3D ovarian cancer models: Imaging and therapeutic combinations

We introduce a new platform to study treatment response in adherent micrometastatic ovarian cancer, combining an in vitro 3D model, with custom quantitative analysis routines to report growth and cytotoxic response in large sets of image data. OVCAR-5 human ovarian cancer cells were grown on a bed of Growth Factor Reduced Matrigel™ (GFR Matrigel™). Using batch analysis routines to analyze longitudinal image data we show that in vitro tumor growth leads to a reproducible log-normal size distribution with two well-defined peaks. These distinct growth modes correspond to a population with approximately constant diameter of 20μm over the time probed, while the other peak corresponds to a more rapidly assembling sub-distribution of micronodules which shifts towards larger peak center positions with mean equivalent diameters of 92μm, 120μm and 150μm at days 7, 10 and 17 following plating. At day 10, 3D and monolayer cultures were treated with a regimen of either carboplatin or photodynamic therapy. Using a quantitative fluorescence imaging approach we report dose response curves and demonstrate that 3D nodules are significantly less sensitive to treatment than the same cells grown in monolayer. 3D cultures subject to 5J/cm² PDT (250nM BPD-MA) exhibited a mean viability of 80% (95% CI = 73% to 82%) relative to no treatment control. 3D cultures subject to carboplatin treatment at 100μM concentration exhibited a mean viability of 92% (95% CI =86% to 97%). A combination treatment of 5J/cm² PDT followed by 100μM carboplatin yielded an enhanced cytotoxic effect with mean viability of 46%, 95% confidence interval (CI) = (35 % to 46%).