# A combined automatic organ at risk contouring and synthetic Computed Tomography solution for pelvic Magnetic Resonance-only radiotherapy

## TL;DR

This study shows that an automatic workflow for MR-only radiotherapy in pelvic cancers produces similar results to manual methods, improving efficiency and consistency.

## Contribution

A combined automatic MR OAR contouring and synthetic CT solution is evaluated for pelvic radiotherapy, showing clinical feasibility.

## Key findings

- Mean dose differences were ≤1Gy for most OARs except the bowel bag.
- Automatic workflow performed similarly to manual methods in terms of inter-observer variability.
- The solution could enable more efficient and consistent pelvic radiotherapy treatment.

## Abstract

Magnetic Resonance (MR)-only radiotherapy improves treatment accuracy and efficiency but requires a synthetic Computed Tomography (sCT) for dose calculations. Automatic MR-based Organs At Risk (OARs) contouring could further improve consistency and efficiency. A combined automatic MR OAR contouring and sCT solution has been developed, with each component separately validated. This study aimed to evaluate this combined MR-only workflow against a manual MR-CT workflow for pelvic cancers.

Radiotherapy MR and CT scans were acquired for 20 patients (10 prostate, 4 rectum, 6 anus cancer). The MR-only workflow used sCTs and automatic contours for the bladder, bowel bag, femoral heads, rectum, penile bulb, prostate, seminal vesicles and urethra along with manual target contours to optimise a MR-only plan. Doses were compared to the current clinical standard of deformably registered CT and manual MR OAR contours (MR-CT).

Mean dose differences were ≤1Gy; 1 percentage point relative volume for all OARs except the bowel bag. Mean absolute dose differences were also ≤1Gy for all OARs except bowel bag in ano-rectal patients and except bowel bag, bladder V50Gy, V40Gy, V30Gy and rectum V30Gy in prostate patients.

The automatic MR OAR contouring and sCT workflow had similar dose differences to manual inter-observer variability for all OARs except the lower dose bladder and rectum volumes (prostate patients only) and bowel bag (prostate and ano-rectal patients). This combined OAR contouring and MR-only workflow could enable more efficient and consistent pelvic radiotherapy treatment.

Graphical abstract Image 1

•Ano-rectal patients: dose differences ≤ 1 Gy for all organs except bowel bag.•Prostate patients: differences ≤ 1 Gy for femoral heads, penile bulb and urethra.•These differences are similar to manual inter-observer variability.•Automatic workflow may perform similarly to manual contouring for these organs.

Ano-rectal patients: dose differences ≤ 1 Gy for all organs except bowel bag.

Prostate patients: differences ≤ 1 Gy for femoral heads, penile bulb and urethra.

These differences are similar to manual inter-observer variability.

Automatic workflow may perform similarly to manual contouring for these organs.