RESUMO
OBJECTIVES: To construct and externally validate machine learning-based nomograms for predicting progression stages of initial prostate cancer (PCa) using biomarkers and clinicopathologic features. METHODS: 362 inpatients diagnosed with PCa at the First Affiliated Hospital were randomly assigned to training and testing sets in a 3:7 ratio, while 136 PCa patients from People's Hospital formed the external validation set. Imaging and clinicopathologic information were collected. Optimal features distinguishing advanced prostate cancer (APC) and metastatic PCa (mPCa) were identified through logistic regression (LR). ML algorithms were employed to build and compare ML models. The best-performing algorithm established models for PCa progression stage. Models performance was evaluated using metrics, ROC curves, calibration, and decision curve analysis (DCA) in training, testing, and external validation sets. RESULTS: Following LR analyses, PSA (P=0.001), maximum tumor diameter (P=0.026), Gleason score (P<0.001), and RNF41 (P<0.001) were optimal features for predicting APC, while ALP (P<0.001), PSA (P<0.001), and GS score (P=0.024) were for mPCa. Among ML models, the LR models exhibited superior performance. Consequently, the LR algorithm was used for the APC-risk-nomogram and mPCa-risk-nomogram construction, with AUC values of 0.848, 0.814, 0.810, and 0.940, 0.913, 0.910, in the training, testing, and external validation sets, respectively. Calibration and DCA curves affirmed nomograms' consistency and net benefits for clinical decision-making. CONCLUSIONS: In summary, ML-based APC-risk-nomogram and mPCa-risk-nomogram exhibit outstanding predictive performance for PCa progression stages. These nomograms can assist clinicians in finely categorizing newly diagnosed PCa patients, facilitating personalized treatment plans and prognosis assessment.
RESUMO
BACKGROUND: Predicting the accurate preoperative staging of bladder cancer (BLCA), which markedly affects treatment decisions and patient outcomes, using traditional clinical parameters is challenging. Nevertheless, emerging studies in radiomics, especially machine learning-based computed tomography (CT) image-based radiomics, hold promise in improving stage prediction accuracy in various tumors. However, the comparative performance and clinical utility of models for BLCA are under investigation. PURPOSE: We aimed to investigate the application value of machine learning-based CT radiomics in preoperative staging prediction by comparing the performance of clinical, radiomics, and clinical-radiomics combined models. METHODS: A retrospective cohort of 105 patients with initial BLCA was randomized into training (70%) and testing (30%) cohorts. Radiomics features were extracted from CT images using the optimal feature filter, followed by the application of the least absolute shrinkage and selection operator algorithm for optimum feature selection. Furthermore, machine learning algorithms were used to establish a radiomics model within the training cohort. Independent risk factors for muscle-invasive BLCA (MIBC) obtained by multivariate logistic regression (LR) analysis were separately used to construct a clinical model. For a clinical-radiomics fusion model, radiomics features were combined with clinical parameters. Performance was evaluated based on receiver operating characteristic curves, calibration curves, decision curve analysis (DCA), and standard performance metrics. RESULTS: Patients exhibited a significantly higher age (p = 0.029), larger tumor size (p = 0.01), and an increased neutrophil-to-lymphocyte ratio (NLR; p = 0.045) in the MIBC group than in the NMIBC group. LR analysis revealed age (p = 0.026), tumor size (p = 0.007), and NLR (p = 0.019) as significant predictors for constructing the clinical model. In the testing cohort, the radiomics model, which used an Support Vector Machine classifier, achieved the highest area under the curve (AUC) value of 0.857. The clinical-radiomics model outperformed the remaining two models, with AUC values of 0.958 and 0.893 in the training and testing cohorts, respectively. DeLong's test indicated significant differences between the three models. Calibration curves showed good agreement, and DCA confirmed the superior clinical utility of the clinical-radiomics model. CONCLUSIONS: Machine learning-based CT radiomics combined with clinical parameters was a promising approach in staging BLCA accurately, which outperformed the individual models. Integrating radiomics features with clinical information holds the potential to improve personalized treatment planning and patient outcomes in BLCA.