Using machine learning to develop preoperative model for lymph node metastasis in patients with bladder urothelial carcinoma.

BACKGROUND: Lymph node metastasis (LNM) is associated with worse prognosis in bladder urothelial carcinoma (BUC) patients. This study aimed to develop and validate machine learning (ML) models to preoperatively predict LNM in BUC patients treated with radical cystectomy (RC). METHODS: We retrospectively collected demographic, pathological, imaging, and laboratory information of BUC patients who underwent RC and bilateral lymphadenectomy in our institution. Patients were randomly categorized into training set and testing set. Five ML algorithms were utilized to establish prediction models. The performance of each model was assessed by the area under the receiver operating characteristic curve (AUC) and accuracy. Finally, we calculated the corresponding variable coefficients based on the optimal model to reveal the contribution of each variable to LNM. RESULTS: A total of 524 and 131 BUC patients were finally enrolled into training set and testing set, respectively. We identified that the support vector machine (SVM) model had the best prediction ability with an AUC of 0.934 (95% confidence interval [CI]: 0.903-0.964) and accuracy of 0.916 in the training set, and an AUC of 0.855 (95%CI: 0.777-0.933) and accuracy of 0.809 in the testing set. The SVM model contained 14 predictors, and positive lymph node in imaging contributed the most to the prediction of LNM in BUC patients. CONCLUSIONS: We developed and validated the ML models to preoperatively predict LNM in BUC patients treated with RC, and identified that the SVM model with 14 variables had the best performance and high levels of clinical applicability.

Construction of a Model for Predicting the Risk of pT3 Based on Perioperative Characteristics in cT1 Renal Cell Carcinoma: A Retrospective Study at a Single Institution.

INTRODUCTION: This study explored the predictors of upstaging and multiple sites of extension, and constructed a predictive model based on perioperative characteristics to calculate the risk of upstaging of cT1 renal cell carcinoma to pT3. METHODS: We retrospectively reviewed 1012 patients diagnosed with cT1 renal cell carcinoma who underwent surgical treatment at the Affiliated Hospital of Qingdao University between June 2016 and August 2021. The continuous and categorical variables were analyzed using the Mann-Whitney U test and Chi-square test, respectively. After randomly dividing patients into a training set and an internal validation set with a ratio of 7:3, univariate and multivariate logistic regression analyses were used to explore the predictors of upstaging and multiple sites of extension. A nomogram model was established based on the predictors of upstaging and was validated. RESULTS: Ninety-one cases (8.99%) of renal cell carcinoma were upstaged to pT3. In the training set, multivariate logistic regression identified the following predictors of upstaging: maximum tumor diameter, hilus involvement, tumor necrosis, tumor edge irregularity, symptoms, smoking, and platelet-lymphocyte ratio. A nomogram model was established based on the predictors. The area under the receiver operating characteristic curve was 0.810 in the training set, and 0.804 in the validation set. A 10-fold internal cross-validation conducted 200 times showed that the mean area under the curve was 0.797. The calibration curve and decision curve analysis suggested that the nomogram had robust clinical predictive power. Analyses showed higher neutrophil-lymphocyte ratio and tumor necrosis were associated with multiple sites of extrarenal extension in patients with pT3a renal cell carcinoma. CONCLUSIONS: We identified 7 predictors of upstaging to pT3 and 2 predictors of multiple sites of extension. A nomogram model was constructed with satisfactory accuracy for predicting upstaging to pT3.

External validation of a four-tiered grading system for chromophobe renal cell carcinoma.

This study aimed to validate the prognostic value of a four-tiered grading system recently proposed by Avulova et al. and to explore the prognostic ability of another four-tiered classification grading system in which there is a separate Grade 3 for tumor necrosis. Grading of chromophobe renal cell carcinoma (ChRCC) by the Fuhrman system is not feasible because of the inherent nuclear atypia in ChRCC. We collected relevant data of 263 patients with ChRCC who had undergone surgery in our hospital from 2008 to 2020. The Kaplan-Meier method was used to calculate the survival rate and Cox proportional hazard regression models to assess associations with cancer-specific survival and distant metastasis-free survival by hazard ratios (HRs) and 95% confidence intervals (CIs). Ten patients died from ChRCC, and 12 developed metastases. The 5 year CSS rates were 95.9%. Grades 2 (HR = 10.9; CI 1.11-106.4; P = 0.04), 3 (HR = 33.6, CI 3.32-339.1; P = 0.003), and 4 (HR = 417.4, CI 35.0-4976.2; P < 0.001) in a four-tiered grading system were significantly associated with CSS in a multivariate setting. However, the difference in CSS between Grades 2 and 3 was not significant (HR = 2.14, 95% CI 0.43-10.63; P = 0.35). The HRs of the associations between an exploratory grading system that includes a separate Grade 3 for tumor necrosis and CSS were as follows: Grade 2, 10.2 (CI 1.06-97.9, P = 0.045); Grade 3, 11.4 (CI 1.18-109.6, P = 0.04); and Grade 4, 267.9 (CI 27.6-2603.3, P < 0.001). Similarly, Grades 2 and 3 did not differ significantly. The four-tiered grading system studied is useful for predicting death from ChRCC and metastasis. However, Grade 3 did not more accurately predict risk of death and metastasis than did Grade 2. This was also true for the novel exploratory grading system that classifies tumors with necrosis into a separate Grade 3.

Identification and validation of a ferroptosis-related signature for prediction of the prognosis and tumor microenvironment in patients with chromophobe renal cell carcinoma.

BACKGROUND: Ferroptosis is a novel form of regulated cell death that is different from other forms, which has an important role in tumor growth inhibition. The purpose of this study was to construct and validate a prognostic signature related to ferroptosis in chromophobe renal cell carcinoma (ChRCC) and to explore its role in immune cell infiltration and systemic therapy. METHODS: The gene expression profiles of ChRCC patients obtained from The Cancer Genome Atlas (TCGA) database were used to identify differentially expressed prognostic ferroptosis-related genes (FRGs) by univariate Cox proportional hazards analyses. Ferroptosis molecular subtypes were obtained by consensus clustering analysis. The FRG-based signature in the training set was established by least absolute shrinkage and selection operator analysis and verified in the testing set. The association between molecular subtypes and the prognostic signature and immune microenvironment was explored to predict responses to immunotherapy. Immunohistochemistry was used to verify expression of the FRG-based signature externally. RESULTS: ChRCC patients were divided into two FRG subtypes. Two FRGs (TFRC and SLC7A11) were identified to construct the prognostic signature. The high-risk group and cluster 2 had worse overall survival than the low-risk group and cluster 1, respectively. The low-risk group and cluster 1 had higher levels of immune cell infiltration and expression of MHC and immune checkpoint molecules than the high-risk group and cluster 2. The risk score was a predictor of overall survival and had a good predictive ability, which was verified in the testing set and evaluated by ROC and calibration curves. The high-risk group had a higher tumor mutation burden. The different sensitivities of targeted drugs in patients with different risks were evaluated. External immunohistochemical analysis showed that TFRC and SLC7A11 were highly expressed in tumor tissues compared with para-cancer normal tissues, and the expression level was significantly associated with a more advanced stage and worse cancer-specific survival. CONCLUSIONS: An FRG signature was identified and validated to predict the clinicopathological features and prognosis of ChRCC. A significant association between the signature and immune cell infiltration, immune checkpoint expression, and drug response is helpful to guide comprehensive treatment of ChRCC.

Systemic therapies for high-volume metastatic hormone-sensitive prostate cancer: a network meta-analysis.

BACKGROUND: We compared the effectiveness of currently available systemic therapies for high-volume metastatic hormone-sensitive prostate cancer (mHSPC) and aimed to establish the optimal treatment regimen. MATERIAL AND METHODS: We searched multiple databases for randomized controlled trials (RCTs) that evaluated the efficacy of systemic therapy in patients with high-volume mHSPC. Bayesian network meta-analysis was used to indirectly compare overall survival (OS) and progression-free survival (PFS) of various systemic therapies. RESULTS: Eleven RCTs (6708 participants) finally met the eligibility criteria. Compared with androgen deprivation therapy (ADT) alone, rezvilutamide (REZ) [hazard ratio (HR) = 0.58, 95% confidence interval (CI): 0.44-0.77], abiraterone (ABI) (HR = 0.61, 95% CI: 0.53-0.71), apalutamide (APA) (HR = 0.70, 95% CI: 0.56-0.88), enzalutamide (ENZ) (HR = 0.65, 95% CI: 0.53-0.80), docetaxel (DOC) (HR = 0.72, 95% CI: 0.63-0.84), darolutamide (DAR) + DOC (HR = 0.49, 95% CI: 0.39-0.62), and ABI + DOC (HR = 0.52, 95% CI: 0.38-0.71) significantly improved OS in patients with high-volume mHSPC. Compared with DOC, no advantages were observed for doublet therapies, including REZ, ABI, APA, and ENZ on the basis of ADT, whereas DAR + DOC (HR = 0.68, 95% CI: 0.57-0.82) and ABI + DOC (HR = 0.72, 95% CI: 0.55-0.95) was associated with better OS. The ranking analysis showed that triplet therapy (DAR + DOC + ADT and ABI + DOC + ADT) had the greatest improvement in OS, followed by REZ + ADT. All the regimens showed improved PFS in patients with high-volume mHSPC. Compared with DOC, significant differences were detected for DAR + DOC, ABI + DOC, ENZ + DOC, REZ, and ENZ. According to the ranking analysis, triplet therapy ranked first, followed by ENZ and REZ. CONCLUSIONS: REZ + ADT were the highest ranked doublet therapy for improvement in OS of patients with high-volume mHSPC, second only to triplet therapy (DAR + DOC + ADT and ABI + DOC + ADT).