Explainable and visualizable machine learning models to predict biochemical recurrence of prostate cancer.

PURPOSE: Machine learning (ML) models presented an excellent performance in the prognosis prediction. However, the black box characteristic of ML models limited the clinical applications. Here, we aimed to establish explainable and visualizable ML models to predict biochemical recurrence (BCR) of prostate cancer (PCa). MATERIALS AND METHODS: A total of 647 PCa patients were retrospectively evaluated. Clinical parameters were identified using LASSO regression. Then, cohort was split into training and validation datasets with a ratio of 0.75:0.25 and BCR-related features were included in Cox regression and five ML algorithm to construct BCR prediction models. The clinical utility of each model was evaluated by concordance index (C-index) values and decision curve analyses (DCA). Besides, Shapley Additive Explanation (SHAP) values were used to explain the features in the models. RESULTS: We identified 11 BCR-related features using LASSO regression, then establishing five ML-based models, including random survival forest (RSF), survival support vector machine (SSVM), survival Tree (sTree), gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), and a Cox regression model, C-index were 0.846 (95%CI 0.796-0.894), 0.774 (95%CI 0.712-0.834), 0.757 (95%CI 0.694-0.818), 0.820 (95%CI 0.765-0.869), 0.793 (95%CI 0.735-0.852), and 0.807 (95%CI 0.753-0.858), respectively. The DCA showed that RSF model had significant advantages over all models. In interpretability of ML models, the SHAP value demonstrated the tangible contribution of each feature in RSF model. CONCLUSIONS: Our score system provide reference for the identification for BCR, and the crafting of a framework for making therapeutic decisions for PCa on a personalized basis.

[Cuproptosis-related lncRNAs to predict biochemical recurrence of prostate cancer].

OBJECTIVE: To construct a cuproptosis-related lncRNA model and obtain some new ideas and methods for predicting the biochemical recurrence (BCR) of PCa. METHODS: We identified cuproptosis-related lncRNAs from the gene expression data, mutation load data and clinical data on PCa patients in the Cancer Genome Atlas (TCGA) database and divided the patients into a training group and a verification group. We constructed a prognostic risk scoring model based on the cuproptosis -related lncRNAs, verified the validity of the model by BCR-free survival analysis, logistic regression analysis and independent prognosis analysis, and visualized the results using ROC curve analysis, Kaplan-Meier survival curves and the correlation heat map. We performed differential analysis and survival analysis of the tumor mutation burden (TMB), and assessed the value of the model and TMB in predicting the BCR of PCa. RESULTS: A prognostic risk scoring model was successfully constructed based on the 6 cuproptosis -related lncRNAs identified from the PCa cases in the training group, which were divided into a high- and a low-risk groups according to the median value. The incidence of BCR rose with the increase of the risk score, and the BCR-free time was significantly shorter in the high-risk group (P < 0.05). The model also exhibited a high differentiation value in different age groups (P < 0.05), which was shown to be a reliable and independent prognostic indicator for predicting the BCR of PCa, even more valuable than other clinicopathological indicators. TMB was differentially expressed in the high- and low-risk groups (P < 0.01) and significantly correlated with BCR. The highest rate of BCR-free survival was found in the patients with low risk scores and low TMB (P < 0.01). CONCLUSION: A cuproptosis -related lncRNA model was successfully constructed, which can accurately predict the risk of BCR in PCa patients. The higher the prognostic risk score, the greater the possibility of BCR. TMB is high in patients with a high risk, and the TMB level has certain suggestive significance for BCR.

An anoikis-related gene signature for prediction of the prognosis in prostate cancer.

Purpose: This study presents a novel approach to predict postoperative biochemical recurrence (BCR) in prostate cancer (PCa) patients which involves constructing a signature based on anoikis-related genes (ARGs). Methods: In this study, we utilised data from TCGA-PARD and GEO databases to identify specific ARGs in prostate cancer. We established a signature of these ARGs using Cox regression analysis and evaluated their clinical predictive efficacy and immune-related status through various methods such as Kaplan-Meier survival analysis, subject work characteristics analysis, and CIBERSORT method. Our findings suggest that these ARGs may have potential as biomarkers for prostate cancer prognosis and treatment. To investigate the biological pathways of genes associated with anoikis, we utilised GSVA, GO, and KEGG. The expression of ARGs was confirmed by the HPA database. Furthermore, we conducted PPI analysis to identify the core network of ARGs in PCa. Results: Based on analysis of the TCGA database, a set of eight ARGs were identified as prognostic signature genes for prostate cancer. The reliability and validity of this signature were well verified in both the TCGA and GEO codifications. Using this signature, patients were classified into two groups based on their risk for developing BCR. There was a significant difference in BCR-free time between the high and low risk groups (P < 0.05).This signature serves as a dependable and unbiased prognostic factor for predicting biochemical recurrence (BCR) in prostate cancer (PCa) patients. It outperforms clinicopathological characteristics in terms of accuracy and reliability. PLK1 may play a potential regulatory role as a core gene in the development of prostate cancer. Conclusion: This signature suggests the potential role of ARGs in the development and progression of PCa and can effectively predict the risk of BCR in PCa patients after surgery. It also provides a basis for further research into the mechanism of ARGs in PCa and for the clinical management of patients with PCa.

Artificial intelligence for the diagnosis of clinically significant prostate cancer based on multimodal data: a multicenter study.

BACKGROUND: The introduction of multiparameter MRI and novel biomarkers has greatly improved the prediction of clinically significant prostate cancer (csPCa). However, decision-making regarding prostate biopsy and prebiopsy examinations is still difficult. We aimed to establish a quick and economic tool to improve the detection of csPCa based on routinely performed clinical examinations through an automated machine learning platform (AutoML). METHODS: This study included a multicenter retrospective cohort and two prospective cohorts with 4747 cases from 9 hospitals across China. The multimodal data, including demographics, clinical characteristics, laboratory tests, and ultrasound reports, of consecutive participants were retrieved using extract-transform-load tools. AutoML was applied to explore potential data processing patterns and the most suitable algorithm to build the Prostate Cancer Artificial Intelligence Diagnostic System (PCAIDS). The diagnostic performance was determined by the receiver operating characteristic curve (ROC) for discriminating csPCa from insignificant prostate cancer (PCa) and benign disease. The clinical utility was evaluated by decision curve analysis (DCA) and waterfall plots. RESULTS: The random forest algorithm was applied in the feature selection, and the AutoML algorithm was applied for model establishment. The area under the curve (AUC) value in identifying csPCa was 0.853 in the training cohort, 0.820 in the validation cohort, 0.807 in the Changhai prospective cohort, and 0.850 in the Zhongda prospective cohort. DCA showed that the PCAIDS was superior to PSA or fPSA/tPSA for diagnosing csPCa with a higher net benefit for all threshold probabilities in all cohorts. Setting a fixed sensitivity of 95%, a total of 32.2%, 17.6%, and 26.3% of unnecessary biopsies could be avoided with less than 5% of csPCa missed in the validation cohort, Changhai and Zhongda prospective cohorts, respectively. CONCLUSIONS: The PCAIDS was an effective tool to inform decision-making regarding the need for prostate biopsy and prebiopsy examinations such as mpMRI. Further prospective and international studies are warranted to validate the findings of this study. TRIAL REGISTRATION: Chinese Clinical Trial Registry ChiCTR2100048428. Registered on 06 July 2021.

[Difference between prostate cancer patients' Gleason scores from preoperative biopsy and those from postoperative pathology].

OBJECTIVE: To evaluate the consistency of the Gleason scores of PCa patients based on preoperative biopsy with those from postoperative pathology, identify the possible factors influencing results of scoring, and construct a risk scoring model. METHODS: We collected the demographic and clinical data on the patients with PCa confirmed by preoperative prostate biopsy or postoperative pathology and treated by radical prostatectomy within 6 months after diagnosis. Using paired sample t-test, we identified the difference between the Gleason scores based on preoperative biopsy and those from postoperative pathology, analyzed the demographic and clinical data on the patients for relevant factors affecting the consistency of the Gleason scores, and calculated and visualized the relative risk values of the factors through Poisson regression. From the continuous variables with statistical significance, we screened independent risk factors for the difference in the Gleason scores by Lasso regression analysis, established a risk scoring model, generated risk coefficients, and evaluated the predictive ability of the model using the ROC curve. Based on the results of imaging examination with statistically significant differences, we constructed a column chart by logistic regression and evaluated the predictive validity of the chart using calibration curves, decision curves and ROC curves. RESULTS: The results of paired sample t-test for 210 PCa patients showed statistically significant differences between the Gleason scores from preoperative biopsy and those from postoperative pathology (P < 0.001). There were significant differences in the body weight, BMI and PSA level as well as in all other factors but prostate calcification between the patients with consistent and those with inconsistent Gleason scores (all P < 0.05). An 8-factor prediction model was successfully constructed, which could predict the consistency of Gleason scores, with a better predicting performance than the single indicator within the model. The nomogram exhibited a C-index value of 0.85, with the calibration curve similar to the standard one, the threshold of the decision curve 0.10－0.92, and the area under the ROC curve higher than other predictive indicators. CONCLUSION: Based on the demographic and clinical data on PCa patients, a risk prediction model and a column chart were successfully constructed, which could effectively predict the difference between the Gleason scores from preoperative prostate biopsy and those from postoperative pathology.

Morphological characteristics of amino acids in wet deposition of Danjiangkou Reservoir in China's South-to-North Water Diversion Project.

Amino acids are an important constituent in organic nitrogen deposition, and changes in the content of their components have a direct impact on the nitrogen input to the ecosystem. From December 2018 to November 2019, 176 precipitation samples were collected at Danjiangkou Reservoir, the source of the middle line of the South-to-North Water Diversion Project, and the variation characteristics of dissolved free amino acids (DFAA) and dissolved combined amino acids (DCAA) were analyzed. The volume-weighted value concentration ranges of DFAA and DCAA were 0.159-1.136 µmol/L and 1.603-7.044 µmol/L, respectively, and amino acids were dominated by DCAA in wet deposition. Our results showed that glutamic acid (Glu), glycine (Gly), and aspartic acid (Asp) were the dominant amino acids in both DFAA and DCAA. The concentration of DFAA was highest in winter, while the concentration of DCAA was in autumn. Dissolved total amino acids (DTAA) were insignificantly correlated with DFAA, whereas they were linearly correlated with DCAA, indicating a significant influence of agricultural activities on DTAA. The analysis of the backward trajectory of air masses showed that amino acids were mainly influenced by proximity inputs around the reservoir. The bioavailability of organic matter was higher in the southeastern of the reservoir than in the northwestern. The wet deposition flux of TDN was 14.096 kg N/ha/year, and the potential ecological impact on water bodies cannot be ignored. This study was conducted to clarify the variation characteristics of amino acids fractions in wet deposition and to provide parameters for regional assessment of amino acids wet deposition. The ecological impact of nitrogen wet deposition on water bodies will be explored to provide a basis for nitrogen pollution control and water quality protection in the middle line of the South-to-North Water Diversion Project.

[miR-129 in the diagnosis, treatment and prediction of clinical prognosis of prostate cancer].

OBJECTIVE: To explore the value of miR-129 in the diagnosis, treatment and prediction of the clinical prognosis of PCa by observing the correlation between miR-129 and the progression of the malignancy. METHODS: This retrospective analysis included 310 male patients who visited the Department of Urology of the General Hospital of Eastern Theater Command from January 2014 to January 2022, 80 as normal healthy men, 80 with BPH, and the other 150 with PCa treated by radical prostatectomy without chemotherapy, radiotherapy or androgen-deprivation therapy. We determined the miR-129 expression in the serum and prostatic tissue of all the subjects by real-time quantitative PCR (RT-qPCR), performed pathological grading of the primary cancerous and pericancerous (≥ 3 cm from the focus) prostate tissues from the 150 PCa patients, collected the demographic and clinical data on all the subjects, and analyzed the correlation of their demographic data with the clinical parameters. We selected and transfected PC-3 and DU-145 cell lines with miR-129 precursor or miR-129 scramble, assessed the proliferation ability of each cell line and detected the expression levels of related proteins by cell proliferation assay and Western blot. RESULTS: The expression of miR-129 was significantly decreased in the serum of the PCa patients compared with that in the normal healthy men and BPH patients (P < 0.01), so was it in the PCa tissue in comparison with that in the pericancerous prostatic tissue (P < 0.01), and it was negatively correlated with the preoperative serum PSA level (P < 0.001), histological grade (P < 0.001), pathological stage (P < 0.001), Gleason score (P < 0.001), lymph node metastasis (P = 0.02), angiolymphatic invasion (P = 0.021) and biochemical recurrence (BCR) (P=0.001). Kaplan-Meier analysis showed a strong correlation of down-regulated miR-129 expression with a lower BCR-free survival rate, while multivariate survival analysis indicated that the expression of miR-129 was an independent prognostic indicator of BCR-free survival of the PCa patients (P < 0.001). Highly expressed miR-129 significantly inhibited the proliferation of PCa cells by regulating the expressions of cell cycle-regulatory proteins. CONCLUSION: The expression of miR-129 is significantly down-regulated in PCa tissues, which plays an important role in inhibiting the proliferation of PCa cells and tumor progression and improving BCR-free survival. Therefore, miR-129 can be considered as a new independent biomarker for the diagnosis, treatment and prediction of the clinical prognosis of PCa.

[Basement membrane-related gene signature to predict biochemical recurrence in patients with prostate cancer after radical prostatectomy].

OBJECTIVE: To construct and verify a key gene signature of the basement membrane of prostate cancer (PCa) to predict the progression and biochemical recurrence of the malignancy after radical prostatectomy. METHODS: Based on the PCa-related transcriptome, gene mutation and clinical data from the Cancer Genome Atlas Project (TCGA) database, we analyzed the differentially expressed genes (DEG) related to the basement membrane in the PCa and adjacent normal prostate tissues, and subjected them to GO function enrichment and KEGG pathway enrichment analyses. We identified prognosis-related genes from the DEGs and analyzed their mutations. According to the follow-up data and biochemical recurrence after prostatectomy, we established a prognostic risk scoring model, verified its accuracy using the Gene Expression Omnibus (GEO) database, and performed survival analysis, principal component analysis (PCA), independent prognostic analysis and ROC curve analysis of the model. We constructed a protein-protein interaction network after verifying the correctness of the model by immunohistochemistry. We also established a nomogram and tested its accuracy using ROC and calibration curves. RESULTS: Totally, 85 DEGs were identified, among which 18 were up-regulated and 67 down-regulated. The prognostic risk scoring model was established with 11 of the genes. The risk of biochemical recurrence PCa was significantly higher in the high-risk than in the low-risk group (HR: 3.51, 95% CI: 2.32－5.32, P < 0.01), which was verified with the GEO database data (P < 0.01). In addition, the patients in the high-risk group were older with higher clinical T-stage, higher Gleason score, higher positive rate, larger numbers of positive lymph nodes, and a larger proportion of residual tumors than those in the low-risk group (P < 0.05). The nomogram constructed with the patients' age, pN, pT and cT stages, Gleason score and prognostic risk score manifested that the area under the ROC curve was higher than the other predictors. The calibration chart showed consistency of the predicted outcomes to the actual results. CONCLUSION: A prognostic risk scoring model of basement membrane-related genes and an effective nomogram were successfully constructed, which can predict the risk of biochemical recurrence in PCa patients after radical prostatectomy.

[Erectile function of renal failure patients after hemodialysis or kidney transplantation: A comparative study].

OBJECTIVE: To study the changes in the erectile function of the male patients with renal failure after hemodialysis (HD) or kidney transplantation (KT) and explore the causes of these changes. METHODS: From January 2015 to January 2021, 160 male patients with renal failure complaining of ED underwent HD (n = 80) or KT (n = 80) in the General Hospital of Eastern Theater Command. The patients were aged 25－45 (31.7 ± 4.8) years, 32 ± 4.5 years in the HD group and 31.4 ± 5.1 years in the KT group. We recorded the levels of serum T, E2, FSH and LH and the scores on IIEF-5, Self-Rating Anxiety Scale (SAS) and Self-Rating Depression Scale (SDS) of the patients, and compared them between the two groups. RESULTS: Compared with the patients in the HD group, those in the KT group showed a significantly higher T level (ï¼»7.45 ± 3.54ï¼½ vs ï¼»17.75 ± 7.32ï¼½ nmol/L, P < 0.01) and a lower E2 level (ï¼»151.37 ± 20.89ï¼½ vs ï¼»94.17 ± 40.79ï¼½ pmol/L, P < 0.01), but no statistically significant difference from the former group in the levels of FSH (ï¼»8.12 ± 5.12ï¼½ vs ï¼»8.97 ± 2.36ï¼½ IU/L, P > 0.05) and LH (ï¼»5.16 ± 3.87ï¼½ vs ï¼»4.69 ± 2.18ï¼½ IU/L, P > 0.05). There were fewer cases of severe ED in the KT than in the HD group (3.75% vs 16.25%, P < 0.05). Different degrees of anxiety and depression were observed in both groups, with fewer severe cases of anxiety (6.25% vs 30.00%, P < 0.05) and depression (6.25% vs 31.25%, P < 0.05) and more mild cases of anxiety (68.75% vs 47.50%, P < 0.05) and depression (70.00% vs 48.75%, P < 0.05) in the KT than in the HD group, but no statistically significant difference in the incidence of moderate anxiety (25.00% vs 22.50%, P > 0.05) and depression (23.75% vs 20.00%, P > 0.05) between the KT and HD groups. CONCLUSION: For male patients with renal failure, kidney transplantation can evidently improve erectile function, while hemodialysis has a poorer effect. The altered hormone levels, anxiety and depression of the patients are important causes of the changes in their erectile function.

[Prostate-specific membrane antigen-targeted drug delivery system for prostate cancer].

Prostate cancer (PCa) is a most common malignancy in the genitourinary system, which imposes a huge burden on the patients and society. Various traditional medication methods fail to achieve satisfactory effects, and therefore it is imperative to develop innovative and effective routes of administration. Drug delivery systems have the characteristics of both delivery and controlled release and are widely used clinically. Prostate-specific membrane antigen (PSMA) is one of the most characteristic and highly selective biomarkers of PCa with a good PCa-targetability. Many drug delivery systems targeting PSMA have been explored, including drug-ligand conjugates and nano-drug delivery systems (polymer nanoparticles, inorganic nanoparticles, liposomes, and biological nanoparticles, etc.). This article reviews the recent studies on the role of the PSMA-targeting drug delivery system in PCa, in order to provide some reference for the precise and effective treatment of the malignancy.

[Trace elements and sperm quality: An update].

With changes in the modes of life and work, infertility is becoming a serious international problem. The incidence of infertility is 10－15% of the couples of childbearing age in China and is increasing year by year, and 30－50% of the cases are related to male factors. Trace elements refer to the chemical elements that account for less than 1/10,000 of the total weight of the human body, of which less than 100 mg is required per person per day. They are important biochemical components in the semen. Trace elements are essential for normal spermatogenesis, sperm maturation and capacitation, and maintenance of normal sperm function, though sperm quality can be affected by genetic, epigenetic, environmental and lifestyle-related factors. This focuses on the latest progress in the studies of the effects of trace elements on sperm quality.