Featured Projects

Featured Projects


MicroCT in Surgical Pathology

Project with the Medical Physics and Pathology Department

Summary: Micro-CT is an emerging technology within the biomedical field and holds great promise for imaging pathology specimens for reconstruction, modeling, and analysis in 3D. Micro-CT’s capacity to create high resolution 3D models of ex vivo tissue therefore creates a unique opportunity to correlate micro-CT image data with other imaging modalities and bring additional information for diagnosis in future Pathology. The objectives of this research are 1) to evaluate micro-CT in tissue specimens in pathology, 2) to develop a scanning protocol and workflow specific to the tissue specimens

Correlating MicroCT of Suspicious Breast Lesion (BI-RADS 4 or 5) Core Biopsy Specimens with Histopathology

Project with the Radiology and Pathology Department

Summary: Micro-CT technology offers high resolution images at the microscopic level (up to a resolution of less than 10 microns), and can allow to identify different components such as fibroglandular tissue and adipose tissue in the breast specimen. In this pilot study we will explore the utility of micro-computed tomography in a clinical setting for the evaluation of breast surgical specimens along with histopathologic correlation /validation. Scanning with the micro-CT is completed within 5 minutes, making it suitable for rapid evaluation of fresh surgical specimens without a substantial delay in fixation. The use of micro-CT, could potentially enhance the definition of the tumor spatial distribution, in particular with regard to tumor multifocality. It could also guide selective sampling of the tumor areas, reducing the number of tissue sections/block and workload in the pathology laboratory. We expect this approach to improve the accuracy of breast tissue assessments e.g. lymph nodes status, especially in axillary lymph node dissection

The Roles of Micro-Computed Tomography (CT) in Breast Pathology 

Project with the Pathology Department

Summary: It is established that Breast cancer spreads by contiguous stromal invasion and also by intraductal proliferation with possible noncontiguous invasion. Intra-operative frozen section evaluation of surgical margin is advocated by some investigators to accelerate detection of margin involvement and reduce the re-excision rate and local recurrence, but sampling error or skip lesions may limit the assessment. Micro-CT enables the study of the 3D structure of the tissue and does not require any tissue sectioning or loss of sample. We evaluated micro-CT images of fresh breast tissue and lymph nodes , and compared the findings with those in micro-CT images of formalin-fixed paraffin embedded (FFPE) tissue blocks with available hematoxylin-eosin (H&E) stained sections.

Mitotic Counting and Classification Study

Project with the FDA and EEDAP Whole Slide Imaging Working Group

Summary: Supervised image analysis algorithms are only as good as the ground-truth on which they are trained. The most practical ground-truth for training an algorithm is a pathologist’s assessment on whole slide images (WSIs). Inter-observer variability may affect the reliability of the algorithm. Annotations from WSIs are subject to other limitations, such as the inability to focus on nearby planes of a section (as can be done on a microscope). In this work, we conducted a preliminary feature analysis study detecting mitoses with WSI and with a microscope. This study provides candidate mitoses for a larger study to be conducted on a 14-head microscope. Data from the larger study will be used to evaluate the performance of an automated mitosis detection algorithm. Detecting and quantifying mitoses is an important pathology task when evaluating tumors in many organs; it is also challenging and burdensome to pathologists. Because such a task is likely to be impacted by scan quality, we also investigate its suitability for evaluating image quality.

Whole Block Imaging in Prostate Cancer

Project with the Pathology Department

Summary: Traditional histologic methods are hindered by glass slide creation through tissue sectioning, staining, and additional scanning. Traditional sectioning microscopy also only provides 2D image representations of 3D structures. MicroCT opens a disruptive imaging opportunity, not only in obviating the need for sectioning and tissue staining, but in rendering 3D images potentially useful for new morphologic insights. Recently microCT has become high performance in image acquisition and volume processing. We have applied microCT in rendering 3D images of prostate whole mounted paraffin embedded slices.

Whole Block Imaging in Bone Tumor

Project with the Radiology and Pathology Departments

Summary: In Orthopedic Pathology, correlation of imaging characteristics to histopathology is important for a meaningful and accurate diagnostic interpretation. This traditional practice has given many insights on tumor characteristics such as growth pattern, relationship to adjacent normal bone, matrix production, tumor aggressiveness etc., While in Radiology practice 3D reconstruction of digital images is an easily available technique, traditional Pathology provides only a 2D image of 3D anatomical structures albeit with much higher resolution than radiological images. While 3D histological reconstruction is possible, it requires automated sectioning of 100s of slides with 3D reconstruction with appropriate software. 3D Whole block imaging (WBI) by MicroCT provides a new imaging modality to create 3D reconstruction of tissue sections with microscopic level resolution potentially up to 10x without the need for tissue sectioning.

Three-Dimensional Assessment of Spread Through Air Spaces in Lung Adenocarcinoma: Insights and Implications

Project with the Pathology and Surgery Department

Summary: Tumor spread through air space (STAS) is a newly recognized form of invasion in lung adenocarcinoma and squamous cell carcinoma and growing evidence shows it is associated with recurrence and survival. The observation that tumor STAS clusters/nests or single cells within air spaces on two-dimensional H&E slides raised the question of how these cells could survive within air spaces without a vascular supply and this has led some to speculate STAS is an artifact. Herein, we perform the high resolution-high quality 3D reconstruction and visualization of normal lung and tumor in a lung adenocarcinoma to investigate the invasive pattern of STAS.

Automated FFPE FISH Signal Scoring using Confocal Scanner’

Project with the Department of Pathology

Summary: Overall comparison and detailed comparison. FISH reports will be made based on the quantification and scoring results and reports from manual and automated analysis will be compared. The overall scoring between manual scoring and automated scoring will also be compared. The false positive and false negative results will be calculated as compared to the manual scoring (as the ground truth.) For detailed comparisons, we will compare some challenging individual nuclei judged by manual with software analysis. The over or under detection of the fluorescence signals by automation in software can be assessed.

Web Based WSI Sharing and Management

Project with the Pathology Department

Summary: Digital pathology systems help us to do telepathology, including consultation. After testing the telepathology web conference capabilities of multiple vendors, we analyzed the merits and demerits of each. Our NDP server 3.0 software (Hamamatsu photonics, Hamamatsu, Shizuoka, Japan) is a digital pathology system that makes in network consultation and conferences easy and accessible in a clinical setting.

Cytology Evaluation Studies 

Project with the Department of Pathology

Summary: Recently, our Digital Pathology System received approval from the FDA. This system enables us to make pathological diagnoses using whole slide imaging techniques. General histological slides are sectioned with 4-5um in thickness, which is enough to scan with single layer scanning. However, cytology specimens vary in thickness and size when compared to Histology. This is because cytology specimens include cell clusters comprised of more than 10um in thickness. We are seeing a variation in single layer scanning for cytology slides based on the Institution. Thus, multilayer scanning is necessary for the accurate diagnosis of cytology specimens.

If we increase the number of layers we scan, scanning time and file size will be increased. We have to find a balance between the value of having layers versus scanning time/file size for analysis and diagnosis.

The purpose of this project is to optimize the scanning protocol per type in cytology, such as smear, liquid based thin prep, cell block, etc. We will focus on the number of layers, the total distance and the step size to establish scanning parameters. We will investigate image quality, scanning duration and file size as well. Then we will validate our optimized protocol for diagnostic purpose.

Technical and Clinical Standardization for Digital Pathology

Project with the Pathology Department

Summary: Working to develop small but useful technologies in digital pathology to enhance workflows.