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

he bmj | BMJ 2018;360:j5757 | doi: 10.1136/bmj.j5757 1 RESEARCH Polygenic haza d sco e o guide sc eening o agg essi e p os a e cance : de elopmen and alida ion in la ge scale coho s Tyle M Seibe ,1,2 Chun Chieh Fan,1,3 Yunpeng Wang,4 Ve ena Zube ,4,5 Roshan Ka unamuni,1,2 J Kellogg Pa sons,6 Rosalind A Eeles,7,8 Douglas F Eas on,9 ZSo ia Ko e-Ja ai,7 Ali Amin Al Olama,9,10 Sa a Benlloch Ga cia,9 Kenne h Mui ,11,12 Hen ik G önbe g,13 F ed ik Wiklund,13 Ma kus Aly,13,14,15 Johanna Schleu ke ,16,17,18 Csilla Sipeky,16,17 Teu o LJ Tammela,19 Bø ge G No des gaa d,20,21 Sune F Nielsen,20,21 Ma en Weische ,21 Rasmus Bisbje g,22 M And eas Røde ,23 Pe e I e sen,20,23 Tim J Key,24 Ru h C T a is,24 Da id E Neal,25,26 Jenny L Dono an,27 F eddie C Hamdy,25 Paul Pha oah,28 No a Pashayan,28,29 Kay-Tee Khaw,30 Ch is iane Maie ,31 Wal he Vogel,31 Manuel Luedeke,31 Ka hleen He komme ,32 Adam S Kibel,33 Ceza y Cybulski,34 Dominika Wokolo czyk,34 Wojciech Kluzniak,34 Lisa Cannon-Alb igh ,35,36 He mann B enne ,37,38,39 Ka a ina Cuk,37 Kai-Uwe Saum,37 Jong Y Pa k,40 Thomas A Selle s,41 Cha da Sla o ,42 Radka Kane a,43 Vanio Mi e ,43 Jyo sna Ba a,44 Judi h A Clemen s,44 Amanda Spu dle,44,45,46 Manuel R Teixei a,47,48 Paula Paulo,47 So ia Maia,47 Ha de Pandha,49 Agnieszka Michael,49 And zej Kie zek,49 Da id S Ka ow,1,50 Ian G Mills,4,25,51 Ole A And eassen,4 Ande s M Dale1,50,52; The PRACTICAL Conso ium* ABSTRACT OBJECTIVES To de elop and alida e a gene ic ool o p edic age o onse o agg essi e p os a e cance (PCa) and o guide decisions o who o sc een and a wha age. DESIGN Analysis o geno ype, PCa s a us, and age o selec single nucleo ide polymo phisms (SNPs) associa ed wi h diagnosis. These polymo phisms we einco po a ed in o a su i al analysis o es ima e hei e ec s on age a diagnosis o agg essi e PCa ( ha is, no eligible o su eillance acco ding o Na ional Comp ehensi e Cance Ne wo k guidelines; any o Gleason sco e ≥7, s age T3-T4, PSA (p os a e speci ic an igen) concen a ion ≥10 ng/L, nodal me as asis, dis an me as asis). The esul ing polygenic haza d sco e is an assessmen o indi idualgene ic isk. The inal model was applied o an independen da ase con aining geno ype and PSAsc eening da a. The haza d sco e was calcula ed o hese men o es p edic ion o su i al ee omPCa. SETTING Mul iple ins i u ions ha we e membe s o in e na ional PRACTICAL conso ium. PARTICIPANTS All conso ium pa icipan s o Eu opean ances y wi h known age, PCa s a us, and quali y assu ed cus om (iCOGS) a ay geno ype da a. The de elopmen da ase comp ised 31 747 men; he alida ion da ase comp ised 6411 men. MAIN OUTCOME MEASURES P edic ion wi h haza d sco e o age o onse o agg essi e cance in alida ion se . RESULTS In he independen alida ion se , he haza d sco e calcula ed om 54 single nucleo ide polymo phisms was a highly signi ican p edic o o age a diagnosis o agg essi e cance (z=11.2, P<10−16). When men in he alida ion se wi h high sco es (>98 h cen ile) we e compa ed wi h hose wi h a e age sco es (30 h-70 h cen ile), he haza d a io o agg essi e cance was 2.9 (95% con idence in e al 2.4 o 3.4). Inclusion o amily his o y in a combined model did no imp o e p edic ion o onse o agg essi e PCa (P=0.59), and polygenic haza d sco e pe o mance emained high when amily his o y was accoun ed o . Addi ionally, he posi i e p edic i e alue o PSA sc eening o agg essi e PCa was inc eased wi h inc easing polygenic haza d sco e. CONCLUSIONS Polygenic haza d sco es can be used o pe sonalised gene ic isk es ima es ha can p edic o age a onse o agg essi e PCa. In oduc ion P os a e cance (PCa) is a majo heal h p oblem, wi h o e a million new cases and o e 300 000 associa ed dea hs es ima ed wo ldwide in 2012.1 An in e na ional andomised con olled ial showed ha sc eening o p os a e speci ic an igen (PSA) esul ed in a 27% Fo numbe ed a ilia ions see end o a icle. Co espondence o: T M Seibe [email p o ec ed] A M Dale [email protected] Addi ional ma e ial is published online only. To iew please isi he jou nal online. Ci e his as: BMJ 2018;360:j5757 h p://dx.doi.o g/10.1136/bmj.j5757 Accep ed: 4 Decembe 2017 WhAT IS AlReAdy knoWn on ThIS TopIC Sc eening o p os a e cance (PCa) wi h es s o p os a e speci ic an igen (PSA) can lead o ea ly de ec ion and allow o cu a i e ea men , bu uni e sal sc eening also has conside able disad an ages o he men who migh ne e de elop agg essi e disease Ideally, physicians would iden i y and sc een pa ien s a high isk o de eloping agg essi e PCa o PCa a a young age A p ac ical clinically use ul ool o p edic age o onse is no ye a ailable WhAT ThIS STudy AddS This s udy p esen s and alida es a no el polygenic haza d sco e ha is an indica o o age a onse o agg essi e PCa The sco e is a ela i ely inexpensi e assessmen o an indi idual man’s age speci ic isk and p o ides objec i e in o ma ion on whe he a gi en pa ien migh bene i om PSA sc eening RESEARCH 2 doi: 10.1136/bmj.j5757 | BMJ 2018;360:j5757 | he bmj educ ion in PCa mo ali y.2 Because o conce ns o e a high a e o alse posi i e esul s, in addi ion o agg essi e ea men o appa en ly indolen disease, howe e , many clinical guidelines do no endo se uni e sal sc eening and ins ead s ess he impo ance o aking in o accoun indi idual pa ien isk ac o s o decide whe he o sc een.3-5 The goal is o a oid unnecessa y sc eening while s ill iden i ying men a high isk o whom sc eening and ea ly de ec ion can educe mo bidi y and mo ali y. A pa ien ’s gene ic p edisposi ion could be c i ical o he decision o whe he and when o o e sc eening. Genome- wide associa ion s udies (GWAS) ha e shown gene ic a ian s associa ed wi h inc eased isk o PCa.6 7 These de elopmen s, combined wi h he ecen accessibili y o geno yping, p o ide an oppo uni y o cance sc eening in o med by gene ic isk.8 Wi h a combina ion o isk in o ma ion om an a ay o single nucleo ide polymo phisms (SNPs), polygenic models can es ima e an indi idual’s gene ic isk o de eloping he disease.9 P edic ed polygenic isk could imp o e clinical decisions such as who o sc een o PCa and a wha age.10 11 We used da a om 31 747 men o Eu opean ances y om he in e na ional PRACTICAL conso ium (h p:// p ac ical.ccge.medschl.cam.ac.uk/) o de elop a polygenic haza d sco e (PHS) o p edic ing age ela ed isk o de eloping agg essi e PCa. This is designed o use be o e he decision o whe he o sc een ( o example, wi h PSA) by p o iding a isk s a i ica ion s a egy o maximise sc eening e iciency. The haza d sco e was es ed in da a om an independen sc eening s udy (UK P o ecT12), wi h he hypo hesis ha i would be an indica o o a pa ien ’s inhe en gene ic isk o de eloping PCa a a ious ages in his li e ime and hus could guide PSA sc eening. Me hods De ini ion o agg essi e disease Conce ns abou o e diagnosis and o e ea men o indolen disease ha e in luenced discussion o PCa sc eening, whe eas he e is consensus ha agg essi e cance wa an s ea men .13 14 When possible, we he e o e ocus alida ion in his s udy on p edic ion o agg essi e disease, de ined as any umou ha would equi e adical ea men o a ypical heal hy man acco ding o guidelines om he Na ional Comp ehensi e Cance Ne wo k (NCCN)— ha is, no eligible o ac i e su eillance.14 This includes cance s wi h any o Gleason sco e ≥7, s age T3-T4, PSA concen a ion ≥10 ng/mL, nodal me as asis, o dis an me as asis. S age T2 umou s we e classi ied wi hou a subca ego y in ou da abase, so a pa ien wi h low Gleason sco e and low PSA concen a ion bu s age T2b o T2c would be conside ed low isk in his analysis e en hough NCCN guidelines would indica e ea men o in e media e isk; his was o ensu e ha no low isk umou s we e included as cases o agg essi e cance . Some addi ional analyses used age o diagnosis o any PCa ( a he han only agg essi e PCa) as complemen a y in o ma ion. Ano he seconda y analysis es ed p edic ion o “ e y agg essi e disease,” de ined as any o Gleason sco e ≥8, s age T3-4, posi i e nodes, o dis an me as ases. Pa icipan s De elopmen se To de elop he polygenic haza d sco e model, we ob ained geno ype and da a on age om 21 s udies om he PRACTICAL conso ium ( able A in appendix 1), ep esen ing 31 747 men (18 868 wi h any PCa, 10 635 wi h agg essi e PCa, 5406 wi h e y agg essi e PCa, 12 879 con ols) o geno ypic Eu opean ances y. Age was ei he a diagnosis o las ollow-up ( o con ols). Geno yping, pe o med wi h a cus om Illumina a ay (iCOGS), and quali y con ol s eps ha e been desc ibed p e iously.6 A o al o 201 043 single nucleo ide polymo phisms we e a ailable o analysis. We could no ca ego ise cance as agg essi e o no in 4803 o he men wi h cance because o incomple e da a on s aging; hese we e excluded om analyses o agg essi e cance . Valida ion se An independen s udy examined pe o mance o he model. The alida ion se came om he P o ecT s udy, which sc eened 82 429 men wi h PSA es ing and ound 8891 men wi h PSA concen a ion o e han he speci ied h eshold o 3.0 ng/L o highe , 2896 o whom ecei ed a diagnosis o PCa.12 Among hose indi iduals, we ob ained da a on geno ype and age o 6411 men (1583 wi h any PCa, 632 wi h agg essi e PCa, wi h 220 e y agg essi e PCa, 4828 con ols). S aging da a we e a ailable o all cases. This da ase was selec ed o alida ion because PSA esul s we e also a ailable o all pa icipan s a ime o ei he diagnosis o in e iew. Fu he de ails in appendix 1. Missing da a Du ing model de elopmen , we excluded single nucleo ide polymo phisms wi h call a es less han 95%. We impu ed missing calls o he emaining polymo phisms wi h he mean geno ype coun o ha allele ac oss all pa icipan s. Polygenic haza d sco e The polygenic haza d sco e was de eloped p e iously as a pa simonious su i al analysis model o p edic he ime o e en ou come (in his case, age o onse o PCa). I has been published elsewhe e,15 and u he de ails o applica ion he e a e desc ibed in appendix 1. The sco e is de ined as he ec o p oduc o a pa ien ’s geno ype ( Xi ) o n selec ed single nucleo ide polymo phisms and he co esponding pa ame e es ima es ( i ) om a Cox p opo ional haza ds eg ession (see equa ion). PHSx = Xiβi n i Σ Gene ic p edic ion speci ically o agg essi e PCa has p o ed elusi e, wi h mos single nucleo ide polymo phisms associa ed wi h agg essi e disease also RESEARCH he bmj | BMJ 2018;360:j5757 | doi: 10.1136/bmj.j5757 3 showing associa ion wi h any PCa.16 The e o e, in he in e es o maximising powe o selec polymo phisms associa ed wi h age o onse , we decided o ini ially include all cases om he de elopmen se ( ha is, any PCa) o gene a ion o he model. We hen es ed an al e na e s a egy ha limi ed gene a ion o cases o agg essi e cance o compa ison. The p ima y me ic o alida ion in bo h ins ances emained p edic ion o agg essi e cance in he independen alida ion se . To e i y whe he he polygenic haza d sco e accu a ely p edic s age a onse o agg essi e PCa, we calcula ed he sco e o all pa ien s in he alida ion se and es ed i as he sole p edic i e a iable in a Cox p opo ional haza ds eg ession model o age o diagnosis. Pa ien s in he alida ion se wi h a diagnosis o low isk disease (Gleason sco e ≤6, PSA concen a ion <10 ng/L, and s age T2N0M0 o lowe ) we e censo ed a ime o diagnosis, e lec ing he ac ha i is unknown i hey would la e ecei e a diagnosis o agg essi e disease o a wha age ha migh ha e occu ed. Signi icance was se a α o 0.01 o his and all subsequen Cox models. As an indica o o e ec size o he model, we calcula ed a haza d a io compa ing men wi h high sco es (>98 h cen ile) wi h hose wi h a e age isk (30 h-70 h cen ile). All haza d a ios p esen ed he e e e o he same pa e n: high e sus a e age isk. Because o e idence ha ini ially low isk disease o en p og esses o equi e ea men ,17-19 and because his migh be pa icula ly impo an o men wi h a diagnosis a a young age, we pe o med a seconda y analogous analysis o es o p edic ion o age o diagnosis o any PCa. We did a seconda y analysis o p edic ion o e y agg essi e disease. To u he assess he clinical signi icance o he polygenic haza d sco e, we looked a he posi i e p edic i e alue (PPV) o PSA es ing wi hin he alida ion se , wi h clinical diagnosis (including biopsy esul ) as he ideal. We posi ed ha isk s a i ica ion by cen iles o he sco e would e lec he unde lying incidence o PCa and he e o e also a ec he posi i e p edic i e alue o PSA es ing. De ails on he calcula ion o posi i e p edic i e alue a e in appendix 1. Ca ego ies o he sco e we e designa ed by cen ile compa ed wi h he young heal hy popula ion wi hin he de elopmen se — ha is, hose con ols aged <70. All cen iles epo ed in his manusc ip e e o his popula ion. To isualize dis ibu ion o he polygenic haza d sco e among cases o agg essi e PCa in he alida ion se , we gene a ed a Lo enz cu e.20-22 Compa ison wi h amily his o y One o he mos impo an isk ac o s used cu en ly o sc eening decisions is amily his o y.3 We compa ed amily his o y and polygenic haza d sco e o p edic ion o onse o agg essi e PCa using he same Cox model app oach as be o e, wi h he 5703 men (1405 wi h any PCa, 554 wi h agg essi e PCa, 4298 con ols) om men in he alida ion se wi h known amily his o y s a us (none o one o mo e a ec ed i s deg ee ela i es). Models we e cons uc ed wi h amily his o y alone, haza d sco e alone, o wi h bo h. These we e compa ed ia log likelihood es s. Pa ien in ol emen No pa ien s we e di ec ly in ol ed in designing he esea ch ques ion o in conduc ing he esea ch. A link o he published esul s will be pos ed on he PRACTICAL conso ium websi e, and he espec i e p incipal in es iga o s o each con ibu ing s udy will be p o ided he esul s o dissemina e o indi idual pa icipan s when possible. Resul s Model de elopmen O he 201 043 single nucleo ide polymo phisms included in he da ase , 2415 we e associa ed wi h inc eased isk o PCa in he end es , wi h P<10−6. The s epwise eg ession amewo k hen iden i ied 54 o hese polymo phisms ha we e inco po a ed in o he Cox p opo ional haza ds model ( able B in appendix 1). The 54 pa ame e es ima es ( o he haza d o de eloping PCa) we e combined wi h indi idual geno ype o gene a e he polygenic haza d sco e. Figu e 1 shows Kaplan-Meie and Cox eg ession es ima es o he inal model. The inal model pe o med well o p edic ion o age a onse o agg essi e PCa in he de elopmen se (z=37.5, P<10−16, haza d a io 2.3, 95% con idence in e al 2.2 o 2.4). We excluded only 43 polymo phisms (0.02%) o low call a e du ing model de elopmen and used impu a ion o missing calls o 0.4% o calls in he inal model. O he 6411 pa icipan s in he alida ion se , he median indi idual polymo phism call a e was 100%, wi h a minimum o 98%. Age (yea s) P obabili y o PCa- ee su i al Kaplan-Meie 40 45 50 55 60 65 70 75 80 85 90 95 0 0.2 0.4 0.6 0.8 1.0 PHS 0-20 h cen ile PHS 30-70 h cen ile PHS 80-97 h cen ile PHS 98-100 h cen ile Cox Fig1 | Kaplan-Meie and Cox es ima es o p os a e cance - ee su i al o pa ien s in de elopmen se by cen ile anges o polygenic haza d sco e. Cen iles a e in e e ence o dis ibu ion o sco e wi hin 11 190 con ols aged unde 70 in de elopmen se . Time o “ ailu e” is age a any diagnosis o p os a e cance . Con ols we e censo ed a age o obse a ion. Fo mal es ing o p opo ionali y is desc ibed in appendix 1 RESEARCH 4 doi: 10.1136/bmj.j5757 | BMJ 2018;360:j5757 | he bmj Risk p edic ion wi h polygenic haza d sco e In he independen alida ion se om he P o ecT s udy, a Cox p opo ional haza ds model showed ha he polygenic haza d sco e was a signi ican p edic o o age a onse o agg essi e PCa (z=11.2, P<10−16). Compa ed wi h a e age isk, he haza d a io o men wi h a high sco e (>98 h cen ile) was 2.9 (95% con idence in e al 2.4 o 3.4). The sco e was also p edic i e o any PCa (z=15.4, P<10−16; haza d a io 2.5, 2.2 o 2.8) and e y agg essi e PCa (z=6.8, P<10−11; 3.0, 2.2 o 4.0). An al e na e model used only cases o agg essi e PCa om he de elopmen se o selec polymo phisms. P edic ion o onse o agg essi e PCa was s ill signi ican (z=9.4, P<10−16, 2.6; 2.1 o 3.1) bu did no ou pe o m he o iginal model, so we used he o iginal o all subsequen analyses as planned. As he polygenic haza d sco e was p edic i e o isk o PCa, we expec ed i o modula e he posi i e p edic i e alue o PSA es ing in he alida ion se . Indeed, he posi i e p edic i e alue o PSA was lowe among pa ien s wi h a low sco e and highe among pa ien s wi h p og essi ely highe sco es ( ig 2). This pa e n held o he posi i e p edic i e alue o any PCa, as well ( ig B in appendix 1). The Lo enz cu e in igu e C in appendix 1 shows he dis ibu ion o he polygenic haza d sco e among cases o agg essi e PCa in he alida ion se . Pa ien s wi h sco es abo e he 50 h cen ile accoun ed o 76% o cases o agg essi e PCa, and he uppe i h accoun ed o 42%. Family his o y Wi h he subse o he alida ion se wi h known amily his o y s a us (1405 cases, 4298 con ols), we epea ed he Cox es wi h adjus men o amily his o y. Family his o y alone was no p edic i e o age o onse o agg essi e PCa (z=0.9, P=0.37; haza d a io 1.1, 95% con idence in e al 0.9 o 1.4), hough he e was a end owa d p edic ion o any PCa (z=2.0, P=0.05; 1.2, 1.0 o 1.3). Inclusion o amily his o y did no imp o e p edic ion o e he polygenic haza d sco e alone o agg essi e PCa (P=0.59) o any PCa (P=0.14), and he sco e emained p edic i e when adjus ed o amily his o y. discussion PCa isk p edic ion wi h polygenic haza d sco e Gene ic in o ma ion can guide he decision o whe he an indi idual pa ien needs PCa sc eening.8 The polygenic haza d sco e desc ibed he e ep esen s a pe sonalised gene ic assessmen o a man’s age ela ed isk ha could in o m bo h whe he and when o o de sc eening es s. When applied o da a om an independen clinical ial, he sco e was a highly signi ican p edic o o age a diagnosis o agg essi e PCa. Men in he op 2% o he sco e had a haza d a io o 2.9 o agg essi e PCa compa ed wi h men wi h a e age isk. As he sco e is ep esen a i e o a man’s ixed gene ic isk, i can be calcula ed once, long be o e onse o PCa, and subs an ially in o m he decision o whe he he should unde go PCa sc eening. Posi i e p edic i e alue is di ec ly dependen on p e alence, so i he polygenic haza d sco e p edic s age o onse o PCa, he posi i e p edic i e alue o PSA should a y wi h he sco e. Figu e 2 shows ha his was ue in he alida ion se . Nea ly a qua e o he posi i e PSA es esul s in men wi h a high sco e po ended a diagnosis o agg essi e PCa. The isk was much lowe o men wi h low sco es wi h a aised PSA concen a ion. The sco e is an indica o o he u ili y o PSA sc eening and could be in luen ial in he decision whe he o o de a PSA es o a gi en pa ien . These esul s also add o exis ing da a as u he e idence ha gene ic ea u es can p edic isk o PCa.6-8 11 23-25 In es iga ion in o he geno ypic ea u es desc ibed he e and elsewhe e could gi e addi ional insigh in o biological a ionales o he associa ion wi h PCa. The polygenic haza d sco e is based on haza d a ios and is he e o e an es ima e o ela i e isk. Absolu e isk can be es ima ed wi hin a gi en popula ion i he unde lying a e age haza d a e is known. This echnique would hen allow es ima ion o an indi idual PCa- ee su i al cu e o any PHS. An example o hese indi idual cu es has been published o Alzheime ’s disease.15 Compa ison wi h amily his o y Family his o y o PCa is one o he mos commonly used isk ac o s in clinic o de e mine sc eening decisions.3 Family his o y, howe e , was no p edic i e o age o onse o agg essi e PCa in he alida ion se , and i did no imp o e p edic ion o e he haza d sco e alone. This could e lec a lack o powe o de ec an associa ion o amily his o y in he ela i ely small alida ion se . Conce n o o e ea men A conce n wi h PSA sc eening is o e diagnosis and o e ea men o indolen disease. As wi h o he gene ic p edic ion ools, he polygenic haza d sco e is no speci ic o agg essi e PCa alone,16 hough he haza d PHS cen ile Posi i e p edic i e alue (PPV) 0 0.10 0.15 0.20 0.25 0.30 0.05 <20 h All >50 h >80 h >95 h Fig2 | Posi i e p edic i e alue o PSA es ing o agg essi e PCa in alida ion se . Cen iles e e o dis ibu ion o polygenic haza d sco e among young con ols in de elopmen se . 95% con idence in e als a e om andom samples o cases in alida ion se (see me hods) RESEARCH he bmj | BMJ 2018;360:j5757 | doi: 10.1136/bmj.j5757 5 a io was sligh ly highe o agg essi e PCa han o any PCa. The p oblem o o e diagnosis is compounded by he obse a ion ha many men wi h an ini ial diagnosis o low isk disease la e ecei e a diagnosis o agg essi e disease.17 19 Ac i e su eillance is one answe o o e ea men ha a oids up on ea men bu s ill allows o moni o ing o de elopmen o indica ions ha ea men is necessa y. Indeed, mos umou s e en ually equi e ea men ,17 18 and ea lie ea men p e en s de elopmen o me as a ic disease.18 Hence, a oiding sc eening al oge he in pa ien s who migh de elop PCa a a young age does ca y isk o conside able mo bidi y. The p esen esul s show ha he polygenic haza d sco e can help o a ge sc eening e o s owa d hose men a highes isk o ea ly onse PCa o agg essi e PCa equi ing ea men . As he sco e is p edic i e o agg essi e PCa in gene al, i migh also be use ul o p edic ing ou comes o men wi h a diagnosis o low isk PCa in P o ecT. The clinical da a necessa y o answe his in e es ing ques ion ha e no ye been made a ailable o he PRACTICAL conso ium, so i will ha e o be explo ed in u u e analyses. P e ious ools P e ious s udies ha e used GWAS-associa ed polymo phisms o p edic isk o PCa wi h a case- con ol design.23-25 Epidemiological da a, howe e , show ha isk o PCa is no a simple dicho omy o cases and con ols bu a he is highly dependen on inc easing age. We he e o e op ed o a su i al analysis app oach op imised o gene ic p edic ion o age o onse o PCa. The polygenic haza d sco e can hen be used in clinical decisions, when age plays a c i ical ole. I a man has a high isk o de eloping PCa a age 95, his is a di e en clinical si ua ion om a man a high isk a age 55. A compa ison o he polygenic haza d sco e wi h a adi ional polygenic isk sco e is desc ibed in appendix 2. O he PCa isk calcula o s use clinical a iables and a e mos use ul o a man who migh al eady ha e PCa.26-28 PSA concen a ion is o en included, meaning he decision o whe he o sc een has necessa ily al eady been made when he ools a e o be used. These a e less use ul o p edic ing his li e ime isk be o e he eaches an age a which he and his physician ha e o decide whe he he should ollow some p og amme o PCa sc eening. The isk s a i ica ion me ic wi h bes suppo i e e idence desc ibed in he li e a u e is an ea ly midli e PSA concen a ion measu ed a a ela i ely young age ( o example, <50). While no cu en ly ecommended in many majo clinical guidelines,3-5 ea ly midli e PSA has been shown o be p edic i e o u u e isk o PCa and le hal PCa.22 29-31 One nes ed case-con ol s udy showed ha jus he op 10% o he dis ibu ion o concen a ions o PSA in es s done in men aged unde 50 accoun ed o 40% o cases o me as a ic PCa.22 This has led o a ecommenda ion o conside PSA es ing as ea ly as age 45 in men hough by hei physician o be a high isk.32 A di ec compa ison o he polygenic haza d sco e and ea ly midli e PSA o p edic ion o age a onse o agg essi e PCa would be wo hwhile. The e migh also be an ad an age o combining he wo p edic o s. Un o una ely, ea ly midli e PSA concen a ion was no a ailable in he da ase s used in he p esen s udy so he ques ion is le o u u e wo k. Limi a ions The de elopmen se was a he e ogeneous composi e o se e al s udies o a ied design ( able A in appendix 1), which p o ides su icien powe o s udy single nucleo ide polymo phisms wi h ela i ely small e ec sizes bu also aises he conce n o unde ec ed bias in a e ospec i e analysis. The alida ion se , howe e , came om an independen la ge p ospec i e ial, and wha e e p oblems migh exis in he de elopmen se , he mos pe inen ques ion is whe he he model allows use ul p edic ions. The sco e was applied he e o PSA sc eening alone. PSA is he mos p e alen sc eening es cu en ly o PCa, bu he haza d sco e could also be expec ed o add alue o o he sc eening s a egies, by p edic ing unde lying isk o PCa o a gi en age and he e o e in luencing p e es p obabili y (and, by ex ension, posi i e p edic i e alue). This migh include PSA eloci y, PSA densi y, o some sc eening ool comple ely independen o PSA. The e idence p esen ed he e sugges s ha he polygenic haza d sco e can help a physician decide whe he o o de PSA, based on he p e es p obabili y and posi i e p edic i e alue o PSA o a gi en pa ien . Ou s udy does no , howe e , deal wi h an al e na e ques ion: how he haza d sco e migh compa e o diagnos ic ools (including isk calcula o s) ha a e pa o he clinical in es iga ions a e a aised PSA concen a ion has been ound. Adequa e da a a e no a ailable in he p esen da ase o answe his ques ion, bu i could be ied in u u e wo k as an addi ional applica ion o he polygenic haza d sco e. The age ange o he alida ion se was limi ed o 50-70; o una ely, his includes he age a which sc eening is belie ed o ha e he mos bene i .33-36 Finally, e hnici y in his model was limi ed o Eu opean ances y. Valida ion o he sco e in o he e hnic g oups—and, i necessa y, cus om models o each—is needed. We plan o in es iga e his impo an ques ion. Conclusions In conclusion, we desc ibe he e he de elopmen o a new polygenic haza d sco e o pe sonalised gene ic assessmen o indi idual age associa ed isk o PCa. This sco e has been alida ed in an independen da ase , showing accu a e p edic ion o onse o agg essi e PCa. Mo eo e , he sco e can p edic he u ili y o PSA es ing o an indi idual man. This gene ic isk model migh play a ole in guiding decisions abou whe he and when o sc een o PCa. In es iga ion in o he ela ion be ween he sco e and ea ly midli e PSA es ing is wa an ed. RESEARCH 6 doi: 10.1136/bmj.j5757 | BMJ 2018;360:j5757 | he bmj AUTHOR AFFILIATIONS 1Cen e o Mul imodal Imaging and Gene ics, Uni e si y o Cali o nia, San Diego, La Jolla, CA, USA 2Depa men o Radia ion Medicine and Applied Sciences, Uni e si y o Cali o nia, San Diego, La Jolla, CA, USA 3Depa men o Cogni i e Science, Uni e si y o Cali o nia, San Diego, La Jolla, CA, USA 4NORMENT, KG Jebsen Cen e, Oslo Uni e si y Hospi al and Uni e si y o Oslo, Oslo, No way 5MRC Bios a is ics Uni , Camb idge Biomedical Campus, Camb idge CB2 0SR, UK 6Depa men o Su ge y, Uni e si y o Cali o nia, San Diego, La Jolla, CA, USA 7Ins i u e o Cance Resea ch, London, SM2 5NG, UK 8Royal Ma sden NHS Founda ion T us , London, SW3 6JJ, UK 9Cen e o Cance Gene ic Epidemiology, Depa men o Public Heal h and P ima y Ca e, Uni e si y o Camb idge, S angeways Resea ch Labo a o y, Camb idge CB1 8RN, UK 10Depa men o Clinical Neu osciences, S oke Resea ch G oup, Uni e si y o Camb idge, R3, Box 83, Camb idge Biomedical Campus, Camb idge CB2 0QQ, UK 11Ins i u e o Popula ion Heal h, Uni e si y o Manches e , Manches e , UK 12Wa wick Medical School, Uni e si y o Wa wick, Co en y, UK 13Depa men o Medical Epidemiology and Bios a is ics, Ka olinska Ins i u e , S ockholm, Sweden 14Depa men o Molecula Medicine and Su ge y, Solna, 171 76 S ockholm, Sweden 15Depa men o U ology, Ka olinska Uni e si y Hospi al, Solna, 171 76 S ockholm, Sweden 16Depa men o Medical Biochemis y and Gene ics, Ins i u e o Biomedicine, Kiinamyllynka u 10, FI-20014 Uni e si y o Tu ku, Finland 17Tyks Mic obiology and Gene ics, Depa men o Medical Gene ics, Tu ku Uni e si y Hospi al, Tu ku, Finland 18BioMediTech, 30014 Uni e si y o Tampe e, Tampe e, Finland 19Depa men o U ology, Tampe e Uni e si y Hospi al and Medical School, Uni e si y o Tampe e, Finland 20Facul y o Heal h and Medical Sciences, Uni e si y o Copenhagen, Denma k 21Depa men o Clinical Biochemis y, He le and Gen o e Hospi al, Copenhagen Uni e si y Hospi al, He le , Denma k 22Depa men o U ology, He le and Gen o e Hospi al, Copenhagen Uni e si y Hospi al, He le , Denma k 23Copenhagen P os a e Cance Cen e, Depa men o U ology, Rigshospi ale , Copenhagen Uni e si y Hospi al, Copenhagen, Denma k 24Cance Epidemiology Uni , Nu ield Depa men o Popula ion Heal h Uni e si y o Ox o d, Ox o d OX3 7LF, UK 25Nu ield Depa men o Su gical Sciences, Facul y o Medical Science, Uni e si y o Ox o d, John Radcli e Hospi al, Ox o d, UK 26Uni e si y o Camb idge, Depa men o Oncology, Box 279, Addenb ooke’s Hospi al, Camb idge CB2 0QQ, UK 27School o Social and Communi y Medicine, Uni e si y o B is ol, B is ol BS8 2PS, UK 28Cen e o Cance Gene ic Epidemiology, Depa men o Oncology, Uni e si y o Camb idge, S angeways Resea ch Labo a o y, Camb idge, UK 29Uni e si y College London, Depa men o Applied Heal h Resea ch, London WC1E 7HB, UK 30Clinical Ge on ology Uni , Uni e si y o Camb idge, Camb idge UK 31Ins i u e o Human Gene ics, Uni e si y Hospi al o Ulm, Ulm, Ge many 32Depa men o U ology, Klinikum ech s de Isa de Technischen Uni e si ae Muenchen, Munich, Ge many 33Di ision o U ologic Su ge y, B igham and Women’s Hospi al, Dana-Fa be Cance Ins i u e, 75 F ancis S ee , Bos on, MA 02115, USA 34In e na ional He edi a y Cance Cen e, Depa men o Gene ics and Pa hology, Pome anian Medical Uni e si y, Szczecin, Poland 35Di ision o Gene ic Epidemiology, Depa men o Medicine, Uni e si y o U ah School o Medicine, Sal Lake Ci y, U ah, USA 36Geo ge E. Wahlen Depa men o Ve e ans A ai s Medical Cen e , Sal Lake Ci y, U ah, USA 37Di ision o Clinical Epidemiology and Aging Resea ch, Ge man Cance Resea ch Cen e (DKFZ), Heidelbe g, Ge many 38Di ision o P e en i e Oncology, Ge man Cance Resea ch Cen e (DKFZ) and Na ional Cen e o Tumo Diseases (NCT), Heidelbe g, Ge many 39Ge man Cance Conso ium (DKTK), Ge man Cance Resea ch Cen e (DKFZ), Heidelbe g, Ge many 40Depa men o Cance Epidemiology, Mo i Cance Cen e , 12902 Magnolia D i e, Tampa, FL 33612, USA 41O ice o he Cen e Di ec o , Mo i Cance Cen e , 12902 Magnolia D i e, Tampa, FL 33612, USA 42Depa men o U ology and Alexand o ska Uni e si y Hospi al, Medical Uni e si y, So ia, Bulga ia 43Depa men o Medical Chemis y and Biochemis y, Molecula Medicine Cen e , Medical Uni e si y, So ia, 2 Zd a e S , 1431 So ia, Bulga ia 44Aus alian P os a e Cance Resea ch Cen e-Qld, Ins i u e o Heal h and Biomedical Inno a ion and School o Biomedical Science, Queensland Uni e si y o Technology, B isbane, Aus alia 45Molecula Cance Epidemiology Labo a o y, Queensland Ins i u e o Medical Resea ch, B isbane, Aus alia 46Aus alian P os a e Cance BioResou ce, Ins i u e o Heal h and Biomedical Inno a ion and School o Biomedical Science, Queensland Uni e si y o Technology, B isbane, Aus alia 47Depa men o Gene ics, Po uguese Oncology Ins i u e, Po o, Po ugal 48Biomedical Sciences Ins i u e (ICBAS), Uni e si y o Po o, Po o, Po ugal 49Uni e si y o Su ey, Guild o d, Su ey, GU2 7XH 50Depa men o Radiology, Uni e si y o Cali o nia, San Diego, La Jolla, CA, USA 51Cen e o Cance Resea ch and Cell Biology, Queens Uni e si y Bel as , Bel as , UK 52Depa men o Neu osciences, Uni e si y o Cali o nia, San Diego, La Jolla, CA, USA *De ails o addi ional membe s om he P os a e Cance Associa ionG oup o In es iga e Cance Associa ed Al e a ions in heGenome conso ium (PRACTICAL, h p://p ac ical.ccge.medschl. cam.ac.uk/), acknowledgmen s, and unding a e p o ided in appendix2. Con ibu o s: TMS, CCF, VZ, DSK, IGM, OAA, and AMD designed he s udy. RAE, DFE, ZSKJ, AAAO, SBG, KM, HG, FW, MA, JS, CSi, TLJT, BGN, SFN, MW, RB, MAR, PI, TJK, RCT, DEN, JLD, FCH, PPh, NP, KTK, CM, WV, ML, KH, ASK, CC, DW, WK, LCA, HB, KC, KUS, JYP, TAS, CSI, RKan, VM, B, JAC, AS, APCBR, MRT, PPa, SM, HP, AM, and AK collec ed he da a. TMS, CCF, IGM, OAA, and AMD pe o med he li e a u e sea ch. TMS, CCF, YW, VZ, RKa , DSK, and AMD pe o med he da a analysis. TMS, CCF, RKa , JKP, DSK, OAA, and AMD in e p e ed he da a. TMS, RKa , JKP, DSK, OAA, and AMD c ea ed he igu es. TMS, CCF, OAA, and AMD w o e he manusc ip . All au ho s e iewed he manusc ip , added app op ia e e isions, ag eed o submission o publica ion, and app o ed he inal e sion. TMS and AMD a e gua an o s. Compe ing in e es s: All au ho s ha e comple ed he ICMJE uni o m disclosu e o m a www.icmje.o g/coi_disclosu e.pd and decla e no suppo om any o ganisa ion o he submi ed wo k excep as ollows: DSK and AMD epo a esea ch g an om he US Depa men o De ense, OAA epo s esea ch g an s om KG Jebsen S i elsen, Resea ch Council o No way, and Sou h Eas No way Heal h Au ho i y, TMS epo s hono a ia om WebMD o educa ional con en , as well as a esea ch g an om Va ian Medical Sys ems, ASK epo s ad iso y boa d membe ships o Sano i-A en is, Dend eon, and P o ound, AK epo s paid wo k o Ce a a Quan i a i e Sys ems Pha macology, DSK epo s paid wo k o Human Longe i y, OAA has a pa en applica ion (US 20150356243) pending, AMD also applied o his pa en applica ion and assigned i o UC San Diego. AMD has addi ional disclosu es ou side he p esen wo k: ounde , equi y holde , and ad iso y boa d membe o Co Techs Labs, ad iso y boa d membe o Human Longe i y, ecipien o non- inancial esea ch suppo om Gene al Elec ic Heal hca e; no inancial ela ionships wi h any companies ha migh ha e an in e es in he submi ed wo k in he RESEARCH No comme cial euse: See igh s and ep in s h p://www.bmj.com/pe missions Subsc ibe: h p://www.bmj.com/subsc ibe p e ious 3 yea s; no o he ela ionships o ac i i ies ha could appea o ha e in luenced he submi ed wo k. E hics app o al: No equi ed. Funding: This s udy was unded in pa by g an s om he US Depa men o De ense (W81XWH-13-1-0391), P os a e Cance Founda ion, he Resea ch Council o No way (223273), KG Jebsen S i elsen, and Sou h Eas No way Heal h Au ho i y. Funding o he PRACTICAL conso ium membe s udies is de ailed in appendix 2. T anspa ency: The lead au ho a i ms ha he manusc ip is an hones , accu a e, and anspa en accoun o he s udy being epo ed; ha no impo an aspec s o he s udy ha e been omi ed; and ha any disc epancies om he s udy as planned ha e been explained. Da a sha ing: No addi ional da a a ailable. This is an Open Access a icle dis ibu ed in acco dance wi h he C ea i e Commons A ibu ion Non Comme cial (CC BY-NC 4.0) license, which pe mi s o he s o dis ibu e, emix, adap , build upon his wo k non-comme cially, and license hei de i a i e wo ks on di e en e ms, p o ided he o iginal wo k is p ope ly ci ed and he use is non- comme cial. See: h p://c ea i ecommons.o g/licenses/by-nc/4.0/. 1 To eLA, B ayF, SiegelRL, Fe layJ, Lo e -Tieulen J, JemalA. Global cance s a is ics, 2012. CA Cance J Clin2015;65:87-108. doi:10.3322/caac.21262 2 Sch öde FH, HugossonJ, RoobolMJ, ERSPC In es iga o s. Sc eening and p os a e cance mo ali y: esul s o he Eu opean Randomised S udy o Sc eening o P os a e Cance (ERSPC) a 13 yea s o ollow-up. Lance 2014;384:2027-35. doi:10.1016/S0140- 6736(14)60525-0 3 Wol AMD, Wende RC, E zioniRB, Ame ican Cance Socie y P os a e Cance Ad iso y Commi ee. Ame ican Cance Socie y guideline o he ea ly de ec ion o p os a e cance : upda e 2010. CA Cance J Clin2010;60:70-98. doi:10.3322/caac.20066 4 Ho wichA, HugossonJ, de ReijkeT, WiegelT, FizaziK, Ka ajaV, Panel Membe s, Eu opean Socie y o Medical Oncology. P os a e cance : ESMO Consensus Con e ence Guidelines 2012. Ann Oncol2013;24:1141-62. doi:10.1093/annonc/mds624 5 QaseemA, Ba yMJ, Denbe gTD, OwensDK, ShekelleP, Clinical Guidelines Commi ee o he Ame ican College o Physicians. Sc eening o p os a e cance : a guidance s a emen om he Clinical Guidelines Commi ee o he Ame ican College o Physicians. Ann In e n Med2013;158:761-9. doi:10.7326/0003-4819-158-10- 201305210-00633 6 EelesRA, OlamaAAA, BenllochS, COGS–Cance Resea ch UK GWAS– ELLIPSE (pa o GAME-ON) Ini ia i e, Aus alian P os a e Cance Bio esou ce UK Gene ic P os a e Cance S udy Collabo a o s/B i ish Associa ion o U ological Su geons’ Sec ion o Oncology, UK P o ecT (P os a e es ing o cance and T ea men ) S udy Collabo a o s, PRACTICAL (P os a e Cance Associa ion G oup o In es iga e Cance - Associa ed Al e a ions in he Genome) Conso ium. Iden i ica ion o 23 new p os a e cance suscep ibili y loci using he iCOGS cus om geno yping a ay. Na Gene 2013;45:385-91, e1-2. doi:10.1038/ ng.2560 7 GohCL, Schumache FR, Eas onD. Gene ic a ian s associa ed wi hp edisposi ion o p os a e cance and po en ial clinical implica ions. J In e n Med2012;271:353-65. doi:10.1111/ j.1365-2796.2012.02511.x 8 Wi eJS. Pe sonalized p os a e cance sc eening: imp o ing PSA es s wi h genomic in o ma ion. Sci T ansl Med2010;2:62ps55. doi:10.1126/sci anslmed.3001861 9 Cha e jeeN, ShiJ, Ga cía-ClosasM. De eloping and e alua ing polygenic isk p edic ion models o s a i ied disease p e en ion. Na Re Gene 2016;17:392-406.. doi:10.1038/n g.2016.27 10 Pha oahPDP, An oniouAC, Eas onDF, Ponde BAJ. Polygenes, isk p edic ion, and a ge ed p e en ion o b eas cance . N Engl J Med2008;358:2796-803. doi:10.1056/NEJMsa0708739 11 PashayanN, Du ySW, Chowdhu yS. Polygenic suscep ibili y o p os a e and b eas cance : implica ions o pe sonalised sc eening. B J Cance 2011;104:1656-63. doi:10.1038/bjc.2011.118 12 LaneJA, Dono anJL, Da isM, P o ecT s udy g oup. Ac i e moni o ing, adical p os a ec omy, o adio he apy o localised p os a e cance : s udy design and diagnos ic and baseline esul s o he P o ecT andomised phase 3 ial. Lance Oncol2014;15:1109-18. doi:10.1016/S1470-2045(14)70361-4 13 LoebS, Bju linMA, NicholsonJ. O e diagnosis and o e ea men o p os a e cance . Eu U ol2014;65:1046-55. doi:10.1016/j. eu u o.2013.12.062 14 Mohle JL, A ms ongAJ, BahnsonRR. P os a e Cance , Ve sion 1.2016. J Na l Comp Canc Ne w2016;14:19-30. doi:10.6004/ jnccn.2016.0004 15 DesikanRS, FanCC, WangY. Pe sonalized gene ic assessmen o age associa ed Alzheime s disease isk: De elopmen and alida ion o a polygenic haza d sco e. PLoS Med2016;14:e1002289.doi. o g/10.1371/jou nal.pmed.1002258 16 EelesR, GohC, Cas oE. The gene ic epidemiology o p os a e cance and i s clinical implica ions. Na Re U ol2014;11:18-31. doi:10.1038/n u ol.2013.266 17 Fleshne NE, LuciaMS, Ege dieB. Du as e ide in localised p os a e cance managemen : he REDEEM andomised, double-blind, placebo-con olled ial. Lance 2012;379:1103-11. doi:10.1016/ S0140-6736(11)61619-X 18 HamdyFC, Dono anJL, LaneJA, P o ecT S udy G oup. 10-Yea Ou comes a e Moni o ing, Su ge y, o Radio he apy o Localized P os a e Cance . N Engl J Med2016;375:1415-24. doi:10.1056/ NEJMoa1606220 19 PashayanN, Pha oahP, NealDE. PSA-de ec ed p os a e cance and he po en ial o dedi e en ia ion--es ima ing he p opo ion capable o p og ession. In J Cance 2011;128:1462-70. doi:10.1002/ ijc.25471 20 Gas wi hJL. A Gene al De ini ion o he Lo enz Cu e. Econome ica1971;39:1037-9doi:10.2307/1909675. 21 Vicke sAJ, C oninAM, Bjö kT. P os a e speci ic an igen concen a iona age 60 and dea h o me as asis om p os a e cance : case-con ol s udy. BMJ2010;341:c4521. doi:10.1136/ bmj.c4521 22 Vicke sAJ, Ulme D, Sjobe gDD. S a egy o de ec ion o p os a e cance based on ela ion be ween p os a e speci ic an igen a age 40-55 and long e m isk o me as asis: case-con ol s udy. BMJ2013;346: 2023. doi:10.1136/bmj. 2023 23 Amin Al OlamaA, BenllochS, An oniouAC, UK Gene ic P os a eCance S udy Collabo a o s/B i ish Associa ion o U ological Su geons’ Sec ion o Oncology, UK P o ecT S udy Collabo a o s, PRACTICAL Conso ium. Risk Analysis o P os a e Cance in PRACTICAL, a Mul ina ional Conso ium, Using 25 Known P os a e Cance Suscep ibili y Loci. Cance Epidemiol Bioma ke s P e 2015;24:1121-9. doi:10.1158/1055-9965. EPI-14-0317 24 PashayanN, Du ySW, NealDE. Implica ions o polygenic isk- s a i ied sc eening o p os a e cance on o e diagnosis. Gene Med2015;17:789-95. doi:10.1038/gim.2014.192 25 Mik opoulosC, GohC, Leongamo nle D, Ko e-Ja aiZ, EelesR. T ansla ing gene ic isk ac o s o p os a e cance o he clinic: 2013 and beyond. Fu u e Oncol2014;10:1679-94. doi:10.2217/ on.14.72 26 ThompsonIM, Anke s DP, ChiC. Assessing p os a e cance isk: esul s om he P os a e Cance P e en ion T ial. J Na l Cance Ins 2006;98:529-34. doi:10.1093/jnci/djj131 27 K anseR, RoobolM, Sch öde FH. A g aphical de ice o ep esen he ou comes o a logis ic eg ession analysis. P os a e2008;68: 1674-80. doi:10.1002/p os.20840 28 Kade AK, SunJ, ReckBH. Po en ial impac o adding gene ic ma ke s o clinical pa ame e s in p edic ing p os a e biopsy ou comes in men ollowing an ini ial nega i e biopsy: indings om he REDUCE ial. Eu U ol2012;62:953-61. doi:10.1016/j. eu u o.2012.05.006 29 LiljaH, Ulme D, Bjö kT. Long- e m p edic ion o p os a e cance up o 25 yea s be o e diagnosis o p os a e cance using p os a e kallik eins measu ed a age 44 o 50 yea s. J Clin Oncol2007;25:431-6. doi:10.1200/JCO.2006.06.9351 30 LoebS, Ca e HB, Ca alonaWJ, MoulJW, Sch ode FH. Baseline p os a e-speci ic an igen es ing a a young age. Eu U ol2012;61: 1-7. doi:10.1016/j.eu u o.2011.07.067 31 P es onMA, Ba is aJL, WilsonKM. Baseline P os a e-Speci ic An igen Le els in Midli e P edic Le hal P os a e Cance . J Clin Oncol2016;34:2705-11. doi:10.1200/JCO.2016.66.7527 32 Ca ollPR, Pa sonsJK, And ioleG. NCCN Guidelines Insigh s: P os a e Cance Ea ly De ec ion, Ve sion 2.2016. J Na l Comp Canc Ne w2016;14:509-19. doi:10.6004/jnccn.2016.0060 33 SiegelRL, Mille KD, JemalA. Cance s a is ics, 2016. CA Cance J Clin2016;66:7-30. doi:10.3322/caac.21332 34 SiegelR, Wa dE, B awleyO, JemalA. Cance s a is ics, 2011: he impac o elimina ing socioeconomic and acial dispa i ies on p ema u e cance dea hs. CA Cance J Clin2011;61:212-36. doi:10.3322/caac.20121 35 JemalA, SiegelR, Wa dE. Cance s a is ics, 2006. CA Cance J Clin2006;56:106-30. doi:10.3322/canjclin.56.2.106 36 G eenleeRT, Hill-Ha monMB, Mu ayT, ThunM. Cance s a is ics, 2001. CA Cance J Clin2001;51:15-36. doi:10.3322/ canjclin.51.1.15 Appendix 1: Supplemen a y me hods, ables, and igu es Appendix 2: De ails o addi ional conso ium membe s, acknowledgemen s, and unding