13 months)) of the non-bevacizumab group (P=0.042) ( Fig.13 B).

[0096] In some embodiments, the method further comprises: predicting the survival rate of the subject based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, obtaining a result that the subject's survival rate is high; if the RAS gene mutation risk is high risk, obtaining a result that the subject's survival rate is low;

[0097] Optionally, the survival period of the subject is predicted based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, a result indicating that the subject has a long survival period is obtained; if the RAS gene mutation risk is high risk, a result indicating that the subject has a short survival period is obtained;

[0098] Optionally, the survival rate or survival period includes any one or more of the following years: 1 year, 3 years, 5 years.

[0099] In some embodiments, the method further comprises: identifying the type of targeted drug suitable for the subject based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, the targeted drug suitable for the subject is cetuximab; if the RAS gene mutation risk is high risk, the targeted drug suitable for the subject is bevacizumab;

[0100] In patients with colorectal cancer, the mutation status of the RAS gene has a crucial impact on the treatment options for patients with advanced colorectal cancer. In patients with advanced colorectal cancer, clarifying the mutation status of the RAS gene is a "must-have" for subsequent treatment. Currently, the detection of the mutation status of the RAS gene usually requires the analysis of tissue specimens obtained by surgery or biopsy, which is not only invasive but also expensive. Especially for patients whose tissue specimens are old and cannot undergo puncture biopsy, non-invasive detection of the RAS gene status through imaging omics is particularly important. In addition, RAS mutations are associated with poor prognosis and resistance to anti-EGFR monoclonal antibodies. The recent CRICKET and CAVE trials highlighted the efficacy of a cetuximab-based rechallenge strategy in the treatment of third-line RAS wild-type metastatic CRC. Circulating tumor DNA (ctDNA) analysis offers hope for guiding treatment decisions, especially for RAS wild-type metastatic CRC. However, the sensitivity of ctDNA NGS detection remains challenging, raising concerns about undetected mutations. Therefore, a noninvasive and accurate method for assessing RAS genetic status is essential to personalize CRLM treatment and optimize outcomes.

[0101] In the present invention, the prognostic indicators include objective response rate, overall survival rate (Overall survival rate, Overall survival, OS), progression-free survival (PFS), objective response rate (Objective Response Rate, ORR), time to progression (time to progress, TTP), disease-free survival (Disease-free survival, DFS), time to treatment failure (time to treatment failure, TTF), response rate (RR), complete response (CR), partial response (PR). The term "prognosis" is recognized in the art and includes predictions about possible disease processes or disease development, especially about the possibility of disease remission, disease recurrence, tumor recurrence, metastasis and death. "Good prognosis" means the possibility that a patient with cancer, especially pancreatic cancer, remains disease-free (i.e., cancer-free). "Poor prognosis" means the possibility of recurrence or recurrence, metastasis or death of a potential cancer or tumor.

[0102] Figure 3 is a schematic diagram of a computer device provided by an embodiment of the present invention, such as Figure 3 As shown, the device may include: one or more processors, and one or more memories; wherein the memories store computer-readable codes, and when the computer-readable codes are run by the one or more processors, the method described above may be executed.

[0103] The processor in this embodiment can be an integrated circuit chip with signal processing capabilities. The above processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, operations and logic block diagrams in the embodiments of the present disclosure can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc., which can be an X86 architecture or an ARM architecture.

[0104] In general, various example embodiments of the present disclosure may be implemented in hardware or dedicated circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor, or other computing device. When various aspects of the disclosed embodiments are illustrated or described as block diagrams, flow charts, or using some other graphical representation, it will be understood that the blocks, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof as non-limiting examples.

[0105] For example, the method or device according to the embodiment of the present disclosure may also be implemented by Figure 4 The architecture of the computing device 3000 shown in FIG. Figure 4 As shown, the computing device 3000 may include a bus 3010, one or more CPUs 3020, a read-only memory (ROM) 3030, a random access memory (RAM) 3040, a communication port 3050 connected to a network, an input / output component 3060, a hard disk 3070, etc. The storage device in the computing device 3000, such as ROM 3030 or hard disk 3070, may store various data or files used for processing and / or communication of the method provided by the present disclosure and program instructions executed by the CPU. The computing device 3000 may also include a user interface 3080. Of course, Figure 4 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 4 One or more components of a computing device are shown.

[0106] The embodiment of the present invention also provides a computer-readable storage medium, such as Figure 5 As shown, it is a schematic diagram of a storage medium provided by an embodiment of the present invention, and a computer readable instruction 4010 is stored on the computer storage medium 4020. When the computer readable instruction 4010 is executed by the processor, the method according to the embodiment of the present disclosure described with reference to the above figures can be executed. The computer readable storage medium in the embodiment of the present disclosure can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. The non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory (DR RAM). It should be noted that the memory of the methods described herein is intended to include, but is not limited to, these and any other suitable types of memory. It should be noted that the memory of the methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0107] The embodiments of the present disclosure further provide a computer program product or system, including a computer program, which implements the steps of the above method when executed by a processor.

[0108] In some embodiments, this embodiment also discloses a system for predicting the survival of CRLM patients, such as Figure 2 As shown, the system comprises:

[0109] A first acquisition module is used to acquire arterial phase and venous phase images of the subject including the target area;

[0110] A score calculation module, used for inputting the arterial phase image into a classifier to calculate the arterial phase DL score, inputting the venous phase image into a classifier to calculate the venous phase DL score, and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score;

[0111] A first prediction module is used to predict the survival period of the subject based on the total DL score.

[0112] In some embodiments, this embodiment also discloses a system for predicting the risk of RAS gene mutation in CRLM patients, such as Figure 7 As shown, the system comprises:

[0113] A second acquisition module is used to acquire CECT images of the subject; the CECT images include venous phase images and arterial phase images;

[0114] The second prediction module is used to input the CECT image into a classifier to obtain the predicted RAS gene mutation risk.

[0115] Materials and methods:

[0116] A retrospective study was performed on 514 patients with histopathologically confirmed CRLM admitted to Beijing Luhe Hospital and Peking University Cancer Hospital from August 2016 to October 2023. Tumor stage and RAS gene status were evaluated according to the American Joint Committee on Cancer (AJCC) TNM staging and pathological examination criteria.

[0117] Inclusion criteria included: (1) histopathological diagnosis of CRC; (2) availability of RAS mutation status; (3) measurable metastatic disease according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1; (4) abdominal contrast-enhanced computed tomography (CECT) scan before systemic treatment for all CRLM; and (5) at least one measurable liver metastasis. Exclusion criteria included: (1) presence of other malignancies; (2) concurrent cardiac or renal insufficiency or severe central nervous system damage; (3) CECT image quality insufficient for reliable assessment; and (4) incomplete follow-up records. A similar protocol was implemented at the Peking University Cancer Hospital during the same period.

[0118] Evaluation and follow-up

[0119] CRC patients were monitored regularly until death or the last follow-up date (October 31, 2023). Follow-up evaluations, including CECT scans and laboratory tests, were performed every 8-12 weeks. These evaluations were performed before each treatment or during scheduled follow-up and at least every 3-4 weeks. Two experienced radiologists specializing in abdominal imaging with more than 5 years of expertise independently reviewed and interpreted the imaging data according to RECIST guidelines. When the evaluations were inconsistent, consensus was reached by discussion.

[0120] Overall survival (OS) was defined as the period from the start of initial systemic therapy for liver metastases (LM) to all-cause death or last follow-up. Progression-free survival (PFS) was determined as the period from the start of initial systemic therapy to the development of progressive disease (PD) (characterized by abnormal nodular, disseminated, and / or atypical peripheral enhancement patterns in the liver parenchyma) or the conclusion of follow-up.

[0121] Clinical and pathological information

[0122] Key patient demographics and clinical variables included age, sex, Karnofsky performance status (KPS), alcohol and smoking history, histological grade, primary tumor location, LM maximum diameter and number, metastatic time, metastatic location, carcinoembryonic antigen (CEA) level, aspartate aminotransferase (AST) level, alanine aminotransferase (ALT) level, albumin level, and TNM stage. In addition, the use of cetuximab during treatment was recorded.

[0123] Deep learning radiomics model construction

[0124] Regions of interest (ROIs) for LM nodules were manually drawn by a radiologist with 10 years of experience in abdominal imaging. ROIs were extracted using ITK-Snap 3.8.0. DL models were built based on the two largest tumors in the CECT arterial phase (AP) and venous phase (VP). The ROI size of each CECT image was resized to 64x64 pixels. DL signatures were extracted from the AP-DL model using ShuffleNet V2 via a voting classifier and from the VP-DL model using VGG19. During training, the parameters of the DL models were iteratively updated via back-propagation. The input of the model was the image ROI, and the output was the prediction of RAS mutation status. Binary cross entropy between the model output and the label was used as the loss function. To mitigate overfitting, data augmentation strategies were adopted, including random sampling from video frames, random flipping, and rotation of the input image. For dimensionality reduction of radiomic features, several methods were used: SelectKBest, correlation comparison, least absolute shrinkage, and selection operator (LASSO). These methods were used to identify the most important radiomic features associated with RAS mutation status. The selected radiomic features were then fed into a machine learning (ML) classifier, and the best performing model output was converted into a probabilistic risk score, termed the radiomic score (R-score). In addition, a deep learning score was generated using a fully connected neural network (FCNN) based on the previously extracted important deep learning features. The DL model was trained in Python 3.8.2 framework using PyTorch 2.2.0. Detailed mathematical definitions, descriptions, and training procedures are provided in the Supplementary Methods.

[0125] Predictive model construction and performance

[0126] In addition to the DL (AP / VP / AP+VP) model, radiomics and DL-Radiomic (DL-R) models were constructed and evaluated. Radiomics features were extracted from tumor ROIs using PyRadiomics, and cross-validation (CV) and least absolute shrinkage and selection operator (LASSO) were applied for feature selection. Then, the radiomics model was constructed using logistic regression with selected key features. The DL and radiomics models were merged to create the DL-R model. The predictive validity of all models was validated and compared using ITC and ETC.

[0127] Statistical analysis

[0128] Quantitative variables were expressed as mean ± standard deviation (SD) or median with interquartile range (IQR) and compared using appropriate statistical methods. Categorical variables were expressed as counts with matching percentages and compared using chi-square test or Fisher's exact test. This investigation used the area under the receiver operating characteristic (AUC) curve to demonstrate predictive ability. The DeLong test was used to compare the AUC curves of several models. The optimal threshold of the DL model was determined using X-tile software. Statistical analysis was performed using SPSS version 23.0 and R software version 3.5.1, and significance was set at P < 0.05.

[0129] Accurate and rapid quantification of RAS mutation status is essential for identifying CRLM patients who are suitable for treatment with cetuximab or bevacizumab. However, compared with imaging-based detection methods, the practicality of RAS detection is limited by the dynamic changes in the proportion of cells expressing RAS mutations and the invasiveness of tissue sampling. Therefore, a non-invasive, accurate, and reproducible method is needed to assess RAS mutation status. In our study, we developed a DL model to predict RAS mutation status using CECT images of LM in patients with CRLM. The AUC values ​​of this model for TC, ITC, and ETC were 0.86, 1740.82, and 0.98, respectively. The practicality of the model was further validated by generating a DL score, which helps identify patients who are most likely to benefit from treatment with cetuximab or bevacizumab. At the same time, the DL score is an independent prognostic factor for patients with colorectal liver metastases. This deep learning model represents a promising non-invasive tool for assessing RAS mutation status in patients with CRLM, with the potential to improve the accuracy of targeted treatment decisions.

[0130] Grad-CAM is used in convolutional neural networks for “visual explanation” of decisions, enabling visualization of the network’s focus areas. It stress-classifies LM regions in CECT based on rough topographic maps. In this study, we observed heat maps of high-weight regions surrounding LM regions in AP and VP, which were consistent with the predicted regions of LM. Furthermore, the weights of the regions surrounding LM played a more important role in the prediction than the weights of the tumor regions themselves. These findings demonstrate the discriminative power and robustness of DL models in identifying and analyzing LM.

[0131] Radiomics is a complex quantitative technique that extracts imaging features that are not discernible to the naked eye. Previous studies on radiomics prediction of CRC RAS ​​gene status have focused on positron emission tomography (PET), magnetic resonance imaging (MRI), and CECT images of the primary tumor site. However, there is still a gap in multimodal studies that combine radiomics with LM CECT images. For example, Zhao et al. used pretreatment 18F-fluorodeoxyglucose (FDG) PET to establish a prediction model in CRC patients, with an area under the curve (AUC) of 0.759 (95% CI: 0.585-0.909) in the training cohort and 0.701 (95% CI: 0.468-0.916) in the cohort. Despite its high accuracy, PET scans are expensive compared with CT and MRI and are less used in clinical practice. Ming et al. found that the diagnostic accuracy of a multiparametric MRI-based radiomics model for predicting KRAS mutations in rectal cancer was 0.651. Similarly, Cui et al. reported a diagnostic accuracy of 0.722 for the prediction of KRAS mutations by an MRI-based radiomics model, which was validated in both internal and external cohorts. Although these studies showed high diagnostic accuracy, they have limitations such as small sample size196 and lack of extensive external validation. In addition, all of these studies used tumor images from the primary colon and rectal sites. This approach may result in inaccurate region of interest (ROI) delineation due to the presence of intestinal contents and gas. Given the high incidence of advanced CRC liver metastases and minimal differences in genetic status between primary and metastatic lesions, our study utilized CECT images of CRLM. By delineating the ROI of the LM, we minimized the inaccuracy of ROI delineation, resulting in more reliable results.

[0132] Our results showed that the DL model constructed using AP and VP ROIs showed excellent performance in both ITC and ETC, with AUC values ​​of 0.82 and 0.98, respectively. These values ​​exceeded those achieved by the DL-R model, the radiomics model, and the clinical model. This superior performance can be attributed to the ability of the deep learning algorithm to extract extensive image feature information from biphasic CECT, thereby capturing key details in the images. Nevertheless, the addition of radiomic features did not provide additional value in our study. However, we recognize that radiomics signature features can help identify the genotype of colorectal cancer. Therefore, combining these two sets of features can complement each other and potentially improve the predictive accuracy of the model. Unfortunately, in this study, we did not observe higher AUC values ​​when combining the features compared to using the DL model alone.

[0133] We also developed a visual nomogram based on the DL model, integrating traditional tumor biomarkers such as CEA, DCR, and RAS mutation risk. This nomogram is designed for personalized and non-invasive cumulative OS rate prediction before treatment in patients with CRLM. When CRLM patients were divided into high-risk and low-risk RAS mutations by our proposed DL model, we observed significant differences in the cumulative OS curves. Our results showed that the 1-year, 3-year, and 5-year cumulative OS rates of patients with high-risk RAS mutations were significantly lower than those of patients with low-risk RAS. These differences are consistent with the actual results of the RAS mutation group and the RAS wild-type group. In addition, a comparison of the benefits of targeted therapy between the high-risk and low-risk groups showed that bevacizumab showed benefits in the high-risk group, while cetuximab was significantly beneficial to the low-risk group. These findings are consistent with the actual clinical results. This indirectly reaffirms the value of the DL model in predicting RAS mutation status and suggests its potential in predicting the effect of targeted therapy.

[0134] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment, or a part of a code, and the module, program segment, or a part of a code contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0135] In general, various example embodiments of the present disclosure may be implemented in hardware or dedicated circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor, or other computing device. When various aspects of the disclosed embodiments are illustrated or described as block diagrams, flow charts, or using some other graphical representation, it will be understood that the blocks, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof as non-limiting examples.

[0136] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0137] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0138] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0139] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0140] The exemplary embodiments of the present disclosure described in detail above are merely illustrative and not restrictive. It should be understood by those skilled in the art that various modifications and combinations may be made to these embodiments or their features without departing from the principles and spirit of the present disclosure, and such modifications should fall within the scope of the present disclosure.

Claims

1. A method for predicting the survival of CRLM patients, characterized in that: The method comprises: S101: Acquire arterial phase and venous phase images of the subject including the target area; S102: inputting the arterial phase image into a classifier to calculate the arterial phase DL score, inputting the venous phase image into a classifier to calculate the venous phase DL score, and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score; S103: Predicting the survival period of the subject based on the total DL score.

2. The method for predicting the survival period of CRLM patients according to claim 1, characterized in that: The method further comprises: obtaining a CEA value and / or a DCR value of the subject; predicting the survival period of the subject based on any one or more of the DL score, CEA value, and DCR value; Optionally, when predicting the survival of a subject based on the DL score, CEA value and DCR value, the survival of the subject is predicted by calculating the total score of the DL score, CEA value and DCR value.

3. The method for predicting the survival period of CRLM patients according to claim 2, characterized in that: When the total score is calculated based on the DL score, CEA value, and DCR value, the calculation method of the total score includes: Y = 2.995 × 10 -5 ×CEA+(-1.581×DCR)+(-6.483×10 -1 ×DL), where Y is the total score.

4. The method for predicting the survival period of CRLM patients according to claim 1, characterized in that: Between S101 and S102, the method further includes: segmenting an arterial phase ROI region from the arterial phase image, and segmenting a venous phase ROI region from the venous phase image; calculating an arterial phase DL score based on the arterial phase ROI region, and calculating a venous phase DL score based on the venous phase ROI region; and integrating the arterial phase DL score and the venous phase DL score to obtain a total DL score.

5. A method for predicting the risk of RAS gene mutation in CRLM patients, characterized in that: The method comprises: S201: Acquire CECT images of a subject; the CECT images include venous phase images and arterial phase images; S202: Inputting the CECT image into a classifier to obtain a predicted RAS gene mutation risk. The method for constructing the classifier comprises: S2021: Obtain training set samples including images of the arterial phase and the venous phase and RAS gene status labels corresponding to the samples; the RAS gene status labels include mutation and wild type; S2022: inputting the arterial phase image into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining an arterial phase classifier based on the arterial phase image; inputting the venous phase image into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a venous phase classifier based on the venous phase image; S2023: Using an integrated classifier to combine the arterial phase classifier and the venous phase classifier to obtain the classifier.

6. The method for predicting the risk of RAS gene mutation in CRLM patients according to claim 5, characterized in that: The method further comprises: predicting the survival rate of the subject based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, obtaining a result that the survival rate of the subject is high; if the RAS gene mutation risk is high risk, obtaining a result that the survival rate of the subject is low; Optionally, the survival period of the subject is predicted based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, a result indicating that the subject has a long survival period is obtained; if the RAS gene mutation risk is high risk, a result indicating that the subject has a short survival period is obtained; Optionally, the survival rate or survival period includes any one or more of the following years: 1 year, 3 years, 5 years.

7. The method for predicting the risk of RAS gene mutation in CRLM patients according to claim 5, characterized in that: The method further includes: identifying the type of targeted drug suitable for the subject based on the RAS gene mutation risk; if the RAS gene mutation risk is low risk, the targeted drug suitable for the subject is cetuximab; if the RAS gene mutation risk is high risk, the targeted drug suitable for the subject is bevacizumab; Optionally, between S2021 and S2022, the construction method also includes: segmenting an arterial phase ROI area from the arterial phase image, and segmenting a venous phase ROI area from the venous phase image; inputting the arterial phase ROI area into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining an arterial phase classifier based on the arterial phase image; inputting the venous phase ROI area into the machine learning model for training to obtain a predicted classification result, comparing the predicted classification result with the classification label, optimizing the model according to the comparison result, and obtaining a venous phase classifier based on the venous phase image.

8. A computer device, characterized in that: The device comprises: a memory and a processor; the memory is used to store a computer program; the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images