Abstract #3497

Mazaheri Y, Koutcher J, Hricak H, Shukla-Dave A

We introduce a novel approach to combine multiple MR parameters for detection and assessment of prostate cancer (PCa). Our technique combines MR diffusion weighted imaging (DWI) and MR spectroscopic imaging (MRSI) by partitioning tissue populations into three clusters: (2 healthy tissues, 1 tumor). The K-means algorithm and priori information of prostatic and tumor metabolic and ADC values are used to define the three classes. Voxels identified as tumor are further partitioned into high, medium, and low degrees of suspicion.