Quantitative pathological study of aggressiveness of giant cell tumor of bone

Clinical behavior of giant cell tumor of bone (GCT) may vary remarkably from latent to very aggressive. Quantitative pathological methods were used to evaluate the aggressiveness of GCT. Fifteen cytometric parameters were measured and computed on routine sections of 40 GCTs which had been treated with curettage. The surgery factor (with or without additional procedures like bone cement packing or freezing after the curettage) was also taken into account and computed. Nineteen of the 40 GCTs were cured and 21 recurred. While no single one from the 15 parameters measured showed significant differences between the cured and recurrent groups according to the Mann-Whitney's U test or Student's t test, a 4-variable function was established with a stepwise discriminant analysis which could correctly identify 70.8% of the predicted cases as cured or recurrent (jackknife procedure). The function also suggested that in addition to the surgery factor, which no doubt had close relation with the prognosis, the most important risk factor in histological parameters was SA40, i.e. the percent of cells with nucleus larger than 40 square microns. Single cells extracted from paraffin-embedded blocks of 38 GCTs were analyzed by DNA-image cytometry. The 2c deviation index (2cDI) showed wide heterogeneity ranging from those consistent with benign tumors to those with apparent malignant ones, which may account for its diversity in clinical behaviour. Sixteen of the 38 cases had been treated with curettage, 8 of them were cured and followed up for at least 3 years and the remaining 8 recurred. The significant difference of 2cDI between the two groups suggested that these DNA parameters are useful in evaluating the aggressiveness of GCT for the selection of an adequate treatment. Quantitative approaches appeared to be more objective and more sensitive in evaluating the aggressiveness and predicting the prognosis of GCT than the subjective grading system used before.