Polygenic hazard score to guide screening for aggressive prostate cancer: development and validation in large scale cohorts

• DOI: 10.1136/bmj.j5757
• Publication Date: 2018-01-11T00:00:55Z
• Authors: Tyler M Seibert; Chun Chieh Fan; Yunpeng Wang; Verena Zuber; Roshan Karunamuni; J Kellogg Parsons; Rosalind A Eeles; Douglas F Easton; ZSofia Kote-Jarai; Ali Amin Al Olama; Sara Benlloch Garcia; Kenneth Muir; Henrik Grönberg; Fredrik Wiklund; Markus Aly; Johanna Schleutker; Csilla Sipeky; Teuvo LJ Tammela; Børge G Nordestgaard; Sune F Nielsen; Maren Weischer; Rasmus Bisbjerg; M Andreas Røder; Peter Iversen; Tim J Key; Ruth C Travis; David E Neal; Jenny L Donovan; Freddie C Hamdy; Paul Pharoah; Nora Pashayan; Kay-Tee Khaw; Christiane Maier; Walther Vogel; Manuel Luedeke; Kathleen Herkommer; Adam S Kibel; Cezary Cybulski; Dominika Wokolorczyk; Wojciech Kluzniak; Lisa Cannon-Albright; Hermann Brenner; Katarina Cuk; Kai-Uwe Saum; Jong Y Park; Thomas A Sellers; Chavdar Slavov; Radka Kaneva; Vanio Mitev; Jyotsna Batra; Judith A Clements; Amanda Spurdle; Manuel R Teixeira; Paula Paulo; Sofia Maia; Hardev Pandha; Agnieszka Michael; Andrzej Kierzek; David S Karow; Ian G Mills; Ole A Andreassen; Anders M Dale
• Keywords: Male; Aging; Biomedical and clinical sciences; Outcome Assessment; IMPACT; Cohort Studies; PSA; Outcome Assessment, Health Care; 2.1 Biological and endogenous factors; Psychology; Aetiology; European Continental Ancestry Group/genetics; Age of Onset; Early Detection of Cancer; Cancer; risk; OUTCOMES; Prostate cancer; OVERDIAGNOSIS; Prostate Cancer; Prostatic Neoplasms/blood; Single Nucleotide; Middle Aged; Polymorphism, Single Nucleotide/genetics; STATISTICS; ddc:; 3. Good health; Public Health and Health Services; Medicine; TRIAL; Kallikreins; /dk/atira/pure/core/keywords/centre_for_surgical_research; Single Nucleotide/genetics; Life Sciences & Biomedicine; Urologic Diseases; SUSCEPTIBILITY LOCI; Genotype; Oncology and Carcinogenesis; Clinical Sciences; European Continental Ancestry Group; Prostate-Specific Antigen/analysis; 610; General & Internal; Polymorphism, Single Nucleotide; Risk Assessment; Disease-Free Survival; White People; Outcome Assessment (Health Care); Medicine, General & Internal; AGE; Predictive Value of Tests; General & Internal Medicine; 616; Genetics; Journal Article; BREAST-CANCER; Humans; Kallikreins/analysis; Genetic Testing; name=Centre for Surgical Research; Polymorphism; Aged; Science & Technology; Biomedical and Clinical Sciences; RISK PREDICTION; Prevention; screening; Research; Health sciences; Prostatic Neoplasms; prediction; /dk/atira/pure/core/keywords/centre_for_surgical_research; name=Centre for Surgical Research; Prostate-Specific Antigen; ta3122; PREVENTION; PRACTICAL Consortium*; Survival Analysis; Health Care; Good Health and Well Being; age; Early Detection of Cancer/methods; genetic

<h3>Abstract</h3> <h3>Objectives</h3> To develop and validate a genetic tool to predict age of onset aggressive prostate cancer (PCa) guide decisions who screen at what age. <h3>Design</h3> Analysis genotype, PCa status, select single nucleotide polymorphisms (SNPs) associated with diagnosis. These were incorporated into survival analysis estimate their effects on diagnosis (that is, not eligible for surveillance according National Comprehensive Cancer Network guidelines; any Gleason score ≥7, stage T3-T4, PSA (prostate specific antigen) concentration ≥10 ng/L, nodal metastasis, distant metastasis). The resulting polygenic hazard is an assessment individual risk. final model was applied independent dataset containing genotype screening data. calculated these men test prediction free from PCa. <h3>Setting</h3> Multiple institutions that members international PRACTICAL consortium. <h3>Participants</h3> All consortium participants European ancestry known age, quality assured custom (iCOGS) array development comprised 31 747 men; the validation 6411 men. <h3>Main outcome measures</h3> Prediction in set. <h3>Results</h3> In set, 54 highly significant predictor (z=11.2, P&lt;10⁻¹⁶). When set high scores (&gt;98th centile) compared those average (30th-70th centile), ratio 2.9 (95% confidence interval 2.4 3.4). Inclusion family history combined did improve (P=0.59), performance remained when accounted for. Additionally, positive predictive value increased increasing score. <h3>Conclusions</h3> Polygenic can be used personalised risk estimates