Interfractional Monitoring of Spacer Hydrogel Morphology in Patients With Prostate Cancer Receiving SBRT Treatment.
Authors
Affiliations (4)
Affiliations (4)
- Department of Radiation Oncology, New York University School of Medicine, New York, New York.
- Department of Population Health, Biostatistics, New York University School of Medicine, New York, New York.
- Department of Population Health, Biostatistics, New York University School of Medicine, New York, New York; Drexel College of Medicine, Philadelphia, Pennsylvania.
- Department of Radiation Oncology, New York University School of Medicine, New York, New York. Electronic address: [email protected].
Abstract
The purpose of this study is to explore whether spacer hydrogel morphology changes during the course of stereotactic body radiation therapy (SBRT) radiation to the prostate. A total of 198 patients at our institution each received spacer gel before SBRT on magnetic-resonance-guided linear accelerator (MR-Linac) between October 2023 and March 2025. Each patient received at least 1 simulation magnetic resonance imaging and 5 on-treatment magnetic resonance imaging scans. Average time between simulation and first treatment and between simulation and final treatment was approximately 19.9 days and 31.1 days, respectively, with 2- to 3-day intervals between treatments. A U-Net deep learning model was implemented using nnUNetv2. This model was iteratively trained using inference images, resulting in a Dice similarity coefficient >0.98, and was then applied to detect changes in gel morphology over time. Longitudinal modeling of spacer gel radiomic shape features demonstrated a statistically significant change in flatness that remained significant over false discovery rate correction and all sensitivity analyses. This metric increased by 0.003 units per 10 days (95% CI, 0.001-0.005; false discovery rate q ≈ 0.02), implying a progressively rounder shape over the observation period. All other features showed minimal or inconsistent change. The hydrogel appeared larger on treatment imaging, with a mean percent change of 10.62% (95% CI, 8.68%-12.56%; P < .001). The corresponding absolute volume difference was 1.19 cm³ (95% CI, 1.00 cm³-1.39 cm³; P < .001). Because prostate SBRT planning assumes gel stability, interfractional changes could affect rectal dose estimates. In this cohort, the spacer hydrogel's overall geometry remained stable across the 5-fraction course, with no meaningful evidence of gel resorption. A small increase in the flatness metric was observed, but without a change in size or compactness. These findings indicate that the gel maintains its rectal-sparing configuration throughout SBRT.