Comparison of different anti-Ki67 antibody clones and hot-spot sizes for assessing proliferative index and grading in pancreatic neuroendocrine tumours using manual and image analysis.

AIMS: Ki67 proliferative index (PI) is essential for grading gastroenteric and pancreatic neuroendocrine tumours (GEP NETs). Analytical and preanalytical variables can affect Ki67 PI. In contrast to counting methodology, until now little attention has focused on the question of clone equivalence and the effect of hot-spot size on Ki67 PI in GEP NETs. Using manual counting and image analysis, this study compared the Ki67 PI achieved using MM1, K2 and 30-9 to MIB1, a clone which has been validated for, and is referenced in, guidelines relating to assessment of Ki67 PI in GEP NETs. METHODS AND RESULTS: Forty-two pancreatic NETs were each immunohistochemically stained for the anti-Ki67 clones MIB1, MM1, K2 and 30-9. Ki67 PI was calculated manually and by image analysis, the latter using three different hot-spot sizes. In manual comparisons using single hot-spot high-power fields, non-MIB1 clones overestimated Ki67 PI compared to MIB1, resulting in grading discordances. Image analysis shows good agreement with manual Ki67 PI but a tendency to overestimate absolute Ki67 PI. Increasing the size of tumour hot-spot from 500 to 2000 cells resulted in a decrease in Ki67 PI. CONCLUSION: Different anti-Ki67 clones do not produce equivalent PIs in GEP NETs, and clone selection may therefore affect patient care. Increasing the hot-spot size decreases the Ki67 PI. Greater standardisation in terms of antibody clone selection and hot-spot size is required for grading GEP NETs. Image analysis is an effective tool for assisting Ki67 assessment and allows easier standardisation of the size of the tumour hot-spot.

HMGA2 is commonly expressed in uterine serous carcinomas and is a useful adjunct to diagnosis.

AIMS: Serous carcinoma is the prototype of type 2 uterine carcinoma. In many cases, establishing a diagnosis is straightforward, but problems can arise in that papillary variants of endometrioid carcinoma may be mistaken for serous carcinoma, and glandular variants of serous carcinoma may be misdiagnosed as endometrioid carcinoma. Markers such as p53, oestrogen receptor and p16 may be of use in problematic cases, but there is overlap and these may not therefore be of value in an individual case. It has been shown recently that high-mobility group AT-hook 2 (HMGA2) is expressed by most ovarian serous carcinomas, and our aim was to ascertain whether it is also expressed in uterine serous carcinoma and of value in its distinction from endometrioid carcinoma. METHODS AND RESULTS: Whole tissue sections of uterine serous (n = 33) and endometrioid (n = 38) carcinoma were immunostained using HMGA2 antibody. As many of the diagnostic problems relate to the distinction between serous carcinoma and grade 3 endometrioid carcinoma, tissue microarrays (TMAs) containing uterine serous (n = 71) and uterine grade 3 endometrioid (n = 68) carcinomas were also stained. Staining was classified as negative (totally negative or occasional nuclei positive), 1+ (<10% of nuclei positive), 2+ (10-49% of nuclei positive), 3+ (50-74% of nuclei positive), or 4+ (≥75% of nuclei positive). On the whole tissue sections, positive staining was also classified as weak, moderate, or strong, and an immunohistochemical composite score, taking into account both extent and intensity of staining, was calculated. On whole tissue sections, there was a statistically significant difference between HMGA2 staining in serous and endometrioid carcinomas with regard to both extent and composite score, with higher expression in serous carcinomas (P < 0.0001). Thirty of 33 (91%) serous carcinomas were positive, usually with diffuse (3+ or 4+) staining. All five cases of serous endometrial intraepithelial carcinoma (EIC) (the postulated precursor of uterine serous carcinoma) were positive, as were 14 of 38 (37%) endometrioid carcinomas, usually with 1+ or 2+ staining. There was a statistically significant difference in HMGA2 staining in the TMAs between the serous and grade 3 endometrioid carcinomas, with higher expression in the former (P < 0.0001). CONCLUSIONS: Immunoreactivity for HMGA2 is diffusely positive in whole tissue sections in most uterine serous carcinomas and negative in most endometrioid carcinomas, although, as with other markers, there is overlap in individual cases. In conjunction with other markers, HMGA2 may be of value in problematic uterine carcinomas where the differential diagnosis includes serous and endometrioid carcinoma. As HMGA2 is expressed in serous EIC, this suggests that it may be implicated in the early development of uterine serous carcinoma.

Papillary syncytial metaplasia associated with endometrial breakdown exhibits an immunophenotype that overlaps with uterine serous carcinoma.

Uterine serous carcinoma (USC) is an aggressive variant of Type 2 endometrial carcinoma, which in most cases exhibits, at least focally, a papillary architecture. Occasionally, especially in small biopsy specimens, it may be difficult to distinguish between USC and a variety of metaplastic or reactive processes. In particular, papillary syncytial metaplasia (PSM), as a result of endometrial breakdown, may be confused with USC or its precursor serous endometrial intraepithelial carcinoma. In such cases, immunohistochemistry is often undertaken, the panel of markers usually including estrogen receptor (ER), p53, p16, and MIB1. The expected immunoprofile of USC is ER negative, p53 and p16 positive, and a high MIB1 proliferation index, although studies have shown that significant numbers of cases deviate from this immunophenotype. With regard to the aforementioned markers, PSM has not been studied extensively, but intuitively, the expected immunophenotype would be ER positive, p53 and p16 negative, and a low MIB1 proliferation index. After 2 index cases in which breaking down menstrual endometrium with florid PSM was misdiagnosed on an endometrial biopsy as USC or suspected USC, in part due to the observed immunophenotype, we studied the expression of ER, p53, p16, MIB1, and HMGA2 (a recently described useful marker of USC) in 10 further cases of PSM associated with endometrial breakdown. We illustrate that compared with a nonbreaking down endometrium, PSM is characterized by a decreased expression of ER and an increased expression of p53 (although still wild-type staining) and p16, the latter marker typically being diffusely positive. HMGA2 is negative, and there is a low MIB1 proliferation index. In cases of PSM, which are morphologically problematic, the immunophenotype may further heighten the suspicion of serous malignancy and potentially result in a misdiagnosis.

Misplaced Skene's glands: glandular elements in the lower female genital tract that are variably immunoreactive with prostate markers and that encompass vaginal tubulosquamous polyp and cervical ectopic prostatic tissue.

So-called ectopic prostatic tissue in the cervix and vaginal tubulosquamous polyps are rare morphologically similar lesions that may show positive immunohistochemical staining with prostatic markers. It has been suggested that they are related to paraurethral Skene's glands that are the female equivalent of prostatic glands in the male. We report a large series of lesions in women aged 23 to 81 years, found within the cervix (n=24), vagina (n=10), and vulva (n=2), which we believe to be a part of a spectrum of lesions derived from Skene's glands, either eutopic or more commonly misplaced during embryonic development. In all cervical cases, the lesion was predominantly situated in the ectocervix and was an incidental finding in specimens procured for a variety of reasons. In the vagina, the lesions usually presented themselves as polyps or cysts, although occasionally they were an incidental finding. The 2 vulval cases were incidental findings in punch biopsies. The basic morphological features were of epithelial elements of both glandular and squamous type; in some cases, the glandular elements formed a double cell layer. Uncommon findings included the presence of sebaceous glands in 2 cases (1 cervix, 1 vagina), basaloid formations resembling hair follicle structures in 4 (2 cervix, 2 vagina), and a microglandular proliferation resembling nephrogenic adenoma in 1 vaginal case. Prostate-specific antigen was positive in 13 of 26 cases and prostatic acid phosphatase in 16 of 26 tested. Six cases were negative with both markers. We propose that these benign lesions in the cervix, vagina, and vulva are derived from eutopic or misplaced Skene's glands.

Mesonephric adenocarcinomas of the uterine cervix and corpus: HPV-negative neoplasms that are commonly PAX8, CA125, and HMGA2 positive and that may be immunoreactive with TTF1 and hepatocyte nuclear factor 1-ß.

Mesonephric adenocarcinomas are rare neoplasms that most commonly arise in the uterine cervix and exceptionally rarely in the uterine corpus. Although the morphologic features of these neoplasms are well described, there has been relatively limited investigation of the immunoprofile. We report a series of 8 mesonephric adenocarcinomas arising in the uterine cervix (7 cases) and corpus (1 case) and undertake a comprehensive immunohistochemical analysis. This includes markers that have not been investigated previously in mesonephric adenocarcinomas but that are commonly used in gynecologic pathology and may be undertaken when other, mainly Mullerian, adenocarcinomas are considered in the differential diagnosis. Linear array human papillomavirus (HPV) genotyping was also performed. Our results broadly confirm the immunohistochemical profile demonstrated in previous studies with the majority of mesonephric adenocarcinomas staining positively with CD10 (6 of 8), epithelial membrane antigen (8 of 8), vimentin (8 of 8), and calretinin (7 of 8). Estrogen receptor was positive in 2, carcinoembryonic antigen in 3, and inhibin in 4 cases. p16 was positive in 5 cases (1 diffuse and strong), despite all being HPV negative (in 1 case, there was insufficient DNA for HPV analysis). Novel findings in our study were the demonstration of nuclear positivity with PAX8 and HMGA2 in 7 cases, CA125 immunoreactivity in all 8 cases, and TTF1 and hepatocyte nuclear factor 1-ß staining in 3 cases. As PAX8, CA125, HMGA2, and hepatocyte nuclear factor 1-ß are commonly positive in a variety of Mullerian adenocarcinomas arising in the female genital tract, this may result in diagnostic confusion. All cases were WT1 negative.

HMGA2 is a sensitive but not specific immunohistochemical marker of vulvovaginal aggressive angiomyxoma.

Aggressive angiomyxoma is a rare mesenchymal neoplasm, which almost always occurs in the vulvovaginal region and which exhibits a marked tendency for local recurrence after excision. A wide range of other mesenchymal lesions occur in this region, which potentially mimic aggressive angiomyxoma to a variable extent. Because rearrangement of the HMGA2 gene has been shown in a significant percentage of aggressive angiomyxomas, we examined the expression of HMGA2 protein in this neoplasm and in a large number of other mesenchymal lesions occurring at this site, including many which were referred with a possible diagnosis of aggressive angiomyxoma. There was positive staining of tumor cell nuclei, mostly with a diffuse distribution, in all but 2 aggressive angiomyxomas (n=12). Blood vessel endothelial cells within the neoplasms, entrapped tissues, and surrounding normal tissues were negative. Ten of 23 leiomyomas were positive, as were occasional other neoplasms. Based on our study, we conclude that HMGA2 is positive in most aggressive angiomyxomas and is useful in diagnosis as most mesenchymal lesions which closely mimic this are negative. However, caution should be exercised as some other vulvovaginal mesenchymal lesions, especially, but not exclusively leiomyomas can be HMGA2-positive. As well as being of value in primary diagnosis, HMGA2 is useful in evaluating margins and in reexcision specimens of aggressive angiomyxoma in identifying foci of residual or recurrent tumor.

CD56 is a sensitive and diagnostically useful immunohistochemical marker of ovarian sex cord-stromal tumors.

Ovarian sex cord-stromal tumors comprise a heterogeneous group of neoplasms with wide morphological diversity, and they can be mistaken for a variety of other tumors. Some types, including granulosa and Sertoli cell tumor, may be confused with a neuroendocrine neoplasm. CD56 is a widely used neuroendocrine marker with a high sensitivity for neuroendocrine tumors and is commonly used as part of a panel to distinguish between a neuroendocrine neoplasm and other tumors in the differential diagnosis. In this study, we investigate CD56 staining in ovarian sex cord-stromal tumors. CD56 staining has not previously been studied in this group of neoplasms. We stained a large series of ovarian sex cord-stromal neoplasms (n = 85) with CD56. Neoplasms studied were adult granulosa cell tumor (n = 40), juvenile granulosa cell tumor (n = 8), Sertoli cell tumor (n = 1), Sertoli-Leydig cell tumor (n = 14), Leydig cell tumor (n = 2), steroid cell tumor, not otherwise specified (n = 2), sclerosing stromal tumor (n = 1), sex cord tumor with annular tubules (n = 2), and fibroma (n = 15). Three uterine tumors resembling ovarian sex cord tumor were also studied. Nonneoplastic ovaries, including 3 cases of pregnancy-related granulosa or Sertoli cell proliferation, were also included. In nontumorous ovaries, granulosa cells of follicular and corpus luteum cysts were consistently negative. The normal ovarian stroma was diffusely positive, as were the 3 pregnancy-related proliferations. All sex cord-stromal tumors except one were positive with CD56; staining ranged from focal to diffuse but was usually diffuse involving more than 50% of tumor cells. Staining was usually membranous with weaker cytoplasmic positivity. CD56 immunoreactivity is almost universal in ovarian sex cord-stromal tumors of all the major morphological types and is of no value in distinguishing a sex cord-stromal and a neuroendocrine neoplasm. Since CD56 is an extremely sensitive marker of ovarian sex cord-stromal tumors, it may be useful in the diagnosis of this group of neoplasms, especially in cases which are alpha inhibin or calretinin negative, and in distinguishing these from mimics which are CD56 negative.

A panel of immunohistochemical stains, including carcinoembryonic antigen, vimentin, and estrogen receptor, aids the distinction between primary endometrial and endocervical adenocarcinomas.

The histological distinction between a primary endometrial and a primary endocervical adenocarcinoma is often difficult, especially in small biopsy specimens. A preoperative distinction is important because primary surgical management differs between the two tumors. Cases of primary endometrioid endometrial (n=30) and primary endocervical (n=26) adenocarcinoma of endocervical type were stained immunohistochemically with the monoclonal antibodies against carcinoembryonic antigen (CEA), vimentin, estrogen receptor (ER), and 34 beta E12. In all cases the origin of the adenocarcinoma was confirmed by examination of the definitive pathology specimen. There was diffuse positive nuclear staining for ER in 28 of 30 (93%) endometrial adenocarcinomas. ER was negative in 16 of 26 endocervical adenocarcinomas, and there was focal weak nuclear staining in the other cases. Vimentin was positive in 29 of 30 (97%) endometrial adenocarcinomas but in only 2 of 26 (8%) endocervical adenocarcinomas. CEA was positive in 25 of 26 (96%) endocervical adenocarcinomas, mostly with diffuse membranous and cytoplasmic staining. Positivity with CEA was present in 21 of 30 (70%) endometrial adenocarcinomas but was largely confined to squamoid areas with only 12 tumors exhibiting focal membranous staining of the glandular component. 34 beta E12 was diffusely positive in all except one cervical adenocarcinoma. In endometrial carcinomas, positivity was strongest in squamoid areas but there was positive staining, either focally or diffusely, of the glandular component in 27 cases. In summary, primary endometrioid endometrial adenocarcinomas are characterized by diffuse, strong, positive staining for vimentin and ER and negative or very focal, positive staining of the glandular component for CEA. In contrast, primary endocervical adenocarcinomas are characterized by CEA positivity, which is usually but not always diffuse, negativity for vimentin, and negativity or focal weak positivity for ER. 34 beta E12 is of no value in the distinction between endometrial and endocervical adenocarcinomas. A panel of immunohistochemical stains, comprising CEA, vimentin, and ER, generally allows confident preoperative distinction between a primary endometrial and endocervical adenocarcinoma.