Feasibility of reusing online-generated treatment plans for adaptive radiotherapy in prostate cancer

: Online adaptive radiotherapy (oART) is underused as generating a treatment plan at every fraction is slow and resource intensive. One method to address this involves reusing plans generated online in previous fractions with similar anatomy. However, manually assessing the suitability of each pre-existing treatment plan is prohibitively time-consuming. To gauge potential impact and motivate the development of software to enable plan recycling, we assessed a strategy whereby all pre-existing plans were considered for subsequent fractions in nine hypofractionated prostate patients treated on the magnetic resonance (MR) linear accelerator.

The verification MR was used to estimate the delivered dose after adaptation to establish a Current Clinical Practice Benchmark. Each structure from the daily MR was propagated backwards onto the reference and daily MRs from previous fractions to calculate the dose to each structure that would have been received had the corresponding plan been delivered. The resulting dose statistics were assessed against: (A) standard target and organ-at-risk objectives, (B) the Current Clinical Practice Benchmark, and (C) circumstances where a pre-existing plan would have matched or outperformed the online plan.

The median [interquartile range] percentage of fractions with at least one acceptable pre-existing plan was 25% [20%], 40% [35%], and 60% [20%] for criteria A, B, and C respectively. Reusing the reference plan was only acceptable in 0%-20% of fractions.

Reusing pre-existing plans is feasible and could accelerate oART and reduce hospital resources in approximately 40% of fractions whilst achieving the same dose-volume metrics as current oART workflows.