Cancer Risk Prediction Models: A Workshop on Development, Evaluation, and Application


Overview

A great deal of research has recently been published or is currently underway to develop risk prediction models to accurately estimate the absolute risk of cancer in average-risk individuals, as well as develop models to estimate genetic susceptibility carrier status in high-risk individuals.

The Cancer Risk Prediction Workshop, held in Washington D.C. on May 20 - 21, 2004, was planned to bring together experts in the emerging field of cancer risk prediction to:

  • Identify cancer risk prediction model applications and their usefulness in planning intervention trials, estimating the population burden of disease, creating benefit/risk indices and clinical decision making processes, and designing prevention strategies.
  • Discuss the strengths and limitations of cancer risk prediction models in current use and under development.
  • Discuss methodological issues relevant to the evaluation, validation, and discrimination of cancer risk prediction models.
  • Identify ways to improve current and future cancer risk prediction models, incorporating new clinical, environmental, and genetic data.
  • Identify research needs and population data resources for future cancer risk prediction modeling and validation and disseminate this information to the scientific community

For questions regarding meeting content, please contact Dr. Andrew Freedman.

Return to Top

Agenda

View Agenda for Day 1
Thursday, May 20 Topic
8:30 a.m. Introduction and Perspectives: Cancer Control and Population Sciences
Robert Croyle
Director, Division of Cancer Control and Population Sciences, National Cancer Institute

Introduction and Perspectives: Cancer Epidemiology and Genetics
Robert Hoover
Director, Epidemiology and Biostatistics Program, Division of Cancer Epidemiology and Genetics, National Cancer Institute

8:45 a.m. Workshop Overview and Objectives
Andrew Freedman
9:00 a.m. Session I: Applications of Cancer Risk Prediction Models
Moderator: Andrew Freedman

Clinical Use of Risk Prediction Models
Laura Esserman

Estimating Population Burden of Disease
Karen Kuntz

Application of Cancer Risk Prediction Models: Intervention Trials
Joe Costantino

10:00 a.m. Session II: Cancer Risk and Susceptibility Gene Prediction Models in Use and Development
Oral Presentation followed by Poster Session

A Breast Cancer Prediction Model Incorporating Familial and Personal Risk Factors
Jack Cuzick

Poster Session
Rooms: Thomas Room and Salon C

11:00 a.m. Session III: Goals and Issues in the Development of Cancer Risk Prediction Models for Various Purposes
Moderator: Graham Colditz
Panelists: Mitchell Gail, Bernard Rosner, Beverly Rockhill, Colin Begg

Topics to be discussed:
Developing Risk Models

  • to incorporate clinical, epidemiologic, and/or biological/genetic markers (e.g., melanoma vs. Gail Model)
  • for rare vs. common cancers (e.g., ovary vs. breast)
  • using different design methodologies and data sources (e.g., cohorts vs. case-control vs. expert opinion)
  • using specialized vs. generalized populations
  • for clinical decision making
  • for individual vs. population-wide prevention strategies
12:30 p.m. Working Lunch
Lessons Learned From Cardiovascular Risk Models: Experience from the Framingham Study
Lisa Sullivan
1:15 p.m. Session IV: Risk Assessment Models for Predicting Cancer Susceptibility Genes and Cancer Risk
Moderator: Daniela Seminara

The BOADICEA model of genetic susceptibility to breast and ovarian cancer: updating, validation and predictions
Antonis Antoniou

Modifiers of Cancer Risk in BRCA1/2 Mutation Carriers: Study Design and Analysis Issues
Tim Rebbeck

Susceptibility Prediction in Familial Colon Cancer
Giovanni Parmigiani

Risk Assessment for HNPCC
Chris Amos

Genetic Susceptibility Risk Models in Clinical Decision Making
Susan Domchek

3:45 p.m. Session V: Preliminary Discussion of Key Objectives and Research Gaps
Moderator: Patricia Hartge

Identify research issues, gaps, priorities, and resources needed to advance the field of cancer risk prediction. Make specific recommendations for implementation.

Breakout Discussion Sessions:

  • Session 1: Salons A&B
    Risk prediction models for clinical decision making, intervention studies, and population-based prevention strategies (focus on breast)
    Breakout leaders: Joe Costantino, Jack Cuzick, Laura Esserman, and Victor Vogel
  • Session 2: Room 3017
    Risk prediction models for clinical decision making, intervention studies, and population-based prevention strategies (focus on lung, colorectal, melanoma, and cancers other than breast)
    Breakout leaders: Peter Bach, Graham Colditz, Ernie Hawk, and Tom Imperiale
  • Session 3: Logan Room
    Risk prediction models for genetic susceptibility
    Breakout leaders: David Euhus, Judy Garber, and Tim Rebbeck
  • Session 4: Room 3016
    Risk prediction model evaluation and validation
    Breakout leaders: Michael Kattan, Dan McGee, Martin Schumacher, and Ewout Steyerberg
5:30 p.m. Poster Session Revisited

View Agenda for Day 2
Friday, May 21 Topic
8:00 a.m. Session VI: Validation and Evaluation Methodology
Moderator: Susan Hilsenbeck

General Talk on Criteria for Model Assessment
Ruth Pfeiffer

Variability Explained: Calibration, Goodness of Fit, and Unbiased Estimation
Dan McGee

Comparing the Accuracy of Prediction Models
Michael Kattan

Assessment of Prediction Error of Risk Prediction Models
Martin Schumacher

9:45 a.m. Presentation on Current Population Resources for the Development and Validation of Cancer Risk and Susceptibility Prediction Models
Daniela Seminara
10:00 a.m. Session VII: Discussion, Summary and Future Research Directions and Wrap-Up
Moderator: Rachel Ballard-Barbash
  • Report from Breakout Session 1: Risk prediction models for clinical decision making, intervention studies, and population-based prevention strategies (focus on breast)
  • Report from Breakout Session 2: Risk prediction models for clinical decision making, intervention studies, and population-based prevention strategies (focus on lung, colorectal, melanoma, and cancers other than breast)
  • Report from Breakout Session 3: Risk prediction models for genetic susceptibility
  • Report from Breakout Session 4: Risk prediction model evaluation and validation
Noon Adjourn

Return to Top

Proceedings

Commentary summarizing the 2004 Workshop and its recommendations:

Return to Top

Planning Committee

  • Andrew Freedman, Ph.D., National Cancer Institute (Co-Chair)
  • Ruth Pfeiffer, Ph.D., National Cancer Institute (Co-Chair)
  • Rachel Ballard-Barbash, M.D., M.P.H., National Cancer Institute
  • Graham Colditz, M.D., Dr.P.H., Brigham and Women's Hospital, Harvard Medical School
  • Mitchell Gail, M.D., Ph.D., National Cancer Institute
  • Patricia Hartge, Sc.D., National Cancer Institute
  • Daniela Seminara, Ph.D., M.P.H., National Cancer Institute

Return to Top

Workshop Sponsors

  • NCI's Division of Cancer Control and Population Sciences
  • NCI's Division of Cancer Epidemiology and Genetics
  • NCI's Office of Women's Health
  • DHHS's Office of Women's Health

Return to Top