The case-control study as data missing by design: estimating risk differences

OBJECTIVE: To determine the DNA content and S-phase fraction (SPF) of pituitary adenomas by image analysis and to correlate them with clinical and morphologic parameters. STUDY DESIGN: The study group consisted of 26 prospectively collected cases of operated pituitary adenomas (3 microadenomas and 23 macroadenomas). The tumors were classified by histology, immunocytochemistry and electron microscopy. DNA measurement was performed on imprints from fresh pituitary tissue. Samples of nontumorous adenohypophysial parenchyma served as normal controls. RESULTS: Overall, 31% of adenomas, all but one functioning one, were aneuploid. The remaining nonfunctioning aneuploid tumor was a null cell adenoma with glycoprotein differentiation. All aneuploid tumors were macroadenomas, mostly at advanced stages, III and IV. Dural invasion, although frequent in macroadenomas (78%), was not correlated with DNA ploidy and SPF. An increased number of hyperpentaploid aneuploid cells was noted primarily in aneuploid tumors. The mean SPF was < 2.50%, with a statistically significant difference between aneuploid and diploid adenomas (3.60% vs. 1.70%). CONCLUSION: The results suggest that quantitative assessment of DNA content may provide important information, particularly in functioning adenomas. In addition, fresh tissue imprints represent excellent material for optimum cytometric measurements by image analysis systems, even for microadenomas.