Negative Age-Dependence of the Polygenic Risk Score Gradient for Colorectal Cancer

As a researcher developing cancer risk models, I read with interest the article by Archambault et al.1Archambault A.N. al.Gastroenterology. 2020; 158: 1274-1286.e12Abstract Full Text PDF PubMed Scopus (27) Google Scholar on difference in association of polygenic score (PRS) colorectal (CRC) between early-onset and late-onset CRC defined threshold for age at diagnosis 50 years. The authors concluded that from “an analysis associations per standard deviation PRS, we found cumulative burden CRC-associated common genetic variants to associate cancer, be more strongly associated than ….” But perhaps they should have said earlier because, their own data, PRS gradient appears change continuously across range. Supplementary Table 4 shows odds ratio went 1.75 1.60 1.52 1.44 as increased 40s 50s 60s 70s (P = 3.44 × 10?10). On log(odds ratio) scale, these gradients are 0.56, 0.47, 0.42, 0.37, which is an almost linear decline approximately 0.06, or 14%, decade. This important observation, not least because it contrary what has been breast where currently best PRSs overall, estrogen receptor negative discernibly dependent (the decreases 0.01 decade; P 0.39).2Mavaddat N. al.Am J Hum Genet. 2019; 104: 21-34Abstract (242) null age-dependency implies that, might predicting death cancer. other hand, CRC, likely this disease, wide Note if disease mortality, average would expected decrease cases, controls, but slower rate latter.3Apicella C. al.Breast Cancer Res. 2003; 5: R206-R216Crossref (35) It even take into account age-dependence state, “Analyses along environmental lifestyle factors, identify younger individuals who benefit preventive measures.” Their data show statement only relevant prevention before years, being diagnosed next 3 decades: there no threshold. indicated strength its value shown previously, vast amount information after years (89% study sample) used commented published paper. great address presenting analyzing sex both cases estimating relative risk, error, using individual-level data. done men women, separately combined, so effect absolute can quantified purposes suggested targeting screening risk. Cumulative Burden Colorectal Cancer–Associated Genetic Variants Is More Strongly Associated With Early-Onset vs Late-Onset CancerGastroenterologyVol. 158Issue 5PreviewEarly-onset (CRC, persons old) increasing incidence; yet, absence family history population lacks harmonized recommendations prevention. We aimed determine whether developed 95 was CRC. Full-Text