KEY POINTS
- Single-institution analysis of 97 patients with 102 peripheral lung tumours treated in free breathing with 67.5 Gy in 3 fractions stereotactic ablative radiotherapy.
- Pre- and post-treatment cone beam computed tomography at the first and second fractions quantified intrafractional tumour shifts; plans were recalculated for worst-case drifts exceeding the 4 mm posterior planning target volume margin.
- Systematic posterior drift was observed: mean vertical shift was –1.45 mm at fraction 1 and –1.24 mm at fraction 2, both p < 0.001; lateral and longitudinal shifts were negligible.
- Based on drift rates, 15.7% of patients would exceed the 4 mm posterior margin within a 20-minute treatment, and 27.5% within 30 minutes; no patient or tumour characteristic predicted drift.
- In the 16 affected patients, worst-case recalculation showed severe gross tumour volume underdosage: mean D90% coverage fell from 99.93% to 74.20%, p < 0.001. No clinical toxicity or local-control endpoint was reported.
CLINICAL TAKEAWAY
For free-breathing peripheral lung stereotactic ablative radiotherapy, posterior tumour drift is not just random noise around setup. This study supports routine post-treatment cone beam computed tomography verification, or another intrafraction strategy, to identify patients who drift beyond margins before later fractions. The evidence is technically important but single-institution and dosimetric, so it should prompt workflow review rather than immediate universal margin expansion.