Modifications to prostate cancer diagnosis following COVID-19 and following models.

Abstract

Introduction: The COVID-19 pandemic has impacted the treatment of prostate cancer (PCa). The study examines any predictions that could point to future models.

Methods: Two interrupted time series analyses were conducted: one for the pre-COVID period (January 2017 to December 2019) and another for the post-COVID period during 2022. Information on age, total prostate-specific antigen (PSA), abnormal digital rectal exam (DRE), prostate volume (PV), previous negative biopsy, number of positive biopsies, Gleason score, and biopsy outcome were collected for all patients. The categories for the results were no cancer, insignificant, low and intermediate, high-risk, and very high-risk PCa. Using a generalized linear model (GLM), the outcomes are modeled. The area under the curve (AUC) and accuracy were used to assess how well multi-class predictions performed.

Results: 244 patients who had biopsies following the COVID-19 pandemic and 832 patients who had biopsies before the pandemic were compared. The accuracy of the GLM model was only 0.635. The AUC for category no cancer, low and intermediate-risk, and very high-risk patients was 0.821, 0.716, and 0.926. With scaled relevance values, PSA was the most critical test. The two features that significantly influenced the treatment model prediction for PCa were biopsy PSA level and DRE, respectively.

Conclusion: Advanced age and a very high-risk group appear to have a detrimental impact on the results of biopsies conducted after the first wave of the COVID-19 era. At the same time, PSA levels and abnormal DRE are the most significant predictors in GLM.