Diagnostic workstations in pathology

Diagnostic workstations in pathology

Human PATHOLOGY VOLUME 21 October ‘1990 NUMBER 10 Editorial Diagnostic Workstations in Pathology their image storage and retrieval are relevant to...

197KB Sizes 0 Downloads 28 Views





Editorial Diagnostic Workstations in Pathology their image storage and retrieval are relevant to the dates of their study. Being a computer user and program developer, however, means to be dated in two ways: in a positive sense, being on the cutting edge. and in a negative sense, using last year’s technology. This is not a criticism of the van Ginneken study. Given the equipment and methods used, their experience demonstrates the effectiveness of’ their svstem as a participant in the diagnostic process. While the program described by van Ginneken et al’ includes text that they have provided in the form of descriptions and differential diagnoses, it is now also possible to have the entire text of the reference books, which were an alternative source of verbal information , available on the computer workstation. hlost reference textbooks are less than 6 megabytes in length, and can readily be stored cm hard discs. It is now possible, using printer’s tapes, to render any textbook into a computer format in less than 1 week. ‘I‘he complete digitized texts of Cotran. Kumar, and Robbins and the Merck Manual are available at Robert Wood Johnson Medical School in HyperCard stacks for medical student and resident education. The full texts of Scientific American Medicine and the Physicians’ Desk Reference are presently available on compact discs. In short, the verbal base of information necessary for the surgical pathology diagnostic process could be available on computers immediately. The central issue for the availability of this information base now lies in the hands of publishers: it is no longer a technical issue. The storage of images for rapid retrieval currently involves two formats: analog records on videodiscs that store tens of thousands of images (34,000 for 8 inch; 100,000 for 12 inch) and digital records stored in magnetic form. The storage of images in digital form with high resolution remains storage intensive and, for adequate color and size, requires approximately 0.1 to 0.2 megabytes each. Stored images can be located in a central computer to which workstations are networked. Digital images can also be easilv manipulated for presentation purposes. The stor-

Rendering .I diagnosis on a surgical specimen usuallv involves the integration of visual, verbal, and written information. Direct visual information is both macroscopic and microscopic. The macroscopic inc~ludes the tissue viewed in situ and/or at the surgical pathology bencll: the microscopic information is derived from \rarious t! pes of preparations made from the speciinen. Indirect visual information consists of t-emembtred observation5 of prior cases, images published as pictures in textbooks, and local archives of histopathology slides and/or photomicrographs. The 1 erbal information is derived from discussions with clinicians and cc)Ileagues. and is rarelv available as a reference sourc’e. The written information comes t rom requisition slips. patient records. letters from ( onsultants, and reference textbooks and articles. Keferenc:e to the prior experience of other paI hologists is m essential part of the diagnostic pro(‘ess. Those who have trained with experts often quote from the oral history,of their training and use those opinions in their declslon-making process. Prior rxperiente is also encoded in the literature and consists primarilk of text, diagrams, tables, and images. It IS COII~N~I~ tc; have reference texts, the UAKEPiAFIP Fascicles, or t.he WHO International Histological Classific.ntion of’ ‘l‘umors on the bookshelves in the surgical pathology sign-out room. ‘I‘he computer is an information storage, mandEemen1:, and presentation device, and is likely to begm to play a role in the process described above. The article b! A.M. van Ginneken et al.’ addresses the advent of the computer in pathologic diagnosis and provides outcome analyses beyond usual user enthusiasm or acceptance, which are encouraging and suggest that computers can be effectively used in the surgical pathology sign-out laboratory as a reference source. I-hose who read this article, in which van Ginneken t’t al conclude that “textbooks are more effective than their (computer) program” and conclude, in turn, that computers will be ineffective in this effort, are. in this writer’s opinion, missing the point. The vigor of‘ their program. their hardware platform, and 987


Volume 21, No. 10 (October 1990)

age of images on videodiscs does not require computer memory or hard disc storage, but does require a separate videodisc player and monitor for each workstation. Videodiscs cannot be networked. Both approaches have been extensively used in pathology, and a variety of different products and promises lie ahead. For teaching undergraduate medical students, first-year residents, and pathologists in training, a library of videodiscs is now available, including (1) the Slice of Life, produced by Dr Suzanne Stensaas, University of Utah, which is in its fifth edition and includes all illustrations from the Atlas of Anatomic Pathology by Cooke and Stewart, published by Churchill Livingstone; (2) seven videodiscs from the National Library of Medicine produced by Drs Robin Jones and James Woods, which deal with topics ranging from cell necrosis to the metastatic process: (3) Lymph Node Pathology from Intellipath on videodisc, produced by Dr Bharat Nathwani, University of Southern California, distributed by the ASCP Press; and (4) Hematology videodisc, produced by Drs -Jim Fine and John Bolles, University of Washington. In digital imaging, Drs Daniel Alonso and Richard Erde at Cornell Medical College have collected nearly 3,000 images to date and plan continued growth of their archive. The article by van Ginneken et al is to be lauded

for showing how close this new technology is to being useful, and reminding us, at the same time, that it is far from perfect. Technical developments have removed doubts about the feasibility of text and image storage and their immediate and simultaneous retrieval. The computer-assisted management of this visual and verbal information and its logical, stepwise presentation in the context of the diagnostic process is a next step. One such example is that illustrated by the van Ginneken article; another excellent effort is from Nathwani et al,’ initially for lymph nodes, but soon to address a number of other subjects. With these and other programs available, scholars and practitioners of surgical pathology can continue to explore and refine the surgical pathology decisionmaking process. Robert

L. Trelstad,




REFERENCE 1. vanGinneken AM, Baak JPA, Jansen W, et al: Evaluation of a diagnostic encyclopedia workstation for ovarian pathology. HU,M PATHOI. 2 1:989-997. 1990 2. Nathwani BN, Heckerman DE, Horvitz EJ, et al: Integrated expert systems and videodisc in surgical pathology: An overview. HIJ~M PATHOL 2 1: 1 l-27, 1990