Table 4.

FPRP values for the 10 SNPs associated with IMPC risk

FPRP based on prior:
SNPFirst GWAS siteMeta-OR (95% CI)aPowerbReported P0.100.010.001
rs578776Nicotine dependence1.15 (1.05–1.26)0.820.0040.030.250.77
rs11249433Breast cancer1.16 (1.04–1.28)0.750.0050.040.290.81
rs3803662Breast cancer1.13 (1.03–1.23)0.920.0060.440.340.84
rs4975616Lung cancer0.89 (0.82–0.98)0.930.0140.150.660.95
rs8042374Lung cancer1.12 (1.02–1.24)0.910.0150.120.610.94
rs4857841Prostate cancer0.90 (0.82–0.98)0.960.0200.130.610.94
rs4785763Melanoma1.11 (1.01–1.22)0.950.0230.220.760.97
rs965513Thyroid cancer0.89 (0.80–0.99)0.890.0260.240.780.95
rs5945619Prostate cancer1.09 (1.01–1.18)0.990.0350.210.750.97
rs3790844Pancreatic cancer1.23 (1.01–1.49)0.400.0350.440.890.99
rs10086908Prostate cancer1.13 (1.01–1.27)0.840.0400.300.830.98
rs4939827Colorectal cancer1.09 (1.00–1.19)0.980.0470.330.850.98

NOTE: The values in bold indicate that the associations remained significant at an FPRP threshold of 0.20 and a prior probability range of 0.1 through 0.001 of detecting an OR of 1.2 or 0.83.

aORs from Table 2, models were adjusted for age at diagnosis for the index cancer (continuous), sex (MEC only), study design (WHI only: clinical trial vs. observational study), the most significant principal components of genetic ancestry (five principal components for the MEC and three principal components for WHI), index cancer site, stage, and diagnosis year (to account for treatment cohort effects).

bStatistical power calculated using the recessive model, except where noted, is the power to detect an OR of 1.2 or 0.83 at a level of 0.05.