The association of TP53 codon 72 polymorphism with cancer susceptibility remains uncertain and varies with ethnicity. Northeast India represents a geographically, culturally, and ethnically isolated population. The area reports high rate of tobacco usage in a variety of ways of consumption, compared with the rest of Indian population. A total of 411 cancer patients (161 lung, 134 gastric, and 116 oral) and 282 normal controls from the ethnic population were analyzed for p53 codon 72 polymorphism by polymerase chain reaction-restriction fragment length polymorphism. No significant difference in genotypic distribution of p53 between cases and controls was observed. Results suggested betel quid chewing as a major risk factor for all the three cancers (odds ratio [OR]=3.54, confidence interval [CI]=2.01-6.25, p<0.001; OR=1.74, CI=1.04-2.92, p=0.03; and OR=1.85, CI=1.02-3.33, p=0.04 for lung, gastric, and oral cancers, respectively). Tobacco smoking was associated with risk of lung and oral cancers (OR=1.88, CI=1.11-3.19, p=0.01 and OR=1.68, CI=1.00-2.81, p=0.04). Interactions between p53 genotypes and risk factors were analyzed to look for gene-environment interactions. Interaction of smoking and p53 genotype was significant only for oral cancer. Interactions of betel quid with p53 genotypes in lung cancer showed significant increase for all the three genotypes, indicating a major role of betel quid (OR=5.90, CI=1.67-20.81, p=0.006; OR=5.44, CI=1.67-17.75, p=0.005; and OR=5.84, CI=1.70-19.97, p=0.005 for Arg/Arg, Arg/Pro, and Pro/Pro, respectively). In conclusion, high incidence of these cancers in northeast India might be an outcome of risk habits; further, tissue- and carcinogen-specific risk modification by p53 gene is probable.