Development of a Population Based Model for Progression of DCIS to Invasion
A Wong, P Macklin, YL Chuang, V Cristini, ME Sanders, PA Schuyler, WD Dupont, DL Page, ME Edgerton. The University of Texas at Austin, Austin, TX; University of Texas Health Science Center, Houston, TX; Vanderbilt University School of Medicine, Nashville, TN; UT MD Anderson Cancer Center, Houston, TX
Background: Understanding the progression of ductal carcinoma in situ (DCIS) to invasive breast cancer (IBC) is critical to effective treatment and early prevention of subsequent invasion. Clinical observations report that nearly 67% of DCIS cases have already progressed to IBC at time of diagnosis. Our goal is to develop a mathematical model that predicts this observation.
Design: We developed a linear compartmental model with transition rates for normal mammary epithelium to DCIS of any grade and then from high grade (HG), intermediate grade (IG) and low grade (LG) DCIS to IBC. The model consists of coupled ordinary differential equations.
Results: We assume time interval between initiation of DCIS and diagnosis is the average time between mammograms based on current screening recommendations, 1.3 years. We fit rate constants for progression to IBC from published data (literature and SEER statistics) and apply it to an assumed distribution of DCIS by grade (40% HG, 40% IG and 20% LG). We predict that 7.1% of DCIS will have co-existing IBC within 1.3 years. We also fit a rate of progression for LG DCIS and then scaled it for IG and HG as a function of the relative proliferation index for the different grades as measured in our laboratory. This results in a prediction of 43% of cases of DCIS having co-existing IBC. If we consider DCIS to be 100% HG, then we predict 65% has invaded after 1.3 years.
Conclusions: We conclude that the progression of DCIS to IBC cannot be fully explained with a simple transition model. Our findings suggest that there are two rates for transition from DCIS to IBC, a fast one that nearly reaches 90% of its steady state value within 1 year and is independent of grades, and a rate constant that is dependent upon grade with LG and IG DCIS having the slowest rates of transition.
Wednesday, March 24, 2010 1:00 PM
Poster Session VI # 15, Wednesday Afternoon