Background: Quantifying the relative contributions of genetic and environmental factors in cancer incidence informs etiology and opportunities for prevention. Studies based on large twin cohorts afford a unique opportunity to directly estimate the relative contribution of these factors. We are updating an analysis within the Nordic Twin Registry of Cancer (NorTwinCan), with the expansion of the original cohort, addition of a national twin registry for Norway, and on average 10 additional years of follow-up for cancer incidence.

Methods: NorTwinCan includes 105,306 monozygotic (MZ) and dizygotic (DZ) twin pairs (N=210,612 total) from population-based registries in Denmark, Finland, Norway and Sweden. The twins were followed prospectively for cancer incidence and mortality by linkage with national cancer registries through 2008, with up to six decades of follow-up. We calculated zygosity- and sex-specific proband concordance rates for 20 unique cancer sites, and estimated the relative risk (RR) of a cancer in an individual given his/her co-twin had cancer compared to individuals where the co-twin did not have cancer.

Results: During follow-up, 29,599 cancers were diagnosed. The co-twin of a twin with cancer had an increased risk of having the same cancer for most malignancies. For prostate cancer, the proband concordance rates were 0.298 among MZ twins and 0.138 among DZ twins. These estimates translate to a 29.8 percent probability that the MZ genetically identical twin of a man with prostate cancer will also be diagnosed with prostate cancer, and a 13.8 percent probability for DZ twins who are genetically alike as siblings. For MZ twins, the relative risk of prostate cancer was 19.8 if the co-twin had cancer, four-fold greater than the relative risk estimate for DZ pairs. For breast cancer among women, proband concordance rates were 0.173 for MZ and 0.110 for DZ twins. The relative risk of breast cancer if the co-twin had breast cancer was higher in MZ (RR=7.3) than DZ (RR=3.3) twins. The relative risk estimates were similarly higher in MZ compared to DZ twin pairs for cancers of the lung (e.g. RRMZ=23.5 versus RRDZ=8.2 for women), stomach (RRMZ=24.8 versus RRDZ=9.1 for women), testes (RRMZ=33.8 versus RRDZ=16.0), bladder (RRMZ=11.3 versus RRDZ=4.2 for men), corpus uteri (RRMZ=10.1 versus RRDZ=3.1), and ovary (RRMZ=10.3 versus RRDZ=2.6). Quantitative genetic modeling to estimate heritability of each cancer site is ongoing, and will account for censoring and competing causes of death.

Discussion: These preliminary analyses indicate a strong genetic component in the etiology of most malignancies, based on the greater relative risk and concordances in MZ versus DZ twin pairs. NorTwinCan is the largest twin study of cancer, and quantitative genetic modeling will quantify the relative contributions of genetic and environmental factors to variation in cancer incidence with greater precision than previously estimated, including for the more rare cancers.

Citation Information: Cancer Prev Res 2011;4(10 Suppl):A63.