Table 7. Akaike information criterion (AIC) calculations for the OS and DSS multivariate models with or without the SLC6A4 genotypes.

Endpoint	−2 log likelihood	Number of parameters	AIC	ΔAIC
OS
Model 1	1545.859	6	1557.859	53.792
Model 2	1514.252	7	1528.252	24.185
Model 3	1490.067	7	1504.067	0
DSS
Model 1	977.655	6	989.655	30.834
Model 2	962.69	7	976.69	17.869
Model 3	944.821	7	958.821	0

Model 1 contains the clinical variables that were significant in the multivariate analyses; namely age, stage, grade, and MSI-H status (for both OS and DSS); Model 2 contains the SLC6A4-rs12150214 genotype as a variable in addition to the clinical variables in Model 1, and Model 3 contains the combined SLC6A4 genotypes (for rs4251417, rs12150214, rs140700) in addition to the variables in Model 1. ΔAIC is the difference in AIC values of a model and the best model, which is Model 3 for both OS and DSS.