Method, device and program product for predicting Gleason score upgrading and degrading after radical prostatectomy

Abstract

The invention relates to the field of intelligent medical treatment, in particular to a method, equipment and a program product for predicting Gleason score upgrading and degrading after radical prostatectomy. Comprising the following steps: S1, acquiring preoperative pathological data and image data of a to-be-tested person subjected to radical prostate resection; s2, obtaining a preoperative puncture score based on the preoperative pathological data; s3, performing prediction based on the preoperative pathological data, the image data and the preoperative puncture score to obtain a postoperative Gleason score and a Gleason upgrading and degrading result, and the application can perform upgrading and degrading prediction of the postoperative Gleason score at the same time, and has good clinical value.

Description

Technical Field

[0001] The present application relates to the field of intelligent medical care, and specifically to a method, device, program product, and computer-readable storage medium for predicting the upgrade or downgrade of Gleason score (GS) after radical prostatectomy. Background Art

[0002] Prostate cancer (PCa) is the most common malignancy in European or American men. Digital rectal examination is the most common method for PCa screening, but it has low sensitivity and specificity for diagnosing PCa. PCa is usually manifested by elevated plasma prostate-specific antigen (PSA) levels; however, PSA testing often produces false-positive results, and men without cancer may also have elevated PSA levels. Magnetic resonance imaging (MRI) is a noninvasive test with excellent soft tissue resolution and is the preferred imaging method for diagnosing PCa. Multiparametric MRI (mpMRI) combines anatomical and functional parameters and has become an important tool for detecting PCa. The Prostate Imaging Reporting and Data System (PI-RADS) was first proposed in 2012 and revised in 2019 (PI-RADSv2.1) to ensure standardization of the acquisition and reporting of mpMRI data.

[0003] PCa treatment usually depends on the biopsy Gleason score (bGS). However, the discordance rate between the Gleason score after radical prostatectomy (pGS) and the corresponding bGS has been reported to be 20%-60%, and accurate patient classification is very important in making treatment decisions. Summary of the invention

[0004] The goal of PCa treatment is to prevent overtreatment of low-risk patients and undertreatment of patients who need active treatment. Determining more reliable preoperative risk factors to predict the upgrade and downgrade of GS after radical prostatectomy (RP) can not only reduce the risk of misdiagnosis of patients, but also provide guidance for patients' treatment decisions. Therefore, in view of the above problems, the present invention proposes a method for predicting the upgrade and downgrade of Gleason score after radical prostatectomy, comprising:

[0005] S1. Obtain preoperative pathological data and imaging data of patients undergoing radical prostatectomy;

[0006] S2. obtaining a preoperative puncture score based on the preoperative pathological data;

[0007] S3. Based on the preoperative pathological data, imaging data, and preoperative puncture score, the postoperative Gleason score and the Gleason upgrade or downgrade result are predicted.

[0008] Furthermore, S3 specifically includes: inputting the preoperative pathological data and imaging data into a prediction and judgment model to obtain a predicted postoperative Gleason score and upgrade and downgrade results, wherein the prediction and judgment model is trained based on a training set data of preoperative pathological data, imaging data and postoperative Gleason score labels of patients undergoing prostatectomy.

[0009] The prediction of S3 is replaced by: judging whether the Gleason score is upgraded by using the preoperative pathological data, imaging data, and preoperative puncture score; when the judgment result is no, judging whether it is downgraded; when the judgment result is yes, judging it is upgraded; in judging whether it is downgraded, when the judgment result is no, judging the result is not upgraded or downgraded, and when the judgment result is yes, judging it is downgraded;

[0010] Optionally, the upgrade and downgrade results of the Gleason score include: upgrade, downgrade, no upgrade or downgrade;

[0011] The method further includes category determination, when the Gleason score judgment result is an upgrade, performing a secondary judgment of upgrade to determine whether to upgrade from Class I, and when the judgment result is no, determining to upgrade from Class II to Class I; when the Gleason score judgment result is a downgrade, performing a secondary judgment of downgrade to determine whether to downgrade from Class III, and when the judgment result is no, determining to downgrade from Class II to Class I;

[0012] Optionally, the second-level upgrade judgment further includes a third-level upgrade judgment. When the second-level upgrade judgment determines whether to upgrade from Class I, the third-level upgrade judgment is performed to determine whether to upgrade to Class III. When the judgment result is yes, the upgrade result is determined to be an upgrade from Class I to Class III. When the judgment result is no, the upgrade result is determined to be an upgrade from Class I to Class II.

[0013] Optionally, the second-level downgrade judgment also includes a third-level downgrade judgment. When the second-level downgrade judgment is performed, it is judged whether to downgrade from Class III. When the judgment result is yes, the third-level downgrade judgment is performed to judge whether to downgrade to Class I. When the judgment result is yes, the downgrade result is determined to be downgraded from Class III to Class I. When the judgment result is no, the downgrade result is determined to be downgraded from Class III to Class II.

[0014] Optionally, the category is given medication treatment, when the category is Class I, active monitoring and regular follow-up are performed; when the category is Class II, internal prostate irradiation therapy, external radiotherapy, radical prostatectomy are performed; when the category is Class III, radical prostatectomy + lymph node dissection, radical prostatectomy + adjuvant therapy are performed, and the adjuvant therapy includes one or more of the following: radiotherapy, endocrine therapy.

[0015] S2 of the method is replaced by S21: feature indicators are obtained by extracting features based on pathological data and imaging data, and prediction is made on the upgrade or downgrade of the Gleason score of the subject after surgery based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, location of PCa, and PI-RADS v2.1 score.

[0016] The S21 is replaced by S22: inputting the pathological data and the imaging data into the prediction model for classification to obtain the upgrade and downgrade results, wherein the upgrade and downgrade prediction model is a model obtained by training the pathological data set, the imaging data set and the upgrade and downgrade feature labels;

[0017] The method further includes upgrading the prediction, where S2 is replaced by:

[0018] Feature extraction is performed based on pathological data and imaging data to obtain feature indicators, and an upgrade group is obtained based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, location of PCa, PI-RADS v2.1 score;

[0019] Optionally, the method further includes downgrade prediction, and S2 is replaced by: extracting features based on pathological data and imaging data to obtain feature indicators, and obtaining a downgrade group based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, PCa location.

[0020] Optionally, the upgrade prediction is replaced by: inputting the pathological data and the imaging data into an upgrade prediction model for prediction to obtain a prediction result, wherein the upgrade prediction model is a model obtained by training with the pathological data set, the imaging data set and the upgrade feature label;

[0021] Optionally, the degradation prediction is replaced by: inputting the pathological data and the imaging data into a degradation prediction model for prediction to obtain a prediction result, wherein the degradation prediction model is a model obtained by training with a pathological data set, an imaging data set and degradation feature labels;

[0022] Optionally, the prediction model includes one or more of the following: Logistic regression, random forest, convolutional neural network, ResNet, Inception V1 / V2 / v3, Vision Transformer, XGBoost, FCNs.

[0023] There are seven categories for upgrades and downgrades, including: Class I upgraded to Class II, Class I upgraded to Class III, Class II upgraded to Class III, Class III downgraded to Class II, Class III downgraded to Class I, Class II downgraded to Class I, and no upgrade / downgrade;

[0024] Optionally, the category corresponds to M Gleason score levels, and the Gleason score level is determined by adding a primary structure score and a secondary structure score of the Gleason score, and M is a natural number greater than or equal to 1;

[0025] Optionally, the class I corresponds to level 1, the class II corresponds to level 2, and the class III corresponds to levels 3-5;

[0026] Optionally, the sum of the primary mode score and the secondary mode score of the 5 levels is 9, including: the primary mode score is 4, the secondary mode score is 5, the primary mode score is 5, the secondary mode score is 4;

[0027] Optionally, the sum of the primary mode score and the secondary mode score of the level 5 is 10, and the primary mode score is 5 and the secondary mode score is 5.

[0028] The object of the present invention is to provide a computer program product having a computer program or instructions thereon, wherein the computer program or instructions are executed by a processor to implement the above-mentioned method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy.

[0029] The object of the present invention is to provide a computer device, including a memory, a processor, and a computer program or instruction stored in the memory, wherein the computer program or instruction is executed by the processor to implement the above-mentioned method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy.

[0030] The object of the present invention is to provide a computer-readable storage medium having a computer program or instruction stored thereon, wherein the computer program or instruction is executed by a processor to implement the above-mentioned method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy.

[0031] Advantages of the present invention:

[0032] 1. A method for simultaneously predicting GS upgrade or downgrade after radical prostatectomy is proposed, which specifically includes dividing the preoperative puncture Gleason score into upgrade and downgrade, and obtaining the process of predicting the upgrade or downgrade of the postoperative Gleason score, that is, upgrading from one category to another. The method of the present invention reduces the risk of misdiagnosis and can also provide treatment decision guidance for patients.

[0033] 2. Classification of upgrade and downgrade by clinical indicators PBCs, PCa location, and PI-RADS v2.1 score can improve the classification accuracy of the upgraded and downgraded groups, thereby improving the accuracy of overall GS upgrade and downgrade;

[0034] 3. The present invention can independently predict the upgrade or downgrade of the Gleason score, meet the actual clinical needs, and has important clinical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] 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.

[0036] Figure 1 A schematic flow chart of a method for predicting the upgrade or downgrade of Gleason score after radical prostatectomy provided by an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of a system for predicting the upgrade or downgrade of Gleason score after radical prostatectomy provided by an embodiment of the present invention;

[0038] Figure 3 A schematic diagram of a device for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy provided by an embodiment of the present invention;

[0039] Figure 4 An upgrade example provided for an embodiment of the present invention;

[0040] Figure 5 A downgrade example provided for an embodiment of the present invention;

[0041] Figure 6 A flow chart of upgrading and downgrading judgment provided by an embodiment of the present invention;

[0042] Figure 7 The AUC value of whether to upgrade provided in the embodiment of the present invention;

[0043] Figure 8 The AUC value of whether it is degraded provided in the embodiment of the present invention;

[0044] Fig. 9 The AUC value for whether to upgrade from Class I in the upgrade judgment process provided in the embodiment of the present invention;

[0045] Fig.10 The AUC value of whether to upgrade to Class III in the upgrade judgment process provided in the embodiment of the present invention. 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 sequence numbers of the operations, such as S101, S102, etc., are only used to distinguish between different operations, and the sequence 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] Figure 1 A schematic diagram of a method for predicting the upgrade or downgrade of Gleason score after radical prostatectomy provided by an embodiment of the present invention specifically includes:

[0049] S1: Obtain preoperative pathological data and imaging data of patients undergoing radical prostatectomy;

[0050] In one embodiment, the image data adopts one or more of the following: CT, MRI, and ultrasound.

[0051] In a specific embodiment, the present invention retrospectively collected clinical and MRI data of 306 patients with prostate disease from January 2015 to December 2021. Patient inclusion criteria include (1) complete MRI examination; (2) complete laboratory tests, including total PSA (tPSA), free PSA (fPSA), free PSA / total PSA ratio (F / T) and PSA density (PSAD); (3) patients were diagnosed with perineal ultrasound-guided prostate puncture biopsy and RP after MRI; (4) no treatment before prostate biopsy and RP. Exclusion criteria include (1) patients received treatment before prostate biopsy and RP (n=32), (2) poor MRI image quality, affecting PI-RADS v2.1 score (n=9) and (3) incomplete clinical or imaging data (n=40). Finally, a total of 225 patients were recruited.

[0052] In a specific embodiment, the following clinical pathological data are recorded in detail: including age, prostate volume, tPSA, fPSA, F / T, PSAD, number of positive puncture needles (PBC), time interval from biopsy to RP, bGS and pGS. A urologist performs transperineal prostate puncture biopsy under 11+X needle ultrasound guidance. The urologist divides the prostate into 11 areas through ultrasound scanning, first performs 11 needle systematic punctures, and finally performs at least one cognitive targeted puncture. The operation steps of cognitive targeted puncture are as follows: the urologist first reviews the MRI examination results; then uses the MRI image information to perform targeted biopsy on the most significant nodules under ultrasound image guidance. The RP specimen is sliced ​​at intervals of 3-5mm from the base to the tip, and the corresponding PCa boundary is outlined on each slice. A pathologist then performs pathological scoring on the tissue specimens of the prostate puncture biopsy and RP. Finally, a radiologist matches the lesions with the highest PI-RADS v2.1 scores on the MRI images with the pathological results.

[0053] In a specific embodiment, MRI scanning was performed using a GE750 3.0T scanner (GE Healthcare, Milwaukee, Wisconsin, USA). The scanning sequence included axial T1WI and T2WI, coronal and sagittal T2WI, diffusion weighted images (DWI), and dynamic contrast enhancement (DCE). When the b value was 100 and 1,000 sec / mm 2 When the present invention is evaluated by the apparent diffusion coefficient (ADC) graph.

[0054] In a specific embodiment, two radiologists independently analyzed the images under double-blind conditions. When disagreements occurred, the two radiologists reached consensus through consultation. Prostate volume, PCa diameter, and PCa location (peripheral zone / transition zone, bottom / middle / apex) were measured on MRI images. Radiologists scored the lesions in each sequence in mpMRI and then calculated the final score of PI-RADS v2.1. When multiple lesions were present, the lesion with the highest PI-RADS score was used for statistical analysis.

[0055] Statistical analysis was performed using Rv.4.3.1. Normally distributed data were expressed as mean ± SD, and the data were tested by independent sample t test. Skewed distributed data were expressed as median (upper and lower quartiles), and the data were tested by Mann-Whitney U test. The chi-square test was used to compare the differences between PCa location and PI-RADS v2.1 score, and logistic regression was used to analyze the influencing factors of GS upgrade and downgrade after RP. The area under the curve (AUC) for predicting GS upgrade and downgrade was calculated using the receiver operating characteristic (ROC) curve. P values ​​< 0.05 were considered statistically significant.

[0056] In a specific embodiment, 225 prostate cases: the age of the patients was 46-80 years old (average age 66 years old), and the main clinical manifestations were elevated PSA, dysuria, frequent urination, and urgency. Among the 225 patients, 91 (40.4%) had an upgraded Gleason score after surgery, 45 (20%) had a downgraded Gleason score after surgery, and 89 (39.6%) had the same bGS as pGS. 45 (20%) patients had bGS ≤ 6, and their Gleason scores rose to 7 or above after surgery. 5 (2.2%) patients had bGS ≥ 7, and their Gleason scores dropped to 6 after surgery. There were 78 patients with a PI-RADS v2.1 score of 5, 105 patients with a score of 4, 29 patients with a score of 3, and 13 patients with a score of 2.

[0057] S2: obtaining a preoperative puncture score based on the preoperative pathological data;

[0058] In one embodiment, S2 of the method is replaced by S21: feature indicators are obtained by extracting features based on pathological data and imaging data, and a prediction is made on the upgrade or downgrade of the Gleason score of the subject after surgery based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, location of PCa, PI-RADS v2.1 score.

[0059] In one embodiment, S21 is replaced by S22: pathological data and imaging data are input into a prediction model for classification to obtain an upgrade or downgrade result, wherein the upgrade or downgrade prediction model is a model trained by a pathological data set, an imaging data set and upgrade or downgrade feature labels; the prediction model classifies the input pathological data and imaging data to obtain an upgraded, downgraded or unchanged classification result.

[0060] In one embodiment, the method further includes upgrading the prediction, and S2 is replaced by:

[0061] Feature extraction is performed based on pathological data and imaging data to obtain feature indicators, and an upgrade group is obtained based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, location of PCa, and PI-RADS v2.1 score.

[0062] In one embodiment, the upgrade prediction is replaced by: inputting the pathology data and the imaging data into the upgrade prediction model for prediction to obtain a prediction result, wherein the upgrade prediction model is a model trained by the pathology data set, the imaging data set and the upgrade feature label. The upgrade prediction model performs prediction by comparing the features in the input case data and imaging data with the features learned by the model to obtain a prediction result, and the prediction result is the input data: yes (upgrade) or no (upgrade).

[0063] In one embodiment, the method further includes downgrade prediction, and S2 is replaced by: extracting features based on pathological data and imaging data to obtain feature indicators, and obtaining a downgrade group based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, PCa location.

[0064] In one embodiment, the downgrade prediction is replaced by: inputting the pathological data and the imaging data into the downgrade prediction model for prediction to obtain a prediction result, wherein the downgrade prediction model is a model trained by the pathological data set, the imaging data set and the downgrade feature label. The downgrade prediction model performs prediction by comparing the features in the input case data and imaging data with the features learned by the model to obtain a prediction result, and the prediction result is the input data: yes (downgrade) or no (downgrade).

[0065] In one embodiment, the prediction model includes one or more of the following: Logistic regression, random forest, convolutional neural network, ResNet, Inception V1 / V2 / v3, Vision Transformer, XGBoost, FCNs.

[0066] S3, predicting the postoperative Gleason score and the Gleason upgrade or downgrade result based on the preoperative pathological data, imaging data, and preoperative puncture score;

[0067] In one embodiment, the Gleason score is composed of two main scores: the primary pattern score and the secondary pattern score. These two scores are based on the most aggressive cancer cell structure in prostate tissue. The primary pattern score describes the cell structure that occupies the largest proportion of the tumor, and the score ranges from 1 to 5. 1 point represents a benign or very similar structure to normal prostate tissue, and 5 points represent a highly invasive structure. The secondary pattern score describes the second most common cell structure in the tumor, and the score ranges from 1 to 5. If the secondary structure is the same as the primary structure, it is only recorded once and represented by the higher score. These two scores are added together to obtain a total Gleason score, which ranges from 2 to 10.

[0068] In a specific embodiment, the GS groups are defined as follows: 3+3, 3+4, 4+3, 8 (4+4 or 3+5 or 5+3), 9 (4+5 or 5+4) and 10 (5+5).

[0069] In a specific embodiment, the Gleason grading system is used to grade PCa histopathologically, and according to the 2014 International Society of Urological Pathology criteria, GS upgrade is defined as a higher Gleason grade of the RP specimen than the Gleason grade of the biopsy. GS downgrade is defined as a lower Gleason grade of the RP specimen than the Gleason grade of the biopsy.

[0070] In one embodiment, S3 specifically includes: inputting the preoperative pathological data and imaging data into a prediction and judgment model to obtain a predicted postoperative Gleason score and upgrade and downgrade results, wherein the prediction and judgment model is trained based on a training set data of preoperative pathological data, imaging data and postoperative Gleason score labels of patients undergoing prostatectomy.

[0071] In one embodiment, the prediction and judgment model adopts one or more of the following: convolutional neural network, residual network, void convolutional neural network, Transformer, XGBoost, AdaBoost, SVM, MLP.

[0072] In one embodiment, the prediction of S3 is replaced by: judging whether the Gleason score is upgraded by using the preoperative pathological data, imaging data, and preoperative puncture score; when the judgment result is no, judging whether it is downgraded; when the judgment result is yes, judging it as an upgrade; in judging whether it is downgraded, when the judgment result is no, judging the result is not upgraded or downgraded, and when the judgment result is yes, judging it as downgrade.

[0073] In one embodiment, the method further includes category determination. When the Gleason score judgment result is an upgrade, a secondary judgment on upgrade is performed to determine whether it is upgraded from Class I. When the judgment result is no, it is determined to be upgraded from Class II to Class I. When the Gleason score judgment result is a downgrade, a secondary judgment on downgrade is performed to determine whether it is downgraded from Class III. When the judgment result is no, it is determined to be downgraded from Class II to Class I.

[0074] In one embodiment, the second-level upgrade judgment also includes a third-level upgrade judgment. When the second-level upgrade judgment determines whether to upgrade from Class I, a third-level upgrade judgment is performed to determine whether to upgrade to Class III. When the judgment result is yes, the upgrade result is determined to be an upgrade from Class I to Class III. When the judgment result is no, the upgrade result is determined to be an upgrade from Class I to Class II.

[0075] In one embodiment, the second-level downgrade judgment also includes a third-level downgrade judgment. When the second-level downgrade judgment is performed, it is judged whether to downgrade from Class III. When the judgment result is yes, the third-level downgrade judgment is performed to judge whether to downgrade to Class I. When the judgment result is yes, the downgrade result is determined to be downgraded from Class III to Class I. When the judgment result is no, the downgrade result is determined to be downgraded from Class III to Class II.

[0076] In one embodiment, the upgrade and downgrade results of the Gleason score include: upgrade, downgrade, no upgrade or downgrade.

[0077] In one embodiment, the category is given medication treatment. When the category is Category I, active monitoring and regular follow-up are performed; when the category is Category II, intraprostate irradiation therapy, external radiation therapy, and radical prostatectomy are performed; when the category is Category III, radical prostatectomy + lymph node dissection, radical prostatectomy + adjuvant therapy are performed, and the adjuvant therapy includes one or more of the following: radiotherapy, endocrine therapy.

[0078] In one embodiment, for a data set with a sample size of 225 cases, the present invention first determines whether the postoperative Gleason score is upgraded based on clinical and MRI information, and uses the ROC curve to mark the predicted statistical effect. The area under the curve AUC value for predicting the postoperative Gleason score upgrade is 0.7727, indicating that the judgment has statistical validity; for the same sample, the area under the ROC curve AUC value for predicting the postoperative Gleason score degradation is 0.7485, which also has statistical validity;

[0079] The present invention obtains the categories of promotion and demotion based on clinical information statistics and makes the following judgment process based on the statistical results, such as Figure 6As shown: First, determine whether the postoperative Gleason score is upgraded. When the judgment result is yes, enter the upgrade category judgment. When the judgment result is no, make a downgrade judgment. In an embodiment with a sample size of 225 cases, the area under the judgment curve AUC value is 0.7897, and the 95% confidence interval is [0.7314, 0.8480], which has statistical validity, as shown in Figure 1. Figure 7 As shown; in the judgment of whether to downgrade, when the judgment result is yes, it enters the downgrade category judgment, and when the judgment result is no, it is determined that no upgrade or downgrade has occurred. In an embodiment with a sample size of 134 cases (a total sample size of 225), the area under the judgment curve AUC value is 0.7970, and the 95% confidence interval is [0.7205, 0.8735], which has statistical validity, as shown in FIG. Figure 8 As shown; further, in the upgrade category judgment, it is judged whether to upgrade from Class I. When the judgment result is no, the judgment result is upgraded from Class II to Class III. When the judgment result is yes, it is judged whether to upgrade to Class III. In an embodiment with a sample size of 91 cases (total sample size is 225), the area under the judgment curve AUC value is 0.8507, and the 95% confidence interval is [0.7718, 0.9296], which has statistical validity, as shown in FIG. Fig. 9 As shown; in the judgment of whether to upgrade to Class III, when the judgment result is no, the judgment result is upgraded from Class I to Class II, and when the judgment result is yes, the judgment result is upgraded from Class I to Class III. In an embodiment with a sample size of 46 cases (total sample size is 225), the area under the judgment curve AUC value is 0.7771, and the 95% confidence interval is [0.6267, 0.9275], which has statistical validity, as shown in FIG. Fig.10 As shown; in the judgment of the downgrade category, whether to downgrade from Class III is judged, when the judgment result is no, the judgment result is downgraded from Class II to Class I, and when the judgment result is yes, a judgment is made whether to downgrade to Class I. In an embodiment with a sample size of 45 cases (a total sample size of 225), the area under the judgment curve AUC value is 1, and the 95% confidence interval is [1.0000, 1.0000], which has statistical power; in the judgment of whether to downgrade to Class I, when the judgment result is no, the judgment result is downgraded from Class III to Class II, and when the judgment result is yes, the judgment result is downgraded from Class III to Class I. In an embodiment with a sample size of 42 cases (a total sample size of 225), the area under the judgment curve AUC value is 1, and the 95% confidence interval is [1.0000, 1.0000], which has statistical power.

[0080] In one embodiment, the postoperative Gleason score is obtained based on the preoperative pathological data and imaging data and the preoperative puncture score, and the upgrade and downgrade results are determined. The upgrade and downgrade results include upgrade, downgrade, no upgrade or downgrade. When the judgment result is an upgrade, an upgrade category judgment is performed, and when the judgment result is a downgrade, a downgrade category judgment is performed.

[0081] In one embodiment, the categories of upgrade and downgrade include seven categories, including: Class I upgraded to Class II, Class I upgraded to Class III, Class II upgraded to Class III, Class III downgraded to Class II, Class III downgraded to Class I, Class II downgraded to Class I, and no upgrade / downgrade.

[0082] The category corresponds to M Gleason score levels, and the Gleason score level is determined by adding the main structure score and the secondary structure score of the Gleason score, and M is a natural number greater than or equal to 1;

[0083] Optionally, the class I corresponds to level 1, the class II corresponds to level 2, and the class III corresponds to levels 3-5;

[0084] Optionally, the categories correspond to 5 levels; the main structure score and the secondary structure score of level 1 are both 3; the main structure score of level 2 is 3, and the secondary structure score is 4; the main structure score of level 3 is 4, and the secondary structure score is 3; the sum of the main structure score and the secondary structure score of level 4 is 8; the sum of the main structure score and the secondary structure score of level 5 is 9 or 10.

[0085] In one embodiment, the sum of the 4-level main structure score and the secondary structure score is 8, including: a main structure score of 4 and a secondary structure score of 4, a main structure score of 3 and a secondary structure score of 5, and a main structure score of 5 and a secondary structure score of 3.

[0086] In one embodiment, the sum of the five-level primary structure score and the secondary structure score of 9 includes: a primary structure score of 4 and a secondary structure score of 5, and a primary structure score of 5 and a secondary structure score of 4.

[0087] In one embodiment, the sum of the five-level major structure score and the minor structure score is 10: the major structure score is 5, and the minor structure score is 5.

[0088] In a specific embodiment, Table 1 shows the differences between the GS upgraded group and the GS non-upgraded group. The number of PBCs in the GS upgraded group was significantly higher than that in the GS non-upgraded group; the proportion of lesions located at the bottom in the GS upgraded group was higher than that in the GS non-upgraded group; the proportion of PI-RADS v2.1 scores of 4-5 in the GS upgraded group was significantly higher than that in the GS non-upgraded group (all P<0.05). Figure 4Shown is a 68-year-old man with an upgraded Gleason score from 6 at biopsy to 7 (4+3) after RP.

[0089] Table 2 shows the differences between the GS downgrade group and the GS non-downgrade group. The number of PBCs in the GS downgrade group was significantly lower than that in the GS non-downgrade group; the proportion of lesions located at the bottom in the GS downgrade group was significantly lower than that in the GS non-downgrade group; and the proportion of PI-RADS v2.1 score of 3 in the GS downgrade group was significantly higher than that in the GS non-downgrade group (all P<0.05). Figure 5 Shown is a 66-year-old man whose Gleason score was downgraded from 9(4+5) at biopsy to 7(3+4) after RP.

[0090] Table 1

[0091]

[0092]

[0093]

[0094] Table 2

[0095]

[0096]

[0097] Table 3

[0098]

[0099]

[0100] Table 4

[0101]

[0102] In one embodiment, univariate analysis showed that PBCs, PCa location (bottom / middle / apex) and PI-RADSv2.1 score were factors affecting postoperative GS upgrade (Table 3). In order to evaluate the predictive performance of the observed predictors for GS upgrade after RP, the present invention constructed ROC curves of clinical factors (PBCs), mpMRI performance (PCa location and PI-RADS v2.1 score) and a joint diagnostic model combining clinical and mpMRI variables. The AUC of clinical factors and mpMRI results were 0.7134 and 0.7067, respectively. The AUC of the joint diagnostic model was 0.7727, which was greater than the AUC of clinical factors alone (P=0.0298).

[0103] Univariate analysis showed that PBCs and PCa location (bottom / middle / apex) were factors driving postoperative GS downgrade (Table 4). To evaluate the predictive performance of the observed predictors for GS downgrade after RP, the present invention constructed ROC curves for clinical factors (PBCs), mpMRI findings (PCa location), and a combined diagnostic model combining clinical and mpMRI variables. The AUCs of clinical factors and mpMRI results were 0.6630 and 0.6443, respectively. The AUC of the combined diagnostic model was 0.7485, which was greater than the AUC value of clinical factors alone (P=0.0202).

[0104] In a specific embodiment, GS is closely related to tumor aggressiveness and prognosis, which affects the treatment options for PCa. However, bGS relies on incomplete sampling, and total prostatectomy specimens can provide complete and accurate tumor pathological grading. The difference between bGS and pGS has always been an unresolved issue. Identifying reliable preoperative risk factors for predicting GS upgrade and downgrade after RP will help reduce the risk of misdiagnosis in PCa patients. The present invention innovatively explores the impact of mpMRI manifestations (PCa diameter, PCa location, and PI-RADS v2.1 score) on GS upgrade and downgrade after RP. The results of the present invention show that patients with a PI-RADS v2.1 score of ≥4 are more likely to have a GS upgrade after RP, while patients with a PI-RADS v2.1 score of 3 are more likely to have a GS downgrade after RP.

[0105] In a specific embodiment, the regression analysis of the present invention study showed that PCa location (bottom / middle / apex) is a predictor of GS upgrade and downgrade after RP. Patients with lesions located at the bottom are more likely to have GS upgrade after RP, and patients with lesions located at the apex are more likely to have GS downgrade after RP.

[0106] PSA, prostate volume, and the time interval from puncture to RP were not significantly correlated with the upgrade or downgrade of GS after RP. PBCs were effective predictors of the upgrade or downgrade of GS after RP. The more PBCs, the greater the possibility of GS increase after RP. On the contrary, the fewer PBCs, the greater the possibility of GS decrease after RP.

[0107] High-risk PCa is highly invasive and the disease often progresses rapidly, leading to increased metastasis and mortality rates in patients. One of the goals of PCa treatment is to prevent overtreatment of low-risk patients, which is more valuable in clinical practice because it can allow patients to avoid RP. Therefore, it is particularly important to accurately predict GS upgrade and downgrade after RP. In the present study, the combination of clinical factors and mpMRI manifestations can significantly improve the diagnostic efficacy.

[0108] In conclusion, mpMRI can help predict GS upgrade and downgrade after RP and minimize the possibility of misdiagnosis of PCa. Combining clinical factors and mpMRI manifestations (PBCs, PCa sites, PI-RADS v2.1 score) can significantly improve the diagnostic performance of GS upgrade and downgrade after RP and minimize the possibility of undertreatment and overtreatment of PCa patients.

[0109] In one embodiment, treatment options for prostate cancer are primarily differentiated by risk stratification, and GS is the basis for risk stratification.

[0110] Prostate cancer treatments can be roughly divided into three categories:

[0111] Category I: Grade 1 (GS is 3+3) low-grade tumors, regular follow-up is recommended under active surveillance.

[0112] Category II: Grade 2 (GS is 3+4) intermediate-grade tumors are recommended for internal prostate irradiation, external beam radiotherapy, and radical prostatectomy.

[0113] Category III: For high-grade tumors of grade 3-5 (GS is 4+3, 8, 9, 10), radical prostatectomy + lymph node dissection or radical prostatectomy + adjuvant therapy (radiotherapy, endocrine therapy) is recommended.

[0114] The upgrade situations are divided into: Class I to Class II, Class I to Class III, and Class II to Class III.

[0115] The downgrade situations are divided into: Class III downgraded to Class II, Class III downgraded to Class I, and Class II downgraded to Class I.

[0116] The embodiments of the present disclosure also provide a computer program product or system, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy.

[0117] Figure 2 The schematic diagram of the system for predicting the upgrade or downgrade of Gleason score after radical prostatectomy provided by the embodiment of the present invention specifically includes:

[0118] Acquisition unit: acquiring preoperative pathological data and imaging data of subjects undergoing radical prostatectomy;

[0119] Preoperative scoring unit: obtain the preoperative puncture GS score based on preoperative pathological data;

[0120] Classification judgment unit: based on the preoperative pathological data, imaging data, and preoperative puncture score, predict the postoperative Gleason score and the Gleason upgrade or downgrade result.

[0121] Figure 3 The schematic diagram of the device for predicting the upgrade or downgrade of Gleason score after radical prostatectomy provided by the embodiment of the present invention specifically includes:

[0122] A memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, any one of the above-mentioned methods for predicting the upgrade or downgrade of Gleason score after radical prostatectomy is performed.

[0123] The present disclosure also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it is any one of the above-mentioned methods for predicting the improvement or deterioration of the Gleason score after radical prostatectomy.

[0124] The verification results of this verification embodiment show that assigning inherent weights to indications can improve the performance of the method relative to the default setting. Those skilled in the art can clearly understand that for the convenience and simplicity 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. The systems, devices and methods shown in several embodiments provided in this application 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, and there may be other division methods in actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between each other 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. The unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist 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. A person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above-mentioned embodiments may be completed by instructing the relevant hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0125] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be implemented by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned medium storage can be a read-only memory, a disk or an optical disk, etc.

[0126] The above is a detailed introduction to a computer device provided by the present invention. For a person skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for predicting the Gleason score upgrade or downgrade after radical prostatectomy, characterized in that: include: S1. Obtain preoperative pathological data and imaging data of patients undergoing radical prostatectomy; S2. obtaining a preoperative puncture score based on the preoperative pathological data; S3. Based on the preoperative pathological data, imaging data, and preoperative puncture score, the postoperative Gleason score and the Gleason upgrade or downgrade result are predicted.

2. The method for predicting Gleason score upgrade or downgrade after radical prostatectomy according to claim 1, characterized in that: Specifically, S3 includes: inputting the preoperative pathological data and imaging data into a prediction and judgment model to obtain a predicted postoperative Gleason score and upgrade and downgrade results, wherein the prediction and judgment model is trained based on a training set of preoperative pathological data, imaging data and postoperative Gleason score labels of patients undergoing prostatectomy.

3. The method for predicting Gleason score upgrade or downgrade after radical prostatectomy according to claim 2, characterized in that: The prediction of S3 is replaced by: judging whether the Gleason score is upgraded based on the preoperative pathological data, imaging data, and preoperative puncture score; if the judgment result is no, judging whether the score is downgraded; if the judgment result is yes, judging it as upgraded; In determining whether to downgrade, when the determination result is no, the determination result is not upgraded or downgraded, and when the determination result is yes, the determination result is downgraded; Optionally, the upgrade and downgrade results of the Gleason score include: upgrade, downgrade, no upgrade or downgrade.

4. The method for predicting Gleason score upgrade or downgrade after radical prostatectomy according to claim 3, characterized in that: The method further includes category determination, when the Gleason score judgment result is an upgrade, performing a secondary judgment of upgrade to determine whether to upgrade from category I, and when the judgment result is no, determining to upgrade from category II to category I; When the Gleason score is judged as downgrade, a second-level downgrade judgment is performed to determine whether it is downgraded from Class III. If the judgment result is no, the downgrade is determined to be from Class II to Class I. Optionally, the second-level upgrade judgment further includes a third-level upgrade judgment. When the second-level upgrade judgment determines whether to upgrade from Class I, the third-level upgrade judgment is performed to determine whether to upgrade to Class III. When the judgment result is yes, the upgrade result is determined to be an upgrade from Class I to Class III. When the judgment result is no, the upgrade result is determined to be an upgrade from Class I to Class II. Optionally, the second-level downgrade judgment further includes a third-level downgrade judgment. When the second-level downgrade judgment is made, it is judged whether to downgrade from Class III. When the judgment result is yes, a third-level downgrade judgment is made to judge whether to downgrade to Class I. When the judgment result is yes, it is determined that the downgrade result is downgraded from Class III to Class I. When the judgment result is no, it is determined that the downgrade result is downgraded from Class III to Class II. Optionally, medication is administered to the category, and when the category is category I, active monitoring and regular follow-up are performed; when the category is category II, internal prostate irradiation therapy, external beam radiation therapy, radical prostatectomy are performed; When the category is III, radical prostatectomy + lymph node dissection or radical prostatectomy + adjuvant therapy is performed, and the adjuvant therapy includes one or more of the following: radiotherapy, endocrine therapy.

5. The method for predicting Gleason score upgrade or downgrade after radical prostatectomy according to claim 1, characterized in that: S2 of the method is replaced by S21: extracting features based on pathological data and imaging data to obtain feature indicators, and predicting the postoperative Gleason score upgrade or downgrade of the subject based on the feature indicators; The characteristic indicators include one or more of the following: number of PBCs, location of PCa, PI-RADS v2.1 score; Optionally, S21 is replaced by S22: inputting the pathological data and the imaging data into a prediction model for classification to obtain an upgrade and downgrade result, wherein the prediction model is a model trained by using a pathological data set, an imaging data set and upgrade and downgrade feature labels.

6. The method for predicting Gleason score upgrade or downgrade after radical prostatectomy according to claim 5, characterized in that: The method further includes upgrading the prediction, where S2 is replaced by: Feature extraction is performed based on pathological data and imaging data to obtain feature indicators, and an upgrade group is obtained based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, location of PCa, PI-RADS v2.1 score; Optionally, the method further includes downgrade prediction, and S2 is replaced by: extracting features based on pathological data and imaging data to obtain feature indicators, and obtaining a downgrade group based on the feature indicators; the feature indicators include one or more of the following: number of PBCs, PCa location. Optionally, the upgrade prediction is replaced by: inputting the pathological data and the imaging data into an upgrade prediction model for prediction to obtain a prediction result, wherein the upgrade prediction model is a model obtained by training with the pathological data set, the imaging data set and the upgrade feature label; Optionally, the degradation prediction is replaced by: inputting the pathological data and the imaging data into a degradation prediction model for prediction to obtain a prediction result, wherein the degradation prediction model is a model obtained by training with a pathological data set, an imaging data set and degradation feature labels; Optionally, the prediction model includes one or more of the following: Logistic regression, random forest, convolutional neural network, ResNet, Inception V1 / V2 / v3, VisionTransformer, XGBoost, FCNs.

7. The method for predicting Gleason score upgrade or downgrade after radical prostatectomy according to claim 1, characterized in that: There are seven categories for upgrades and downgrades, including: Class I upgraded to Class II, Class I upgraded to Class III, Class II upgraded to Class III, Class III downgraded to Class II, Class III downgraded to Class I, Class II downgraded to Class I, and no upgrade / downgrade; Optionally, the category corresponds to M Gleason score levels, and the Gleason score level is determined by adding a primary structure score and a secondary structure score of the Gleason score, and M is a natural number greater than or equal to 1; Optionally, the class I corresponds to level 1, the class II corresponds to level 2, and the class III corresponds to levels 3-5; Optionally, the sum of the primary mode score and the secondary mode score of the 5 levels is 9, including: the primary mode score is 4, the secondary mode score is 5, the primary mode score is 5, the secondary mode score is 4; Optionally, the sum of the primary mode score and the secondary mode score of the level 5 is 10, and the primary mode score is 5 and the secondary mode score is 5.

8. A computer program product having a computer program or instructions thereon, characterized in that: The computer program or instructions are executed by a processor to implement the method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy according to claims 1-7.

9. A computer device comprising a memory, a processor and a computer program or instruction stored in the memory, characterized in that: The computer program or instructions are executed by a processor to implement the method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy according to claims 1-7.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: The computer program or instructions are executed by a processor to implement the method for predicting the upgrade or downgrade of the Gleason score after radical prostatectomy according to claims 1-7.

Images