Quantitative integration of imaging and non-imaging data: application to integrating multi-parametric MRI for prostate cancer diagnosis, grading and treatment evaluation

Abstract

OF THE DISSERTATION Quantitative Integration of Imaging and Non-imaging Data: Application to Integrating Multi-parametric MRI for Prostate Cancer Diagnosis, Grading and Treatment Evaluation by Pallavi Tiwari Dissertation Director: Anant Madabhushi The problem of data integration involving imaging and non-imaging modalities is largely unexplored in the biomedical field, mainly due to the challenges in quantitatively combining such heterogeneous modalities existing in different dimensions and scales. Although several methods have been proposed in the literature involving quantitative integration of multi-protocol imaging, there has been a paucity of similar biomedical tools for quantitative integration of imaging and non-imaging data. In this work, we present novel data integration schemes to overcome the aforementioned challenges limiting the integration of imaging and non-imaging modalities, and hence improve disease characterization. Our novel data integration methods are applied to integration of multi-parametric Magnetic Resonance (MR) imaging (MP-MRI)-structural MR imaging with metabolic spectroscopic information (non-imaging) for improved prostate cancer (CaP) diagnosis, grading, and treatment evaluation post-radiation therapy (RT). To this end, we have developed novel data integration schemes such as, Multimodal Wavelet Embedding Representation for data Combination (MaWERiC), and