This special focus is jointly sponsored by the Center for Discrete Mathematics and Theoretical Computer Science (DIMACS), the Biological, Mathematical, and Physical Sciences Interfaces Institute for Quantitative Biology (BioMaPS), and the Rutgers Center for Molecular Biophysics and Biophysical Chemistry (MB Center).
The practice of medicine depends on the science of prediction. Prediction depends on clinical observations or laboratory variables or factors that are linked to outcome. These factors can be anatomic, histological, and/or molecular. Found in all specialties of medicine, predictive factors take on significant clinical meaning when treatment options are available, and they become more important if treatment options are limited and not always effective.
The most common predictors in cancer medicine are the three variables of the TNM (Tumor, Lymph Node, Metastasis) that define the extent of disease anatomically . The TNM is essentially a classification of the severity of disease and is useful for guiding therapy. The TNM staging system has been universally used for more than 40 years. Many cancer patients have been classified and treated according to the TNM. The design of clinical trials, strategies for clinical management and evaluation of outcome have depended on the TNM.
However, there are many reasons to expand/revise the current TNM. Advances in molecular medicine, imaging, and therapeutics, and the role of early detection and screening are now forcing a reconsideration of the current TNM in order to accommodate additional predictive factors. TNM is a bin model . As such it can incorporate a very limited number of predictive factors. Other models need to be explored that can incorporate the additional factors to improve prediction of outcome and guide treatment. Based on the availability of large cancer patient datasets (e.g., SEER ), expanding/revising the staging system has become a major challenge. Such a dataset typically includes records from a very large number of patients with each record containing measurements made on many factors or variables (continuous or categorical). Cancer patient data can contain a high percentage of censored survival times and efficiently dealing with censored times is one of the keys [4, 5]. Integrating multiple predictive factors is complex and will require experts from many different fields, such as medicine, mathematics, statistics, computer science, and machine learning.
The purpose of the workshop will be to consider new computational models for validating predictive factors and combining them into predictive systems. The workshop will bring together investigators and practitioners to exchange research ideas and interests, as well as to discuss new directions and identify open problems in the development and application of cancer predictive systems for personalized medicine.