SOX2 and PIK3CA Amplification as Prognostic Biomarkers in Advanced Lung Squamous Cell Carcinoma

IASLC TTLC 2026

Feb 18, 2026
Oncology
Presentation

Fawzi Abu Rous, Michelle Weitz, Radhika Gutta, Kayla Layng, Jen Godden, Stamatina Fragkogianni, Shannon Carskadon, Geena Mody, Sudha Sadasivan, Bindu Potugari, Shirish Gadgeel, Nallasivam Palanisamy

Introduction
Chromosome 3q amplification is a frequent genomic alteration in lung squamous cell carcinoma (LUSC), implicated in tumor initiation, progression, and immune evasion. We identified a minimal common region (MCR) spanning 25 genes with potential prognostic and predictive value. Given the strong correlation between MCR amplification and SOX2/PIK3CA amplification, we used these genes as practical surrogates for MCR status. We characterized genomic, immunologic, and clinical features of MCR-amplified tumors and their association with outcomes following immune checkpoint inhibitor (ICI)-based therapy.

 

Methods
We used Tempus Lens (Tempus AI, Inc., Chicago, IL) to query the Tempus multimodal de-identified database and identify patients with stage III/IV LUSC who underwent both xT (DNA) and xR (RNA) testing with tumor purity ≥30% (N = 1,408). Patients were classified as MCR-amplified based on the co-amplification (copy number ≥4) of SOX2 and PIK3CA. Demographic and clinical characteristics, somatic and germline alterations, quanTIseq-based immune cell infiltration estimates, immune checkpoint proteins, and biomarkers (TMB, PD-L1, MSI) were compared. The objective response rate (ORR) was defined as complete or partial response at 90 and 180 days from the start of first-line treatment (1L). Real-world time to next treatment (rwTTNT) and overall survival (rwOS) were measured from the start of 1L and analyzed using risk-set-adjusted Kaplan-Meier curves, log-rank tests, and Cox models.

 

Results
643 patients (46%) were MCR-amplified. Compared to the non-amplified group, they were significantly more often male (68% vs. 61%, p=0.006), had a smoking history (98% vs. 92%, p<0.001), and had Stage III disease (35% vs. 26%, p<0.001).

 

MCR-amplified tumors revealed a significant enrichment of alterations in NFE2L2 (21.9% vs. 9.3%), TP53 (94.4% vs. 85.2%), CDKN2A (43.2% vs. 33.6%), CCND1 (10.7% vs. 5.6%), FGF3 (11.7% vs. 6.8%), and FGF4 (11.8% vs. 7.1%) compared to non-amplified tumors (all q < 0.05). Conversely, KRAS (1.7% vs. 8.8%), TERT (1.4% vs. 7.6%), and SMARCA4 (1.1% vs. 5.2%) alterations were enriched in non-amplified tumors (q < 0.05).

 

Immune analysis revealed an immune-excluded phenotype in MCR-amplified tumors, characterized by lower infiltration of B cells, NK cells, CD8+ T cells, Tregs, neutrophils, and M1 macrophages (all p < 0.01). PDCD1 (PD-1) and CTLA-4 were reduced in MCR-amplified tumors (both p < 0.001), while PD-L1, TMB, and MSI were similar.

 

Treatment exposure was comparable (63% received chemo-immunotherapy). ORR did not differ (65% each). The median time to tumor recurrence (rwTTNT) was prolonged in MCR-amplified tumors (11.8 vs. 10.6 months). rwOS was significantly longer in MCR-amplified patients (17.0 vs. 13.7 months, p = 0.04; HR, 1.17; 95% CI, 1.01–1.37).

 

Conclusions
MCR amplification, characterized by SOX2 and PIK3CA, defines a LUSC subtype that is enriched for TP53, NFE2L2, and CDKN2A alterations and depleted in KRAS and TERT. Although immunologically "cold," patients with MCR-amplified tumors showed superior survival with ICI-based therapy. This positions MCR amplification as a potential biomarker to improve patient selection in LUSC, warranting further clinical validation.