Identification of single nucleotide polymorphisms associated with survival of lung cancer patients using logistic regression for grouped survival data.


Session type:

Nicosha De Souza1, Olaide Raji1, Martin Ledson2, Richard Page2, Russell Hyde1, John K Field1
1University Of Liverpool, Liverpool, UK, 2Liverpool Heart & Chest Hospital, Liverpool, UK


Genome wide association analysis in lung cancer were previously carried out to identify markers that confer susceptibility (1, 2) and predict survival (3). The current study aims to identify genetic polymorphisms associated with mortality from lung cancer.


Histologically or cytologically consented cases in the Liverpool Lung Project between 20-80 years of age were recruited and genotyped using the Illumina Human Hap 300k beadchip array. Data quality controls were applied. Cause of death was established using the ONS registry. 308617 SNPs were univariately tested in a logistic regression model for grouped survival data (4, 5) adjusting for age at diagnosis, gender and smoking pack years.The effect of SNPs on mortality was tested in an additive mode.


The genome wide analysis identified 22 significant SNPs at a P<10-4.The most interesting SNP (rs13095440) with an OR (odds ratio) of 3.38 (95%CI: 1.85 - 6.18) is located in the BBX gene and has previously been identified in a genome wide study of cigarette smoking behaviours (6). Another SNP (rs8063107) located in the CDH13 gene depicted an OR of 2.13 (95%CI:1.48 - 3.07) and has previously been reported to be hypermethylated in tumour and serum of non small lung cancer patients (7) while the SNP (rs9424148) located in the DHTKD1 gene associated with cancer (8) produced an OR of 2.34 (95%CI:1.53 - 3.58).


The analysis identified important SNPs that increased the probability of observing death due to lung cancer. They maybe of value with other clinical and epidemiological factors to determine treatment regime and improve the clinical outcome. These SNPs need to be further validated in an independent survival study.