Automation of pattern recognition analysis of dynamic contrast-enhanced MRI data to characterize intratumoral vascular heterogeneity

Purpose: To automate dynamic contrast-enhanced MRI (DCE-MRI) data analysis by unsupervised pattern recognition (PR) to enable spatial mapping of intratumoral vascular heterogeneity. Methods: Three steps were automated. First, the arrival time of the contrast agent at the tumor was determined, including a calculation of the precontrast signal. Second, four criteria-based algorithms for the slice-specific selection of number of patterns (NP) were validated using 109 tumor slices from subcutaneous flank tumors of five different tumor models. The criteria were: half area under the curve, standard deviation thresholding, percent signal enhancement, and signal-to-noise ratio (SNR). The performance of these criteria was assessed by comparing the calculated NP with the visually determined NP. Third, spatial assignment of single patterns and/or pattern mixtures was obtained by way of constrained nonnegative matrix factorization. Results: The determination of the contrast agent arrival time at the tumor slice was successfully automated. For the determination of NP, the SNR-based approach outperformed other selection criteria by agreeing >97% with visual assessment. The spatial localization of single patterns and pattern mixtures, the latter inferring tumor vascular heterogeneity at subpixel spatial resolution, was established successfully by automated assignment from DCE-MRI signal-versus-time curves. Conclusion: The PR-based DCE-MRI analysis was successfully automated to spatially map intratumoral vascular heterogeneity