Dosimetric evaluation of variation in Library of plans selection for cervical cancer radiotherapy

Purpose: The aim was to determine whether differences in treatment plan selection methods from a Library-ofPlans (LoP) on cone-beam computed tomography (CBCT) leads to significant discrepancies in total accumulated dose. Methods: Ten cervical cancer patients were retrospectively analyzed on plan selection and dose distribution. Dose accumulation was performed using two different plan selection methods. 1) The clinical plan selections (i.e. plans as selected during actual treatment) and 2) automatic plan selection, which was based on the largest overlap between LoP clinical target volume (CTV) and (auto-)delineated CTV on CBCT. To evaluate the influence of daily plan selection on the treatment dose, the dose of selected plans was calculated on the daily CBCTs for all fractions and accumulated on the CT. The predefined clinical goals were used to determine if there was a relevant dose difference between the clinical LoP-selection, automatic LoP-selection, and selecting solely the default plan (i.e. full bladder plan with similar margins). Results: Different methods of LoP-plan selection did not result in significant dose differences. However, the default non-LoP plan resulted in significant dose differences compared to LoP-selection methods, with a reduced target coverage (99 % vs 100 %) and higher V40Gy to the rectum (55 % vs 44 %) compared to the clinical selection, respectively. Conclusions: LoP is robust to variations in plan selection and LoP outperforms non-LoP approaches in terms of target coverage and organs-at-risk (OAR) sparing. Therefore, LoP is a safe and efficient alternative to online adaptive replanning, particularly for longer fractionation schemes.