A5087300921- AI in cancer detection
- Cervical Cancer and HPV Research
- Radiomics and Machine Learning in Medical Imaging
- Endometrial and Cervical Cancer Treatments
- Multiple Sclerosis Research Studies
- Ovarian cancer diagnosis and treatment
- Digital Imaging for Blood Diseases
- Cell Image Analysis Techniques
- Privacy-Preserving Technologies in Data
- Aquaculture disease management and microbiota
- Reproductive biology and impacts on aquatic species
- Mathematical Biology Tumor Growth
- Aquaculture Nutrition and Growth
- Breast Lesions and Carcinomas
- Artificial Intelligence in Healthcare and Education
- Image Processing Techniques and Applications
- Medical Imaging and Analysis
- Colorectal and Anal Carcinomas
Roche (United States)
2021-2024
University of British Columbia
2011-2021
Occupational Cancer Research Centre
2015-2018
BC Cancer Agency
2014-2015
This paper addresses the problem of quantifying biomarkers in multi-stained tissues based on color and spatial information microscopy images tissue. A deep learning-based method that can automatically localize quantify regions expressing biomarker(s) any selected area a whole slide image is proposed. The learning network, which we refer to as Whole Image (WI)-Net, fully convolutional network whose input true RGB tissue output map showing locations each biomarker. WI-Net relies different...
The correlations between T1-hypointense lesion ('black hole') volume and clinical measures have varied widely across previous studies. degree of hypointensity in black holes is associated with the severity tissue damage, but impact on correlation disability unknown.To determine how variations intensity level used for classification can correlation, specifically Expanded Disability Status Scale (EDSS), whether using a restricted range improve correlation.A highly automated image analysis...
Cervical cancer remains a major health problem, especially in developing countries. Colposcopic examination is used to detect high-grade lesions patients with history of abnormal pap smears. New technologies are needed improve the sensitivity and specificity this technique. We propose test potential fluorescence confocal microscopy identify lesions. examined quantification ex vivo differentiate among normal cervical tissue, low-grade Intraepithelial Neoplasia (CIN), CIN. sought (1) quantify...
This paper addresses the problem of quantifying biomarkers in multi-stained tissues, based on color and spatial information. A deep learning method that can automatically localize quantify cells expressing biomarker(s) a whole slide image is proposed. The network fully convolutional (FCN) whose input true RGB tissue output map different biomarkers. FCN relies neural (CNN) classifies each cell separately according to biomarker it expresses. In this study, images immunohistochemistry (IHC)...
Despite several studies focusing on the validation of whole slide imaging (WSI) across organ systems or subspecialties, use WSI for specific primary diagnosis tasks has been underexamined.To assess pathologist performance histologic subtyping individual sections ovarian carcinomas using a light microscope and WSI.A panel 3 experienced gynecologic pathologists provided reference subtype diagnoses 212 from 109 based optical microscopy review. Two additional attending also identified presence...
We examined and established the potential of <i>ex-vivo </i>confocal fluorescence microscopy for differentiating between normal cervical tissue, low grade Cervical Intraepithelial Neoplasia (CIN1), high CIN (CIN2 CIN3). Our objectives were to i) use Quantitative Tissue Phenotype (QTP) analysis quantify nuclear cellular morphology tissue architecture in confocal microscopic images fresh biopsies ii) determine accuracy detection via microscopy. biopsy specimens colposcopically abnormal tissues...
This paper addresses the problem of classifying cells expressing different biomarkers. A deep learning based method that can automatically localize and count each biomarkers is proposed. To classify cells, a Convolutional Neural Network (CNN) was employed. Images Immunohistochemistry (IHC) stained slides contain these were digitally scanned. The images taken from digital scans IHC cervical tissues, acquired for clinical trial. More than 4,500 RGB used to train CNN. evaluate our method, first...
Background: Observer studies in pathology often utilize a limited number of representative slides per case, selected and reported nonstandardized manner. Reference diagnoses are commonly assumed to be generalizable all case. We examined these issues the context pathologist concordance for histologic subtype classification ovarian carcinomas (OCs). Materials Methods: A cohort 114 OCs consisting 72 cases with single slide (Group 1) 42 multiple (148 slides, 2–6 sections Group 2) was...
New imaging technologies are changing the field of digital pathology. This faces numerous challenges and there is a pressing need for standardization, calibration protocols, quality control quantitative assessment. We have designed new slide (Cancer Imaging Slide), specifically to measure characteristics old or systems scanners. The layout consists 138 boxes with side length 1.6 mm, containing objects known morphologic photometric characteristics. Among them, 112 contain different...
Many previous studies in multiple sclerosis (MS) have focused on the relationship between white matter lesion volume and clinical parameters, but few investigated independent contribution of spatial dispersion lesions to patient disability. In this study, we examine ability four different measures including one connectedness-based measure (compactness), region-based (ratio convex hull brain volume) two distance-based (Euclidean distance from a fixed point pair-wise Euclidean distances) act...