1278 - Internally Validated Survival Nomograms for Recurrence, Metastasis, and Overall Survival after Salvage Prostate Brachytherapy
Presenter(s)
M. Bilski1,2, G. Salfi3, A. Gomez De Iturriaga4,5, P. Saniewski6, M. Stankiewicz7, D. Buchser8, F. Mastroleo9,10, T. Soeterik11, T. Surowiak12, B. A. Jereczek-Fossa13,14, T. Zilli15,16, and M. Pedrani17; 1Affidea Nu-med Center of Oncology Diagnostics and Therapy, Zamosc, Poland, 2Department of Radiotherapy, Medical University of Lublin, Lublin, Poland, 3Oncology Institute of Southern Switzerland, Ente Ospedaliero Cantonale, Bellinzona, Switzerland; Institute of Oncology Research, Bellinzona, Switzerland, 4Radiation Oncology Department, Hospital Universitario Cruces, BioBizkaia, Basque Health Research, Baracaldo, Spain, 5Cruces Unisersity Hospital / Biocruces Health Research Institute, Spain, Bilbao, Spain, 6Radiotherapy Department. Maria Sklodowska-Curie Institute-Oncology Centre, Warsaw, Poland, 7Brachytherapy Department, Maria Sklodowska-Curie National Research Institute of Oncology, Gliwice, Poland, 8Hospital Universitario de Cruces, Barakaldo, Spain, 9Division of Radiation Oncology, IEO, European Institute of Oncology, IRCCS, Milan, Italy, 10Department of Radiation Oncology, Mayo Clinic, Rochester, MN, 11Department of Radiation Oncology, University Medical Center Utrecht, Utrecht, Netherlands, 12Radiotherapy Department, Affidea Nu-med Center of Oncological Diagnostics and Therapy, Zamosc, Poland, 13Radiation Oncology, IEO European Institute of Oncology IRCCS, Milan, Italy, 14Oncology and Hemato-Oncology, University of Milan, Milan, Italy, 15Department of Radiation Oncology, Oncology Institute of Southern Switzerland, Ente Ospedaliero Cantonale, Bellinzona, Switzerland, 16Radiation Oncology, Geneva University Hospital, Geneva, Switzerland, 17Oncology Institute of Southern Switzerland, Ente Ospedaliero Cantonale, Bellinzona, Switzerland
Purpose/Objective(s):
Salvage brachytherapy is a potentially curative option for locally radiorecurrent prostate cancer, but long-term outcomes such as overall survival, recurrence-free survival, and metastasis-free survival remain highly heterogeneous. Most existing prognostic models have been developed for primary external beam radiotherapy or other salvage modalities and are not tailored to salvage brachytherapy. There are currently no internally validated, survival models, limiting individualized risk stratification.Materials/Methods: We analyzed 236 patients treated with salvage brachytherapy after prior radiotherapy in a multiinstitutional cohort. Time-to-event endpoints included OS, RFS, and MFS. A prespecified set of candidate predictors comprised PET status, androgen deprivation duration (ADT), ISUP category, age, time from first radiotherapy to relapse, log PSA at recurrence, postprimary PSA nadir =0.2, and radiotherapy descriptors (primary fractionation/type, volume category, CTV, cumulative EQD2). Missing data were handled using multiple imputation. For each endpoint, standard and LASSO Cox models were evaluated in a repeated, nested 5×5-fold cross-validation framework at 48, 60, and 80 months for RFS/MFS, and at 60, 80, and 120 months for OS. Discrimination (C-index, time-dependent AUC), Brier scores, and calibration were quantified, and both apparent and cross-validated (“validated”) estimates were Rubin-pooled across imputations. Nomograms were derived from the best-performing models.
Results: Median follow-up for OS, estimated by the reverse Kaplan–Meier method was 43.7 months (IQR: 23.2–66.2; maximum 193.1 months); median RFS, MFS, and OS were 64.1, 105.3, and 124.4 months, respectively. For RFS, the final LASSO model included ISUP, time from first radiotherapy to relapse, ADT duration, log PSA at recurrence, age, and primary EBRT administered as moderately hypofractionated versus conventional EBRT. Apparent Rubin pooled AUCs at 48 and 60 months were 0.72 and 0.78, while validated AUCs were 0.68 and 0.74. For MFS, the model retained the same core predictors, with apparent AUCs of 0.74 and 0.74, and validated AUCs of 0.68 and 0.69at 48 and 60 months, respectively. For OS, ISUP, time from first radiotherapy to relapse, ADT duration, log PSA at recurrence, age, and primary RT type were most influential. Apparent AUCs at 60 and 80 months were 0.78 and 0.85, and validated AUCs were 0.71 and 0.79, with Brier scores indicating acceptable absolute risk prediction.
Conclusion: In this cohort, we developed and internally validated the first RFS-, MFS-, and OS-focused prediction models and nomograms specific to radiorecurrent prostate cancer. Using multiple imputation and Rubin’s pooled nested cross-validation, we demonstrated good apparent and validated discrimination with consistent calibration. These tools address an important unmet need for individualized longterm prognostication after salvage brachytherapy and warrant external validation.