Diagnostic systematic review and meta-analysis of machine learning in predicting biochemical recurrence of prostate cancer

Abstract

Prostate cancer (PCa) is the most prevalent malignant tumor in males, and many patients remain at risk of biochemical recurrence (BCR) following initial treatment. Accurate prediction of BCR is vital for effective clinical management and treatment planning. This study evaluates the effectiveness of machine learning (ML) models in predicting BCR among prostate cancer patients, comparing their performance to traditional prognostic methods. We systematically searched four databases (PubMed, Web of Science, Embase, and Cochrane) for studies employing ML techniques to predict prostate cancer BCR. Data extraction included model type, sample size, and the area under the curve (AUC). A meta-analysis was conducted using AUC as the primary performance metric to assess predictive accuracy and heterogeneity across models. Sixteen studies comprising a total of 17,316 prostate cancer patients were included. The pooled AUC for ML models was 0.82 (95% CI: 0.81-0.84). Deep learning and hybrid models outperformed traditional models (AUC = 0.83). Models using imaging data showed improved performance (AUC = 0.82). ML models were most effective in predicting 1-year BCR (AUC = 0.86), with performance slightly decreasing for longer time intervals. ML models outperform traditional methods in predicting BCR, especially in the short term. Incorporating multimodal data, such as imaging, enhances predictive accuracy. Future studies should optimize and validate these models through large-scale clinical trials.

Fig. 1

Literature screening process.

Fig. 2

(a) The chart shows the prevalence of different model types, with Non-Deep Learning Models at 76.8%, Deep Learning Models at 18.3%, and ML + DL Models at 4.9%. (b) This chart highlights data type usage, with Hybrid data types at 85.4%, Single data types at 14.6%, and a mix of other types. (c) The chart displays time intervals used in the study, with 1-Year at 15.9%, 2-Year at 14.6%, 3-Year at 35.4%, 5-Year at 29.3%, and NA at 4.9%. ID imaging data, CLD clinical data, PD pathological data.

Fig. 3

This figure shows the predictive performance (AUC) of machine learning models from various studies. The table lists the study details, and the forest plot on the right illustrates the AUC values with 95% confidence intervals.

Fig. 4

This figure shows the summarized area under the curve (AUC) values with 95% confidence intervals for single and hybrid data types across various studies.

Fig. 5

This figure shows the summarized area under the curve (AUC) values with 95% confidence intervals for studies using imaging data compared to other data types.

Fig. 6

This figure shows the analysis results of the area under the curve (AUC) values with 95% confidence intervals for non-deep learning models, deep learning models, and ML + DL models across various studies.

Fig. 7

This figure shows the predictive performance (AUC with 95% CI) of various models across different time intervals (1-year, 2-year, 3-year, and 5-year).

Fig. 8

This funnel plot visualizes the distribution of studies included in the meta-analysis.