Tissue Segmentation Improves Prostate Cancer Detection with Artificial Neural Networks Analysis of 1H MRSI

Lukasz Matulewicz¹, Jacobus F. Jansen², Herbert A. Vargas Alvarez², Oguz Akin², Samson Fine², Amita Shukla-Dave², James Eastham², Hedvig Hricak², Jason A. Koutcher², Kristen L. Zakian²

Artificial Neural Network (ANN) model was introduced for automatic detection of tumor voxels in the prostate from 1H-MRSI datasets with additional information about tissue segmentation. The ANNs accuracy for automatic detection of tumor voxels in the prostate MRSI datasets was demonstrated. Applying tissue segmentation from MRI as an additional input to ANN improves the accuracy of detecting tumor voxels from MRSI.