Risk-Based Optimization of Intensity-Modulated Radiation Therapy Guideline for Post-Operative Oral Cavity Cancer

Abstract

To optimize previously established national intensity-modulated radiation therapy (IMRT) guideline for post-operative oral cavity squamous cell carcinoma (OSCC) in 2008, the follow-up results focusing on pattern and risk of failure were analyzed. Patients with OSCC treated with primary curative-intent surgery followed by postoperative IMRT according to the guideline with or without concurrent chemotherapy in our institute from 2010 to 2014 were retrospectively analyzed. Kaplan-Meier, univariate (UVA), and multivariate (MVA) Cox proportional hazard regression analyses were performed to identify predictors for locoregional control (LRC), distant metastasis-free survival (DMFS), disease-free survival (DFS), and overall survival (OS). Locoregional recurrences (LRR) were reconstructed on the treatment planning computed tomography images by deformable image co-registration. Recurrences were classified as in-field, marginal or out-of-field if >95%, 20-95%, and <20% of the recurrence volume was within the 95% of the prescribed isodose, respectively. A total of 180 patients were enrolled in this study with a median follow-up of 49.6 months (range, 4.4-91.5 months). Among these patients, 133 (73.9%) remained disease-free during the follow-up period. There were 30 (16.7%) patients with LRR, 25 (13.9%) with distant metastasis, and 29 (16.1%) with second primary cancers. The 5-year LRC, DMFS, DFS, and OS were 82.6%, 73.3%, 67.6%, and 75.4%, respectively. On UVA, extracapsular extension (ECE) was the only factor significantly associated with poorer LRC (hazard ration, 2.49; 95% CI, 1.14-5.41; P = .022). On MVA, only pN2-3 was associated with worse DMFS (HR, 2.29; 95% CI, 1.02-5.15; P = .046), DFS (HR, 2.57; 95% CI, 1.18-5.56; P = .017), and OS (HR, 2.90; 95% CI, 1.13-7.45; P = .027). Of the 30 patients with LRR, 22 failed at the primary tumor site, 3 at the regional lymph nodes, and 5 at both the primary and lymph node areas. There were 28 LRR in high-risk planning target volume (PTV), 6 in intermediate-risk PTV, and 1 in low-risk PTV. Most LRR were in-field failures, except 11 marginal failures in high-risk PTV, and 1 in intermediate-risk PTV. Most common marginal failures occurred in infratemporal fossa (n=6) and ipsilateral parotid gland (n=4), all of which belonged to primary cancers originating from the buccal and gingival mucosa. Two out-of-field failures were noted over level VI neck and skin of the ipsilateral neck. The median time to LRR was 9.9 months (range, 2.0-68.2 months). The 5-year follow-up outcome of previous IMRT guideline for postoperative OSCC might be feasible for good LRC and OS. Factors such as ECE and pN2-3 warrant further consideration for intensified adjuvant therapy. The marginal recurrence rates in infratemporal fossa and ipsilateral parotid gland were relatively high in buccal and gingival cancers, indicating a need for more generous margin coverage of related spreading routes to further optimize the guideline.