Generation and validation of a predictive model for estimating survival among patients with EGFR-mutant non-small cell lung cancer

Although epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) have become the standard therapy for patients with EGFR-mutant non-small cell lung cancer (NSCLC), treatment outcomes vary significantly. Previous studies have indicated that concurrent mutations may compromise the effectiveness of first-line EGFR-TKIs. However, given the high cost of next-generation sequencing, this information is often inaccessible in routine clinical practice. A prediction model based on pre-treatment clinical characteristics may thus offer a more practical solution. This study established a nomogram based on pretreatment clinical characteristics to stratify patients according to optimal treatment strategies. We retrospectively reviewed 761 patients with EGFR-mutant NSCLC who received first- or second-generation EGFR-TKIs at a tertiary referral center between 2010 and 2019. The pretreatment clinical characteristics and progression-free survival data were collected. Using COX proportional hazard regression analysis, we constructed a nomogram based on seven clinically significant prognostic factors: sex, Eastern Cooperative Oncology Group performance status, histology subtype, mutation subtype, stage, and metastasis to the liver and brain. Our nomogram could stratify patients into three groups with different risks for disease progression and was validated in a patient cohort from other hospitals. This risk stratification can provide additional information for determining the optimal first-line treatment strategy for patients with EGFR-mutant NSCLC.