Lymph Node Ratio as a Predictor of Distant Metastases in Major Salivary Gland Carcinomas

Abstract

To investigate the prognostic value of lymph node ratio (LNR, number of positive lymph nodes[LN]/total number of excised LN) on distant metastases (DM) and overall survival (OS) in patients (pts) with major salivary gland carcinoma (SGC). An REB-approved retrospective review was conducted for SGC patients treated at our institution with curative surgery and neck dissection (>=6 dissected LN)+/−adjuvant treatment in 2000-2015. Patients, treatment and outcomes data were collected from our institutional maintained databases. Staging was reviewed according to the AJCC-UICC 8th edition. High risk pathology was defined with histologic grade and WHO histologic criteria, and included: adenoid cystic carcinoma (ACC), salivary duct carcinoma, SCC, G2/3 adenocarcinoma, G2/3 mucoepidermoid carcinoma (MEC), G2/3 carcinoma ex-pleomorphic adenoma, carcinosarcoma, undifferentiated (small-, large-cell or lymphoepithelial) carcinoma and G3 of other histologic subtypes. Distant control (DC) and OS were analyzed with competing risk and Kaplan-Meier methods respectively. LNR (continuous variable) was subjected to multivariable analysis (MVA) for DM and OS (adjusted for age, gender, primary SGC subsite, pathologic stage, high-risk pathology, lymphovascuar invasion [LVI], perineural invasion [PNI], extranodal extension [ENE] and surgical margin status). The optimal cutpoint of LNR that maximized the difference in outcomes was determined using maximally selected rank statistics. Subgroup analysis was performed for patients with pN+. A total of 204 pts were identified: median age: 56 yr (16-91); median follow-up: 5.2 yr (0.4- 17.6); parotid gland primary tumor location: 168 (82%); high risk pathology: 151 (74%); pT3-4: 132 (65%), pN+: 99 (44%); LVI: 49 (26%); positive microscopic surgical margin: 103 (52%); ENE: 37 (19%). PORT was used in 195 pts (96%); and adjuvant concurrent chemotherapy in 11 (5%). Of 2,725 LNs evaluated, 328 (12%) were pN+. The median number of dissected LN was 23 (6–101). For pN+ pts, the median number of involved LN was 3 (1-65) and median LNR was 14% (1%–100%). High-risk pathology and LVI were associated with high LNR (p<0.001 for both). On MVA LNR was independently correlated with DM (HR:1.18, 95%CI: 1.07-1.30, p<0.001) and OS (HR:1.16; 95%CI: 1.06-1.28, p=0.002) and so did LVI+ (DM: HR:2.54, 95%CI: 1.38-4.70, p=0.003) and OS (HR:2.50; 95%CI: 1.33-4.70, p=0.004), positive margins (DM: HR:2.47, 95%CI: 1.31-4.65, p=0.005) and OS (HR:2.59; 95%CI: 1.35-4.95, p=0.004),. The optimal cut-off point for LNR was 8%; 5-yr DC: 42% vs 88%, p<0.001; 5-yr OS: 44% vs 89%, p<0.001 for LNR ≥8% vs <8%. In a subgroup analysis of patients with pN+, LNR remained predictive for DM (HR:1.24, 95%CI: 1.14-1.35, p<0.001) and OS (HR 1.27, 95%CI: 1.16-1.38, p<0.001). High LNR is associated with a higher risk of DM and lower OS. LNR should be evaluated in future prospective studies for intensified therapy or surveillance schedule in patients with SGC.