Abstract
Immunohistochemistry has rapidly established a significant role in diagnostic pathology. We use immunohistochemistry as an adjunct to morphological diagnosis and employ a “panel approach” to the classification of poorly differentiated tumors. This retrospective analysis was conducted to examine the efficacy of such an approach, using, as an example, the two most common categories of poorly differentiated tumors, namely, the poorly differentiated round cell tumors and spindle cell tumors. Five hundred and fifty-seven consecutive cases of such tumors, collected over a two-yr period, had been subjected to immuno-histochemical staining because specific or definitive categorization of the tumor was not possible on the basis of the examination of hematoxylin and eosin stained sections. The clinical history, gross and microscopic findings, as well as the results of immunohistochemical stains were analyzed. Immunohistochemistry allowed a definitive diagnosis to be assigned in 420 cases (75.4%). It was particularly useful in identifying malignant melanoma of both epithelioid and spindle types and distinguished between melanoma, lymphoma and poorly differentiated carcinoma in 126 cases of such lesions occurring in adult lymph nodes. It was also useful in identifying tumors in small biopsies where poor cytomorphological preservation or small size precluded accurate categorization. The application of appropriately chosen panels of antibodies tailored to a narrow list of differential diagnoses helped to identify myogenous, vascular, nerve sheath and fibrocystic lesions among the group of spindle cell tumors. Immunohistochemistry provided definitive diagnoses in 70% of round cell tumors and 92% of spindle cell tumors. Immunostaining was also useful in determining the source of metastatic carcinomas, particularly those of prostate, breast, thyroid and germ cell origin, all of which are “treatable” malignancies. Immunostaining provided contributory but non-diagnostic information in 71 cases (12.8%). While most of these were confirmed to be metastatic carcinoma, their primary source remained undetermined because of the currently limited repertoire of tissue-associated markers. Immunohistochemistry was not contributory in 66 instances (11.8%) and the major reason for this was the suboptimal preservation of tissue antigens, especially in consultation material.
Various combinations of a total of 48 antibodies were employed to produce the diagnostic information. For the 557 cases examined, a total of 2,352 sections were stained at an average cost of about $71 per case inclusive of $5.00 for staining reagents.
We conclude that the efficacy and cost of this important adjunct to morphological diagnosis is largely dependent on two factors, viz, optimal preservation of tissue antigens and the experience and knowledge of the characteristics of the antibodies to allow the selection of appropriate reagents for the separation of a narrow list of differential diagnoses developed from morphological examination.