903P - Patterns and predictors of recurrence-free survival (RFS) in salivary gland cancers (SGCs): A Spanish multicenter study

Background

SGCs include multiple histologic subtypes with variable prognosis. Despite curative therapy, up to 40% will recur. Lukovic's score (LS) based on clinic-pathological variables predicted the risk of distant metastasis (DM) in a pan-American SGC cohort (Lukovic J. et al. Ann Onc 2020). This study evaluates the patterns and predictive factors for RFS in a Spanish multicenter cohort of SGC patients (pts).

Methods

Retrospective multicenter cohort analysis of newly diagnosed SGCs treated with surgery +/- adjuvant (adj) radiotherapy (RT)/systemic treatment (stage I-IVB by pathologic TNM AJCC 8th ed.) between 2000 and 2020 at 5 Head and Neck Cancer institutions. We evaluated recurrence rates (RR) by histologic subtype and median recurrence-free survival (mRFS) using Kaplan-Meier method. Adjusted hazard ratios (aHR) for distant-metastasis recurrence-free survival (DMRFS) were calculated with competing risk including stage, histologic subtype and LS.

Results

A total of 142 SGC pts were included. Cohort characteristics: median age was 60 years (y) [50-70], male n=74, adenoid cystic (AC) n=43, mucoepidermoid (MEC) n=35, ductal carcinoma (DC) n=32, acinic cell (ACC) n=32, stage III-IVB n=75, linfovascular invasion n=34, perineural invasion n=60, R0/R1/unknown n=61/72/9, ENE n=13, androgen-receptor+ n=26, HER-2+ n=9, adj RT n=80, adj chemotherapy n=12, adj targeted therapy n=5. Risk group by LS (n=138): High n=89, Low n=49. RR: 45% (Locoregional 29%, Distant 56%, Both 12%). RR differed by histologic subtype (AC/DC vs MEC/ACC, p<0.01), mostly due to DM (57/27% vs 5/10%). Median follow-up was 11y (9-NR). Median RFS significantly differed by histologic subtype: AC 5y (3-6) vs DC 4y (2-NR) vs MEC 11y (7-NR) vs ACC 5y (3-6), p<0.01; and by LS: High 3y (3-5) vs Low 11y (10-NR), p<0.01. In the multivariate analysis, LS (ref. low) and stage (ref. I/II) were the only factors negatively associated with RFS (aHR 3.9 [1.7;8.8], p<0.01; aHR 2.7 [1.4-5], p<0.01) and DMRFS (aHR 4.9 [1.3;18], p<0.01; aHR 2.9 [1.2-7.1], p<0.01) respectively.

Conclusions

In this Spanish multicentre SGC cohort, we confirmed the differential pattern of relapse and RR of the four main histologic subtypes. Lukovic et al. score was validated as the major predictor for DMRFS.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Institut Català d'Oncologia (ICO).

Funding

Spanish Group for Head and Neck Tumors (TTCC).

Disclosure

N. Baste Rotllan: Non-Financial Interests, Advisory Role: Eisai, MSD, Merck Serono, BioNTech, Roche, BMS, Exelixis. C. Lazaro: Financial Interests, Personal, Advisory Board, Advisory boards on BRCA and HRD testing: AstraZeneca; Financial Interests, Personal, Advisory Board, IVDR advisory group: Illumina; Financial Interests, Institutional, Funding: AstraZeneca. M. Oliva Bernal: Financial Interests, Personal, Invited Speaker: Merck, MSD, BMS; Financial Interests, Personal, Advisory Board: Merck, MSD; Financial Interests, Personal, Writing Engagement: MSD; Financial Interests, Personal, Invited Speaker, Teaching activities: MSD, Merck; Financial Interests, Personal, Other, IDMC: Transgene; Financial Interests, Personal and Institutional, Funding: Roche; Financial Interests, Institutional, Local PI: ALX Oncology, MSD, ISA Therapeutics BV, Roche, Ayala Therapeutics, Abbvie, Bayer, Boehringer Ingelheim, Merck, Debiopharm, Seagen, Gilead, Beigene, Nykode; Financial Interests, Institutional, Funding: GSK; Non-Financial Interests, Institutional, Product Samples: Roche. All other authors have declared no conflicts of interest.