572P - Multivariable five-year survival prediction model for prognosing patients with EGFR-mutated NSCLC treated with EGFR-TKIs

Background

Lung cancer was the leading cause of mortality in 2022, with 38.4% of patients having non-small cell lung cancer (NSCLC) with epidermal growth factor receptor (EGFR) mutations. EGFR tyrosine kinase inhibitors (TKIs) are the first-line treatment for EGFR-mutated NSCLC. A limited number of patients achieved five-year survival. This retrospective study aimed to provide real-world evidence and explore the factors affecting the five-year survival of patients with EGFR-mutated NSCLC.

Methods

This study included patients with EGFR-mutated NSCLC treated in Chang Gung Memorial Hospitals between 2011 and 2016, ensuring at least five years of follow-up or mortality. Odds ratios (ORs) were calculated using univariate and multivariate analyses. A scoring system based on a logistic regression model was designed to evaluate factor weightings on patients’ five-year survival. Their overall survival probability was estimated using the Kaplan-Meier model. The model’s accuracy was assessed using the area under the receiver operating characteristic curve (AUC).

Results

Of 1,873 enrolled patients, 185 were lost to follow-up within five years, leaving 1,787 for analysis. Of patients achieving five-year survival, more were female, age 65 years, had performance scores of 0∼1, no metastases, and adequate objective responses and disease control. A scoring system was developed by assigning points to each prognostic factor associated with increased risk of not achieving five-year survival: age > 65 years (1 point); performance score of 2∼4 (2 points); stage IV disease (1 point); liver (2 points), bone (1 point), or pleura (1 point) metastasis; and poor disease control (2 points). In the Kaplan-Meier model, the estimated five-year survival rate was 39.4%, 13.0%, 7.2%, and 2.2% for the low-risk (0∼1 point), intermediate-risk (2 points), high-risk (3 points), and very-high-risk (4∼10 points) groups, respectively. The prediction model’s AUC was 0.787 (95% CI: 0.752∼0.821), indicating fair accuracy.

Conclusions

We proposed a scoring system based on real-world data for predicting the five-year survival of patients with EGFR-mutated NSCLC treated with EGFR-TKIs.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Chang Gung Memorial Hospital (Linkou Branch).

Disclosure

All authors have declared no conflicts of interest.