Efficient and Effective Personalization of PTV Margins During Radiation Therapy for Bladder Cancer

The magnitude of bladder filling variation during bladder cancer radiation therapy varies considerably between patients. Population-based approaches for planning target volume (PTV) margin calculation may be suboptimal for this disease site, and a strategy for personalizing PTV margins for each patient may be particularly beneficial. The purpose of this study was to identify the optimal number and sampling pattern of cone beam CT image data sets that are required when generating personalized PTV margins for whole bladder (WB) and partial bladder (PB) radiation therapy.

Personalized PTV margins were generated for 24 bladder cancer patients (15 WB and 9 PB) using nine experimental strategies that varied in the number and pattern of images incorporated into the margin generation process. These PTVs were compared to the standard-of-care (SoC) PTV at our institution (15 daily fractions included) using PTV volume (cohort-based and individual patient ranking), superior and posterior expansion, and clinical target volume (CTV) coverage.

For WB CTV, strategies ES4 (first five fractions), ES7 (every other fraction), and ES8 (first 10 fractions), provided CTV coverage equivalent to, or better than the SoC (first 15 fractions). Of these three strategies, ES4 resulted in the smallest superior and posterior borders, the smallest volume and the lowest intrapatient volume ranking, all achieved with the smallest number of fractions. For the PB CTV, strategies ES4 (first five fractions), ES7 (every other fraction), ES8 (first 10 fractions), and ES9 (last 10 fractions), provided CTV coverage equivalent to, or better than SoC (first 15 fractions). There were no statistically significant differences in the superior and posterior borders between these strategies, but ES4 resulted in the smallest volume and the lowest intrapatient volume ranking, all achieved with the smallest number of fractions.

This study suggests that using contours from images taken during the first five daily fractions generated a personalized "patient-specific" PTV that provided CTV coverage equivalent to the 15-fraction SoC but decreased the irradiated volume, reduced delineation workload, and reduced the superior and posterior borders for WB. It is now the SoC for whole and PB radiation therapy at our institution.