Comparison of nuclear shape in aspirated and histologic specimens of prostatic carcinoma.

Nuclear shape analysis of histologic sections of operative specimens has separated patients with clinically localized prostatic carcinoma with good and poor prognoses. In order to become more useful clinically, nuclear shape should be evaluable preoperatively but is distorted by core biopsy. We used nuclear morphometry to compare histologic and cytologic specimens obtained from 20 patients who underwent radical prostatectomy for clinically localized prostatic carcinoma. Neoplastic nuclei from cytologic aspirates prepared by the Papanicolaou and Diff-Quik techniques and standard histologic sections stained with hematoxylin and eosin were digitized. Air-dried nuclei of the Diff-Quik preparation had a nuclear area of 104.2 microns 2 +/- 28.2 SD. These nuclei were nearly twice as large (paired t test, P < .001) as in the Papanicolaou (55.6 +/- 13.0 microns 2) and histologic (55.8 +/- 12.7 microns 2) preparations. However, nuclear shape was not affected. Nuclear roundness factor (deviation from circularity) and boundary curvature were similar (paired t tests, P > .05). All nuclear shape descriptors of cytologic smears fell within intraobserver variations of measurements in histologic sections. Nuclear shape is not altered in Papanicolaou- and Diff-Quik-stained cytologic specimens and should be tested for preoperative assessment of outcome in clinically localized prostatic carcinoma.