Evidence Generation for Genomic Medicine...•Intermediate: Prescribing beta-blocker (process) and...
Transcript of Evidence Generation for Genomic Medicine...•Intermediate: Prescribing beta-blocker (process) and...
Questions:
1. WhatmethodscantheeMERGEnetworkdevelopand/oradopttoassessutility,validity,cost-effectiveness,qualityoflife,etc.ofgenetic/genomictesting?
2. HowcaneMERGEintegrateotherinformation(e.g.,familyhistory,physicaland/orpsycho-socialenvironmentalfactors,etc.)withgenetic/genomictestingresultstoimproveourunderstandingofgenomicmedicine?
EvidenceGenerationforGenomicMedicine
7
Deliverable:DevelopmentofaneMERGEseqPlatformGE
NES SITE'S TOP 6
SNVS
CONSENSU
SAC
TIONAB
LELIS
T Genes&
SNVs
Returnedbyallsites
1,551
136
19
- Fingerprint+Ancestry- HLAImpute- PGx- Automatedclassifier
pipeline(LMM)removedLB/B
- 117variantsdeemednon-actionablebyNetworkconsensus
14
- CSGHarmonizationofvariantclassification
- Consultationwithexperts
53
56
+
68
68
14109
Clinicallyactionable
Totalproposed
Clinicallyactionable
• Clinical reports are generated on the “Consensus Actionable List” and any specific genes or SNVs requested by individual sites
• To date: 14,077 samples sequenced and 3,716 reports issued
0
1
2
3
4
5
ConsensusList AdditionstoConsensusList ExclusionstoConsensusList AdditionsandExclusionstoConsensusList
KPW/UW
Harvard
Columbia
CHOP
CCHMC-adolesc.
NW
Vanderbilt
Mayo
GeisingerCCHMC-biobank
Networkactionability andsitereportingpreferences
4
SEQUENCINGandREPORTING:Timelines
BatchOneBatchTwo
GHC/UW-1
Geisinger-1
CCHMC-1
Broadsequencing Baylorsequencing
ClinicalreportingcompletedDec2018
GHC/UW-1
Geisinger-1
CCHMC-1
Harvard-1/2 Harvard-1/2
CCHMC-2 CCHMC-2
Geisinger-2
GHC/UW-2
Partnersreporting Baylorreporting
Geisinger-2
GHC/UW-2
August201
6
Octob
er2016
Decembe
r201
6
February201
7
April201
7
June
2017
August201
7
Octob
er2017
Decembe
r201
7
February201
8
April201
8
June
2018
August201
8
Octob
er2018
Decembe
r201
8
Mayo Mayo
Columbia
Vanderbilt
Northwestern
CHOP CHOP
Northwestern
Columbia
Vanderbilt
ResulttoSite
Reviewbycommittee
Contactparticipant
ReturnbyGeneticCounselor
UploadtoEMR
InformPCP
ProcessofReturnNotwositesarethesame
MOST(7/9)followthisprotocol
Minority(2/9)followthisprotocol
ResulttoSite
Reviewbycommittee
Contactparticipant
ReturnbyGCorspecialist
UploadtoEMR
InformPCP
ALL(9/9)returnthe68commonsetofgenesplusinstitutionalgenes/SNVs
2/9siteshavereturnedresults
Rephenotyping byphysicalexam/lab/testsinformPathogenicityandPenetrance:SeattleIFs(CRCprimary)
Gene Disorder Nparticipants
MYBPC3 hypertrophiccardiomyopathy 8(4LP)
HFE* hemochromatosis 7
BRCA2 breast/ovariancancer 4
SCN5A Brugada,Romano-Ward,dilatedcardiomyopathy
3(3LP)
MYH7 cardiomyopathy 2(2LP)
RYR1 malignant hyperthermia 2(2LP)
PALB2* breastcancer 2
DSC2 Arrhythmogenic rightventricularcardiomyopathy
1(1LP)
LDLR Familialhyperlipidemia 1(1LP)
BRCA1 breast/ovariancancer 1->0
MYL3 hypertrophiccardiomyopathy 1
*Not ACMG recommended
15 cardiomyopathy(10 LP) /1163
Either wrong or low penetrance
Clinically treated as P
Environmentalmeasures:eMERGEGeocodingsupplement
Factors Source Resolution National/ Local
Demographics Coordinating Center/Site EDW Patient Level National
SES Census/ACS Block Group Level NationalBuilt
EnvironmentRUCA (rural-urban-commuting-
area-codes Tract Level National
Traffic Volume Google?
Road Density ArcGIS shapefiles Block Group Level National
Food Accessibility Food Environment Atlas (USDA Economic Research Service) County Level National
Water Quality NURE-HSSR database; Enviromapper? Various
Density of Parks ArcGIS shapefiles Block Group Level National
Walkability Walk Score Professional Zip Code National
Entropy Index Census/ACS Block Group Level National
Crime Local
Hospital Utilization AHRF, HHS, HRSA County Level National
SlidecourtesyofeMERGECC
Familyhistorydata
§Very useful for stratifying analyses, identifying pathogenic variants, etc.
§Not captured well or systematically in most medical records
§Some sites may have clinical patient survey data on family history
§A standardized format would be useful
Familycascadetestingandcommunication
§ Used in pathogenicity assessment; important tool for estimating variant penetrance
§ A major driver of cost effectiveness of genetic testing is follow-up testing of relatives
§ How do we spread the word?
§ Need results early to be successful
ChallengesandOpportunitiesChallenge Opportunities
Knowvariantpathogenicityandpenetrance(evenforACMGgenes!)
• Standardizewhatisreturned(aspossible)• Rephenotypingby EHRANDnewPE• Familycascadetestingforcosegregation• Pooldataacrosssites• Reanalysisofsequenceforpathchanges• Methodstosharevariantreclassifications
Addfamilyhistorytoanalyses • StandardizetoolacrosssitesAdddemographic datatoanalyses • GeocodingCost-effectwhenfamilygetsinformation/tested
• Family communicationtools(Psycho-socialdata)• Cascadetesting• Moreefficientreturnofresults/counseling• Sharenegativereports
Datatoolateformuchfollow-up • Generate sequenceearlier:frontloadsequencingbudget,useexistingplatform(medicalexome,exome,genome)
eMERGEOUTCOMESWORKGROUP
Co-Chairs: Hakon Hakonarson(CHOP)JoshPeterson(Vanderbilt)MarcWilliams(Geisinger)
12
eMERGEOUTCOMESWORKGROUP:Charter
Missionstatement:TheOutcomesworkgroupwilldevelopcross-siteoutcomestotrackimplementationandimpactofeMERGEIIIsequencing. Theworkgroupwillfocusonansweringtheoverarchingquestionofwhetherreturned eMERGE III-generatedgenomicresultsimpacthealthcareutilizationandoutcomesofimportancetopatientsandfamilies.
• ProcessOutcomes• potentialchangesinhealthcareutilizationrelatedtoreturninggeneticinformation• Example:Colonoscopyordered
• IntermediateorSurrogateOutcomes• abiomarkerindicatingbenefitorharmismorelikely
• Example:PositiveFOBT• adherencetoarecommendation
• Example:Colonoscopyperformed• ClinicalOutcomes
• thebenefitsorharmstoapatientwhoreceivesanintervention• Example:Adenomatouspolypremoved
OutcomeTypes(examplepathogenicvariantinMLH1 associatedwithLynchsyndrome)
Chainofevidence
• Evidencethataprocessorintermediateoutcomehasadirectimpactonhealthoutcomesofinterest
• Examples:• Strong:Colonoscopy(intermediate)andcolorectalcancer;LDLc <100mg/dl(intermediate)andCAD
• Intermediate:Prescribingbeta-blocker(process)andsuddencardiacdeath(ifadherencemeasuredthisisintermediateoutcome)
• Weak:CEA125(intermediate)andovariancancer;TotalbodyMR(intermediate)andLi-Fraumeni associatedcancermortality
eMERGEOUTCOMESWORKGROUP–StandardDataCollectionForms
Challenges
• RelianceonprocessandintermediateoutcomesduetolengthofeMERGE 3
• Onetimepointforoutcomesassessment(6monthspost-RoR)• Timingofsequencingandreporting• AttributionofoutcometoRoR (relyonassertionbysite)
Opportunities-Measurehealthoutcomes• PotentialtofollowsomepatientswithRoR ineMERGE 4
• LessstraightforwardthatphenotypeandGWASeffortsacrosseMERGE 1-3
• Identifyconditionsorgenomicresultswherehealthoutcomesaremorelikelytoaccrueinafouryeartimeframe(orstrongchainofevidence)
• Pharmacogenomicsforcommondrugs• Unrecognizedgeneticdisorders(e.g.atypicalCysticFibrosis,metabolicdisorders,renaldiseaseindialysispatients)
• FamilialHypercholesterolemia
• Getsequencingresultsfastertoallowlongerfollow-up• DevelopandtestmethodstoattributeoutcomestotheReturnofResults
Challenges
• Outcomecollectionapproachessite-specific(incontrasttophenotypes)
• Manualprocessesrequiredforcascadetesting
Opportunities-ImplementationandDissemination• Studyvariationinimplementationandtheimpactonoutcomes
• R01DisseminationandImplementationLynchsyndromescreening(Rahm-GeisingerandHCSRN)
• IfcompleteineMERGE 3canusetostandardizeimplementationofRoR ineMERGE 4• Studyvariationinimplementationandtheimpactonoutcomes
• R01DisseminationandImplementationLynchsyndromescreening(Rahm-GeisingerandHCSRN)
• CollaborationwithpragmatictrialsinIGNITE2aroundcertainconditions(2approachestoevidencecollection)
• Needtousestandardoutcomemeasure• Givenpublichealthimpactofcascadetestingmakethisapointofemphasistodevelopandtestmethods
• Couldincludelegalandpolicyemphasistoinformnovelapproachestocontactingatriskrelatives
Opportunities-Economic/CostEffectiveness• Addineconomicoutcomes
• R01(Vanderbilt,UWashington,Geisinger)developingandtestingmodelstounderstandwhichoutcomesdrivecost-effectivenessandotheroutcomesofsequencing
• UsethisworktoprioritizeoutcomestocollectineMERGE 4