James Madison University

Functional Mapping of Cancers

An innovative mathematical method, called functional mapping predicts where cancers will spread.  A functional map is a computer's rearrangement of human anatomy such that diseases spread to adjacent sites in the map.  Places that are likely to share disease are close together in a functional map, even though they may be physically distant. So far, significant functional maps of breast cancer, head & neck cancer, melanoma, stomach cancer, brain tumors, and tuberculosis have been made.  These diseases spread in complex patterns, not simply to anatomically adjacent sites.  Complex patterns are easily seen in a functional map.  It makes sense to think of regions that are likely to share disease as “adjacent.”  Breast cancer goes to bone, melanoma to liver, for example.  Our maps are the tumors’ “view” of our body, and succinctly depict the phenotype of disease.

Interactive visualizations are easily made from functional maps.  Visualizations of western breast cancer, eastern breast cancer, and oral cancer are available. In the breast cancer visualization, pick where a tumor might have started from the menu on the left.  Select how far you think the tumor might have spread by moving the slider. Predicted metastases are seen as changing colors.  Advance the disease and see metastases increase - more from some places than others.  These predictions are >93% accurate (based on data from > 300 Asian patients, Gray et al. 2004).  Click the "Abstract Map" tab at the top of the applet and repeat the selections of primary site and extent of disease to see how metastases are modeled as expanding ellipses in the functional map.   Patterns in a similar visualization of oral cancers are over 95% accurate and are based on over 1400 patients (Gray et al. 1993).

We have identified several meaningful subsets of patients where patterns of metastases are different.  We observe different functional maps of breast cancers in different countries and from different types of primary tumors.  In head and neck cancers, patterns of metastases in the USA and the UK are different than those in South Asia (likely due to different habits – tobacco and Betel nut - that cause the disease).  Recurrences of breast cancer will also like vary with age and with different patterns of genetic alterations.  We want to continue the search for different patterns of disease and use this information to make improved, interactive, personalized, educational visualizations.

Maps are of value at several different levels of medical expertise.   Functional mapping can find meaningful patterns in the myriad details available in electronic medical records and this process can be automated.  Functional time can be run backwards to successfully predict the locations of hidden head and neck tumors that have caused metastases.   Functional maps successfully capture patterns that specialists take years to learn.  On-line maps will help other health-care professionals such as nurses and researchers learn more about these complex diseases.  Patients also seek more and better information.  Functional maps provide an interactive and personalized way for patients to learn about their disease.

More information about this idea can be found in

  • Gray, L., Vaidya, J., Baum, M., Badwe, R., Mittra, I., Siddiqui, T., & Wiarda, D. 2004 Functional Maps of Metastases from Breast Cancers: Proof of the Principle that Multidimensional Scaling Can Summarize Disease Progression.  World J Surg. 28: 646-651.
  • Gray, L., Vaidya, J., Houghton, J., Baum, M. 2003.  Functional Maps: A novel way of visualizing patterns of recurrence from breast cancer.  Breast Cancer Research and Treatment, 82: S90. 2003.  San Antonio Breast Cancer Symposium, #372.
  • Gray, L., Saddique, T., Wiarda, D. Functional mapping of metastases.  San Antonio Breast Cancer Symposium, 2001, #416.
  • Gray L, Woolgar J, Brown J, 2000, A functional map of cervical metastases from oral squamous cell carcinoma, Acta Otolaryngologica 120:885-890.
  • Gray L, 1997, Functional mapping of glomus tumors reveals underlying embryology International Journal of Medical Informatics, 44:163-168.
  • Gray L, Klasky S, Byers, R, 1993, Visualizing complex patterns in the spread of head and neck cancers, International Journal of Supercomputer Applications 7:167-178.
  • Gray L, Robbins KT, Byers R, 1992, Multidimensional scaling of head and neck metastases.  International Journal of Bio-medical Computing 31:177-187.

Cray Research Inc. and the U.S. Government, through a NASA contract and a senior Fulbright Fellowship, have funded this research in the past. Java programming was completed by Rick Lutowski, Dr. Dorothea Wiarda and Yunsong Zhu.

 

        To run a Visualization of Breast Cancer in Western Women, click here.

       To run a Visualization of Head & Neck Cancer, click here.

 

Lincoln Gray, Ph.D.,    Professor of Communication Sciences and Disorders
College of Integrated Science and Technology
James Madison Univesity
540-568-8154 (voice)                           graylc@jmu.edu