# Evaluation of therapeutic radiographer target volume contouring for magnetic resonance image guided online adaptive bladder radiotherapy

## TL;DR

Trained radiographers can contour bladder tumors as accurately as specialists, enabling more efficient MRI-guided radiotherapy.

## Contribution

Demonstrates that trained radiographers can produce high-quality contours for bladder radiotherapy, reducing reliance on specialists.

## Key findings

- RTT contours showed excellent agreement with radiation oncologist contours using metrics like DSC and MDA.
- Dosimetric analysis showed that 65% of adaptive plans met optimal target coverage criteria.
- RTT contouring could reduce the workload of radiation oncologists in MRI-guided bladder treatments.

## Abstract

•Trained RTT bladder target volume contours were found to have excellent agreement with radiation oncologist contours.•Intra-professional variability between radiation oncologists and between trained RTTs demonstrated similar variance.•Adaptive plans optimised on RTT contours result in clinically acceptable target coverage of ‘gold standard’ tumour volume.•Implementation of RTT contouring could release radiation oncologists from online adaptive bladder radiotherapy workflows.•Training RTTs now, equips the workforce with skills to verify auto-segmentation in the future.

Trained RTT bladder target volume contours were found to have excellent agreement with radiation oncologist contours.

Intra-professional variability between radiation oncologists and between trained RTTs demonstrated similar variance.

Adaptive plans optimised on RTT contours result in clinically acceptable target coverage of ‘gold standard’ tumour volume.

Implementation of RTT contouring could release radiation oncologists from online adaptive bladder radiotherapy workflows.

Training RTTs now, equips the workforce with skills to verify auto-segmentation in the future.

One barrier to wider clinical implementation of online MRI-guided radiotherapy (MRIgRT) on the MR Linac (MRL) is resource intensity. Specifically, the requirement for a clinical oncologist/radiation oncologist (CO/RO) to perform online contouring each fraction. We report an evaluation of therapeutic radiographer (RTT) online contouring for patients receiving whole bladder MRIgRT.

RTTs undertook a contouring training programme. RTT and CO/RO clinical target volume (CTV) contours from 95 fractions were assessed using dice similarity coefficient (DSC), hausdorff distance (HD), mean distance to agreement (MDA), sensitivity and specificity volume metrics on the Raystation treatment planning system (TPS) (RaySearch Laboratories). Additionally, CTV DSC was evaluated with respect to a simultaneous truth and performance level estimation (STAPLE) generated in ADMIRE (Elekta AB, Stockholm, Sweden). In dosimetric analysis (Monaco, Elekta AB), online adaptive treatment plans, which had been generated using RTT-defined contours, were evaluated using contours delineated offline by CO/ROs.

Comparison of RTT versus CO/RO contours found the CTV median (interquartile range) (IQR) for DSC was 0.92 (0.91–0.94), MDA was 0.11 (0.09–0.12) cm, and HD was 0.63 (0.53–0.72) cm, sensitivity and specificity were 0.94 (0.90–0.96) and 0.95 (0.92–0.97) respectively. In dosimetric analysis, 65 % (30/46) plans met optimal PTV coverage of V52.25 Gy > 98 % and all plans met mandatory PTV coverage of V52.25 Gy > 95 %.

Following effective training, evaluation results demonstrate RTT whole bladder CTV contours to be comparable to CO/RO contours. Clinical implementation will release CO/ROs from MRL bladder treatments, reducing resource intensity of online workflows.