SU-FF-I-101: Combining Magnetic-Resonance Spectroscopy and Ultrasound Spectral Parameters to Improve Tissue-Type Imaging of Prostate Cancer

Purpose: To improve prostate‐cancer imaging by making use of independent prostate tissue properties sensed by multiple imaging modalities such as the choline‐to‐citrate ratio detected by magnetic‐resonance (MR) spectroscopy in conjunction with parameters of ultrasound(US) echo‐signal spectra and clinical variables. Method and Materials:Ultrasonic(US) radio‐frequency (RF) echo‐signal data, and clinical variables including prostate‐specific antigen (PSA), were acquired during biopsy examinations. A neural‐network classifier was trained using spectral parameters computed from the RF data. Biopsy‐core histology was used as the gold standard. Classifier performance was assessed using ROC analysis. A look‐up table that returned cancer‐likelihood scores for spectral parameter and PSA combinations was used to generate tissue‐type images (TTIs). 3‐D renderings of the prostate were generated from transverse MR and US scans to determine feasibility of combining the two imaging modalities. The MR 3‐D volume was warped in 3‐D using the US volume as a reference to compensate for the distortions introduced by the larger endorectal probe used in MR imaging.Results: The ROC‐curve area for US neural‐network‐based classification was 0.84 compared to the ROC‐curve area of 0.64 for B‐mode based classification. Cancerous regions that were not visible in conventional USimages were revealed in the TTIs. Co‐registration of 3‐D prostate phantom US and MR images after warping was successful. Conclusions: TTIs based on neural‐network classification of US and clinical parameters show promise for improving the detection of prostate cancer. Warping 3‐D renderings of the prostate to compensate for the deformations introduced by different probes appears to be feasible. Consequently, we will be able to produce accurately co‐registered 3‐D data derived from US and MR to improve methods for visualizing prostate‐cancer foci. Success in combining the two imaging modalities will be valuable for depicting cancerous regions in the prostate for biopsy and treatment planning and treatment monitoring.