A narrative review from Tata Memorial Centre, published August 20 in Medical Oncology, reports objective response rates of 37% to 43% and median progression-free survival of roughly six to seven months for sotorasib and adagrasib in previously treated KRAS G12C non-small cell lung cancer, against 45% to 55% for the five next-generation inhibitors behind them. It groups acquired resistance into four categories and places the near-term gains in resistance-directed combinations, biomarker-guided selection, non-G12C mutants and earlier disease settings.
- Publication: Medical Oncology 2026;43(9):251, posted online August 20, 2026, doi 10.1007/s12032-026-03318-2.
- Authors: 14, led by Akhil Kapoor with Anuj Gupta as corresponding author; four are listed under Molecular Pathology at Tata Memorial Hospital in Mumbai and one under Cancer Genetics in Varanasi.
- Approved agents: sotorasib and adagrasib at 37% to 43% ORR, median PFS about six to seven months, with better tolerability than docetaxel.
- Next generation: olomorasib, divarasib, glecirasib, D-1553 and IBI351 at 45% to 55% ORR.
- Resistance: on-target KRAS alterations, RTK-RAS bypass signaling, MAPK and PI3K reactivation, and co-mutation-driven phenotypic remodeling.
- Disclosures: six of the 14 authors report institutional research funding, naming AstraZeneca, Novartis, Lilly, Biocon, Bayer, Johnson and Johnson, Bristol Myers Squibb and Pfizer among others.
Four ways a G12C inhibitor stops working
The review sorts acquired resistance into on-target changes in KRAS itself, bypass signaling through receptor tyrosine kinases back into RAS, reactivation of the MAPK and PI3K arms downstream, and phenotypic remodeling driven by co-mutations. Each of those has a different consequence for what a laboratory would need to detect at progression: a secondary KRAS variant is a sequencing call, a MET or EGFR bypass is a copy-number or expression call, and phenotypic remodeling may show up only as histology. The abstract does not assign frequencies to any of the four.
Olomorasib and the four rivals beside it
The gap the authors report between first- and next-generation G12C inhibitors is about 10 percentage points of objective response, with progression-free survival exceeding the first-generation agents in some studies. Five compounds carry that claim: olomorasib, divarasib, glecirasib, D-1553 and IBI351. Mutant-specific G12D agents get a much weaker verdict, described only as showing early proof-of-concept activity, with no response figure attached. The visible reference list includes the G12D compound MRTX1133 and the G12D-selective degrader ASP3082, both cited from preclinical and chemistry literature, with no clinical readout behind either.
Combinations carrying the argument
Four pairings are singled out as producing higher response rates and deeper suppression in early trials: KRAS inhibition with PD-1 blockade, with SHP2 or SOS1 inhibitors, with EGFR or MET inhibitors, and with platinum doublets. The abstract reports the efficacy direction and stops there. It quantifies no toxicity for any of them, and gives no figure for the hepatic or immune-related events that have shaped how G12C inhibitors and checkpoint blockade are sequenced in practice. Anyone hoping the review settles that question will have to buy the full text.
Who paid for the authors
Six of the 14 authors declare institutional research funding, and the list runs long. Kapoor names AstraZeneca, Novartis, Lilly, Biocon, Bayer, Johnson and Johnson, Bristol Myers Squibb and Pfizer. Kumar Prabhash, the senior author, names Biocon, Dr Reddy's Laboratories, Fresenius Kabi, Alkem, NATCO Pharma, BDR Pharmaceutics, Roche, Aurigene and Johnson and Johnson. Vanita Noronha adds Merck, Roche India and Nanobiotix. All of it is institutional, and the declarations sit at the foot of the record where a reader can check them against the compounds discussed.
Specimen choice goes unaddressed
Springer Nature keeps the full review behind a subscription, so what is described here comes from the published abstract, the author and disclosure blocks, and the visible portion of the bibliography. Nothing in the abstract distinguishes tissue from plasma genotyping for KRAS, or says at what point in a treatment course either should be repeated. One methods citation on liquid biopsy appears in the reference list, a 2024 Clinical Cancer Research paper by Koga and colleagues on plasma detection of KRAS mutations and its limits.
Why This Matters to the APO|APE Reader
The review's own bibliography dates olomorasib's US breakthrough therapy designation to a Lilly announcement of September 4, 2025, which puts a second-generation G12C inhibitor within reach of a filing while G12D agents are still described as proof-of-concept. A panel that reports G12C as a yes-or-no call, without the co-mutation context the review ties to phenotypic resistance, will be answering last generation's question. What the abstract will not settle is whether that context should come from tissue or from plasma at progression, and its single liquid-biopsy citation is a methods paper on the limits of the plasma assay rather than a recommendation.
