Cervix cancer regression measured using weekly MR imaging during fractionated radiotherapy: Radiobiologic modeling and correlation with tumor hypoxia

Abstract

5547 Background: To utilize cervix cancer volumetry, as measured with MR imaging during definitive chemoirradiation (RT-CT), to derive radiobiological parameters using a mathematical model of tumor regression, and compare them to pre-treatment measurements of tumor hypoxia. Methods: Twenty-eight patients receiving RT-CT for cervix cancer underwent weekly magnetic resonance imaging (MRI) scans. Tumor volume was assessed on each of these scans and the rate of regression plotted. A mathematical model of tumor regression was formulated to simulate the relationship between three independent radiobiological parameters, 1) surviving fraction of cells after 2 Gy, SF2, 2) the cell clearance constant Tc, and 3) the cellular proliferation constant Tp. Non-linear regression analysis was applied to fit the MR-derived tumor volumes to the mathematical model and to derive SF2 and Tc values for each patient. These were compared to pre-treatment hypoxia measurements. Results: Initial tumor volume ranged between 8 and 209 cm3. Relative reduction in volume during treatment was 0.02 to 0.79. Simulations using representative values of the independent biologic variables derived from the literature showed SF2 and Tc to strongly influence the shape of the volume response curves. Non-linear regression analysis yielded a median SF2 of 0.71 and median Tc of 10 days. Radioresistant tumors (SF2 >0.71) were significantly more hypoxic at diagnosis (p=0.02). Conclusion: Based on serial MR imaging during treatment, a marked variation in cervix tumor regression is seen from patient to patient. Through our radiobiological model, tumors can be classified as radioresistant or radiosensitive, which correlates with hypoxia [Table: see text] No significant financial relationships to disclose.