Neoadjuvant Osimertinib Meets Pathology Endpoint (~25% vs 2%) in a Phase III Resectable EGFR-Mutant Lung Cancer Trial — Survival Signal Not Yet Significant

Giving the EGFR inhibitor osimertinib (Tagrisso) before surgery — not just after — increased the share of resectable EGFR-mutant lung cancers that came back to the pathologist nearly cleared of viable tumor. In the Phase III NeoADAURA trial, the drug raised the major pathological response (MPR) rate to roughly a quarter of patients, against about 2% on chemotherapy alone. Two cautions belong up front: MPR is a pathology surrogate, not a survival result, and the patient-relevant survival endpoint was not statistically significant at this interim. And neoadjuvant (before-surgery) osimertinib is investigational — it is not approved for this use. Osimertinib is currently approved only after surgery (adjuvant, per ADAURA) and in metastatic EGFR-mutant disease.

This trial applies to a narrow group: patients with resectable, early-stage (Stage II–IIIB, including N2) EGFR-mutant (exon 19 deletion or L858R) non-small cell lung cancer who are surgical candidates. It says nothing about metastatic EGFR-mutant disease or non-EGFR lung cancer.

NeoADAURA (NCT04351555), sponsored by AstraZeneca, randomized 358 such patients 1:1:1 to neoadjuvant osimertinib plus platinum chemotherapy (121 patients), osimertinib monotherapy (117), or placebo plus chemotherapy (120), each followed by surgery. The trial is active but no longer recruiting.

The endpoint that was met (a pathology surrogate)

The primary endpoint was MPR — defined as ≤10% viable tumor cells in the resected lung primary tumor with an R0 margin — by blinded central pathology review. The MPR rate was 26% (95% CI, 18 to 34) with osimertinib plus chemotherapy, 25% (95% CI, 17 to 34) with osimertinib alone, and 2% (95% CI, 0 to 6) with placebo plus chemotherapy. Both osimertinib arms met statistical significance, with odds ratios versus control of 19.82 (95.002% CI, 4.60 to 85.33; p<0.0001) for the combination and 19.28 (99.9% CI, 1.71 to 217.39; p<0.0001) for monotherapy. Those odds ratios are large but imprecise — the near-zero event rate on chemotherapy alone leaves wide confidence intervals, one spanning two orders of magnitude. Crucially, MPR is a pathology measure, not a measure of how long patients live or stay cancer-free.

A quarter of patients reached major pathological response, against roughly 2% on chemotherapy alone — but that is a pathology surrogate. The long-term survival readout is barely begun, and the early signal has not yet cleared the trial’s own bar for significance.

Pathological complete response, a secondary endpoint, reached 9% (95% CI, 4 to 15) with osimertinib alone and 4% (95% CI, 1 to 9) with the combination, versus 0% (95% CI, 0 to 3) on control.

The caveat sits in the survival data. Event-free survival (EFS) — the secondary endpoint that matters most for whether neoadjuvant timing changes outcomes — was analyzed at an interim with only 15% data maturity. The EFS hazard ratio for osimertinib plus chemotherapy versus control was 0.50 (99.8% CI, 0.17 to 1.41; p=0.04), but the paper states this did not reach significance at the prespecified two-sided threshold of 0.002, and the confidence interval crosses 1.0. For osimertinib monotherapy the HR was 0.73 (95% CI, 0.40 to 1.35), and no formal statistical comparison was performed. Twelve-month EFS rates were 93%, 95%, and 83%, but these are early. A pathological surrogate met today does not guarantee a survival benefit tomorrow.

Two structural caveats matter for reading the EFS data. First, the paper itself flags MPR-as-surrogate as a limitation: MPR was chosen as the primary endpoint to keep the study small and read out early, not because it substitutes for overall survival. Second, the trial is effectively perioperative — about 80–86% of patients who underwent surgery in each arm went on to adjuvant osimertinib (82% combination, 86% osimertinib monotherapy, 80% control; standard of care per ADAURA), which the authors say is expected to affect the final EFS analysis. The neoadjuvant EFS signal should not be read in isolation.

On safety, grade ≥3 adverse events during the neoadjuvant period occurred in 36% of osimertinib-plus-chemotherapy patients, 13% of osimertinib-alone patients, and 33% of placebo-plus-chemotherapy patients; investigators reported no new safety concerns. Notably, osimertinib monotherapy delivered comparable MPR with the lowest grade ≥3 toxicity — a result worth watching as the EFS data mature.