ESTRO 2025 - Abstract Book

S1771

Clinical – Upper GI

ESTRO 2025

treatment start. After visual review, the physician decided to treat or perform occasional offline adaptation (OOA) and treat another day. Here, we simulated a daily online adaptation (DOA) based on all available fraction images. We compared the dosimetry properties of four scenarios per fraction: 1) baseline plan, 2) simulated no adaptation, 3) actual clinical OOA, 4) simulated DOA. Results: Median PTV coverage was 97% at baseline but decreased significantly in the no adaptation (median 87.6%, p = 0.0001) and OOA scenarios (median 87.8, p = 0.0001). Conversely, DOA maintained a high PTV coverage (median 95.7, p = 0.17).

Figure 1 : PTV coverage at baseline and per fraction for different adaptation scenarios. Frequent violations of OAR dose constraints occurred similarly in the no adaptation and OOA scenarios (both 20/52 fractions, 38.5%). Bowel dose constraints were violated in three patients with tumors close (< 1 cm) to the bowel. Heart and esophagus dose constraints were violated in seven and two patients, respectively. Here, tumors were located in the liver dome and “beam overshoots” created dose maxima up to 6 cm away from the PTV. With DOA, fewer violations occurred (6/52 fractions, 11.5%), all being moderate and affecting the heart of three different patients.

Figure 2 : Case presentation. No adaptation and occasional offline adaptation (OOA): Caudal liver shift in the third fraction led to a “beam overshoot” towards the esophagus (orange) and heart (pink). Daily online adaptation (DOA): Heart and esophagus are spared while maintaining reasonable PTV (blue) and excellent GTV (green) coverage.