Computational Breast Imaging Group - Members

Shonket Ray, Ph.D.

Shonket Ray, Ph.D.

Postdoctoral Researcher

Contact Information:

University of Pennsylvania
School of Medicine
Department of Radiology
3600 Market Street, Suite 360
Philadelphia PA, 19104

Education & Training:

  • Postdoctoral Researcher, Present (University of Pennsylvania)
  • Ph.D. Biomedical Engineering, 2012 (University of California at Davis)
  • M.S. Biomedical Engineering, 2008 (University of California at Davis)
  • B.S. Electrical Engineering, 2003 (University of California at Davis)

Shonket Ray is a Postdoctoral Researcher in the Computational Breast Imaging Group (CBIG) at the University of Pennsylvania School of Medicine. His doctoral project was a preliminary investigation of quantitative imaging analysis using unique image sets acquired from dedicated breast x-ray cone-beam CT patient scans with clinically-diagnosed lesions. This research included designing and implementing a prototype computer-aided diagnosis system consisting of image preprocessing, lesion segmentation, structural/texture analysis of segmented object(s), and finally classification using neural networks classifiers. In addition, computer-generation of breast phantoms consisting of inserted simulated realistic lesions was performed for various observer visual acuity studies. His MS project was implementation of computer-aided segmentation of hepatic lesions from clinical CT abdominal image sets. Major tasks included lesion segmentation, performance and validation studies, and measuring observer variability using gold-standard manual outlinings of hepatic lesions by radiologists.

Research Interests:

Examining and developing novel techniques using advanced biomedical image analysis and processing. These include segmentation and classification, feature extraction measurements, and other related quantitative imaging topics using various breast imaging modalities (i.e. x-ray CT, digital mammography, breast tomosynthesis).

Current Research Topics:

  • Develop and test applications that will implement various texture feature extraction methods from breast parenchyma seen in digital breast imaging data for integration within a future proposed Breast Imaging Pipeline.
  • Examine how to combine imaging biomarkers including parenchymal texture features and breast percent density (PD), an established breast cancer risk factor, in order to develop a new comprehensive descriptor called Breast Complexity Index (BCI).
  • Through the Penn Center for Innovation in Breast Cancer Screening (PCIPS) project, determine the predictive value of BCI in terms of improving individual breast cancer risk estimation from screening and assessing the risk of false-positive or false-negative screening exams.

Professional Affiliations:

  • SPIE (The Society for Photo-optical Instrumentation Engineers)
  • AAPM (American Association of Physicists in Medicine)

Peer-Reviewed Journals:

  • Shonket Ray, Rosalie Hagge, Marijo Gillen, Miguel Cerejo, Shidrokh Shakeri, Laurel Beckett, Tamara Greasby and Ramsey Badawi "Comparison of two-dimensional and three-dimensional iterative watershed segmentation methods in hepatic tumor volumetrics," Medical Physics, 35(12):5869-81. (2008)
  • Nicolas Prionas, Karen Lindfors, Shonket Ray, Shih-Ying Huang, Laurel Beckett, Wayne Monsky, and John Boone "Contrast-enhanced Dedicated Breast CT: Initial Clinical Experience," Radiology, 256(3):714. (2010)
  • Nicolas Prionas, Shonket Ray and John Boone "Volume assessment accuracy in computed tomography: a phantom study," Journal of Applied Clinical Medical Physics, 11(2):3037. (2010)

Conference Proceeding and Abstract:

  • Shonket Ray, Nicolas Prionas, Karen Lindfors and John Boone "Analysis of breast CT lesions using computer-aided diagnosis: an application of neural networks on extracted morphologic and texture features," Proc. SPIE 8315, Medical Imaging 2012: Computer-Aided Diagnosis, 83152E (February 23, 2012)
  • Shonket Ray, Rosalie Hagge, Shidrokh Shakeri, Miguel Cerejo, Marijo Gillen and Ramsey Badawi "Accuracy of manual and semiautomatic liver tumor segmentation in contrast-enhanced and PET/CT-based noncontrast-enhanced CT", RSNA 93RD Scientific Assembly & Annual Meeting, Radiological Society of North America, Mccormick Place Convention Center, November 25 30 (2007)