Identification of Salivary Gland Tumors (SGT) at Risk for Local or Distant Recurrence After Postoperative Radiation Therapy (PORT) With or Without Systemic Therapy (ST) Using Clinical Risk Grouping

Purpose/Objective(s) To identify risk groups and patterns of recurrence after curative treatment of primary SGT treated with PORT +/- ST. Materials/Methods We retrospectively identified consecutive pts with stage I-IVA SGT, excluding adenoid cystic carcinoma (ca), treated with curative-intent surgery and postoperative IMRT with or without radiosensitizing chemotherapy from 12/2008-02/2020 at a tertiary care center. Pts were classified into clinical risk groups based on histologic diagnoses and pathologic features: (1) low-risk = morphologically low grade (LG) tumors (e.g. acinic cell, secretory, LG mucoepidermoid (MEC), epithelial-myoepithelial, basal cell adenocarcinoma), pN0, without lymphovascular invasion (LVI) or perineural invasion (PNI), including minimally invasive/non-invasive carcinoma ex-pleomorphic adenoma (CXPA); (2) intermediate-risk = intermediate grade morphology (e.g. grade 2 MEC) or LG morphology with PNI or LVI or pN+; (3) high-risk = high grade histology (e.g. salivary duct, high-grade MEC, poorly differentiated ca including invasive CXPA). Kaplan Meier estimates were used for recurrence-free survival (RFS) and univariable predictors. Results 72 pts were included. Median age at surgery was 59.6 years [interquartile range (IQR) 49.7,67.7 yrs]. Sites were: 86% parotid (n = 62), 8% submandibular (n = 6), 4% minor salivary (n = 3), and 1% unknown primary. SGT were classified into clinical risk groups: 26% low (n = 19), 25% intermediate (n = 18), and 49% high-risk (n = 35). Concurrent chemotherapy (CRT) was given to 3% of low-risk (n = 1), 13% of intermediate risk (n = 4), and 71% of high-risk pts (n = 25). Regimens included carboplatin + paclitaxel, cisplatin, and concurrent/adjuvant trastuzumab (in 10 pts with HER2-positive SGT). With a median follow-up of 40.5 months, 2-year OS was 95% (85-98%); there were 11 recurrences (9 distant metastasis (DM), 2 in-field local failure (LF) both in pts with gross residual after surgery), with 82% (9 of 11) having received CRT. All recurrences were in the high-risk group with 2-year RFS of 73% (54%-85%) and 2-year cumulative incidence of DM 21% and LF 6%. Factors associated with RFS included positive margins (P = 0.004) and low-neck positive lymph nodes (level III/IV/V vs. level I/II, P = 0.04). Thirteen pts (18%) had targeted genomic sequencing (4 with recurrence); common mutations included TP53 (n = 4), HRAS (n = 3), NFKBIZ (n = 3), and PIK3CA (n = 3). Mutations in NF1 (n = 3) and HNFA1 (n = 2) were observed only in pts that recurred. Conclusion DM are the primary mode of recurrence in pts with resected SGT treated with PORT +/- ST and occurred only in high-risk pts. After randomized trials like RTOG 1008 determine the added benefit of chemo to PORT, DM is likely to remain an unmet need. Further studies incorporating molecular characterization of high-grade SGT may identify potentially actionable targets to prevent distant recurrence.