Purpose: To assess the role of image parameters derived from dynamic contrast enhanced magnetic resonance imaging (DCEMRI) in bladder cancer staging, and to investigate the potential use of such parameter images in biological image-adapted radiotherapy (RT).
Materials and methods: High-resolution volumetric interpolated breath-hold (VIBE) DCEMRI of 26 patients diagnosed with bladder cancer was performed. DCEMRI parameters derived from tumor and muscle contrast uptake curves were extracted and subjected to correlation analysis with tumor volume as well as clinical, pathological, histological and T2-weighted MR tumor stage. For parameters showing a significant correlation with tumor stage, 3D malignancy maps were generated. As an initial step towards delivery of biologically adapted intensity modulated radiotherapy (IMRT) it was hypothesized that the malignancy map could be used as a RT dose prescription map. Simulating IMRT delivery with multi-leaf collimators (MLCs), idealized dose distributions, constituted by dose cubes, were adapted to the prescription map. The size of the dose cubes were varied to mimic MLCs of varying leaf width. The difference between the adapted and prescribed dose distributions was quantified by the root mean square deviation (RMSD).
Results: No significant relationships were found between tumor volume and extracted DCEMRI parameters. The normalized area between tumor and muscle contrast uptake curves (nABC) evaluated from 0-180 seconds (nABC(180)) and 0-480 s (nABC(480)) correlated significantly with tumor stage (p=0.047 and p=0.035, respectively). Dose prescription maps for 10 patients were generated from the nABC(480). The RMSD between the prescribed and adapted dose distribution decreased with decreasing size of the dose cubes. Large interpatient variations in the RMSD and in the dependence of the RMSD on different dose cube sizes were found.
Conclusions: The nABC(180) and nABC(480) may provide added value in staging of bladder cancer. High-resolution IMRT is required for some patients for optimal adapted RT.