Design: multicentre, open-label, phase 3 randomised trial, stopped at first interim analysis for futility
Population: 402 eligible patients with T1-T3N0M0 NSCLC 7 cm or less, inoperable or declining surgery; median age 72.8
Primary outcome: overall survival
Effect: hazard ratio 1.04 (95% CI 0.69-1.58); 2-year survival 82% in both groups; grade 3+ events 12% vs 3%

Stereotactic body radiation therapy is standard for early-stage non-small-cell lung cancer in patients who cannot have surgery. A phase 3 cooperative group trial across 146 US institutions randomised 417 patients with T1-T3N0M0 disease of 7 cm or less, medically inoperable or declining surgery and carrying at least one recurrence risk factor, to radiotherapy alone or radiotherapy plus up to eight cycles of atezolizumab given before, during and after.

Accrual closed at the first interim analysis for futility. On updated analysis with 92 deaths and a median 24.8 months of follow-up, the overall survival hazard ratio was 1.04 (95% CI 0.69 to 1.58). Two-year overall survival was 82% in both groups. Grade 3 or higher adverse events occurred in 12% with the combination against 3% with radiotherapy alone, and there were two grade 5 respiratory events in the combination arm.

This is the first fully reported phase 3 cooperative group test of immunotherapy in this setting, and it is clearly negative. The result matters beyond oncology because the reasoning it defeats is common everywhere: a drug that works in advanced disease is assumed to work earlier, and the assumption is usually allowed to run ahead of the trial. Here the harm arrived and the benefit did not.

Do not extrapolate immunotherapy benefit from advanced to early-stage disease without trial evidence in that stage.
Quote 82% two-year survival with radiotherapy alone when counselling inoperable early-stage patients.
Note the fourfold increase in grade 3 or higher toxicity, including two fatal respiratory events.
Futility stopping means the trial could not have shown benefit had it continued, not that it was underpowered by accident.
For any patient asking about adding immunotherapy to curative radiotherapy in early disease, the answer now has evidence behind it.

Why it matters

It defeats the common assumption that a drug effective in advanced disease will help earlier, when the toxicity arrives all the same.

The statistics, in plain English

A hazard ratio of 1.04 with a confidence interval from 0.69 to 1.58 is compatible with a modest benefit or a substantial harm, so this does not prove the drug is useless - it proves the trial could not find the benefit it was designed to detect, which is why it stopped. The identical 82% two-year survival in both arms is the more persuasive figure, because it is a direct observation rather than a model estimate.