Lung-MAP (SWOG S1400 and S1900)

Lung-MAP opened in June 2014 as SWOG S1400, the first master protocol run jointly by the National Cancer Institute's cooperative groups, the FDA, Friends of Cancer Research, the Foundation for the NIH and several drug companies. Patients with previously treated stage 4 squamous non-small-cell lung cancer had their tumour sequenced once with Foundation Medicine's panel; those with a matching alteration were assigned to a targeted sub-study and everyone else to a non-match sub-study, each sub-study a phase 2 that could expand seamlessly into phase 3. Roy Herbst and Vassiliki Papadimitrakopoulou led it, with Mary Redman as statistician. In 2019 it became S1900 (NCT03851445), open to all non-small-cell histologies and to patients who had already had immunotherapy.

The first generation of sub-studies is summarised in the Lancet Oncology in 2020: 1,864 patients were screened and the genotype-matched arms, taselisib for PIK3CA (S1400B), palbociclib for cell-cycle gene amplification (S1400C) and AZD4547 for FGFR alterations (S1400D), all closed for futility with few responses, while durvalumab (S1400A) and durvalumab with tremelimumab after progression on anti-PD-1 therapy (S1400F) gave modest activity. A MET-directed arm of rilotumumab with erlotinib closed early, talazoparib for homologous recombination repair alterations (S1400G) did not meet its bar, and telisotuzumab vedotin for c-MET-positive tumours (S1400K) reported a low response rate. S1400I, the platform's first phase 3, found that adding ipilimumab to nivolumab did not improve survival. The one clearly positive sub-study so far is S1800A, in which ramucirumab plus pembrolizumab improved overall survival against standard of care in patients whose cancer had progressed on both chemotherapy and immunotherapy; the pragmatic phase 3 Pragmatica-Lung (S2302) was opened to confirm it.

What Lung-MAP changed was less any one drug than the machinery: a single FDA-agreed screening assay, sub-studies designed to support approval, drug companies entering and leaving one protocol, and, from 2024, published genomic landscapes of squamous lung cancer from the screening data. What it also showed, alongside NCI-MATCH and the National Lung Matrix Trial, is that matching one drug to one mutation in squamous lung cancer rarely produces large responses, which is why the newer sub-studies pair targeted agents with immunotherapy or test combinations.