Prognostic relevance of cell proliferation in major salivary gland carcinomas

Several proliferation markers, such as DNA ploidy, Ki67, MiB1 and proliferating cell nuclear antigen have been shown to correlate with clinical course and prognosis in several epithelial tumours and lymphomas. In the present study, the prognostic relevance of these markers was evaluated in major salivary gland carcinomas. A sample of 36 cases out of 85 patients submitted to surgery for major salivary gland carcinomas at our institution between 1987 and 1997 were studied. The sample comprised 8 adenoid-cystic carcinomas, 6 ductal carcinomas, 11 mucoepidermoid carcinomas and 11 acinic cell carcinomas. Follow-up ranged from 1 to 12 years (mean 6.2). In some patients, DNA ploidy (euploid or aneuploid) was studied by flow cytometry. In others, proliferation activity was studied by means of monoclonal antibody MiB1, identifying cells in the proliferative cycle. In some patients, both techniques were used. Follow-up was related to these indices, TNM and stage. Even if ploidy suggested a favourable outcome in diploid cancer (13 favourable vs. 2 unfavourable) and poor outcome in aneuploid cancer (4 favourable vs. 5 unfavourable), the difference was not statistically significant with p = 0.06 in Fisher's exact test. Instead, the proliferative tumour cell fraction, evaluated by MiB1, was statistically correlated with prognosis. Comparing survival curves by Log rank Test it yielded p = 0.007 using an MiB1 cut-off of 5. Applying a cut-off of 20 yielded p = 0.001. Of particular interest were MiB1 values in acinic cell carcinomas for which grading is challenging and lacks consensus. In our group of acinic cell carcinomas, survival correlated with values of MiB1 > or < 15 with p = 0.009 in Log rank test. In conclusion, despite a trend towards correlation between ploidy and prognosis, the present study yielded p = 0.06, whereas the proliferative fraction assessed by MiB1 was significantly correlated with outcomes. Indeed, "growth fraction" in acinic cell carcinomas may stratify different classes of risk.