Personalized deep learning auto-segmentation models for adaptive fractionated magnetic resonance-guided radiation...

Manual contour corrections during fractionated magnetic resonance (MR)-guided radiotherapy (MRgRT) are time-consuming. Conventional population models for deep learning auto-segmentation might be suboptimal for MRgRT at MR-Linacs since they do not incorporate manual segmentation from treatment planning and previous fractions.

In this work, we investigate patient-specific (PS) auto-segmentation methods leveraging expert-segmented planning and prior fraction MR images (MRIs) to improve auto-segmentation on consecutive treatment days.

Data from 151 abdominal cancer patients treated at a 0.35 T MR-Linac (151 planning and 215 fraction MRIs) were included. Population baseline models (BMs) were trained on 107 planning MRIs for one-class segmentation of the aorta, bowel, duodenum, kidneys, liver, spinal canal, and stomach. PS models were obtained by fine-tuning the BMs using the planning MRI ( $$ ). Maximal improvement by continuously updating the PS models was investigated by adding the first four out of five fraction MRIs ( $$ ). Similarly, PS models without BM were trained ( $$ and $$ ). All hyperparameters were optimized using 23 patients, and the methods were tested on the remaining 21 patients. Evaluation involved Dice similarity coefficient (DSC), average ( $$ ) and the 95^th percentile (HD₉₅) Hausdorff distance. A qualitative contour assessment by a radiation oncologist was performed for BM, $$ , and $$ .

$$ and $$ networks had the best geometric performance. $$ and BMs showed similar DSC and HDs values, however $$ models outperformed BMs. $$ predictions scored the best in the qualitative evaluation, followed by the BMs and $$  models.

Personalized auto-segmentation models outperformed the population BMs. In most cases, $$ delineations were judged to be directly usable for treatment adaptation without further corrections, suggesting a potential time saving during fractionated treatment.