A pathologist-designed imaging system for anatomic pathology signout, teaching, and research.

Pathology images are derived from gross surgical specimens, light microscopy, immunofluorescence, electron microscopy, molecular diagnostic gels, flow cytometry, image analysis data, and clinical laboratory data in graphic form. We have implemented a network of desktop personal computers (PCs) that allow us to easily capture, store, and retrieve gross and microscopic, anatomic, and research pathology images. System architecture involves multiple image acquisition and retrieval sites and a central file server for storage. The digitized images are conveyed via a local area network to and from image capture or display stations. Acquisition sites consist of a high-resolution camera connected to a frame grabber card in a 486-type personal computer, equipped with 16 MB (Table 1) RAM, a 1.05-gigabyte hard drive, and a 32-bit ethernet card for access to our anatomic pathology reporting system. We have designed a push-button workstation for acquiring and indexing images that does not significantly interfere with surgical pathology sign-out. Advantages of the system include the following: (1) Improving patient care: the availability of gross images at time of microscopic sign-out, verification of recurrence of malignancy from archived images, monitoring of bone marrow engraftment and immunosuppressive intervention after bone marrow/solid organ transplantation on repeat biopsies, and ability to seek instantaneous consultation with any pathologist on the network; (2) enhancing the teaching environment: building a digital surgical pathology atlas, improving the availability of images for conference support, and sharing cases across the network; (3) enhancing research: case study compilation, metastudy analysis, and availability of digitized images for quantitative analysis and permanent/reusable image records for archival study; and (4) other practical and economic considerations: storing case requisition images and hand-drawn diagrams deters the spread of gross room contaminants and results in considerable cost savings in photographic media for conferences, improved quality assurance by porting control stains across the network, and a multiplicity of other advantages that enhance image and information management in pathology.

Medullary cancer of the breast revisited.

Common as well as unusual, heretofore unmentioned histopathologic features observed in 336 typical and 273 atypical medullary breast cancers from 6404 patients enrolled in various stage I and II protocols of the National Surgical Adjuvant Breast and Bowel Projects (NSABP) are presented. Both medullary types exhibited comparable pathologic findings, except for the infiltrative border and/or slight or absent tumor lymphoid infiltrate which by definition characterize the atypical form. Both also demonstrated a similar, high proclivity to be aneuploid, and to lack estrogen and progesterone receptors and nodal metastases. After appropriate statistical adjustments, survival (analyzed for 198 patients with typical and 149 with atypical medullary cancers) was found to be better for untreated, node-negative and node-positive patients treated with L-PAM + 5Fu who had typical medullary cancers than those with the NOS histologic type. The magnitude of this difference was 6% at 5 and 17% at 10 years post-operatively (cumulative odds = 1.81 with a 95% confidence interval of 1.08 - 3.3) for the former group, and 4% at 5 and 16% at 10 years (cumulative odds = 1.56 with a 95% confidence interval of 1.08 - 2.23) for the latter. Survival was comparable for patients with atypical medullary and NOS types in both situations. No clear difference in survival was found in untreated, positive node patients with the 3 histologic types examined, although the sample sizes in this subset were relatively small. This information as well as other pertinent considerations indicate that the prognosis of typical medullary cancer is not as 'good' as previously perceived. It is also concluded that there is insufficient evidence at present to exclude the atypical medullary variant as a histologic type of breast cancer.

A three-dimensional reconstruction of metastatic adenocarcinoma in lymph node.

Lymph nodes are a common site of metastatic cancer. The ability to view the three-dimensional configuration of complex biological structures, rendered in solid model form, might help to further elucidate the pathophysiology of metastatic disease. This paper presents a method for three-dimensional reconstruction from serial sections of a lymph node containing metastatic adenocarcinoma. The reconstruction and subsequent animation were carried out using the P3D graphics software (developed at the Pittsburgh Supercomputing Center) and rendered by the Dore high-speed renderer on an Ardent Titan graphics workstation. As different views of the model were produced, they were recorded a frame at a time on U-matic video tape. A three-dimensional solid-modeled object portrays metastatic, neoplastic elements within a lymph node.

Value of assessment of ploidy in rectal cancers.

Flow cytometry was performed on available archival material from 232 patients with rectal cancer enrolled in the National Surgical Adjuvant Breast and Bowel Project protocol R-01. Tumor ploidy was not found to be significantly related to such pathologic and clinical parameters as Dukes' stage; nodal status; nuclear or histologic grades; patient's age, sex, or overall survival rate with an average study time of 79 months. A trend was evident that patients with poor histologic grade, or those with Dukes' B and C tumors that were aneuploid fared worse than those with diploid cancers. However, measurements of survival were found to be more strongly and consistently related to such conventional prognostic parameters as tumor differentiation, Dukes' stage, and nodal status. Further, numbers of nodes with metastases (ie, 1 to 4 or 5+) more significantly discriminated Dukes' C cases than estimation of tumor ploidy. Although tumor ploidy may reflect some features of rectal cancers, their natural history and prognosis are explained better by assessment of conventional parameters used for these purposes.