The Experts below are selected from a list of 324 Experts worldwide ranked by ideXlab platform
Saeed Hassanpour - One of the best experts on this subject based on the ideXlab platform.
-
Pathologist level classification of histologic patterns on resected lung adenocarcinoma slides with deep neural networks
Scientific Reports, 2019Co-Authors: Laura J Tafe, Yevgeniy A Linnik, Louis J Vaickus, Naofumi Tomita, Saeed HassanpourAbstract:Classification of histologic patterns in lung adenocarcinoma is critical for determining tumor grade and treatment for patients. However, this task is often challenging due to the heterogeneous nature of lung adenocarcinoma and the subjective criteria for evaluation. In this study, we propose a deep learning model that automatically classifies the histologic patterns of lung adenocarcinoma on surgical resection slides. Our model uses a convolutional neural network to identify regions of neoplastic cells, then aggregates those classifications to infer predominant and minor histologic patterns for any given whole-slide image. We evaluated our model on an independent set of 143 whole-slide images. It achieved a kappa score of 0.525 and an agreement of 66.6% with three Pathologists for classifying the predominant patterns, slightly higher than the inter-Pathologist kappa score of 0.485 and agreement of 62.7% on this test set. All evaluation metrics for our model and the three Pathologists were within 95% confidence intervals of agreement. If confirmed in clinical practice, our model can assist Pathologists in improving classification of lung adenocarcinoma patterns by automatically pre-screening and highlighting cancerous regions prior to review. Our approach can be generalized to any whole-slide image classification task, and code is made publicly available at https://github.com/BMIRDS/deepslide .
-
Pathologist level classification of histologic patterns on resected lung adenocarcinoma slides with deep neural networks
arXiv: Computer Vision and Pattern Recognition, 2019Co-Authors: Laura J Tafe, Yevgeniy A Linnik, Louis J Vaickus, Naofumi Tomita, Saeed HassanpourAbstract:Classification of histologic patterns in lung adenocarcinoma is critical for determining tumor grade and treatment for patients. However, this task is often challenging due to the heterogeneous nature of lung adenocarcinoma and the subjective criteria for evaluation. In this study, we propose a deep learning model that automatically classifies the histologic patterns of lung adenocarcinoma on surgical resection slides. Our model uses a convolutional neural network to identify regions of neoplastic cells, then aggregates those classifications to infer predominant and minor histologic patterns for any given whole-slide image. We evaluated our model on an independent set of 143 whole-slide images. It achieved a kappa score of 0.525 and an agreement of 66.6% with three Pathologists for classifying the predominant patterns, slightly higher than the inter-Pathologist kappa score of 0.485 and agreement of 62.7% on this test set. All evaluation metrics for our model and the three Pathologists were within 95% confidence intervals of agreement. If confirmed in clinical practice, our model can assist Pathologists in improving classification of lung adenocarcinoma patterns by automatically pre-screening and highlighting cancerous regions prior to review. Our approach can be generalized to any whole-slide image classification task, and code is made publicly available at this https URL.
Laura J Tafe - One of the best experts on this subject based on the ideXlab platform.
-
Pathologist level classification of histologic patterns on resected lung adenocarcinoma slides with deep neural networks
Scientific Reports, 2019Co-Authors: Laura J Tafe, Yevgeniy A Linnik, Louis J Vaickus, Naofumi Tomita, Saeed HassanpourAbstract:Classification of histologic patterns in lung adenocarcinoma is critical for determining tumor grade and treatment for patients. However, this task is often challenging due to the heterogeneous nature of lung adenocarcinoma and the subjective criteria for evaluation. In this study, we propose a deep learning model that automatically classifies the histologic patterns of lung adenocarcinoma on surgical resection slides. Our model uses a convolutional neural network to identify regions of neoplastic cells, then aggregates those classifications to infer predominant and minor histologic patterns for any given whole-slide image. We evaluated our model on an independent set of 143 whole-slide images. It achieved a kappa score of 0.525 and an agreement of 66.6% with three Pathologists for classifying the predominant patterns, slightly higher than the inter-Pathologist kappa score of 0.485 and agreement of 62.7% on this test set. All evaluation metrics for our model and the three Pathologists were within 95% confidence intervals of agreement. If confirmed in clinical practice, our model can assist Pathologists in improving classification of lung adenocarcinoma patterns by automatically pre-screening and highlighting cancerous regions prior to review. Our approach can be generalized to any whole-slide image classification task, and code is made publicly available at https://github.com/BMIRDS/deepslide .
-
Pathologist level classification of histologic patterns on resected lung adenocarcinoma slides with deep neural networks
arXiv: Computer Vision and Pattern Recognition, 2019Co-Authors: Laura J Tafe, Yevgeniy A Linnik, Louis J Vaickus, Naofumi Tomita, Saeed HassanpourAbstract:Classification of histologic patterns in lung adenocarcinoma is critical for determining tumor grade and treatment for patients. However, this task is often challenging due to the heterogeneous nature of lung adenocarcinoma and the subjective criteria for evaluation. In this study, we propose a deep learning model that automatically classifies the histologic patterns of lung adenocarcinoma on surgical resection slides. Our model uses a convolutional neural network to identify regions of neoplastic cells, then aggregates those classifications to infer predominant and minor histologic patterns for any given whole-slide image. We evaluated our model on an independent set of 143 whole-slide images. It achieved a kappa score of 0.525 and an agreement of 66.6% with three Pathologists for classifying the predominant patterns, slightly higher than the inter-Pathologist kappa score of 0.485 and agreement of 62.7% on this test set. All evaluation metrics for our model and the three Pathologists were within 95% confidence intervals of agreement. If confirmed in clinical practice, our model can assist Pathologists in improving classification of lung adenocarcinoma patterns by automatically pre-screening and highlighting cancerous regions prior to review. Our approach can be generalized to any whole-slide image classification task, and code is made publicly available at this https URL.
Jonathan I. Epstein - One of the best experts on this subject based on the ideXlab platform.
-
protocol for the examination of specimens from patients with carcinoma of the urethra
Archives of Pathology & Laboratory Medicine, 2010Co-Authors: Jesse K Mckenney, Jonathan I. Epstein, Mahul B Amin, David J Grignon, Esther Oliva, Victor E Reuter, John R Srigley, Peter A HumphreyAbstract:The College of American Pathologists offers these protocols to assist Pathologists in providing clinically useful and relevant information when reporting results of surgical specimen examinations. The College regards the reporting elements in the ‘‘Surgical Pathology Cancer Case Summary (Checklist)’’ portion of the protocols as essential elements of the pathology report. However, the manner in which these elements are reported is at the discretion of each specific Pathologist, taking into account clinician preferences, institutional policies, and individual practice. The College developed these protocols as an educational tool to assist Pathologists in the useful reporting of relevant information. It did not issue the protocols for use in litigation, reimbursement, or other contexts. Nevertheless, the College recognizes that the protocols might be used by hospitals, attorneys, payers, and others. Indeed, effective January 1, 2004, the Commission on Cancer of the American College of Surgeons mandated the use of the checklist elements of the protocols as part of its Cancer Program Standards for Approved Cancer Programs. Therefore, it becomes even more important for Pathologists to familiarize themselves with these documents. At the same time, the College cautions that use of the protocols other than for their intended educational purpose may involve additional considerations that are beyond the scope of these documents. PROTOCOL FOR THE EXAMINATION OF SPECIMENS FROM PATIENTS WITH CARCINOMA OF THE URETHRA
-
Pathologists and the judicial process: how to avoid it.
The American Journal of Surgical Pathology, 2001Co-Authors: Jonathan I. EpsteinAbstract:This review article covers the full range of issues concerning malpractice as it relates to Pathologists. Following a brief summary as to the incidence and general statistics on the outcome of lawsuits as well as common pathology misdiagnoses resulting in lawsuits, the definition of malpractice is discussed. These include duty, breech of standard of care, proximal cause, and damage. Details are provided as to what a Pathologist should do from the initial threat of a lawsuit, to the initial lawsuit, and through the initial physician/lawyer meeting. An in-depth analysis as to how Pathologists should handle themselves through the discovery process and, in particular, deposition is provided. Plaintiff attorneys' goals at deposition are covered in depth. These goals include: 1) education about the Pathologist's case and strategies; 2) impeachment of the Pathologist's credibility; and 3) judgment as to how effective a witness the Pathologist will be at trial. Various types of plaintiff's attorney at deposition are summarized. Also discussed is the post-deposition meeting with the legal representative, whether to settle, and specific issues relating to trial. Finally, general tips on how to avoid a lawsuit in pathology are reviewed.
Dale C Snover - One of the best experts on this subject based on the ideXlab platform.
-
Modeling complexity in Pathologist workload measurement: the Automatable Activity-Based Approach to Complexity Unit Scoring (AABACUS)
Modern Pathology, 2015Co-Authors: Carol C Cheung, Emina E Torlakovic, Hung Chow, Dale C SnoverAbstract:Pathologists provide diagnoses relevant to the disease state of the patient and identify specific tissue characteristics relevant to response to therapy and prognosis. As personalized medicine evolves, there is a trend for increased demand of tissue-derived parameters. Pathologists perform increasingly complex analyses on the same ‘cases’. Traditional methods of workload assessment and reimbursement, based on number of cases sometimes with a modifier (eg, the relative value unit (RVU) system used in the United States), often grossly underestimate the amount of work needed for complex cases and may overvalue simple, small biopsy cases. We describe a new approach to Pathologist workload measurement that aligns with this new practice paradigm. Our multisite institution with geographically diverse partner institutions has developed the Automatable Activity-Based Approach to Complexity Unit Scoring (AABACUS) model that captures Pathologists’ clinical activities from parameters documented in departmental laboratory information systems (LISs). The model’s algorithm includes: ‘capture’, ‘export’, ‘identify’, ‘count’, ‘score’, ‘attribute’, ‘filter’, and ‘assess filtered results’. Captured data include specimen acquisition, handling, analysis, and reporting activities. Activities were counted and complexity units (CUs) generated using a complexity factor for each activity. CUs were compared between institutions, practice groups, and practice types and evaluated over a 5-year period (2008–2012). The annual load of a clinical service Pathologist, irrespective of subspecialty, was ∼40 000 CUs using relative benchmarking. The model detected changing practice patterns and was appropriate for monitoring clinical workload for anatomical pathology, neuropathology, and hematopathology in academic and community settings, and encompassing subspecialty and generalist practices. AABACUS is objective, can be integrated with an LIS and automated, is reproducible, backwards compatible, and future adaptable. It can be applied as a robust decision support tool for the assessment of overall and targeted staffing needs as well as utilization analyses for resource allocation.
Lee R. Silverman - One of the best experts on this subject based on the ideXlab platform.
-
Industry–Contract Research Organization Pathology Interactions:A Perspective of Contract Research Organizations in Producing the Best Quality Pathology Report
Toxicologic Pathology, 2011Co-Authors: Sylvie J. Gosselin, Bernard Palate, Jeffery A. Engelhardt, Pierre A. Tellier, Kevin S. Mcdorman, Jerry F Hardisty, George A. Parker, Lee R. SilvermanAbstract:This article provides observations on the features of sponsor-contract research organization communication that will achieve the best quality pathology report based on our collective experience. Information on the test article and any anticipated findings should be provided, and initial slide examination should be done with knowledge of treatment group (but may be followed by blinded review of target tissues to determine no-effect levels). Only a Pathologist should write or revise the pathology report or the pathology section of the overall study report. To address concerns related to undue sponsor influence, comments by sponsors should be presented as suggestions rather than directives. Adversity should be defined for each finding by the study Pathologist, but the no-observed adverse effect level should not be discussed in the pathology report. Board-certified Pathologists are recommended, but are not essential. Sponsors that have a particular format or report preferences should make them known well in advance. Histologic processing "to glass" of protocol-specified tissues from all dosage groups is recommended for rapid evaluation of target tissues. Telepathology is beneficial in certain situations, but it is usually more efficient for the study Pathologist and reviewing Pathologist to be in the same physical location to review differences of opinion and reach a consensus.
