ESTRO 2024 - Abstract Book

S2494

Clinical - Urology

ESTRO 2024

(1) Attard, Gerhardt et al. “Abiraterone acetate and prednisolone with or without enzalutamide for high-risk non metastatic prostate cancer: a meta-analysis of primary results from two randomised controlled phase 3 trials of the STAMPEDE platform protocol.” Lancet (London, England) vol. 399,10323 (2022): 447 -460.

2033

Digital Poster

Systematic trends in T2w and ADC in hypofractionated prostate cancer treatments on a 1.5T MR-Linac

Peter J Koopmans, Rene Monshouwer, Jeroen Findhammer, Robert Jan Smeenk, Marcel Verheij, Erik van der Bijl

Radboud University Medical Center, Department of Radiation Oncology, Nijmegen, Netherlands

Purpose/Objective:

MR-Linac workflows generate high-quality MR images for every individual fraction used for contouring, planning and monitoring. Workflow timing also provides an opportunity to acquire additional data, in our case diffusion weighted imaging (DWI) during online delineation/planning. Here we investigate whether such a systematic data acquisition process can be leveraged to study treatment effects, potentially ultimately enabling efficacy prediction.

Material/Methods:

Data of 97 patients were used with their consent and approval by the local ethics board. Patients all underwent hypofractionated treatment consisting of five fractions over two weeks. In each fraction, four T2w images were acquired along with DWI (b0, b150, b500) immediately after the first T2w scan. For various practical reasons diffusion data and/or GTV delineations were not always available for every patient The T2w images were used to investigate relative changes in prostate gland volume and intensity. To improve measure stability, we trained a nnUNet[1] to consistently delineate the prostate. This reduced observer variance and moreover it allowed us to use all four T2w images of each fraction as opposed to only the first one which had the manual delineations. For the ADC analysis we looked at changes in the ADC values averaged over the prostate gland and GTV separately. To identify the GTV, its delineations from the pretreatment scans were transferred to the fraction scans through deformable image registration using Elastix[2].

Volume analysis (N=92): the volume was extracted for the entire prostate gland and normalised to the volume of the first fraction of each patient (w.r.t. the average volume in four scans).

Intensity analysis (N=92): the average T2w intensity of the entire prostate gland was extracted for all scans of each patient. As MRI does not provide an absolute measure, intrapatient stability was improved by normalising the prostate intensity w.r.t. that of the femurs. Interpatient scaling was achieved by normalising to the first fraction of each patient.

Diffusion analysis (N=76): The average of voxelwise ADC values was calculated for two masks: 1) the intersection of GTV and the prostate gland (to exclude GTVs outside the prostate), and 2) the prostate without the GTV. To avoid