An update on breast cancer multigene prognostic tests-emergent clinical biomarkers

REVIEW published: 04 Sep embe 2018 doi: 10.3389/ med.2018.00248 F on ie s in Medicine | www. on ie sin.o g 1Sep embe 2018 | Volume 5 | A icle 248 Edi ed by: Venancio A ancini Al es, Uni e sidade de São Paulo, B azil Re iewed by: Da io De Biase, Uni e si à Degli S udi di Bologna, I aly Luca Quaglia a, Uni e si ä sspi al Basel, Swi ze land *Co espondence: Fe nando Schmi [email p o ec ed] Special y sec ion: This a icle was submi ed o Pa hology, a sec ion o he jou nal F on ie s in Medicine Recei ed: 26 Ap il 2018 Accep ed: 15 Augus 2018 Published: 04 Sep embe 2018 Ci a ion: Viei a AF and Schmi F (2018) An Upda e on B eas Cance Mul igene P ognos ic Tes s—Eme gen Clinical Bioma ke s. F on . Med. 5:248. doi: 10.3389/ med.2018.00248 An Upda e on B eas Cance Mul igene P ognos ic Tes s—Eme gen Clinical Bioma ke s And é Filipe Viei a1,2 and Fe nando Schmi 1,2,3* 1IPATIMUP - Epi helial In e ac ions in Cance G oup, Ins i u o de Pa ologia e Imunologia Molecula , Uni e sidade do Po o, Po o, Po ugal, 2Ins i u o de In es igação e Ino ação em Saúde, Uni e sidade do Po o, Po o, Po ugal, 3Faculdade de Medicina, Uni e sidade do Po o, Po o, Po ugal Mul igene signa u es gene a e c ucial p ognos ic in o ma ion pa icula ly use ul o cance pa ien s whe e clinical pa ame e s and adi ional immunohis ochemical ma ke s alone lead o equi ocal p ognosis. Clinicians a e now p o ided wi h molecula ools ha assis in he ou line o adju an he apies, namely helping decide on he ex ension o adju an endoc ine he apy o on supp essing adju an chemo he apy in pa ien s we e oxic e ec s a e pa icula ly dele e ious o when his ea men is undamen ally no needed. The impo ance o cance mul igene p ognos ic signa u es is well elucida ed in he guidelines o adju an sys emic he apy in ea ly-s age b eas cance and he guidelines on disease s aging ha a e p og essi ely in eg a ing gene exp ession assays as classi ica ion bioma ke s. In addi ion o he p edic i e and p ognos ic alue, some gene ic es s p o ide in insic sub yping classi ica ion. He ewi h, we compa e he molecula es s Onco ypeDX, MammaP in , P osigna, EndoP edic , B eas Cance Index, Mammos a , and IHC4 and epo he eligibili y o each one in he sui able se ing. Th ough o now, he e is no a comme cially a ailable mul igene es ha makes ecommenda ions ega ding adju an ea men o HER-2 and iple nega i e b eas cance s. Thus, hese pa ien s s ill ecei e adju an chemo he apy. Impo an ly, iple nega i e ca cinomas a e e y he e ogeneous ega ding p ognosis and new molecula signa u es ha deciphe his e y he e ogeneous subg oup o b eas cance may imp o e he clinical managemen o he disease. Keywo ds: molecula signa u es, gene ic assays, p ognos ic es s, b eas cance , bioma ke s INTRODUCTION The clinical cou se o b eas cance may be di icul o p edic as his malignancy is composed o many biological sub ypes ha in u n exhibi in a umo he e ogenei y, and pa ien s o en p esen a di e en s ages o pa hological de elopmen . In spi e o his, a limi ed numbe o p ognos ic ac o s ha e a c ucial ole oday o assess po en ial ecu ence o dea h om b eas cance . Pa ien age, umo size, como bidi y, umo g ade, and numbe o me as asized axilla y lymph nodes a e he s onges p ognos ic ac o s. One alida ed algo i hm-based model o es ima e o e all su i al (OS) and 10-yea disease- ee su i al (DFS) ha inco po a es mos o he a o emen ioned p ognos ic ac o s is Adju an ! Online (www.adju an online.com) (1,2). Clinicians a e occasionally aced by high- isk b eas cance pa ien s in an ea ly s age o disease, Viei a and Schmi Molecula Signa u es in B eas Cance wi h es ogen- ecep o (ER)-posi i e b eas cance , wi hou axilla y lymph node in ol emen , o in ol emen o up o 3 lymph nodes. The decision o adminis e adju an chemo he apy o ex ended adju an endoc ine he apy o hese pa ien s is equi ocal. Thus, bioma ke s o imp o e he clinical bene i o adju an he apies in pa ien s wi h la e ecu ence a e clinically aluable. Today, b eas cance managemen is al eady changing in ligh o he new molecula analysis ha is becoming mo e accessible in day- o-day pa hology labs. Gene ic p ognos ic es s a e bioma ke s comme cially a ailable in he o m o medical de ices/ es s. The p ognosis o b eas cance disease, he assessmen o pa ien s whe e chemo he apy will be bene icial, as well as he iden i ica ion o he molecula sub ype can be p o ided by such molecula es s. This e iew ocuses on se en majo p ognos ic signa u es o b eas cance (Onco ypeDX, Mammap in , P osigna, EndoP edic , B eas Cance Index, Mammos a and IHC4) alida ed h ough clinical ials, some o which a e al eady app o ed by FDA and ecommended by Ame ican [Na ional Comp ehensi e Cance Ne wo k (NCCN), Ame ican Socie y o Clinical Oncology (ASCO)] and Eu opean [Eu opean Socie y o Medical Oncology (ESMO)] guidelines commi ees. Fu he , we explo e he u u e in he de elopmen o no el molecula p ognos ic es s, namely in subg oups o mixed beha io b eas cance s, such as he iple-nega i e ca cinomas. BREAST CANCER MOLECULAR SUBTYPES DNA mic oa ays and Nex Gene a ion Sequencing (NGS) ep esen a g ea ad ance in molecula echnology. The cha ac e iza ion o b eas cance by DNA mic oa ay analysis has e ealed c ucial classi ica ion sys ems by gene exp ession p o ile (3). Fi e majo sub ypes o b eas ca cinomas we e iden i ied: ER-posi i e/HER2-nega i e (luminal A and luminal B sub ypes); ER-nega i e/HER2-nega i e (basal sub ype); HER2- posi i e; and ca cinomas ha ha e ea u es simila o no mal b eas issue (4–6). Di e ing elapse- ee su i al (RFS) and OS ha e been ound o hese in insic molecula sub ypes in se e al e ospec i e s udies. Fu he , o he b eas ca cinomas ha e been iden i ied as a molecula ly dis inc disease, as is he case o claudin-low cance s (7), me aplas ic (8), molecula apoc ine (9), and in asi e lobula ca cinomas (10). T iple nega i e ca cinomas, which a e ERα-nega i e, PgR-nega i e and HER2 nega i e by immunohis ochemis y, ha e been shown o be he e ogeneous ega ding esponse o ea men (11) and we e ecen ly subdi ided in o molecula sub ypes (12). Abb e ia ions: AJCC, Ame ican join commission o cance ; ASCO, Ame ican socie y o clinical oncology; DFS, disease- ee su i al; DMFS, dis an me as asis- ee su i al; ER, es ogen ecep o ; ESMO, Eu opean socie y o medical oncology; FDA, Food and d ug adminis a ion; FFPE, Fo malin- ixed pa a in embedded; HER2, Epide mal g ow h ac o ecep o 2; HR, ho mone ecep o ; NCCN, Na ional comp ehensi e cance ne wo k; NGS, Nex gene a ion sequencing; OS, o e all su i al; PCR, Pa hologic comple e esponse; PgR, p oges e one ecep o ; RFS, elapse- ee su i al; ROR, isk o ecu ence sco e; RS, ecu ence sco e; RT-PCR, e e se ansc ip ion polyme ase chain eac ion; TNM, umo , node, me as asis (s aging sys em). DEVISING MULTIGENE PROGNOSTIC SIGNATURES FROM GENE EXPRESSION ANALYSIS Big da a p o ided by DNA mic oa ay echnologies and RNA sequencing o b eas ca cinomas in combina ion wi h bioin o ma ics p o ide unpa alleled oppo uni ies o s udying b eas ca cinomas. In he pas decade, algo i hms ha e been gene a ed es ima ing he a es o cance ecu ence and/o su i al ha comp ise a educed se o genes ha cons i u es he gene signa u e. The gene ic signa u e is ob ained by compu e -based models, alida ed in clinical s udies and, in some cases, ansla ed o comme cial p ognos ic assays (13). To de elop a molecula gene signa u e, gene exp ession a ia ions a e de e mined wi hin he candida e exp ession da ase . High a iance genes o genes ha a e ound di e en ially exp essed be ween selec ed pheno ypically di e se g oups (e g., good p ognosis s. bad p ognosis) a e selec ed. Pa ien s a e g ouped in o he ca ego ies based on so ed gene exp ession p o iles, usually ansla ed in o a sco e. The Kaplan-Meie es ima o , logis ic eg ession, and Cox p opo ional haza ds model a e well-es ablished s a is ical app oaches o es he su i al unc ions ha measu e dis an ecu ence- ee su i al (DRFS), DFS o OS du ing ime in la ge-scale da a se s o gene exp ession wi h clinical da a, such as su i al o he apeu ic esponse. When assembling a p edic i e signa u e, he exp ession alues o he genes p esen in he signa u e a e weigh ed o imp o e i s p edic i e success. A ma hema ical equa ion is buil ha p edic s 5-yea o 10-yea pos diagnosis isk o ecu ence/dea h. These s a egies allow o ind a small subse o gene al e a ions ha a e mos in o ma i e o su i al p edic ion (14). In he Kaplan Meie me hod, an es ima ion o he su i al unc ion du ing ime is analyzed by log- ank s a is ics and he ac ion o pa ien s su i ing a each ime a e su ge y and/o ea men is plo ed (15,16). In logis ic eg ession, a s a is ical eg ession me hod is applied whe e he independen a iable de e mines an ou come in which he e a e only wo possible ou comes (ali e o deceased; elapse, o no elapse). I was demons a ed ha he p edic i e accu acy could be ma kedly enhanced by he applica ion o logis ic eg ession analysis, in compa ison o con en ional Kaplan Meie app oaches, which a e o en based upon incomple e o g ea ly igh -censo ed clinical da a (17,18). The Cox p opo ional haza ds model is a s a is ical eg ession model ha allows in es iga ing he e ec o se e al a iables ha may impac pa ien ou come and i assumes a cons an isk o dea h/ elapse du ing ime, known as he haza d/odds a io (19), in con as o he Kaplan Meie es ima o ha assesses he impac o a single ac o in a a ying p opo ion o deceased/ elapsed pa ien s du ing ime (16). The gene signa u e has o be alida ed in clinical assays, whe e a isk sco e can s a i y pa ien s, acco ding wi h he p obabili y o su i al gi en by he su i al unc ions. In gene al, pa ien s a e assigned o a low- isk g oup when he isk o ecu ence F on ie s in Medicine | www. on ie sin.o g 2Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance is abou 10% (10-yea su i al p obabili y a ound 90%) (13). The gene signa u e can be es ed o he in e ac ion be ween he ea men bene i (e.g., chemo he apy) and he isk sco e in Cox p opo ional haza ds and/o Kaplan-Meie models. In addi ion o he assessmen o he ea men e ec in su i al, models can be c ea ed adding he isk sco e o clinical a iables, such as umo size, age, and g ade. The pe o mance o su i al ma hema ical models can g ea ly imp o e wi h he combina ion o bo h he gene signa u e plus clinical and pa hological da a. Some molecula signa u es ha e been ansla ed o a sho lis o p o ein bioma ke s ha can be eadily es ed h ough immunohis ochemis y o immuno luo escence (20,21). This is use ul in pa hology labo a o ies whe e ou ine molecula echniques a e being slowly implemen ed. An algo i hm is calcula ed based on he su i al eg ession models o assess he coe icien s and he es ablishmen o a p ognos ic sco e/index in issues. The implemen a ion o high- h oughpu echniques like ee ci cula ing DNA p o iling and mic o-RNA analysis will pe mi he de elopmen o mul i a ia e models and open a enues o new gene exp ession signa u es o be o mula ed and implemen ed. THE CLINICAL APPLICATION OF GENETIC TESTS: PROGNOSTIC INFORMATION, THERAPY DECISION, MOLECULAR SUBTYPING, AND PATIENT STAGING Ideally, gene ic es s mus be able o accu a ely measu e he gene p o ile o in e es in di e en ce i ied labo a o ies. This equisi e is known as analy ic alidi y o he assay and i mus be main ained as p ognos ic es s become decen alized, i.e., adap ed in house o each labo a o y, p e e ably using o malin– ixed pa a in embedded (FFPE) issue samples. Gene ic es s mus also p o ide clinical alidi y, as hey should be able o clea ly s a i y a popula ion in o wo o mo e g oups o pa ien s ha ha e di e en clinical beha io ega ding pa ien ou come— usually RFS, DRFS, o OS. Finally, he clinical u ili y o gene ic es s is shown in app op ia ely designed clinical ials ha dic a e whe he using gene ic es s leads o op imized clinical decision-making wi h a con iden deg ee o e idence. Gene ic es s pa ien p ognos ic assessmen should be demons a ed in e ospec i e o p ospec i e s udies. Fu he mo e, gene ic es s should be assessed o p edic i e alue h ough he e alua ion o ea men bene i , ideally in p ospec i e s udies (13). Today, he applica ions o gene exp ession signa u es in he clinic a e di e se and hese assays ha e he p opensi y o assume a p ominen o e en c i ical signi icance in e e y pa hology labo a o y (Figu e 1). In 2017, AJCC ecognized he need o inco po a e gene exp ession p ognos ic panels in o he TNM s aging sys em (eigh h edi ion) (22). Al hough he expe panel does no endo se any pa icula assay, i is clea ha genomic assay ecu ence sco es can al e p ognosis and s age. Low isk sco es gi en by Onco ypeDX, Mammap in , Endop edic , PAM50/P osigna, o B eas Cance Index (BCI) can be used ega dless o he umo size, o downs age ho mone ecep o -posi i e, HER2 nega i e and lymph node- nega i e umo s, placing hem in o he same p ognos ic ca ego y as T1a-T1b N0 M0 ca cinomas. As o his ime, no ups aging is ecommended based on mul igene panel es ing (22). In 2016, ASCO guidelines (addi ional upda e in 2017) made se e al ecommenda ions ega ding he decision o supp essing adju an sys emic he apy o women wi h ea ly-s age in asi e luminal b eas cance (23,24). Impo an ly, his he apy has a c ucial impac in educing umo g ow h and pa ien cance mo ali y, hus clinicians should be cau ious when deciding o op ou o his he apy. In addi ion o he clinician, he pa ien is also in ol ed in heal h ca e decision-making and he la e should be in o med o he isks and bene i s o any ea men modi ica ion. S ill, o some women he oxici y associa ed wi h adju an chemo he apy may no jus i y he clinical bene i s ob ained wi h such he apy. None o he gene exp ession o p o ein assays a e ecommended by ASCO, NCCN, o ESMO ega ding decision- making on HER2-posi i e b eas cance o iple nega i e b eas cance s (23–27). Ano he po en ial applica ion o mul ipa ame e gene exp ession assays is he decision o ex end endoc ine he apy in pa ien s wi h ER/PgR-posi i e, HER2-nega i e, node-nega i e b eas cance and wi h 5 yea s o endoc ine he apy wi hou disease ecu ence. No ewo hy, ASCO does no make any ecommenda ion as o a speci ic assay o make decisions on ex ended endoc ine he apy (23). Fu he mo e, ce ain gene signa u es we e de ised o iden i y b eas cance molecula sub ypes, which may help in accessing p ognosis. PAM50/P osigna (28) and BlueP in (29) a e an example o such es s. O no e, mul igene es s and molecula sub ype classi ica ion do no in o m us abou he mu a ions and epigene ic e en s ha ha e impac in cance p og ession. Some au ho s a gue in a o o assays ha a e based in he combina ion o mu a ion p o iling wi h he gene exp ession analysis (30) because he p esence o speci ic d i e gene ic abe a ions can p edic he esponse o speci ic a ge ed he apies (30,31). Thus, one impo an app oach is o assess he comple e spec um o cance mu a ions and ind he speci ic ac ionable molecules ha a e c ucial in o de o pe o m ailo ed he apy. Today, wo companies comme cialize molecula es s ha e alua e dis inc gene mu a ional p o iles and p edic ac ionable a ge s o he clinic. These es s a e no ocused in he iden i ica ion o a simple gene o p o ein signa u e. Speci ically, Founda ion One CDx (Founda ion Medicine, Camb idge, Massachusse s, US and Roche, Basel, Swi ze land) is a FDA app o ed NGS- based in i o diagnos ic assay ha gi es an in o ma i e comp ehensi e genomic p o ile o he pa ien ’s umo , analyzing all classes o gene mu a ions known o be soma ically al e ed in solid ca cinomas and allowing he ma ching wi h a ge ed he apies. I de ec s base subs i u ions, inse ion and dele ion e en s (indels), copy numbe al e a ions and selec gene ea angemen s in 324 genes, as well as genomic pheno ypes including mic osa elli e ins abili y and umo mu a ional bu den, using DNA isola ed om FFPE umo issue specimens. Ano he F on ie s in Medicine | www. on ie sin.o g 3Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance FIGURE 1 | Clinical applica ions o mul igene/p o ein signa u es in b eas cance . Di e en mul igene/p o ein assays may ha e dis inc i e applica ions. Molecula signa u es can be used o es p ognosis, p edic ea men bene i , de e mine umo sub ype o downs age selec pa ien s. es p o iding p ecision medicine is Ca is Molecula In elligence (Ca is Li e Sciences, Phenix, A izona, US) ha uses mul iple umo p o iling echnologies o e ie e in o ma ion om pa ien ’s DNA (base subs i u ions, indels and copy numbe al e a ions), RNA (gene usions and a ian ansc ip s) and p o ein (immunohis ochemis y). Like he a o emen ioned assay, Ca is Molecula In elligence umo p o iling includes umo mu a ional bu den and mic osa elli e ins abili y es ing ia NGS (32). Up o now, he e is no molecula es ha p edic s he si e o dis an me as asis o ma ion. Fo example, luminal b eas umo s equen ly dissemina e o he bone, whe eas, he b ain is an o gan p e e en ially a ge ed by HER2 and basal-like umo s cells. A gene signa u e ha p edic s he si e o elapse could lead o close igilance o po en ially implica ed o gans. MOLECULAR SIGNATURES AND THEIR VALIDATION IN EVIDENCE-BASED CLINICAL TRIALS ASCO and NCCN make speci ic ecommenda ions based on clinical s udies and he le el o e idence hey p o ide on he use o gene ic assays o help clinicians decide on adju an he apy o women wi h ea ly s age in asi e b eas cance (Table 1). Onco ype DX Onco ype DX (Genomic Heal h, Redwood, CA) is a 21-gene signa u e ha is one o he bes - alida ed b eas cance mul igene es s. I is inco po a ed in he s aging guidelines o AJCC 8 h edi ion (22), as well as in ASCO he apy guidelines o ea ly s age b eas cance ea men (23,24), NCCN clinical p ac ice F on ie s in Medicine | www. on ie sin.o g 4Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance TABLE 1 | Clinical ials implica ed in he de elopmen o mul igene p ognos ic signa u es. Numbe o genes/p o eins Candida e pa ien s o adju an chemo he apy assessmen ( ecommended by guidelines) ASCO / NCCN ecommenda ions AJCC s aging Molecula sub yping Clinical ial Combina ion wi h clinical pa ame e s in sco e assessemen Re e ences Onco ypeDX 21 genes (16 genes +5 e e ence genes) ER/PgR+, HER2, node – ER/PgR+, HER2, node + Yes (s ong) Yes No NSABP B14 ( e opec i e) NSABP B20 ( e ospec i e) SWOG 8814 ( e ospec i e) T ansATAC ( e ospec i e) TAILORx (p ospec i e) RxPonde (p ospec i e) No (33) (34) (35) (36) (37,38) (39) MammaP in 70 genes ER/PgR+, HER2, node – ER/PgR+, HER2, node + Yes (s ong) Yes Yes (BlueP in ) TRANSBIG ( e ospec i e) RASTER (p ospec i e) MINDACT (p ospec i e) Yes (Adju an ! Online) (40) (41) (42) P osigna/PAM50 50 genes (+5 e e ence genes) ER/PgR+, HER2-, node - Yes (mode a e) Yes Yes ABCG8 ( e ospec i e) ATAC ( e ospec i e) Yes (P oli e a ion sco e, umo size) (43) (36) EndoP edic 12 genes (8 genes +3 RNA e e ence genes +1 DNA e e ence gene) ER/PgR+, HER2-, node - Yes (mode a e) Yes No GEICAM 9906 ABCSG6 ( e ospec i e) ABCSG8 ( es ospec i e) Yes ( umo size and nodal s a us (EPclin)) (44) (45) (46) B eas Cance Index 7 genes (5 genes +2 genes a io) ER/PgR+, HER2-, node - Yes (mode a e) Yes No T ansATAC ( e ospec i e) S ockholm ial ( e ospec i e) No (47) (48) Mammos a 5 p o eins – No No No NSABP B14 ( e ospec i e) B20 ( e ospec i e) No (49) IHC4 4 p o eins – No No No T ansATAC ( e ospec i e) Yes (nodal s a us, umo size, g ade, and age) (21) (50) Molecula assays can be used o de e mine he bene i s o ea ly s age b eas cance chemo he apy (ASCO and NCCN guidelines) and o op imize he s aging o pa ien s (AJCC TNM s aging 8 h edi ion guidelines). F on ie s in Medicine | www. on ie sin.o g 5Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance guidelines in oncology (26), ESMO clinical p ac ice guidelines o diagnosis, ea men , and ollow-up o p ima y b eas cance (25) and S . Gallen consensus panel guidelines (51). Onco ype DX is based on RNA isola ion om FFPE b eas cance issue ollowed by RT-PCR, p o iding a s a i ica ion o he 5-yea o 10-yea isk o dis an elapse in o isk g oups: low isk whe e he clinical bene i o chemo he apy is expec ed o be small [ ecu ence sco e (RS <18)], in e media e isk whe e i is unce ain whe he he bene icial e ec o chemo he apy ou balance he isks and complica ions media ed by i s oxic la e al e ec s (RS 18–31), and high isk whe e he e is a high p obabili y o cance o ecu ence, and he bene i s o chemo he apy a e should su pass he isks o side e ec s (RS >31). O no e, in he la es clinical ials isk sco es cu o s ha e been op imized, e lec ing a o hcoming adjus men in he assay (37). The assay is FDA clea ed and i was ini ially es ed in node nega i e pa ien s using samples om he NSABP B14 clinical ial (33). The assessmen o chemo he apy bene i was done in NSABP B20 s udy (34) and in he la ge s udies SWOG 8814 (35) and T ansATAC (36). In he e ospec i e analysis pe o med on he SWOG 8814 s udy, a andomized clinical ial in pos -menopausal, axilla y lymph node-posi i e, ER-posi i e b eas cance women, Onco ype DX deli e ed p edic i e e idence o chemo he apy bene i in amoxi en ea ed pa ien s (35). The TAILORx s udy is a p ospec i e phase III ial designed o HR-posi i e, HER-2 nega i e and node nega i e b eas cance (38,52). The RS bounda ies ha we e ini ially de e mined o Onco ype DX we e modi ied in his s udy o a oid unde ea men , wi h he lowe limi going om 18 o 11 and he uppe end was ede ined om 31 o 25. The ini ial esul s om TAILORx showed ha women wi h HR posi i e, HER2 nega i e, and node-nega i e b eas cance in he low RS g oup ha e a e y low isk o ecu ence a 5 yea s (<10%) wi h endoc ine he apy alone, and he e o e, can sa ely omi chemo he apy (37). Recen ly, Spa ano and colleagues epo ed he de ini i e esul s om TAILORx, cla i ying he e ec o chemo he apy o women conside ed o be a in e media e isk o ecu ence. Pa ien s in his g oup we e andomized o ecei e endoc ine he apy wi h o wi hou chemo he apy. The au ho s es ablished ha chemo he apy may be spa ed in all women olde han 50 wi h RS esul s o 11 o 25 and all women age 50 o younge wi h RS esul s o 11–15 (52). In e e ence o he 21-gene signa u e, ASCO guidelines epo ha “chemo he apy is indica ed in ea ly s age pa ien s ha ha e ER/PgR–posi i e, HER2-nega i e, node-nega i e b eas cance wi h a high RS and i is no indica ed in pa ien s wi h a low RS.” In pa ien s wi h an in e media e RS, he assessmen is no di ec and ecommenda ions may be de e mined by TAILORx. In hese cases, he likelihood o dis an ecu ence and bene i om chemo he apy inc eases wi h an inc ease in he RS esul . Fo ER/PgR–posi i e node-posi i e b eas cance , ASCO guidelines a e cau ious abou using Onco ype DX assay. Addi ional s udies a e equi ed o iden i y pa ien s wi h di e en ex en o axilla y nodal s a us and RS whe e chemo he apy is in ac bene icial (23). An ongoing ial (RxPONDER, ClinicalT ials.go iden i ie : NCT01272037) is ying o iden i y he cu o o he RS o which adju an chemo he apy is ad an ageous o axilla y lymph node posi i e pa ien s. Acco ding wi h NCCN guidelines, “ he 21-gene RT-PCR assay can be conside ed in pa ien s wi h 1–3 in ol ed ipsila e al axilla y lymph nodes o guide he addi ion o combina ion chemo he apy o s anda d ho mone he apy” (26). A e ospec i e analysis o p ospec i e andomized ials (NSABP B14 and B20, SWOG 8814 and T ansATAC) sugges s ha Onco ype DX as a simila p edic ion abili y in hese pa ien s as in he pa ien s lacking lymph node in ol emen (26). Bo h ecommenda ion guidelines indica e ha he 21-gene RS should no be used o guide ea men decision in HER2-posi i e b eas cance o iple-nega i e b eas cance (23,24,26). Mammap in MammaP in is a 70-gene signa u e endo sed in he s aging guidelines o AJCC 8 h edi ion (22), as well as in ASCO guidelines o ea ly s age b eas cance ea men (23,24), NCCN clinical p ac ice guidelines in oncology (26), ESMO clinical p ac ice guidelines o diagnosis, ea men , and ollow-up o p ima y b eas cance (25) and S . Gallen consensus panel guidelines (51). MammaP in is a p ognos ic es clea ed by he FDA o s a i y pa ien s wi h ER-posi i e o ER-nega i e b eas ca cinomas in o a high s. low isk o elapse (53). In he TRANSBIG conso ium s udy, he Mammap in gene sco e p o ed o be be e a s a i ying low isk s. high isk pa ien s han he clinical isk assessed wi h he Adju an ! Online ool (40). P ospec i e alida ion in node nega i e pa ien s was ob ained in he RASTER ial, whe e clinical high isk pa ien s bu wi h a low MammaP in gene ic isk wi hou chemo he apy did no nega i ely impac in DMFS (41). Recen ly, p ospec i e indica ion o he p edic i e abili y o MammaP in in ea ly-s age luminal b eas cance o adju an chemo he apy became a ailable in he MINDACT ial (le el1A e idence). The MINDACT s udy included 6,693 women wi h ea ly s age b eas cance (lymph node nega i e o 1-3 lymph node posi i e). This s udy showed ha chemo he apy could be spa ed in women who had a low genomic isk o ecu ence acco ding o MammaP in and who we e a high clinical isk o elapse de ined using Adju an ! (42). In a subse o same clinical ial, a s udy p esen ed a ESMO 2017 showed ha he 70-gene signa u e MammaP in could de ec agg essi e small umo s [ umo size <1 cm (pT1abpN0)]. The au ho s ound ha a ound 25% o small umo s we e agg essi e and pa ien s bene i ed om chemo he apy (54). The MINDACT s udy u he p o ided he pla o m o MammaP in o be included in ASCO guidelines o clinicians o di ec chemo he apy in node posi i e ea ly s age b eas cance pa ien s (23,24). Acco ding wi h he ecen ly upda ed ASCO guidelines o b eas cance ea men , i a pa ien has HR–posi i e, HER2-nega i e, bea ing a node nega i e ca cinoma, bu wi h high clinical isk, MammaP in can be used o guide he apy decisions. ASCO s a es ha “MammaP in assay may also be used in pa ien s wi h one o h ee posi i e nodes and a high clinical isk o in o m decisions on wi hholding adju an sys emic chemo he apy. Howe e , such pa ien s should be in o med ha a bene i om chemo he apy canno be excluded.” Fu he , “i a pa ien has iple nega i e b eas cance , he clinician should no use he MammaP in assay F on ie s in Medicine | www. on ie sin.o g 6Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance o guide decisions on adju an sys emic chemo he apy” (24). O no e, he MammaP in index is posi i ely associa ed wi h he likelihood o PCR, i e., high index pa ien s bene i om neoadju an chemo he apy (29). In e es ingly, MammaP in could p o ide p ognos ic alue in HER2-posi i e b eas cance . Howe e , he 10-yea dis an DFS is 84%, a alue ha is no a o able o supp ess adju an chemo he apy. Thus, he use o he MammaP in p ognos ic es o decide he adminis a ion o adju an chemo he apy in HER2-posi i e b eas cance pa ien s is no ecommended and addi ional s udies a e equi ed (55). Likewise o ASCO guidelines, he 2017 S . Gallen In e na ional B eas Cance panel (ESMO) expanded i s guidelines o ecommend he use o MammaP in o help guide chemo he apy decision-making o pa ien s wi h ea ly-s age b eas cance , wi h HR-posi i e and lymph-node posi i e b eas cance . Onco ype DX gene-exp ession es was also ecommended o guiding ea men decisions in hese pa ien s (27). Rega ding he ecen NCCN guidelines, he la es esul s om he MINDACT s udy ha e no been included. None heless, NCCN s a es ha p ognos ic mul igene assays a e o be conside ed o es ima e isk o ecu ence o dea h and bene i s o adju an chemo he apy in hese pa ien s (26). BlueP in is a molecula classi ica ion sys em based on 80 genes ha allows b eas cance sub yping classi ica ion in o low- isk luminal- ype, high- isk luminal- ype, HER-2- ype and basal- like- ype. I enables pa ien selec ion o ei he chemo he apy o endoc ine ea men (29). Also, he e is a good associa ion be ween BlueP in sub yping and chemosensi i i y PCR, wi h basal-like and HER2 sub ypes ha ing a highe PCR a es (29,56). BlueP in di e s om he PAM50 classi ie as only 9 genes a e p esen in bo h gene se s: ESR1, PGR, ERBB2, GRB7, BCL2, NAT1, FOXA1, FOXC1, MLPH bu he classi ica ion o pa ien s in o luminal, HER2, and basal subg oups by PAM50 o BlueP in is expec ed o ha e g ea simila i y, since he ag eemen wi h he o iginal in insic gene se om Pe ou and colleagues is >90%. O no e, oday he e is no s anda dized me hod o molecula sub yping o b eas cance , hence, i is unce ain which me hodology is ideal a classi ying b eas cance molecula sub ypes (29). P osigna/Pam50 P osigna/PAM50 (P osigna B eas Cance P ognos ic Gene Signa u e Assay; NanoS ing Technologies, Sea le, WA) is a 50 genes molecula signa u e ha was de eloped in p emenopausal and pos menopausal women ea ed wi hou any adju an sys emic he apy (36,43). I encompasses he NanoS ing nCoun e echnology in pa ien analysis. This es p o ides a isk o ecu ence sco e (ROR) ha akes in o accoun he PAM50 p o ile desc ibed by Pa ke e al. (28) and clinical ea u es o he pa ien , such as umo size and p oli e a ion sco e (33). ROR is s a i ied in o low (10-yea dis an ecu ence <10%), in e media e (10-yea dis an ecu ence 10–20%) and high sco es (10-yea dis an ecu ence >20%). Analogously o Mammap in /BlueP in , P osigna/PAM50 p o ides b eas cance in insic sub ype classi ica ion. ASCO guidelines indica e ha he clinician may use his signa u e “in conjunc ion wi h o he clinicopa hologic a iables o guide decisions on adju an sys emic he apy in ER/PgR-posi i e, HER2-nega i e, node-nega i e b eas cance : chemo he apy should be conside ed o pa ien s in he PAM50 high- isk g oup and i is no indica ed o pa ien s in he low- isk g oup.” Addi ional s udies a e needed o suppo ecommenda ions abou adju an chemo he apy in pa ien s wi h an in e media e P osigna/PAM50 ROR sco e. Rega ding node posi i e ER/PgR-posi i e, HER2-nega i e b eas cance , ASCO wa ns ha “mo e da a a e equi ed o de e mine whe he PAM50-ROR can be used wi h con idence in guiding he use o adju an sys emic he apy.” Fu he , no da a suppo he use o PAM50-ROR in HER2-posi i e b eas cance o in iple nega i e b eas cance (23). The ABCG8 s udy and he ATAC ials p o ided e idence o he p ognos ic use o his molecula es , wi h subse s comp ising he e ospec i e analysis o he PAM50 signa u e in endoc ine- ea ed pa ien s wi h ER-posi i e, node-nega i e disease (36,43). The P osigna/PAM50 ROR sco e added s a is ically signi ican p ognos ic in o ma ion beyond he s anda d clinical ea men sco e, which was de i ed om s anda d clinical co a ia es, including age, g ade, umo size, nodal s a us, and adju an he apy (43). In he s udy by Dowse and colleagues ha compa ed P osigna/PAM50 wi h Onco ype DX in he same FFPE samples o endoc ine- ea ed pa ien s wi h ER-posi i e, node- nega i e disease, he au ho s showed ha mo e in o ma ion was added by P osigna/PAM50 ROR han by Onco ype DX RS, i.e., mo e pa ien s we e sco ed as high isk and ewe as in e media e isk by ROR han by RS (50). Al hough he p ognos ic alue o P osigna/PAM50 has been cla i ied, he e is a lack o p ospec i e clinical s udies ha show he p edic i e alue o his signa u e. Endop edic EndoP edic (My iad Gene ics, Inc) is a wel e gene molecula signa u e. I comp ises he measu emen o he exp ession o eigh cance ela ed genes, h ee RNA e e ence genes and one DNA e e ence gene. Endop edic calcula es a isk sco e (EP, endop edic sco e), which can be used oge he wi h umo size and nodal s a us o allow he calcula ion o a comp ehensi e isk sco e (EPclin) (45). I s applica ions include guiding ea men decisions o chemo he apy as well as ex ended an i-ho monal he apy. Clinical e idence o he use o his signa u e came om he GEICAM ial ha showed ha EP is an independen p ognos ic pa ame e in node-posi i e, ER+/HER2– b eas cance pa ien s ea ed wi h adju an chemo he apy ollowed by ho mone he apy (44). EP and EPclin we e also endo sed in wo andomized phase III ials (ABCSG6 and ABCSG8) ha comp ised mo e han 1700 pos menopausal b eas cance pa ien s ea ed wi h endoc ine he apy alone, indica ing ha bo h EP and EPclin could be employed o s a i y subg oups displaying no able di e ences in 10-yea dis an ecu ence su i al in pa ien s wi h node-nega i e and node-posi i e disease (45). Al hough he s eng h o ecommenda ion is lowe han ha wi h Onco ype DX o MammaP in s udies, ASCO guidelines indica e ha EP sco e may also be employed in he decision- making p ocess ega ding he adminis a ion o adju an F on ie s in Medicine | www. on ie sin.o g 7Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance sys emic chemo he apy in pa ien s wi h ER/PgR–posi i e, HER2- nega i e, node-nega i e b eas cance . Fo node posi i e pa ien s, he le el o e idence o he p esen EndoP edic s udies is insu icien o ASCO o make a s ong ecommenda ion (23). Fu he , numbe s a e no a ailable backing he use o EP o EPclin in HER2-posi i e b eas cance o iple nega i e b eas cance . B eas Cance Index B eas Cance Index (BCI, Bio he anos ics, Inc.) combines he exp ession o 5 p oli e a ion genes known as molecula g ade index (MGI) wi h he 2-gene a io HOXB13:IL17BR (H:I) in a linea model. This sco e was de eloped in pos menopausal pa ien s wi h ER-posi i e, lymph node–nega i e b eas cance as a p edic i e es o he likelihood o bene i om ex ended adju an endoc ine he apy (57). The T ansATAC and he S ockholm ials p o ided he clinical alida ion and he indica ion o p ognos ic u ili y o his molecula signa u e (47,48,58). Re ospec i e analysis o umo samples om hese andomized ials allowed he BCI assay (H:I+MGI) o independen ly iden i y pa ien s on 5- o 10 yea endoc ine he apy wi h isk o la e-dis an ecu ence (58). Fu he , al hough s udies demons a e ha BCI has clinical use ega ding decision making abou he ex ension o adju an endoc ine he apy beyond i e yea s in pa ien s wi h ER/PgR-posi i e, HER2-nega i e, node-nega i e b eas cance (58), he applica ion o his es is no ecommended by ASCO guidelines, due o insu icien e idence. Fu he , da a a e no a ailable o suppo he use o BCI in luminal node posi i e b eas cance , in HER2-posi i e o in iple nega i e b eas cance o guide decisions on adju an sys emic he apy (23). Mammos a Mammos a es (Cla ien Diagnos ic Se ices, a GE Heal hca e company, CA) is a i e-p o ein based assay ha p o ides a sco e o low- isk, mode a e- isk, and high- isk pa ien s. Mammos a is no a ue gene ic es . Ra he , i is a i e- an ibody immunohis ochemis y p ognos ic es . This es uses he ma ke s CEACAM5, HTF9C, NDRG1, SLS7A5, and TP53 o g oup pa ien s on amoxi en he apy in o isk g oups o in o m abou p ognosis and pu a i e ea men choices, namely ega ding he likelihood o bene i o adju an sys emic he apy (20). Mammos a was de eloped o ER/PgR-posi i e, ea ly s age in asi e b eas cance pa ien s. S udies showing he suppo he use o he i e-p o ein assay in HER2-posi i e b eas cance o TN b eas cance a e lacking. The e is e idence based s udies showing ha Mammos a has p ognos ic alue in amoxi en- ea ed pa ien s, being able o ecognize hose pa ien s who ha e supe io bene i om adju an chemo he apy (20,59). S ill, he p opo ion o pa ien s who we e ecu ence ee a 10 yea s was only 85% in he low- isk subg oup (59). Thus, ASCO does no make a s ong ecommenda ion o he assessmen o adju an chemo he apy bene i in ea ly s age pa ien s wi h luminal b eas cance (node posi i e o node nega i e) (23). IHC4 IHC4 is an index de i ed om e alua ion o ER, PgR, HER2, and Ki67 by immunohis ochemis y, which a e ansla ed using an algo i hm in o a disease ecu ence isk. These ou ma ke s a e al eady b oadly used in he clinical se ing o de ine su oga e molecula sub ypes. The clinical ial T ansATAC was e ospec i ely e alua ed o he ou p o ein ma ke s (21). Since he alida ion and es ing o IHC4 is limi ed o e y ew s udies and i has no been shown o be su icien ly ep oducible, ASCO does no ecommend i s gene al clinical applica ion. IHC4 is no ecommended o be used in iple nega i e o HER2 b eas ca cinomas (21,23). COMPARISON OF MULTIGENE PROGNOSTIC TESTS Despi e p ognos ic es s ha ing di e en se s o genes/p o eins, some o hem a e sha ed be ween di e en signa u es (Table 2 and Supplemen a y Table 1). Onco ype DX sha es he highes amoun o genes/p o eins wi h o he signa u es, 9 genes being common wi h P osigna/PAM50 (BIRC5, CCNB1, MYBL2, MMP11, GRB7, ESR1, PGR, BCL, BAG1), 1 gene wi h EndoP edic (BIRC5), and 4 genes/p o eins wi h IHC4 (ESR1, PGR, HER2, Ki67). MammaP in signa u e (60) sha es one gene wi h Onco ype DX (SCUBE2), h ee genes wi h P osigna/PAM50 (KNTC2, MELK, ORC6L) and one gene wi h BCI (CENPA). P osigna/PAM50 signa u e has wo genes ha a e p esen in EndoP edic signa u e (BIRC5, UBE2C), one gene ha is p esen in BCI (RRM2), and wo genes/p o eins ha a e p esen in IHC4 (ESR1, PGR). Mammos a does no ha e any genes/p o ein in common wi h he o he signa u es. In ligh o he abo e, he numbe o sha ed genes/p o eins be ween molecula signa u es is e y small, bu hey a e able o s a i y pa ien DFS none heless. In ac , al hough he gene se s being es ed in he di e en signa u es is di e en , some biological unc ions and molecula pa hways a e sha ed. In e es ingly, p oli e a ion genes and ho mone ecep o ela ed genes a e epea edly ound in di e en b eas cance p ognos ic signa u es. In ac , i has been sugges ed ha mos o he su i al eadou ob ained om b eas cance signa u es is de i ed om a p oli e a ion pheno ype (61). O no e, despi e ha cu en gene ic es s ha e been op imized o p o ide he bes p ognos ic in o ma ion, wo k by Vene e al. sugges ed ha a signa u e made om a andom selec ion o genes om he genome has a high chance o be signi ican ly associa ed wi h ou come, mo e so when he andom signa u es a e made o mo e han 100 genes (61). The ac ha molecula p ognos ic es s a e no alike means ha wo es s pe o med on he same sample may gi e dis inc esul s. Fu he , es s ha ha e an “in e media e” isk sco e lead o an inconclusi e esul . When applying mo e han one gene ic es on he same sample, we ha e o cla i y which assay is be e o guide ea men decisions. In a s udy by Fan and colleagues he conco dance o MammaP in and Onco ype DX assays in e ms o pa ien s assigned o he same isk ca ego y was 77% o hose wi h ER-posi i e disease (62). Da a p esen ed in Miami B eas con e ence in 2014 showed ha he conco dance be ween F on ie s in Medicine | www. on ie sin.o g 8Sep embe 2018 | Volume 5 | A icle 248 Viei a and Schmi Molecula Signa u es in B eas Cance TABLE 2 | Sha ed genes/p o eins be ween molecula signa u es. Onco ype DX 21 genes Mammap in 70 genes P osigna/PAM50 50 genes Endop edic 12 genes BCI 7 genes Mammos a 5 p o eins IHC4 4 p o eins Onco ype DX SCUBE2 BIRC5, CCNB1, MYBL2, MMP11, GRB7, ESR1, PGR, BCL, BAG1 BIRC5 ESR1, PGR, HER2, Ki67 MammaP in – – KNTC2, MELK, ORC6L CENPA P osigna/PAM50 – – BIRC5, UBE2C RRM2 ESR1, PGR EndoP edic – – – – – – – BCI – – – – – – – Mammos a – – – – – – – IHC4 – – – – – – – Fo he comple e gene/p o ein lis , see Supplemen a y Table 1. MammaP in isk g oups and Onco ype DX ca ego ies was 85.7% wi hin he pa ien s classi ied as high isk by Onco ype DX and 38.1% wi hin he pa ien s classi ied as high isk by MammaP in (63). In ano he compa a i e s udy, Nunes and colleagues es ed bo h gene signa u es in 29 pa ien s: wo had high isk bo h by RS and MammaP in ; eigh had in e media e RS, wi h ou high isk by MammaP in ; 19 had a low RS, wi h eigh high isk by MammaP in . They concluded ha RS and MammaP in o e di e en p ognos ic in o ma ion (64). Possible explana ions o he obse ed a ia ion in isk s a i ica ion include di e ences in baseline cha ac e is ics o he s udy coho s, di e ences in umo biology and/o di e ences in assay echnology. In e es ingly, among pa ien s who unde go he Onco ype DX 21-gene assay, 39–67% ecei e an in e media e isk esul . Recen ly, in he PROMIS clinical ial i was shown ha 45% o in e media e isk pa ien s ha e a low isk esul wi h MammaP in and 55% had a high isk esul . The MammaP in 70-gene signa u e led o change in physicians’ ea men decisions in his popula ion wi h ea ly b eas cance (65). Dowse and colleagues compa ed Onco ype DX wi h P osigna/PAM50 and he la e assay p o ided supe io p ognos ic esul on elapse isk, showing imp o ed s a i ica ion in he in e media e and high- isk g oups o pa ien s (50). Recen p elimina y esul s we e epo ed o he p ospec i e phase III clinical ial OPTIMA, which will compa e Onco ype DX, MammaP in , P osigna, IHC4, IHC4-Aqua (Nex Cou se B eas ), and MammaType gene signa u es in he same g oup o pa ien s (66). In his s udy, di e gences we e de ec ed in pa ien s a ibu ed in o isk s a i ica ion g oups and molecula sub ypes. The low- isk g oup o Onco ype DX showed a highe numbe o pa ien s han he low/in e media e isk g oup in he P osigna, Mammap in o IHC4 assays and disco dan molecula sub yping was obse ed in 40.7% o umo s. The main OPTIMA ial ini ia ed pa ien en olmen in Janua y 2017. A deepe in es iga ion o T ansATAC s udy compa ing he pe o mance o he p ognos ic mul igene signa u es Onco ype DX, P osigna/PAM50, BCI, EPClin, IHC4, and he Clinical T ea men Sco e showed ha in pa ien s wi h node nega i e disease he P osigna/PAM50, BCI, and EPClin signa u es p o ide supe io p ognos ic accu acy, whe eas BCI and EPClin p o ided supe io p ognos ic accu acy o pa ien s wi h node posi i e disease. O no e, pa icula ly in women wi h node-posi i e disease, mul igene p ognos ic es s when combined wi h clinical ea u es signi ican ly imp o ed p ognos ic alue o dis an elapses and isk s a i ica ion. These esul s poin o he impo ance o combining clinical and pa hological in o ma ion wi h he use o genomic signa u es (13,67). A MULTIGENE SIGNATURE FOR TRIPLE NEGATIVE CARCINOMAS Al hough gene ally seen as clinically e y agg essi e, iple nega i e b eas ca cinomas ep esen a e y he e ogeneous g oup o umo s ega ding p ognosis. Recen ly, iple nega i e ca cinomas we e s a i ied in o molecula sub ypes. Lehmann and colleagues subdi ided his g oup o b eas cance s in o Basal-like (BL)-1, BL2, Mesenchymal-like (M), Mesenchymal s em-like (MSL), Luminal and ogen ecep o (LAR), and Imunnomodula o y (IM) (12,30). BL-1 and IM umo s in gene al ha e a be e p ognosis, as hey espond well o an acyclin, axanes, and cyspla in chemo he apy and immune checkpoin ac i a ion he apy, espec i ely. F om he se e al a emp s o design p ognos ic molecula signa u es o iple nega i e cance s, i s ill emains o be cla i ied i ob aining a “good” p ognosis g oup o iple nega i e ca cinomas will make hem eligible o suspending o emo ing he bene i s o chemo he apy. Ring and colleagues de eloped a new classi ica ion signa u e based on 101 genes using Lehmann’s gene exp ession da ase . Ring e al. (68) and a gene ic assay is being de eloped on his algo i hm building on p edic ing p ognosis and esponse o he apy (69). Al hough dis an ecu ence high- isk and low- isk g oups can be o mula ed based upon da a abou immune cell, in lamma ion esponse and DNA damage/ epai mechanisms, he p ac ical u ili y o his obse a ion is someway es ic ed because e en pa ien s ca ego ized as “good p ognosis” ha e abou 20% isk o dis an ecu ence in he absence o sys emic adju an he apy (70). O no e, a highe han 10% isk o ecu ence in he low isk g oup is oo high o mos pa ien s and physicians do no suppo ha F on ie s in Medicine | www. on ie sin.o g 9Sep embe 2018 | Volume 5 | A icle 248