Background: Algorithmic scoring approaches for evaluating early cases of mycosis fungoides (MF) provide a degree of diagnostic standardization. At the UWMC, biopsies from clinically concerning cases for MF are individually reviewed by a panel of pathologists and an average score is assigned to each biopsy reflecting the degree of suspicion for a diagnosis of MF; however, such an approach may not be practical outside of an academic center.
Methods: 78 cases characterized in our institution, with the diagnosis confirmed by clinicopathologic correlation/clinical follow-up were evaluated with two different algorithms, based entirely on histologic evaluation (Guitart algorithm) and partial implementation of the ISCL algorithm evaluating histology, immunohistochemistry and T-cell clonality.
Results: A receiver operating characteristic curve comparing the results of the two approaches in early MF cases showed no statistical difference between the areas under the two curves. Increased stages of MF showed variable loss of T-cell antigens by immunohistochemistry and higher rates of detectable clonality.
Conclusion: We could not document a statistically significant advantage of adding ancillary immunohistochemical and molecular testing to careful histologic evaluation in the workup of suspected cases of early MF. A systemic approach to histologic diagnosis by a single pathologist correlated favorably to the MF panel diagnosis.
Copyright © 2009 John Wiley & Sons A/S.