Image cytophotometric DNA histogram heterogeneity in adenocarcinoma of the breast.

In 36 breast carcinomas, DNA histograms derived from the edge of an invasive primary tumor were compared with those of a second area in the invasive tumor (n = 10) or of corresponding in situ (n = 22) and/or (nodal) metastatic (n = 6) components using image cytophotometry on formalin-fixed, paraffin-embedded sections. DNA aneuploidy (defined by comparison with diploid host stroma cells) was identified in the invasive areas of 26/36 (72%), the noninvasive foci of 16/22 (73%) and the metastases of 5/6 (83%). Ploidy status of different areas of invasive tumor were concordant in all (10/10) cases; however, the in situ and invasive areas exhibited different DNA content determinations in 5/22 (23%), and metastases differed from the primary neoplasm in 3/6 (50%). Moreover clonal DNA content heterogeneity, defined as the presence of more than one discrete G0/G1 population, was observed in 20/36 (56%) tumors examined. Clonal heterogeneity was generally manifest as bimodal DNA histograms in one (n = 11) or both (n = 4) components of a tumor. Also, most DNA histogram heterogeneity was due to the presence of diploid range (n = 11) or near-tetraploid (n = 6) neoplastic populations accompanying DNA aneuploid clones. The remaining heterogeneous cases were unimodal in both components analyzed but exhibited aneuploid stemlines with different DNA indices (n = 2) or a diploid range population in one histogram and an aneuploid population in the other (n = 3). We conclude that DNA content heterogeneity is frequent in breast carcinoma and, in large part, may be accounted for by near-diploid or near-tetraploid stemlines, which are difficult to resolve from benign host elements in flow cytometric histograms.

Prognostic significance of image morphometric microvessel enumeration in breast carcinoma.

Using image morphometry, microvessel density was analyzed in 58 stage-heterogeneous breast carcinomas by counting vascular cross-sections in type IV collagen (basal lamina)-immunostained cryostat sections (methanol fixed, avidin-biotin method). Vascular density (mean, 19 vessels per square millimeter) did not correlate with metastatic disease at presentation (localized: 17.6 vessels per square millimeter vs. node/distant metastases: 20 vessels per square millimeter, P = NS). Short-term disease-free survival, however, was significantly related to intratumoral vascularity overall (no recurrence: mean = 14 vessels per square millimeter vs. recurred: mean = 26 vessels per square millimeter, P = .001) and in node-negative patients (no recurrence: mean = 12.1 vessels per square millimeter vs. recurred: mean = 27.8 vessels per square millimeter, P = .005; mean follow-up, 52 months). Further, 87% (7/8) of patients with microvascular counts greater than 30 vessels per square millimeter had recurrence vs. only 15% (2/13) with counts less than 10 vessels per square millimeter. We conclude that image morphometric quantitation of vascular basal lamina represents an objective means of assessing angiogenic capacity in breast tumors, which correlates strongly with disease aggressiveness in short-term follow-up.

Measurement of cellular proliferation in human prostate by AgNOR, PCNA, and SPF.

Tumor differentiation and proliferative activity are important predictors of biological behavior. While routine histological evaluation is fairly adequate to assess differentiation, tumor proliferative activity is difficult to measure. Silver staining for nucleolar organizer regions (AgNORs) is reported to be helpful for assessing tumor proliferation. We investigated the AgNOR counts in 20 formalin fixed, paraffin embedded human prostate tissues in three microscopic fields of 330X, using an image analysis system. A total of 200-700 nuclei were evaluated on histologically controlled areas of nonneoplastic prostate tissue, prostatic intraepithelial neoplasia (PIN), and invasive carcinoma. The values were compared to flow cytometrically obtained synthesis phase fractions (SPF) and immunohistochemically semi-quantitated, proliferative cell nuclear antigen (PCNA) patterns. AgNOR counts were also compared to tumor stage and Gleason's score. The pattern of PCNA staining in formalin fixed specimens was widely variable, probably due to differences in preservation of antigen. The positive counts varied from 0 to 55%, with a mean value of 8.55 +/- 15.9. The SPF values ranged from 5 to 13% with a mean value of 8.50 +/- 2.37. Two of 20 tumors were aneuploid and 18 were of diploid range. The mean AgNOR values in nonneoplastic nuclei (1.836 +/- 0.299), PIN (3.129 +/- 0.295), and invasive tumor cell nuclei (4.737 +/- 0.369) were highly significant (P < 0.0001) when paired differences were compared. AgNOR counts correlated significantly with tumor Gleason's score (P < 0.0145). However, the correlation coefficient for SPF and AgNOR values was not significant (P > 0.24), possibly because of the small number of samples examined. The highest AgNOR counts were found in the two aneuploid tumors. We conclude that AgNOR count may be a potential indicator of cellular proliferation, and possibly a marker of tumor differentiation.

Fine needle aspiration diagnosis of hyperplastic and neoplastic follicular nodules of the thyroid. A morphometric study.

The differentiation of hyperplastic nodules, follicular adenomas and follicular carcinomas from fine needle aspiration (FNA) cytology smears may be difficult. To better define the diagnostic criteria, we studied the morphometric parameters of nuclear area (NA), nuclear:cytoplasmic ratio and nuclear roundness (NR) in single cells and cell aggregates. In addition, we quantitated the percentage of touching or overlapping nuclei (NO) and the percentage of extent of nuclear area of overlap (NAO) in cellular aggregates. We measured cellular samples from FNA aspirates obtained from 20 hyperplastic nodules, 21 follicular adenomas, 5 encapsulated follicular carcinomas and 22 invasive follicular carcinomas, all of which were subsequently confirmed by histologic examination. Cellular aggregates provided the maximum diagnostic information. Stepwise discriminant analysis revealed that nuclear size, nuclear roundness and the percentage NAO allow optimum differentiation of hyperplasia, adenomas and carcinomas. Clearly, all of the poorly differentiated carcinomas (large NA, low NR, high NO and NAO) could be reliably diagnosed. Discriminant analysis allowed the differentiation of carcinoma from adenoma in 20/22 carcinomas (91%) and all 21 adenomas (although 2 adenomas were called hyperplasias and 3 hyperplasias were called adenomas).

Carcinoma in situ and microinvasive squamous carcinoma of the laryngeal glottis.

Twenty-five patients with severe squamous intraepithelial neoplasia, grade 3 (SIN III) of the glottis were identified and followed up; three patients (12%) eventually developed invasive carcinoma and nine patients (36%) were noted to have coexisting microinvasive carcinoma. Two of the patients with microinvasive carcinoma were treated by hemilaryngectomy, three with stripping and radiation therapy, two with stripping alone, and two by radiation therapy. None of the patients developed recurrent disease after one to five years of follow-up. This observation supports the approach of treating "true" microinvasive carcinoma similar to carcinoma in situ. The 16 remaining patients were treated by stripping (ten patients), stripping and radiation therapy (four patients), and radiation therapy alone (two patients). Three patients progressed to invasive squamous carcinoma at 35, 39, and 84 months after the diagnosis of SIN III. Patients in this study displaying SIN III with or without microinvasive carcinoma were successfully treated by surgery or radiation therapy alone or by combined therapy. The finding of 12% progression to invasive carcinoma and coexisting microinvasive carcinoma in 36% confirms the preinvasive nature of SIN III of the laryngeal glottis.