# Benefits of daily online plan adaptation with reduced margins in neoadjuvant chemoradiotherapy for esophageal cancer

## TL;DR

Daily online adaptive radiotherapy for esophageal cancer reduces radiation doses to healthy organs and improves target coverage.

## Contribution

Demonstrates the feasibility and benefits of AI-assisted daily adaptive radiotherapy with reduced margins in esophageal cancer treatment.

## Key findings

- Adaptive treatment prevented underdosing in 31 out of 230 fractions.
- Adaptive treatment reduced heart and lung doses by 10% and 11%, respectively.
- Adaptive treatment enabled a 24.6% median reduction in planning target volume.

## Abstract

Unlabelled Image

•Emulated adaptive treatment for esophageal cancer was feasible in all 230 fractions.•Adaptive treatment prevented internal target underdosing in 31/230 fractions.•Adaptive treatment enabled a median 24.6% reduction in planning target volume.•Adaptive treatment reduced mean heart and lung dose by 10% and 11%, respectively.

Emulated adaptive treatment for esophageal cancer was feasible in all 230 fractions.

Adaptive treatment prevented internal target underdosing in 31/230 fractions.

Adaptive treatment enabled a median 24.6% reduction in planning target volume.

Adaptive treatment reduced mean heart and lung dose by 10% and 11%, respectively.

Image-guided radiotherapy (IGRT) for esophageal cancer requires large planning target volume (PTV) margins to account for interfraction variations, increasing radiation-associated side-effects. Cone-beam computed tomography-based (CBCT) online adaptive radiotherapy (oART) enables daily contour and plan adaptation, addressing anatomical changes and allowing for reduced PTV margins while ensuring target coverage. This study compared the dose-volume parameters of daily oART to non-adaptive IGRT in patients with esophageal cancer.

Ten patients with distal esophageal or gastroesophageal junction cancer who received neoadjuvant chemoradiotherapy (41.4 Gy/23 fractions) were included retrospectively. Daily IGRT CBCTs were used to emulate oART with artificial intelligence (AI)-assisted target and organs-of-interest contouring, manual edits when necessary, and adaptive re-optimization. PTV margins (Anterior-Posterior, Left-Right, Caudal-Cranial) were reduced from 5, 7, 10 mm (IGRT) to 3, 5, 5 mm (oART), respectively. Mean organs-of-interest and fraction-equivalent (FE) dose-volume metrics between delivered and adaptive plans were compared.

Compared with IGRT, oART plans significantly reduced the number of fractions with internal clinical target volume underdosing (V95% < 98%: n = 32 vs. n = 1 fraction; p = 0.002) and median hotspot dose (D0.1cm3: 106.2% vs. 103.9%; p = 0.002). Mean heart dose decreased by 10% (p = 0.037), and heart FE-V30Gy by 42% (7.3% vs. 4.3%; p = 0.002). Mean lung dose was reduced by 11% (p = 0.002), lung FE-V20Gy by 39% (4.2% vs. 2.6%; p = 0.002), and lung FE-V10Gy by 22% (24.6% vs. 19.3%; p = 0.002).

CBCT-based oART improved dose distribution in esophageal cancer by enabling PTV margin reduction, improved target coverage and superior organs-of-interest sparing. These findings encourage clinical implementation to reduce radiation-associated side-effects.