Postoperative recurrence defines a 3-year high-risk window in patients with EGFR-mutant lung adenocarcinoma: a...

The temporal distribution of postoperative recurrence in patients with resected epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer remains incompletely characterized. Most prior studies have focused on aggregate survival endpoints without explicitly modeling time-dependent recurrence hazards, thereby potentially obscuring underlying temporal patterns. A formal characterization of recurrence hazard over time may provide insight into the natural history of EGFR-mutant disease and inform postoperative management strategies, particularly in the current era of adjuvant EGFR-tyrosine kinase inhibitors (TKIs). Here, we aimed to evaluate the temporal dynamics of postoperative recurrence in a treatment-naïve multicenter cohort of patients with completely resected stage II-III lung adenocarcinoma. Understanding the intrinsic timing of recurrence is also essential for interpreting outcomes of recent adjuvant EGFR-TKI trials and for designing future time-adapted postoperative strategies.

We conducted a multicenter retrospective cohort study including 787 patients (295 EGFR-mutant and 492 EGFR-wild-type) who underwent complete resection between 2010 and 2020 at four Japanese institutions, all in the pre-adjuvant EGFR-TKI era. Cumulative incidence of recurrence and lung cancer-related death was estimated using competing-risk methods, treating non-cancer-related death as a competing event. Between-group differences were assessed using Gray's test, and subdistribution hazard ratios were estimated using the Fine-Gray model. Temporal changes in recurrence hazard were evaluated using Joinpoint regression, and interval-specific competing-risk analyses were performed. Propensity score matching using preoperative variables was applied as a sensitivity analysis.

With a median follow-up of 62 months, the 5-year cumulative incidence of recurrence (53% vs. 44%; P=0.20) and lung cancer-related death (29% vs. 32%; P=0.40) did not differ significantly between EGFR-mutant and EGFR-wild-type tumors. However, recurrence hazard in patients with EGFR-mutant disease showed a distinct nonlinear temporal pattern. Joinpoint regression identified a significant inflection point at approximately 3 years after surgery (P=0.002), beyond which recurrence hazard declined markedly (annual percent change: -34.64%), following a relatively sustained risk phase during the earlier postoperative period. In contrast, no clear inflection point was observed in patients with EGFR-wild-type tumors. Interval-specific analyses revealed no significant differences in recurrence hazard between the groups at any postoperative interval, suggesting that the observed pattern reflects an intrinsic temporal structure rather than differences in overall risk magnitude. This time-dependent pattern was consistently observed across clinically relevant subgroups and remained unchanged after propensity score matching.

Postoperative recurrence in patients with EGFR-mutant stage II-III lung adenocarcinoma exhibits a nonlinear, time-structured hazard pattern characterized by an early concentration of risk followed by attenuation beyond 3 years. These results are particularly relevant in the current era of adjuvant EGFR-TKI therapy, where the optimal duration of treatment and surveillance remains uncertain in clinical practice. A clearer understanding of intrinsic recurrence timing may help refine future time-adapted postoperative strategies.