Dr. Jacobs is involved in breast, brain, prostate, and metastatic cancer imaging research. His research examines new analysis tools, such as, radiomics with machine and Deep Learning for magnetic resonance imaging (MRI) and PET/CT. He is particularly focused on multiparametric MRI/PET/CT assessment and image analysis of radiological imaging. He holds a number of patents related to using and manipulating radiological images.
Technology Expertise Keywords
Diffusion; ADC; computer science; cancer; PET; CT; Radiomics
Clinical Trial Keywords
PET/CT; Whole Body Imaging; Dynamic Contrast Imaging; Diffusion Imaging; Magnetic Resonance Imaging, Radiomics, Deep Learning, COVID-19, Computer Science, Medical Physics
Pre-Operative Radiation and Veliparib for Breast Cancer
Pulsed Radiofrequency vs. Steroid Injections for Occipital Neuralgia
View all on PubMed
Parekh VS, Jacobs MA, Integrated radiomic framework for breast cancer and tumor biology using advanced machine learning and multiparametric MRI. npj Breast Cancer (2017) 3:43 ; doi:10.1038/s41523-017-0045-3
Leung DG, Carrino JA, Wagner KR, Jacobs MA. Whole-body MRI evaluation of facioscapulohumeral muscular dystrophy. Muscle and Nerve. 2015; epub: Jan 16
Parekh V, Jacobs MA. Radiomics: a new application from established techniques. Expert Review of Precision Medicine and Drug Development. 2016;1(2):207-226
Parekh VS, Macura, KJ, Harvey S, Kamel I, EI-Khouli R, Bluemke DA, Jacobs MA. Multiparametric Deep Learning Tissue Signatures for a Radiological Biomarker of Breast Cancer: Preliminary Results. Medical Physics, 2020; 47(1):75-88. doi: 10.1002/mp.13849
Jacobs MA, Umbricht C, Parekh VS, EI-Khouli R, Cope L, Macura KJ, Harvey S, Wolff AC. Advanced machine learning informatics modeling and multiparametric radiomics for characterizing breast tumor characteristics with the OncotypeDX gene array. Cancers. 2020, 12(10), 2772; https://doi.org/10.3390/cancers12102772
Advanced Treatment Response Prediction Using Clinical Parameters and Advanced Unsupervised Machine Learning: The Contribution Scattergram
Patent # US Patent 10,388,017 B2 | 08/20/2019
This patent provides methods for a method for detection of different ontologies using advanced unsupervised machine learning which will be used to visualize factors not visible to the human observer such as unknown characteristics between imaging datasets and other factors to provide insights into the structure of the data. We have termed this the Contribution Scattergram, For example, using radiological images, we can determine the relationship in dimension, structure, and “distance” between each parameter using advanced graph theoretic measures. This information can be used to determine if the changes within the images have occurred and these changes can be used for treatment response in different types od diseases.
Multiparametric Non-linear dimension reduction methods for segmentation and classification of radiological images.
Patent # US Patent 9,256,966 | 02/09/2016
Visualization of anatomical structures is an important tool in medicine to differentiate normal tissue from pathological tissue using radiological imaging methods, which can generate large amounts of data for a radiologist to read. Integrating these large data sets is difficult and time-consuming. A new approach has used both supervised and unsupervised advanced machine learning techniques for visualizing and segmenting radiological data. This study presents the application of a novel machine learning method based upon Non-Linear Dimensionality Reduction (NLDR) methods using an objective (unsupervised) computer algorithm implementing modified versions of Isomap, Local Linear Embedding (LEE), and Diffusion Maps. We test the utility of these models using magnetic resonance imaging (MRI) data in patients.
Methods and systems for registration of radiological images
Patent # US Patent 9,008,462 | 04/14/2015
We developed a novel 3D registration method for application to radiological images. Typically, global motion seen in imaging is modeled by an affine transformation (rigid), while local motion and slice matching are estimated using some type of non-uniform resampling of the pixel intensities of the target image volume along a chosen reference plane. The reference plane is defined as the plane that the reference image volume was scanned, for example, axial, coronal, or sagittal. The target plane can be any of anatomical planes obliterating the need for the reference and target scan planes to be the same. Our method of registration is performed by searching for a reslicing angles and reslicing planes that maximizes similarity between the reference image volume and the resliced target image volume using novel multi-resolution wavelet transform. This wavelet transform provides estimated reslicing angles could lie between 0 and 360 degrees. By angular reslicing of the target image volume, the number of slices of the reference and target volumes will be the same and it deforms the target volume to morph it with the reference image volume. Such a reslicing-based nonrigid registration of the radiological images according to this aspect of the present invention is a useful tool to match slices location and thickness, and 3D registration of the image or object slices.
Compression device for enhancing normal/abnormal tissue contrast in MRI including devices and methods related thereto
Patent # US patent 8,380,286 | 02/19/2013
Compression device for enhancing normal/abnormal tissue contrast in MRI including devices and methods related thereto.
Patent # US patent 8,380,281 | 02/19/2013