Fig. 7: Differential survival of erlotinib-treated profile 1 and profile 2 NSCLC patients. | Nature Communications

Fig. 7: Differential survival of erlotinib-treated profile 1 and profile 2 NSCLC patients.

From: Biochemical and structural basis for differential inhibitor sensitivity of EGFR with distinct exon 19 mutations

Fig. 7: Differential survival of erlotinib-treated profile 1 and profile 2 NSCLC patients.The alternative text for this image may have been generated using AI.

Data for survival outcomes on erlotinib treatment were obtained from the GENIE+ database54 (a, b) and from an institutional cohort (c, d) as described in the main text for patients with profile 1 (blue) or profile 2 (red) exon 19 mutations. a Using GENIE+ data, a profile 1 mutation predicts a significantly shorter median PFS (2.3 months) than a profile 2 mutation (8.5 months) on erlotinib (P < 0.0001). We also performed the same analysis excluding the common ΔE746-A750 deletion, since we know this is erlotinib sensitive, and the prediction still holds (b), where a profile 1 mutation predicts median PFS of 2.3 months and profile 2 mutation 11.2 months (P < 0.0001). Predicted profile 1 mutations in the GENIE+ dataset were restricted to: ΔL747-A750InsP (4 patients) plus ΔE746-L747InsIP (1 patient). Predicted profile 2 mutations were restricted to: ΔE746-A750 (43), ΔL747-P753InsS (5), ΔL747-T751 (3), ΔE746-S752InsV (3), ΔS752-I759 (2), ΔL747-S752 (2), ΔE746-S752InsI (1), ΔE746-T751InsA (1), ΔL747-T751InsP (1), and ΔE746-T751InsIP (1). c Similarly, among patients treated with erlotinib at the Yale Cancer Center, a profile 1 exon 19 mutation predicts a significantly shorter median PFS (1.8 months) compared to a profile 2 mutation (9.8 months), with P = 0.0002. After adjusting for baseline covariates of age, sex, race, and smoking history using the Cox proportional hazards model78, the hazard for progression was nine times greater for patients with tumors with profile 1 compared to profile 2 mutations (95% confidence interval [2.6–31.4], P = 0.0005). d This holds true for overall survival (OS) on erlotinib as well, with median OS of 24.5 months for profile 1 and 41.3 for profile 2 (P = 0.03). The adjusted hazard ratio for death was 3.0 (95% confidence interval [1.2–7.7], P = 0.02) for patients with tumors harboring profile 1 mutations compared to profile 2 mutations. Predicted profile 1 mutations in the Yale dataset were restricted to: ΔL747-A750InsP (6 patients). Predicted profile 2 mutations were restricted to: ΔE746-A750 (55), ΔL747-P753InsS (5), ΔL747-T751 (3), ΔL747-T751InsP (2), ΔE746-S752InsV (1), ΔL747-T751InsA (1), ΔL747-S752 (1), ΔE746-T751InsVP (1), ΔT751-I759InsN (1).

Back to article page