A multiscale agent-based model of ductal carcinoma in situ
It is estimated that 325,000 new cases of breast cancer will be diagnosed in the United States in 2020. Of these, roughly 15% will be non-invasive in situ carcinomas, the most common of which is ductal carcinoma in situ (DCIS), a stage-zero cancer that occurs within the mammary gland ductal tree. DCIS is a cancer of the luminal epithelial cells, and is associated with subsequent invasive disease. However, the mechanistic link between DCIS and invasive disease remains elusive, and identification of DCIS that will likely transition to invasive disease remains an unmet clinical need. To investigate this problem, we have developed a multiscale, lattice-free agent based model of early-stage DCIS, consisting of both molecular and cellular scales that are explicitly linked through mathematical feedback. Model parameter values were obtained from published experimental work, and simulated growth rates were validated against DCIS growth rates reported in the literature, demonstrating that the model correctly captures bulk-scale behavior based on the underlying biological mechanisms included. We then conducted extensive sensitivity analysis by perturbing model parameters in order to test in silico the effects of molecular scale changes in key signaling pathways, including estrogen, amphiregulin, and FGF, as well as cell-scale phenotypic effects including receptor expression, changes in cell cycle length, cell density-induced quiescence, and oxygen metabolism and resultant hypoxia, necrosis, and calcification. We found that at the molecular scale, DCIS growth rates were most sensitive to estrogen signaling, while at the cell scale cell, DCIS growth was most sensitive to cell proliferation rates, oxygen metabolism rates, and cell sensitivity to hypoxia. The ability to simulate DCIS growth under the effects of molecular processes allows for studying how clinical interventions may be developed to best treat DCIS at the pre-invasive stage, suggesting model-driven strategies to reduce unnecessary invasive clinical interventions while maximizing therapeutic success.