ESTRO 2024 - Abstract Book

S5132

Physics - Radiomics, functional and biological imaging and outcome prediction

ESTRO 2024

4 Oslo University Hospital, Department of Medical Physics, Oslo, Norway. 5 University of Florida College of Medicine, Department of Radiation Oncology, Jacksonville, FL, USA

Purpose/Objective:

Normal tissue complication probability (NTCP) models can be useful to estimate risk of morbidity after radiotherapy. Photon-based NTCP models are based on dose-volume parameters, while proton therapy specific NTCP models have so far been based on (physical) dose-volume parameters combined with the constant relative biological effectiveness (RBE =1.1) assumption. However, pre-clinical data suggests that RBE is associated with dose-weighted linear energy transfer (LETd). Most rectal and bladder NTCP models for prostate cancer include a dose-volume predictor from the high-dose range. In the search for an LET-inclusive NTCP model, a high-dose related LET predictor would therefore be most likely. The aim of this study was thus to quantify the difference in mean LET for the high dose region of the rectum and bladder between patients with vs. without morbidity following proton therapy, when also considering the dose-volume exposure.

Material/Methods:

This study was based on 205 prostate cancer patients treated with 78-82 Gy (RBE=1.1) to the prostate in 2 Gy (RBE=1.1) fractions, using double scattered proton therapy. Rectal and bladder morbidities were scored prospectively according to CTCAE v3 in 6-month intervals up to five years after treatment and annually up to eight years. We calculated dose and LETd distributions for each patient using a validated FLUKA Monte Carlo implementation of the treatment nozzle (Fig 1). The mean LETd was then computed in the region of the rectum and bladder having received dose above different high-dose cut-offs (in the range 60-75 Gy (RBE=1.1), in 5 Gy (RBE=1.1) intervals). Furthermore, we calculated the fractional volumes with LETd above 2 keV/um within the aforementioned high-dose cut-offs. For each organ, we grouped the patients into three equally sized groups to account for differences in treatment and dose-volume exposure. This grouping was done after ranking by organ- and endpoint-specific NTCP values, which were calculated using published proton therapy-specific Lyman-Kutcher-Burman NTCP models for the rectum (Grade >=2) and bladder (Grade 3). Mann-Whitney U tests were used with a significance level of 5% to evaluate the differences in mean LETd distributions and fractional volume distributions between the groups.

Results:

Across the whole cohort, the mean LETd values in the high dose region of both the rectum and bladder ranged between 0.6 keV/um and 2.4 keV/um, with the majority being distributed around 2 keV/um. Overall, the fractional volume of both the rectum and the bladder within the high-dose region having LETd > 2 keV/um increased with increasing NTCP (Fig 2). Across all NTCP groups and for both studied endpoints, the fractional volume of the high dose region having LETd > 2 keV/um was equal or slightly higher for patients with morbidity compared to those without (Fig 2), yet with no Mann-Whitney U test p values below 0.05.