Yachna Sharma

For detailed resume click here

Education

  • Current Status: PhD candidate serving as graduate research assistant in Biomedical Informatics and Bio-imaging Laboratory (Bio-MIBLab) Georgia Institute of Technology and Emory University, Atlanta
  • Master of Science (MS) in Electrical Engineering from University of Tennessee, Knoxville (UTK), USA.
  • Bachelor of Technology (B.Tech.) in Electronics and Communications from Indira Gandhi Institute of Technology, IGIT (Formerly known as Mahila Institute of Technology, MIT), GGSIP University, Delhi (India).
  • Bachelor of Science (BS) in Botany from Hansraj College (HRC), Delhi University, India.

Research Interests and Projects

  • 3D Visualization of QD stained images (in collaboration with Center for Cancer Nanotechnology Excellence CCNE): Quantum Dot (QD) imaging allows multiplexing of different biomarker signatures in one image. 3D visualization of a series of tissue sections stained with multiple QDs can provide details on the distribution and progression of cancerous cells in the bulk of the tissue. 3D modeling can reveal details not otherwise visible in 2D images. For multiplexed QD images, we can use various layers of the model to study the distribution of different QDs. Some features such as inclined progression of cancer can only be identified by slicing the volume at an angle. Several challenges are involved in 3D reconstruction of QD stained image slices. Tissue sections undergo deformation during slicing; thus it is essential to align the individual images before volume rendering. However, there are no feature points since QD staining procedure involves slicing before staining. We resolve this issue by automatic feature point detection using DAPI (4'6-diamidino-2-phenylindole-2HCl) stained images that provide the nuclei location. With only 20 images of prostate gland sections, we perform intensity interpolation to render a smoother volume while preserving the morphological and intensity differences between adjacent slices. Our rendering approach allows interactive 3D visualization (3D rotation, translation, zoom in and zoom out), visualization of any slice along any plane from any orientation in 3D, customized color maps to locate interesting details and decomposition of the multiplexed volume into constituent QD volumes.
  • Cancer Image Processing (in collaboration with Winship Cancer Institute, Emory University): At present, tissue biopsies are analyzed and graded manually by expert pathologists and thus can be time consuming and challenging due to variations in tissue morphology, inconsistencies in preparation of tissue specimen and errors in the image acquisition process. We developed a generalized tool [1] to mark and preprocess cancerous regions in an image. Our tool is designed to automatically standardize the variations in different images due to changing illumination and experimental conditions. Segregating cancerous regions from non-cancerous areas is a mandatory step before extracting relevant information from cancer images such as the number and size of nuclei and subsequently using it for classification and quantitative analysis. We tested our tool for two completely different cancers: Head and Neck Cancer (HNC) and Renal Cell Carcinoma (RCC). The tool enables the user to successfully segment the cancerous areas for both types of cancers and our results match with the manual validation by a pathologist.
  • Computer Assisted Grading of Squamous Cell Carcinoma of Head and Neck (SCCHN) Using Folate Receptor (FR) Expression in IHC Images (in collaboration with Winship Cancer Institute, Emory University): Folate receptors are viewed as potential target for nanotherapeutic drugs in squamous cell carcinoma of head and Neck (SCCHN) However, to date; FR expression has not been correlated with head and neck cancer grade. Other CAD (Computer Assisted Diagnostics) endeavors have only classified HNC based on differentiation level (poorly differentiated, moderately differentiated and well differentiated) based on nuclei to cytoplasmic ratio. This method may not be feasible for images in which nuclei and cytoplasm are not clearly distinct. A robust grading algorithm should be able to grade the images irrespective of the discretion between tissue components, level of illumination, stain color variation and heterogeneity of tissue samples. In this work, we are developing several features that can be used to grade the IHC images of SCCHN. FR expressed as brown stain in the images, is quantified and used as one of the features to study the correlation between cancer grade and FR expression. At present 88% classification accuracy is achieved using only two features.

Publications

Manuscripts in preparation:

  1. Sharma Y, Chaudry Q, Moffitt R, Fox B, Liu J, Nie S, and Wang MD.; 3D Visualization of QD stained images,.(to be submitted for MIA)
  2. Sharma, Y.; Raza, S. H.; Kong, K. Y.; Chaudry, Q.; Muller, S.; Young, A.N.; Chen, Z.; Wang, M. D.; Feature Modeling and Analysis for Computer Assisted Diagnosis of Head and Neck Cancer(to be submitted for TMI)
  3. Stokes T.H., Sharma Y., Ahrens M., Hang S., Sanders T.; TissueWiki: A Collaborative Public Repository for Biomarkers, Immunohistochemistry and Antibody Performance Analysis. (to be submitted for Journal of Biomedical Informatics).

Current and previous students

  • Ji Hun Oh
  • Vishnu K Mishra

Experience

  • Teaching assistant and lab instructor for digital signal processing at Georgia Institute of Technology.
  • Teaching assistant for Electronic Circuits course at University of Tennessee, Knoxville.
  • Research Assistant in Oak Ridge National Laboratory (ORNL), Oak Ridge, TN and Wireless Communications Research Lab (WCRG) at UTK.
  • Internship at Council for Scientific and Industrial Research (CSIR), New Delhi, India and Global Soft Inc., India.