Background. Extranodal extension (ENE) is the most adverse feature in resected oral cavity squamous cell carcinoma (OCSCC), with high distant-metastasis risk despite adjuvant cisplatin-based chemoradiotherapy. The MACE-Postop phase III trial (CTRI/2017/02/007777) showed a marked treatment-by-ENE interaction; we report the pre-specified ENE-positive subgroup analysis. Methods. Of 410 patients randomised post-adjuvant to observation (OBS) or oral metronomic methotrexate 15 mg/m² weekly plus celecoxib 200 mg twice daily for 18 cycles (MACE), 191 were ENE-positive (OBS 104; MACE 87). The primary endpoint was 2-year overall survival (OS), tested by unadjusted log-rank mirroring the parent trial. Secondary endpoints included progression-free survival (PFS) and first distant metastasis (cause-specific Cox). Adjusted sensitivity analyses controlled for subsite, pT, pN, margin, PNI, NACT and adjuvant regimen. Results. Baseline pathological characteristics were balanced between arms (all p > 0.10, χ²/Fisher's exact); MACE had a higher proportion of depth-of-invasion >10 mm (62.1% vs 50.0%), pN2–3 disease (98.9% vs 95.2%) and close/positive surgical margins (12.6% vs 7.7%). With 83 deaths (OBS 53; MACE 30) at a median survivor follow-up of 45.7 months, 2-year OS was 53.7% (OBS) vs 71.3% (MACE) — HR 0.52 (95% CI 0.33–0.81), log-rank p=0.003. 2-year PFS was 53.0% vs 69.3% — HR 0.51 (0.33–0.78), p=0.002. The 2-year cumulative incidence of first distant metastasis was 20.8% vs 9.5% (cause-specific HR 0.37, 95% CI 0.17–0.80, p=0.011); locoregional recurrence did not differ between arms. Treatment effects were unchanged after multivariable adjustment (OS aHR 0.51; PFS aHR 0.49). Conclusions. In ENE-positive resected OCSCC, MACE significantly improved overall and progression-free survival, driven by reduced distant metastasis. These hypothesis-generating findings support a definitive trial of MACE enriched for ENE-positive disease in low- and middle-income settings.