Machine learning technology is revolutionizing the field of medical research, with a recent study focusing on predicting Gleason grade group upgrades in prostate cancer patients. The study, published in Scientific Reports, analyzed clinical variables of patients who underwent MRI-targeted in-bore biopsy followed by radical prostatectomy to identify factors influencing Gleason grade upgrades in final pathology reports.
The retrospective study, conducted at American Hospital in Istanbul, Turkey, included data from 400 men who underwent mp-MRI and subsequent MRI-targeted biopsies between 2012 and 2022. Among the patients diagnosed with prostate cancer, 95 underwent radical prostatectomy as a definitive treatment, with 20 initially diagnosed with Gleason grade group 1 (GG1) prostate cancer.
The researchers aimed to determine if machine learning methods could predict Gleason grade group upgrades based on individual patient data. Clinical variables such as family history of prostate cancer, International Prostate Symptom Score, tumor positivity, and lesion size were considered in the analysis. The time interval between biopsy and surgery was less than 6 months for most patients, and none had received prior radiotherapy or hormone therapy.
MRI examinations were conducted using a 3.0 Tesla scanner, with radiologists interpreting tumor characteristics and ADC values. In-bore biopsies were performed using MRI guidance, with cores obtained from high-likelihood lesions identified on mp-MRI. The study focused on index lesions, with cores sampled based on pre-biopsy imaging findings.
Univariate statistical tests and multivariate machine learning analyses were performed to identify predictive clinical parameters for Gleason grade group upgrades. Baseline accuracy was established through comparison of biopsy and surgical pathology results, with machine learning models including support vector machine, LASSO regression, and ridge regression.
Overall, the study provides valuable insights into the potential of machine learning in predicting Gleason grade group upgrades in prostate cancer patients. The findings highlight the importance of integrating advanced technology into personalized treatment planning for improved patient outcomes.
