Head and neck radiotherapy after reconstructive flap surgery: Results of the multicentric XFLAP1 study

Abstract

Background and purpose: Postoperative target delineation after reconstructive surgery in head and neck cancer (HNC) is heterogeneous; whether to include the flap or focus on the native tissue-flap junction is debated. We quantified patterns of relapse relative to flaps and summarized practice, toxicity and function.

Materials and methods: in the retrospective XFLAP1 cohort (2018-2023), patients with HNC underwent tumor resection with flap reconstruction and postoperative radiotherapy (PORT) ± concurrent chemotherapy. Competing-risks (Fine-Gray) estimated locoregional relapse (LRR) and metastases; overall survival (OS) used Kaplan-Meier. Flaps were contoured a posteriori on planning CTs, when available.

Results: Of 355 patients across eight centres, free flaps were used in 239 (67%) and regional pedicled flaps in 69 (33%). The entire flap body was encompassed in the CTVs in 66% of plans; median flap-body dose was 65.3 Gy and pedicle Dmax 67.5 Gy for the delineated flaps (N = 153). Median follow-up was 32.9 months, 120/355 (34%) patients relapsed, including 68 (19%) LRR and 71 (20%) metastases. Median time to LRR was 7.35 months; only 3 (1%) of relapses arose within the flap body. Two-year cumulative incidence was 15.8% for LRR and 16.7% for metastases; two-year OS was 74.6%. On multivariable analysis, pN2-3 predicted metastases; LRR, metastases, and ECOG ≥ 1 were associated with worse OS.

Conclusions: relapses in the flap-body epicentre were rare; most local failures involved the native tissue-flap junction or non-flap sites. These data support junction-focused CTVs with reduced emphasis of the flap body to limit morbidity, pending prospective validation.

Keywords: Free flaps; Head and neck cancer; Patterns of failure; Pedicled flaps; Postoperative radiotherapy; Reconstructive surgery; Target delineation; Toxicity.