2001 CTOS Annual Meeting Posters— Pathology

DIAGNOSTIC IMPACT OF HISTOLOGIC PARAMETERS IN DIFFERENTIATING ENCHONDROMA FROM GRADE I CENTRAL CHONDROSARCOMA
D. Eefting¹,  M. J. Geirnaerdt²,  S. Le Cessie³,  A. H. Taminiau⁴,  P. C. Hogendoorn^1
1Department of Pathology, Leiden University Medical Center,  ²Department of Radiology, Leiden University Medical Center,  ³Department of Medical Statistics, Leiden University Medical Center,  ⁴Department of Orthopedic Surgery, Leiden University Medical Center

METHODS: Cytologic and tissue-architectural features of 57 patients, with 20 enchondromas and 37 central grade I chondrosarcomas, were assessed. During the microscopically assessment, the observer was aware of the localization of the tumor, the age of the patient and clinical symptoms (like swelling, pain or fracture). A final diagnosis was used as gold standard obtained after at least 10 years of follow-up and assessed in a multidisciplinary meeting with full clinical, radiological and pathology data available. Tumors of the digits and toes were excluded because of their specific (benign) clinical behavior, despite aggressive morphologic features. The chi-square test, classification trees and logistic regression were used to assess the discriminating power of the individual parameters and to find an optimal set of parameters.

RESULTS: Entrapment, cellularity, nuclear polymorphism and distribution of cells were the individual parameters with the most discriminating strength (p<0.001). The combination of high cellularity, present host bone entrapment, open chromatin, mucoid matrix quality >= 20% and an age above 45 years gives a perfect prediction. Also, the matrix quality proved to be a strong discriminating feature. A statistical classification tree shows that if mucoid matrix degeneration is more then 20% or entrapment is present (or both of them), the tumor will be almost certain malignant and otherwise almost certain benign.

CONCLUSION: The application of a combination of 5 histologic parameters showed a optimal ability to differentiate between enchondromas and central grade I chondrosarcomas. With a simple classification tree, based on 2 parameters, 54 of the 57 (94.7%) cases were assessed correct (sensitivity 95% and specificity 95%). Using this diagnostic approach combined with an tailored radiological assessment will lead to an optimal therapeutic classification.