1. Predictive model of castration resistance in advanced prostate cancer by machine learning using genetic and clinical data: KYUCOG-1401-A study
- Author
-
Masaki Shiota, Shota Nemoto, Ryo Ikegami, Shuichi Tatarano, Toshiyuki Kamoto, Keita Kobayashi, Hideki Sakai, Tsukasa Igawa, Tomomi Kamba, Naohiro Fujimoto, Akira Yokomizo, Seiji Naito, and Masatoshi Eto
- Subjects
Neoplasms. Tumors. Oncology. Including cancer and carcinogens ,RC254-282 - Abstract
Abstract Background The predictive power of the treatment efficacy and prognosis in primary androgen deprivation therapy (ADT) for advanced prostate cancer is not satisfactory. The objective of this study was to integrate genetic and clinical data to predict castration resistance in primary ADT for advanced prostate cancer by machine learning (ML). Methods Clinical and single nucleotide polymorphisms (SNP) data obtained in the KYUCOG-1401-A study (UMIN000022852) that enrolled Japanese patients with advanced prostate cancer were used. All patients were treated with primary ADT. A point-wise linear (PWL) algorithm, logistic regression with elastic-net regularization, and eXtreme Gradient Boosting were the ML algorithms used in this study. Area under the curve for castration resistance and C-index for prognoses were calculated to evaluate the utility of the models. Results Among the three ML algorithms, the area under the curve values to predict castration resistance at 2 years was highest for the PWL algorithm with all the datasets. Three predictive models (clinical model, small SNPs model, and large SNPs model) were created by the PWL algorithm using the clinical data alone, and 2 and 46 SNPs in addition to clinical data. C-indices for overall survival by the clinical, small SNPs, and large SNPs models were 0.636, 0.621, and 0.703, respectively. Conclusion The results demonstrated that the SNPs models created by ML produced excellent prediction of castration resistance and prognosis in primary ADT for advanced prostate cancer, and will be helpful in treatment choice.
- Published
- 2024
- Full Text
- View/download PDF