Key Ideas

Resistance to KRAS inhibitors should be understood as three distinct categories, primary/intrinsic, adaptive/non-genetic, and acquired/genetic, each with different mechanisms and different implications for how to sequence or combine therapy. A growing body of preclinical evidence, still awaiting broader clinical confirmation, points toward lineage plasticity (including histologic transformation) as an underappreciated, non-mutational layer of resistance that deserves more attention as this field matures.

Why Resistance Is Inevitable, and Why It's Complicated

Given how extensively KRAS and its paralogs interact with downstream and parallel signaling, through the MAP kinase cascade, PI3 kinase and metabolic survival signaling, RHO/RAC-mediated cytoskeletal and endocytic pathways, and various points of pathway crosstalk, resistance to KRAS-targeted therapy is essentially guaranteed. Blocking KRAS alone is very unlikely to produce durable, long-term benefit on its own. Much of what follows is based on preclinical evidence with only partial clinical confirmation so far, an important caveat given how quickly this field is evolving.

Three Categories of Resistance

Primary or intrinsic resistance occurs when a KRAS mutation is present but isn't actually the tumor's main driver, so the tumor doesn't meaningfully depend on it. Genomic predictors of this include KEAP1 loss, STK11 loss (more controversial as a standalone predictor), and alterations in SMARCA4 or CDKN2A. A low KRAS mutant allele fraction has also been associated with minimal benefit, suggesting some tumors may have pre-existing, intrinsic insensitivity from the outset.

Acquired or genetic resistance was initially studied primarily through a KRAS-centric lens: secondary mutations that alter drug binding, KRAS amplification (enough total KRAS protein that it simply outcompetes the inhibitor), compensatory activation of KRAS paralogs (HRAS, NRAS) when KRAS itself is strongly blocked, and downstream MAP kinase or PI3 kinase pathway mutations, along with broader cell cycle and signaling changes. Deep mutational sequencing has identified specific resistance "hotspot" mutations that predict resistance to adagrasib, sotorasib, or both; some non-pocket mutations may also affect the GDP:GTP ratio directly. This kind of sequencing, at least in a clinical trial context, can help assess an individual patient's resistance profile. Acquired resistance can also arise through bypass activation or fusion events, amplification of EGFR, MET, FGFR1/2, or aurora kinase, or various fusion events, and it's rarely a single mechanism at once: in one published report, over 40% of patients progressing on a KRAS G12C inhibitor had multiple concurrent bypass resistance alterations.

Adaptive or non-genetic resistance develops relatively quickly after treatment starts and is largely driven by signaling feedback in response to inhibition rather than new mutations. This includes receptor tyrosine kinase upregulation, MEK/ERK feedback reactivation (where a downstream MEK inhibitor may be effective), PRMT5 inhibition in MTAP deletion, with one press release reported a 92% response rate with a KRASi/PRMT5 combination in pancreatic cancer), and other adaptive mechanisms including YAP/TAZ activation, epithelial-mesenchymal transition (EMT), and changes in protein localization. This category is fundamentally about plasticity rather than fixed mutation.

Lineage Plasticity: A Genuine, Growing Phenomenon

Lineage plasticity, tumor cells shifting their identity or differentiation state in response to therapeutic pressure, is becoming an increasingly real and recognized resistance mechanism, even if it still sounds abstract. This includes adenosquamous transition and other identity shifts; one specific pattern I learned about while preparing for this talk involves a shift between "type 2" and "type 1" cell states specifically in G12D models. Histologic transformation is the more dramatic, simpler form of lineage plasticity, and was identified early on as a resistance mechanism to KRAS inhibitors, mirroring what's seen with EGFR-targeted therapy (though transformation to small cell histology specifically has not yet been reported with KRAS inhibitors, to my knowledge, though it likely will be eventually). This is generally an adenocarcinoma-to-squamous transition without an obvious underlying genetic driver, suggesting an epigenetic reprogramming process, which is exactly why tissue biopsy at progression matters so much: this kind of transformation can't be identified any other way.

Drug efflux (multidrug resistance) is a less applicable mechanism for the small-molecule inhibitors we use most commonly today, but is more relevant to the newer PROTAC (targeted protein degrader) class of KRAS-directed drugs, an early rationale for considering multidrug-resistance-directed combination strategies with that specific drug class.

Toward a Resistance-Informed Treatment Algorithm

Understanding these resistance categories should genuinely inform combination strategy, not just because a given drug exists, but because there's a specific mechanistic rationale for combining it in a specific resistance context: receptor tyrosine kinase-directed combinations, PRMT5 inhibitors if MTAP deletions are present, SHP2 inhibitors for feedback reactivation, YAP or EMT-directed agents when EMT markers are present, and either continuing targeted therapy or adding chemotherapy when histologic transformation is identified. A comprehensive baseline molecular workup, and reassessment for high-risk co-mutations at the time of initial treatment selection, should already inform this kind of thinking. Whether ctDNA monitoring adds meaningful value here is, in my view, still more a question for the future than something to act on routinely today.

For Patients

When a KRAS-targeted therapy stops working, there isn't one single reason why, and the specific reason matters for choosing what comes next. Some tumors were never truly dependent on the mutation; some develop a new genetic change that gets around the drug; and some change without any new mutation at all, sometimes shifting what type of cancer cell they resemble, a phenomenon called lineage plasticity. This last category can only be caught through repeat tissue biopsy, not a blood test. Ask your care team whether a repeat biopsy is being considered if your KRAS-targeted treatment stops working.

Key Takeaways

  • Resistance to KRAS inhibitors falls into three categories: primary/intrinsic (the tumor was never truly KRAS-dependent), acquired/genetic (new mutations, amplification, paralog switching, or bypass activation), and adaptive/non-genetic (rapid signaling feedback without new mutations).

  • Over 40% of patients progressing on a KRAS G12C inhibitor in one published series had multiple concurrent bypass resistance mechanisms simultaneously, underscoring that resistance is rarely explained by a single alteration.

  • Lineage plasticity, including histologic transformation (most often adenocarcinoma to squamous), is an increasingly recognized, non-mutational resistance mechanism that can only be identified through repeat tissue biopsy.

  • Resistance mechanism should ideally inform combination strategy directly (for example, SHP2 inhibitors for feedback reactivation, chemotherapy addition for histologic transformation, etc.) rather than defaulting to a single next-line approach.

  • Comprehensive baseline molecular testing, including for high-risk co-mutations, should inform initial treatment selection, not only decisions made after resistance develops.

References

  1. Isermann T, Sers C, Der CJ, Papke B. KRAS inhibitors: resistance drivers and combinatorial strategies. Trends in Cancer 2025; 11: (Vol 2). 

  2. Tanaka M, Ebi H. Mechanisms of Resistance to KRAS Inhibitors: Cancer Cells' Strategic Use of Normal Cellular Mechanisms to Adapt. Cancer Science 2025; 116: 600 612. 

  3. Awad MM, Liu S, Rybkin II, et al. Acquired Resistance to KRAS(G12C) Inhibition in Cancer. N Engl J Med 2021; 384: 2382 2393. 

  4. Ash LJ, Busia Bourdain O, Okpattah D, et al. KRAS: Biology, Inhibition, and Mechanisms of Inhibitor Resistance. Curr. Oncol. 2024; 31: 2024 2046.