ESTRO 2025 - Abstract Book

S2120

Clinical - Urology

ESTRO 2025

of the time to reach PSA rare on overall survival was examined, it was found to be 43.7 months in those with less than 6 months, and 63.3 months in those with 7 months and above (p=0.017).

Conclusion: Our study suggests that prostate radiotherapy has important contributions to the survival of patients with oligometastatic prostate cancer. Our results are compatible with the literature.

Keywords: Prostate cancer, oligometastasis, radiotherapy

References: 1.Fleming, C. W., Broughman, J. R., & Tendulkar, R. D. (2021). Treatment options in oligometastatic disease in prostate cancer: thinking outside the box. Current Treatment Options in Oncology , 22 (1), 2. 2.Siva, S., Bressel, M., Murphy, D. G., Shaw, M., Chander, S., Violet, J., ... & Foroudi, F. (2018). Stereotactic abative body radiotherapy (SABR) for oligometastatic prostate cancer: a prospective clinical trial. European urology , 74 (4), 455-462.

4302

Digital Poster Dosiomics-driven machine learning model for predicting late genitourinary toxicity in prostate cancer SBRT Constanza Hormazábal González 1 , Isabella Liedtke Grau 1 , Gabriel Lazcano Álvarez 1 , Tomás Walter Martin 2 , Darlett Folch Mora 2 , Francisco Pérez Peña 2 , José Solís Campos 1 1 Oncology, Hospital Carlos van Buren, Valparaíso, Chile. 2 Radiation Oncology, Universidad de Valparaíso, Valparaíso, Chile Purpose/Objective: To develop a classification machine learning (ML) model based on dosiomics to predict late genitourinary (GU) toxicity using dose distributions from prostate SBRT plans. Material/Methods: A retrospective cohort study was conducted with prostate cancer patients treated with SBRT at Hospital Carlos Van Buren (Valparaíso, Chile) between 2020 and 2023. The analysis included patients who had at least one follow-up appointment scheduled more than 90 days post-treatment. Adverse GU events were assessed using the RTOG scale, and urinary symptoms were evaluated with the International Prostate Symptom Score (IPSS). SBRT treatment consisted of 36.25 Gy in 5 fractions to the entire prostate and 27.25 Gy to the seminal vesicles, with no elective nodal irradiation. A classification ML model was developed using Scikit-learn [2] to predict late genitourinary (GU) toxicity. Patients were divided into two groups: those who experienced late GU toxicity and those who did not. A positive toxicity was defined by meeting at least one of the following criteria: (1) a GU RTOG score greater than zero at least once during follow-up, or (2) a 5-point or greater increase in the IPSS score compared to the pretreatment baseline. Dosiomics are radiomic features obtained from dose distributions. Dosiomics from the bladder, trigone and urethra were extracted using PyRadiomics v3 [1]. Basic clinical information and dose-volume histogram (DVH) metrics were also included as input to the model. Dimensionality reduction involved:

1. Removing highly correlated features (Pearson > 0.99). 2. Feature selection via iterative Random Forest algorithm [2].

Results: This study included 194 patients with a median age of 71 years (IQR 67-76) and an ECOG performance score of 0 (IQR 0-0, range 0-2). 15.3% reported active smoking. 58.8% of patients experienced late toxicity.