1	NIH Symposium: Why Do We Get It Wrong? Methods of Oncology BioMarker Validation: Trying to Get it Right! Daniel F. Hayes, M.D.
2	ASCO Tumor Marker Guidelines Panel ER, PgR Select Endocrine Therapy HER2 Select Trastuzumab CA15-3, CA27.29, CEA Monitor Selected Pts with Metastatic Disease
3	ASCO Tumor Marker Guidelines Why Are the Guidelines So Conservative? Only recommended markers for which results would change clinical decisions Evidence-based Lack of Level of Evidence I or II studies: A Tumor Marker Utility Grading Scale
4	TMUGS: Levels of Evidence Level Definition I Prospective, Marker Primary Objective, Well-powered OR Meta-analysis II Prospective, Marker Secondary Objective III Retrospective, Outcomes, Multivariate Analysis IV Retrospective, Outcomes, Univariate V Retrospective, Correlation with Other Marker, No Outcomes
5	TMUGS: Levels of Evidence Level Definition I Prospective, Marker Primary Objective, Well-powered OR Meta-analysis II Prospective, Marker Secondary Objective III Retrospective, Outcomes, Multivariate Analysis IV Retrospective, Outcomes, Univariate V Retrospective, Correlation with Other Marker, No Outcomes
6	When is a Marker Clinically Useful? It is either prognostic or predictive The magnitude of effect is sufficiently large that clinical decisions based on the data result in outcomes that are acceptable Greater chance for benefit Smaller toxicity risk The estimate of magnitude of effect is reliable Assay is reproducible Clinical trial/marker study design is appropriate Results are validated in subsequent well-designed studies
7
8	Prognostic Factors: Clinical Utility How much added benefit is required to justify treatment? Depends on toxicity of treatment Depends on perspective of patients
9
10
11	Tumor Marker Validation Examples: Breast Cancer ASSAY SETTING Multi-Gene Expression Adjuvant Rosetta/DNA Array GHI/Multi-gene RT-PCR Circulating Tumor Cells Metastatic Immunicon/Immunomagnetic Separation
12	Tumor Marker Validation Examples: Breast Cancer ASSAY SETTING Multi-Gene Expression Adjuvant Rosetta/DNA Array GHI/multi-gene RT-PCR Circulating Tumor Cells Metastatic Immunicon/Immunomagnetic Separation
13	Oncotype DX 21 Gene Recurrence Score (RS) Assay
14	NSABP: Clinical Validation Study of Oncotype DX NSABP 20: Node Neg, ER+, Tam treated patients 10 yr Distant Recurrence
15	NSABP B-14 Clinical Validation Study of Oncotype DX NSABP 14: Node Neg, ER+, Tam treated patients Tam treated patients: 10 year Distant Recurrence
16
17	Conclusions Many proposed tumor markers Most studies are LOE III or worse Cooperative Groups are now performing LOE II associated with prospective therapeutic clinical trials New Cooperative Group studies will be LOE I