Establishing a Prediction Model for the Efficacy of Platinum-Based Chemotherapy in NSCLC Based on a Two Cohorts GWAS Study

Platinum drugs combined with other agents have been the first-line treatment for non-small cell lung cancer (NSCLC) in the past decades. To better evaluate the efficacy of platinum-based chemotherapy in NSCLC, we establish a platinum chemotherapy response prediction model. Here, a total of 217 samples from Xiangya Hospital of Central South University were selected as the discovery cohort for a genome-wide association analysis (GWAS) to select SNPs. Another 216 samples were genotyped as a validation cohort. In the discovery cohort, using linkage disequilibrium (LD) pruning, we extract a subset that does not contain correlated SNPs. The SNPs with p < 10(-3) and p < 10(-4) are selected for modeling. Subsequently, we validate our model in the validation cohort. Finally, clinical factors are incorporated into the model. The final model includes four SNPs (rs7463048, rs17176196, rs527646, and rs11134542) as well as two clinical factors that contributed to the efficacy of platinum chemotherapy in NSCLC, with an area under the receiver operating characteristic (ROC) curve (AUC) of 0.726.

Automated summary

A platinum chemotherapy response prediction model was established to evaluate the efficacy of platinum-based chemotherapy in non-small cell lung cancer (NSCLC). A discovery cohort of 217 samples and a validation cohort of 216 samples were selected for genome-wide association analysis (GWAS). SNPs with p < 10(-3) and p < 10(-4) were selected for modeling. The final model includes four SNPs and two clinical factors, with an AUC of 0.726.