Hans-Jonas Meyer1, Anne Kathrin Höhn2, Katja Woidacki3, Mihailo Andric4, Maciej Powerski5, Maciej Pech5, Alexey Surov5. 1. Department of Diagnostic and Interventional Radiology, University of Leipzig, Leipzig, Germany. Electronic address: Hans-jonas.meyer@medizin.uni-leipzig.de. 2. Department of Pathology, University of Leipzig, Leipzig, Germany. 3. Section Experimental Radiology, Department of Radiology and Nuclear Medicine, Otto-von-Guericke-University of Magdeburg, Magdeburg, Germany. 4. Department of Surgery, Otto-von-Guericke-University of Magdeburg, Magdeburg, Germany. 5. Department of Radiology and Nuclear Medicine, Otto-von-Guericke-University of Magdeburg, Magdeburg, Germany.
Abstract
PURPOSE: Histogram analysis can better reflect tumor heterogeneity than conventional imaging analysis. The present study analyzed possible correlations between histogram analysis parameters derived from Intravoxel-incoherent imaging (IVIM) and histopathological features in rectal cancer (RC). METHODS: Seventeen patients with histopathologically proven rectal adenocarcinomas were retrospectively acquired. In all cases, pelvic MRI was performed. Diffusion weighted imaging was obtained using a multi-slice single-shot echo-planar imaging sequence with b values of 0, 50, 200, 500 and 1000 s/mm2. Simplified IVIM analysis was performed using the IntelliSpace portal, version 10 and the following images were generated: f (perfusion fraction) map, D (true diffusion coefficient) map, and ADC map utilizing all b-values. Histogram based analysis of signal intensities was performed for every IVIM map using an in-house matlab tool. Histopathology was investigated using Ki 67 specimens with calculation of Ki 67-index and cellularity. CD31 stained specimens were used for calculation of microvessel density (MVD). RESULTS: There were statistically significant correlations between Ki 67 index and mode derived from ADC as well as entropy from f, r=-0.50, p=.04 and r=-0.55, p=.02, respectively. MVD correlated well with parameters derived from f. CONCLUSION: IVIM histogram analysis parameters can reflect histopathology in RC. ADC and D values are associated with proliferation potential. Perfusion fraction f is associated with MVD.
PURPOSE: Histogram analysis can better reflect tumor heterogeneity than conventional imaging analysis. The present study analyzed possible correlations between histogram analysis parameters derived from Intravoxel-incoherent imaging (IVIM) and histopathological features in rectal cancer (RC). METHODS: Seventeen patients with histopathologically proven rectal adenocarcinomas were retrospectively acquired. In all cases, pelvic MRI was performed. Diffusion weighted imaging was obtained using a multi-slice single-shot echo-planar imaging sequence with b values of 0, 50, 200, 500 and 1000 s/mm2. Simplified IVIM analysis was performed using the IntelliSpace portal, version 10 and the following images were generated: f (perfusion fraction) map, D (true diffusion coefficient) map, and ADC map utilizing all b-values. Histogram based analysis of signal intensities was performed for every IVIM map using an in-house matlab tool. Histopathology was investigated using Ki 67 specimens with calculation of Ki 67-index and cellularity. CD31 stained specimens were used for calculation of microvessel density (MVD). RESULTS: There were statistically significant correlations between Ki 67 index and mode derived from ADC as well as entropy from f, r=-0.50, p=.04 and r=-0.55, p=.02, respectively. MVD correlated well with parameters derived from f. CONCLUSION: IVIM histogram analysis parameters can reflect histopathology in RC. ADC and D values are associated with proliferation potential. Perfusion fraction f is associated with MVD.