Introduction: Prostate cancer diagnosis (PCa) has potential for improvement through algorithms combining clinical and paraclinical information. This study aims to develop a PCa diagnostic tree based on data from men over 50 with biopsy indications.
Methods: A cross-sectional study was conducted at the MEDIC Medical Center from 2020 to 2023. Retrospective records of men ≥50 with lower urinary tract symptoms and biopsy indications were analyzed. Patients who were indicated for biopsy due to suspected PCa via digital rectal examination (DRE) or tPSA≥4ng/ml according to Vietnam Ministry of Health guidelines were selected, excluding cases with concurrent other cancers.
Results: The diagnostic tree based on tPSA, DRE, TRUS achieved stable performance on both training and internal validation sets with sensitivity over 90%, specificity over 70%, positive predictive value over 70%, negative predictive value over 95%, and accuracy over 80%. ư
Conclusion: The developed diagnostic tree ensures reliability and practical applicability in clinical settings. Furthermore, TRUS makes a greater contribution than MRI in the diagnostic tree of PCa, and the simultaneous use of TRUS and MRI may not be necessary.