PURPOSE: The data of Korean men with prostate cancer from a single institute were analyzed to construct nomograms predicting the pathological stage and to compare the outcomes with pre-existing nomograms. MATERIALS AND METHODS: A total of 254 Korean men, with clinically localized prostate cancer, who underwent radical retropubic prostatectomy at Asan Medical Center, between June 1990 and April 2002, were included in this study. A multinomial log-linear regression analysis was performed for the simultaneous prediction of organ-confined disease(OC), seminal vesicle invasion(SVI) or lymph node metastasis(LN) using serum PSA, Gleason scores and clinical stages. Nomograms representing the percentage probabilities were constructed, and compared with the preexisting nomograms presented in the work of Partin et al. and Egawa et al., by calculating the area under the receiver operating characteristics(ROC) curves. RESULTS: Nomograms predicting the likelihood of OC, SVI and LN were derived from the combination of the aforementioned preoperative variables. When the nomograms were compared using the ROC curves, with the Partin table, the areas under the curves were 0.758, 0.762 and 0.766 for OC, SVI and LN, respectively, and with the Egawa table, 0.766 and 0.669 for OC and SVI, respectively. In the multiple measures analysis, which tested the differences between each corresponding data with respect to each preoperative variable, all the tested differences were revealed to be statistically significant. CONCLUSIONS: Comparison of the prediction nomograms revealed notable differences, especially in the OC and SVI. Therefore, it is recommended that each table should be applied to its corresponding population.
Chungcheongnam-do ; Humans ; Lymph Nodes ; Male ; Nomograms* ; Prognosis ; Prostate* ; Prostatectomy ; Prostatic Neoplasms* ; ROC Curve ; Seminal Vesicles

PURPOSE: The aim of this study was to evaluate the applicability of machine learning methods that combine data on age and prostate-specific antigen (PSA) levels for predicting prostate cancer. MATERIALS AND METHODS: We analyzed 943 patients who underwent transrectal ultrasonography (TRUS)-guided prostate biopsy at Chungnam National University Hospital between 2014 and 2018 because of elevated PSA levels and/or abnormal digital rectal examination and/or TRUS findings. We retrospectively reviewed the patients’ medical records, analyzed the prediction rate of prostate cancer, and identified 20 feature importances that could be compared with biopsy results using 5 different algorithms, viz., logistic regression (LR), support vector machine, random forest (RF), extreme gradient boosting, and light gradient boosting machine. RESULTS: Overall, the cancer detection rate was 41.8%. In patients younger than 75 years and with a PSA level less than 20 ng/mL, the best prediction model for prostate cancer detection was RF among the machine learning methods based on LR analysis. The PSA density was the highest scored feature importances in the same patient group. CONCLUSIONS: These results suggest that the prediction rate of prostate cancer using machine learning methods not inferior to that using LR and that these methods may increase the detection rate for prostate cancer and reduce unnecessary prostate biopsy, as they take into consideration feature importances affecting the prediction rate for prostate cancer.
Biopsy ; Chungcheongnam-do ; Digital Rectal Examination ; Forests ; Humans ; Logistic Models ; Machine Learning ; Medical Records ; Prostate ; Prostate-Specific Antigen ; Prostatic Neoplasms ; Retrospective Studies ; Support Vector Machine ; Ultrasonography