The management of non-small cell lung cancer (NSCLC) is evolving, as researchers increasingly focus on advanced biomarker testing that extends beyond genomic analysis. This shift involves the integration of protein-based and computationally derived markers, spurred by the approval of innovative therapies, particularly antibody-drug conjugates (ADCs). Dr. Soo-Ryum (Stewart) Yang, an assistant attending pathologist and co-director of Clinical Biomarker Development at Memorial Sloan Kettering Cancer Center, detailed these trends during his presentation at the 20th Annual New York Lung Cancers Symposium on November 15, 2025.
Dr. Yang highlighted four pivotal trends reshaping NSCLC management. These include the rise of protein-based immunohistochemistry (IHC) biomarkers for ADCs, the actionable potential of tumor suppressor genes, advances in synthetic lethality applications, and the increasing role of computational pathology. A significant challenge persists, namely tissue scarcity, underscoring the urgent need for multiplex IHC development and the integration of broad panel next-generation sequencing (NGS) with artificial intelligence (AI) to enhance personalized therapy for a broader NSCLC patient population.
Traditionally, cancer diagnostics focused primarily on identifying genetic mutations. However, current practices are shifting towards assessing the expression levels of specific proteins on cancer cells, which can reveal new treatment avenues. While PD-L1 IHC testing has guided checkpoint inhibitor therapies, IHC testing is now being utilized to inform ADC treatment decisions. Dr. Yang emphasized two critical protein biomarkers in NSCLC: HER2 and c-MET overexpression.
HER2 overexpression is observed in up to 20% of patients, with the highest IHC score (3+) present in approximately 3% of cases. Dr. Yang clarified that there is no direct correlation between HER2 mutation status and its overexpression. Most NSCLC cases displaying high-level gene amplification will exhibit IHC 3+ staining; however, this is not universally applicable, as not all 3+ cases are gene amplification-driven.
The FDA’s approval of fam-trastuzumab deruxtecan-nxki (T-DXd; Enhertu) for HER2-positive solid tumors, including previously treated NSCLC patients, was underpinned by the phase 2 DESTINY-Lung01 study (NCT03505710). Dr. Yang advocated for applying gastric cancer HER2 scoring guidelines to NSCLC testing.
In addition to HER2, c-MET overexpression is prevalent in NSCLC. An actionable c-MET-high status, defined by over 50% of tumor cells exhibiting 3+ staining, is found in up to 17% of EGFR wild-type cases. The FDA granted accelerated approval to telisotuzumab vedotin-tllv (teliso-V; Emrelis) for this demographic, supported by data from the phase 2 LUMINOSITY trial (NCT03539536).
The integration of IHC screening for HER2 and c-MET presents diagnostic challenges. Dr. Yang proposed a flexible approach with standardized options, allowing healthcare institutions to tailor workflows based on their specific resources and multidisciplinary insights. Emerging biomarkers are currently under investigation, with the potential to refine personalized treatment protocols for NSCLC patients.
Dr. Yang noted that KRAS mutations occur in up to 40% of lung adenocarcinomas, specifically within codons G12, G13, and Q61. The KRAS G12C mutation is the most common among these, followed closely by the KRAS G12V and G12D mutations. The latter is particularly associated with patients who have a history of never or light smoking and correlates with lower tumor mutational burdens and PD-L1 expression, which complicates responses to chemoimmunotherapy.
Therapies targeting the KRAS G12C mutation, including sotorasib (Lumakras) and adagrasib (Krazati), are now established treatments. Research is ongoing for therapies targeting other KRAS mutations, such as multi-RAS and RAS(ON) inhibitors. Notably, the KRAS G12D inhibitor zoldonrasib (RMC-9805) demonstrated an overall response rate of 61% and a disease control rate of 89% in a phase 1 study (NCT06040541). The multi-RAS inhibitor daraxonrasib (RMC-6236) also shows potential for use in KRAS G12V-mutant NSCLC and pancreatic cancer.
Dr. Yang explained that while KRAS mutations are easily detected using existing NGS and PCR technologies, they do not present the same workflow challenges seen with IHC testing for HER2 and c-MET. He further addressed the implications of STK11 and KEAP1 mutations, which occur in approximately 20% of lung cancers and are often co-mutated with KRAS. These mutations promote an immunosuppressive tumor microenvironment, leading to resistance against single-agent PD-1/PD-L1 inhibitors.
Analysis from the phase 3 POSEIDON trial (NCT03164616) indicated that adding a CTLA-4 inhibitor to PD-L1 inhibitor therapy plus chemotherapy could enhance progression-free and overall survival for patients with STK11/KEAP1 mutations. This positions these mutations as potential biomarkers for escalating checkpoint therapies.
Dr. Yang discussed MTAP deletions, which occur in about 18% of lung cancers and are linked to poor treatment outcomes, particularly with immunotherapy. The identification of MTAP deletions can be achieved through NGS, which detects homozygous deletions without additional tissue requirements. He outlined a diagnostic workflow that starts with NGS for initial screening, followed by confirmatory IHC testing when MTAP status is retained or borderline.
Addressing the limitations of tissue availability, Dr. Yang stated, “Despite this progress, I think tissue will still be the issue. We have the same small biopsy, and we’re required to test for an expanding list of biomarkers that may require additional separate tissues.”
Another promising target in NSCLC is TROP2, a cell surface protein widely expressed in these tumors. The anti-TROP2 ADC datopotamab deruxtecan-dlnk (Dato-DXd; Datroway) is being evaluated as a second-line treatment. The phase 3 TROPION-Lung01 study (NCT04656652) indicated a progression-free survival benefit with Dato-DXd compared to docetaxel, although no statistically significant overall survival advantage was observed.
To enhance predictive power, researchers developed an AI-driven method that employs computational pathology to assess TROP2 expression levels. This method quantifies TROP2 staining on IHC slides, translating the results into a continuous score. Retrospective analysis showed that TROP2 scoring could predict higher response rates and longer progression-free survival with Dato-DXd.
While these findings are promising, Dr. Yang emphasized the need for prospective validation in independent cohorts. He also raised concerns about the current reliance on proprietary digital pathology systems, which may limit broader accessibility and implementation across different platforms.
As the field of lung cancer management progresses, it is clear that the focus is shifting from a purely genomic perspective to a more comprehensive approach that incorporates protein analysis, AI insights, and innovative therapeutic strategies like synthetic lethality. “We’re at a point where we should be starting to explore the feasibility of multiplex IHC similar to what we did with molecular markers and NGS,” Dr. Yang concluded. “In the next few years, broad-panel NGS and IHC, along with AI, are going to be the cornerstones of comprehensive biomarker testing in lung cancer.”
