Extent of resection and 12-month mortality in primary intra-axial brain tumors: bounded estimates from a 20-year cohort

Background Maximal safe resection is central to neuro-oncological surgery, but how robust its association with short-term survival is to incomplete outcome ascertainment is rarely examined. We assessed this association in a 20-year cohort of primary intra-axial brain tumors, bounding the uncertainty from patients whose 12-month status could not be established. Methods We studied consecutive adults undergoing resection for histologically confirmed primary intra-axial brain tumors at one tertiary center, 2002–2024; biopsy-only and extra-axial cases were excluded. The outcome was death within 12 months, ascertained only when death occurred within that window or contact was documented beyond it. A logistic model of age, extent of resection (gross total [GTR] versus subtotal) and WHO grade 4 was fitted in the complete-case set, then refitted assuming all unascertained patients alive or dead and across the intervening range. Sensitivity analyses addressed tumor location, immortal time bias in adjuvant therapy through a two-month landmark, and the glioblastoma (GBM)-spectrum subgroup. Results Of 203 operated patients, 8 biopsy-only and 6 extra-axial cases were excluded, leaving 189. Status was ascertained in 116 patients (61.4%), of whom 31 (26.7%) died. GTR was independently associated with lower 12-month mortality in the complete-case model (odds ratio 0.17, 95% CI 0.06–0.51) and under best-case (0.27) and worst-case (0.45) assumptions, remaining below 0.5 across the intervening range (all p < 0.03). Age was associated with mortality throughout; WHO grade 4 reached significance only in the worst case (2.09, 1.05–4.18). The association persisted after adjustment for multilobar involvement and laterality, and within the GBM-spectrum subgroup ( n = 121). The apparent association of adjuvant therapy (0.08) did not survive the landmark analysis (0.26, p = 0.244). Conclusions GTR and younger age were associated with lower 12-month mortality, the GTR association being stable across all assumptions about unascertained outcomes. The contribution of WHO grade could not be resolved at this sample size. These are observational, not causal, associations from a single center with incomplete vital status ascertainment and no molecular classification.