Quantification of epithelial area by image processing applied to endometrial carcinomas: a comparison with ovarian tumors

In endometrial carcinomas, the epithelial area measured by interactive morphometry is an important feature in the classification of tumors of varying histologic grades. This report describes an image analysis technique for the fully automated estimation of the area percentages of epithelium and stroma in tissue sections of the endometrium obtained from hysterectomy specimens. The method is evaluated using endometrial carcinomas with varying degrees of malignancy. From standard paraffin sections stained with pararosanilin Feulgen and naphthol yellow, a blue-yellow image pair was recorded. The blue image was used to determine the total tissue area and the yellow image was used to determine the epithelial area. Image processing of the blue image was comprised of correction for shading, segmentation of the tissue area, and restoration of the segmented image by removing small artefacts and closing small tears in the tissue. Image processing of the yellow image was based on the fact that epithelial nuclei are generally more tightly packed than stromal nuclei and consists of the following steps: correction for shading, gaussian blurring, segmentation of nuclei, and editing the segmented image by removing small objects and closing small spaces between the epithelial nuclei. These image processing steps are compared with those used for quantification of the epithelial area in ovarian tumors. The performance of the method was evaluated using 120 image pairs from 30 endometrial carcinomas of varying histologic grades. The epithelial area percentages, as assessed by digital image processing, strongly correlate to control percentages that were established by interactive morphometry (r = .987).     

Evaluation of automated estimation of epithelial volume and its prognostic value in ovarian tumors

The paper describes an improved segmentation method to measure the percentages of epithelium and stroma in ovarian (tumor) tissue with automated image analysis and evaluates its prognostic value. In the image processing method, a blue-yellow image pair is recorded from standard paraffin sections and stained with pararosanilin Feulgen and naphthol yellow. The blue image is used for automated determination of the total tissue area and the yellow image for the epithelial area. Results are obtained with 114 ovarian tumors of the common epithelial types (14 borderline tumors and 100 invasive carcinomas with varying degrees of differentiation). The fraction of epithelium in the total tissue shows a strong correlation with the epithelial percentage resulting from interactive morphometry (r = 0.991) for 15 tumors of varying histological grades. The prognostic value is evaluated on the 100 invasive carcinomas. Survival analysis implies that the epithelial percentage is of prognostic importance (Mantel-Cox 7.4, p = 0.0064). Multivariate analysis shows that the estimated fraction of epithelium is the strongest factor and that the FIGO stage has additional prognostic value (Mantel-Cox 12.5, p = 0.0004). It can be concluded that epithelial volume, as automatically estimated by image processing, seems useful in predicting the prognosis of patients with ovarian cancers.     

Comparison of commercially available cytokeratin antibodies in normal and neoplastic adult epithelial and non-epithelial tissues.

Five commercially available cytokeratin antibodies (lu-5, AE1/AE3, CAM 5.2, MFN116 and anti-cytokeratin 18) were used to stain a wide range of normal and neoplastic epithelial and non-epithelial tissues to assess their potential value in diagnostic histopathology. All five showed good specificity, with some cross-reactivity in smooth muscle cells. The wider reactivity of AE1/AE3, lu-5, and MFN 116, which includes cytokeratins 8,18 (Moll's catalogue) expressed in simple epithelia and their tumours, as well as cytokeratins expressed in complex stratified squamous epithelia, permits identification of a wider range of epithelial derived tumours. This wider spectrum of reactivity may allow these antibodies to be used in a diagnostic panel for the identification of poorly differentiated tumours.     

Immunohistochemical study of 158 lung carcinomas.

Lung carcinomas were studied immunohistochemically and the results were related to type of tissue sample (bronchoscopic biopsies, surgical specimens, autopsies). All cytokeratins (CAM 5.2, PKK-1, AE1/AE3) reacted with virtually all adenocarcinomas, most squamous, and 65% of the large cell carcinomas, while CAM 5.2 was most efficient with the small cell carcinomas. CEA stained 33% and 60% of the small and large cell carcinomas, respectively, most adenocarcinomas, and 84% of the squamous cell carcinomas, among which staining decreased with dedifferentiation and was often focal. EMA reacted with 90%, and NSE with 20% of all histological types. There was no staining for NF. All antibodies, except EMA, were more efficient with surgical specimens. Our study implies that the cytokeratins we used work better with surgical material, but are generally comparable to monospecific cytokeratin antibodies. Also, EMA is a reliable marker for epithelial differentiation with all types of tissue samples. Moreover, CEA negativity in several poorly differentiated lung carcinomas might have implications in the differential diagnosis against pleural mesothelioma.     

Benign lymph node inclusions mimicking metastatic carcinoma.

AIMS--To draw attention to non-neoplastic inclusions in axillary lymph nodes removed from women with primary breast cancer which may be mistaken for metastases. METHODS--Five examples of non-malignant inclusions were detected in axillary lymph nodes removed from women with mammary carcinoma. Immunohistochemical staining for CAM 5.2 and S100 markers, as well as morphological assessment were performed. RESULTS--Three of the five cases comprised naevus cells and two heterotopic epithelial elements. One of each was initially mistaken for metastatic carcinoma. Naevus cells in the capsule of axillary nodes are often arranged as lines of small, round cells and may readily be misinterpreted as metastatic lobular carcinoma. Heterotopic epithelial elements, in the form of tubules, can easily be mistaken for metastasis from an infiltrating ductal carcinoma. CONCLUSIONS--Awareness of the occasional occurrence of non-neoplastic nodal inclusions will help avoid misdiagnosis. If immunohistochemical markers for epithelial cells are used to screen lymph nodes for metastasis, such lesions could be detected more frequently. It is important to combine immunohistochemistry with morphological assessment and the use of a panel of antibodies in histological diagnosis.     

Mesonephric adenocarcinoma of the uterine cervix: a case report with immunohistochemical and ultrastructural studies

A very rare case of mesonephric adenocarcinoma with lobular mesonephric hyperplasia in the uterine cervix of a 46-year-old female is presented. The lesion was a 4cm, exophytic, almost circumferential, whitish yellow, friable mass in the uterine cervix. Microscopically, the tumor was composed predominantly of atypical round to polygonal cells arranged in a ductal, tubular, or papillary pattern. The tumor involved the entire cervix with varying depths of penetration. Lobular mesonephric hyperplasia was also observed in the lateral cervical wall and adjacent to the tumor. Immunohistochemically, the tumor was positive for CAM5.2, CK7, epithelial membrane antigen, calretinin, and chromogranin A, and was negative for vimentin, carcinoembryonic antigen, estrogen and progesterone receptors, and CD10. An ultrastructural analysis showed telolysomes, which were characteristic features of mesonephric epithelium. The patient was alive without disease at 4 months after surgery.     

An immunoperoxidase study of renal cell carcinomas: correlation with nuclear grade, cell type, and histologic pattern

We applied a panel of antibodies to formalin-fixed, paraffin-embedded sections of 55 renal cell carcinomas using a three-stage immunoperoxidase technique. The antibody panel included two anti-keratins, AE1 and CAM5.2, anti-epithelial membrane antigen (EMA), anti-vimentin, anti-S100 protein, and the anti-leukocyte marker PD7/26. Forty-eight of 55 renal cell carcinomas expressed keratins. CAM5.2 stained 46 tumors (84%) and AE1 stained 37 neoplasms (67%). AE1 reacted with two CAM5.2-negative tumors. EMA was expressed by 35 carcinomas (64%), including three of the CAM5.2-negative neoplasms. Therefore, using all three antibodies, 50 neoplasms (91%) expressed antigens of epithelial differentiation. Anti-EMA and AE1 were complementary to each other; the combination stained 46 of the carcinomas, comparable with CAM5.2 alone. Vimentin was expressed by 26 tumors (47%), and S100 was expressed by one. PD7/26 did not stain any of the cases. Vimentin expression correlated with nuclear grade; low nuclear grade neoplasms infrequently expressed vimentin, while the converse was true for high nuclear grade tumors. Keratin expression was related to tumor cell type and histologic pattern, as fewer neoplasms of clear cell type and with a solid pattern expressed keratins. In contrast, all papillary and eight of nine (89%) spindled carcinomas expressed keratins.     

Diagnosing tumours on routine surgical sections by immunohistochemistry: use of cytokeratin, common leucocyte, and other markers

Tumours of uncertain tissue of origin were investigated by immunohistochemistry on formalin fixed paraffin embedded sections. Two antibodies--PD7/26, an anti common leucocyte antigen, and CAM5.2, an anticytokeratin--recognised most lymphomas and carcinomas, respectively: 88% of these tumours were identified by the two antibodies alone. These antibodies permitted the separation of the cases into groups: positive with CAM5.2, positive with PD7/26, and a third comprising those negative with both. The negative group contained other tumours and a small number of carcinomas and lymphomas; many of the lymphomas were, apparently, of histiocytic origin. Comparison of CAM5.2 with other epithelial markers showed that it was the most effective. Some further classification of the tumours was carried out with a panel of organ and cell specific antibodies: mesotheliomas were recognised by their pattern of reactivity with epithelial markers. Overall, the tumour type was determined in 90% of cases. Immunohistochemistry performed as described can be a potent aid to the diagnostic histopathology of tumours.     

基于对称区域生长算法的超声医学图像分割方法

提出了一种基于对称区域生长算法的超声医学图像的分割方法。该方法分为三步。首先,通过采用自适应加权中值滤波抑制超声医学图像本身固有的Speckle噪声,然后从图像的第一行开始扫描整个图像,并应用生长准则进行区域的生长与合并,生长完成之后应用种子准则标定感兴趣区域,从而得到最后的分割结果。通过图像的分割实验确定了一套对于超声医学图像适用的生长和合并准则。对心脏B型超声医学图像分割的实验结果显示,该方法具有良好的性能。     

Heterogeneity of epithelial marker expression in routinely processed, poorly differentiated carcinomas.

The application of immunohistochemical markers against epithelial antigens has proved useful for studying tumor differentiation and in aiding tumor diagnosis. However, the reactivity of various epithelial markers with poorly differentiated carcinomas (the situation in which they are most often used) has not been well established. As a result, it is unclear how negative results should be interpreted and how often more than one antibody may be needed to document the epithelial nature of poorly differentiated neoplasms. We studied 98 poorly differentiated epithelial tumors with AE1, CAM 5.2, and EMA to assess the use of these markers in their diagnosis. Both CAM 5.2 and EMA provided support for epithelial differentiation in 71% (70/98) of the cases, while AE1 stained 50% (49/98) of the tumors; CAM 5.2 was the single most useful marker in the subset of poorly differentiated neuroendocrine carcinomas, staining 20 (77%) of 26 tumors. Use of these markers in pairs increased the recognition of epithelial differentiation (at least one marker showing positive staining) as follows: AE1/CAM 5.2, 80% (78/98); AE1/EMA, 87% (85/98); and CAM 5.2/EMA, 99% (97/98). Thirty carcinomas stained with all three markers, 34 with two markers, and in 34 cases only one antibody supported epithelial differentiation. Twelve (21%) of 58 tumors showed evidence of S100 reactivity. None of the 71 cases to which PD7 was applied showed staining This study indicates that poorly differentiated carcinomas are heterogeneous in their expression of antigens recognized by AE1, CAM 5.2, and EMA. Moreover, these results quantitate the probability of reactivity with poorly differentiated carcinomas for each marker and support the use of one or more antibodies in a "backup" panel when a negative result is obtained with a single antibody and the diagnosis of carcinoma is still suspected.     

Cytokeratin expression in adrenal phaeochromocytomas and extra-adrenal paragangliomas.

AIM: To examine whether adrenal phaeochromocytomas and extra-adrenal paragangliomas are immunoreactive for commercially available and routinely used cytokeratin antibodies. METHODS: 18 extra-adrenal paragangliomas and seven adrenal phaeochromocytomas were stained with CAM 5.2, AE1/3, and 34 beta E12 following microwave antigen retrieval of formalin fixed tissue. RESULTS: A single case from the cauda equina was positive for both CAM 5.2 and AE1/3. In addition, two other cases--an intravagal and an orbital paraganglioma--also showed strong immunopositivity with CAM 5.2 and AE1/3. All phaeochromocytomas were negative with all epithelial markers. CONCLUSIONS: Cauda equina paragangliomas are known to stain with cytokeratins; however, occasional paragangliomas from other sites may also be immunoreactive with cytokeratins. If the results of immunohistochemistry are not interpreted in the clinical and morphological context, the failure to recognise that extra-adrenal paragangliomas may on occasion react with anticytokeratin antibodies may lead to their being confused with metastatic carcinomas.     

Intermediate filament aggregates in mitoses of primary cutaneous neuroendocrine (Merkel cell) carcinoma

Primary cutaneous neuroendocrine carcinomas express different kinds of intermediate filaments and frequently in a 'paranuclear globular' pattern. We have observed the same pattern not only in interphase but also in mitotic cells, which are very frequent in these tumours. We report a quantitative and morphological study of eight primary cutaneous neuroendocrine carcinomas stained with different antibodies against cytokeratins (CAM 5.2 and anti-cytokeratin 20), neurofilaments (70 kDa and 200 kDa) and peripherin. We have found a predominance of CAM 5.2 expression in interphase cells and of neurofilament proteins in mitotic cells; 87.02% of the interphase cells were positive with CAM 5.2 whereas only 6.08% were positive for neurofilaments ( P <0.01); 35.41% of the mitotic cells were positive with CAM 5.2, whereas 50% were positive for neurofilaments ( P <0.01). A correlation between a globular pattern of intermediate filament proteins and prognosis has not been found. We describe for the first time the division of neoplastic cells with a globular pattern; the presence of intermediate filament proteins with a globular pattern in all mitotic stages; and the uneven distribution of this formation between the two daughter cells.     

Development of intrahepatic bile ducts in humans. Immunohistochemical study using monoclonal cytokeratin antibodies

Cytokeratins (CKs) are major structural proteins of intermediate filaments of epithelia. Recent availability of monoclonal antibodies (MoAbs) against various CK polypeptides has made it possible to study their development during cellular differentiation. We analyzed the expression of CKs in the human liver during development. Twenty-four liver specimens were tested by the avidin-biotin complex immunohistochemical method by using three MoAbs against different CK polypeptides (CAM 5.2 against CKs 50, 43, and 39 kd; AE1 against acidic CKs 56.5, 50/50', 48, and 40 kd; and 34 beta E12 against CKs 58, 56.5, and 56 kd). Liver parenchymal cells in fetuses as early as 4 weeks of gestational age reacted with MoAbs CAM 5.2 and AE1, but the expression of AE1-positive CK polypeptides in hepatocytes disappeared by 24 weeks of gestational age. Small cells, presumably ductal plate cells, in direct contact with mesenchyme around the portal vein and along the branches of portal veins, showed strong staining with MoAbs CAM 5.2, AE1, and 34 beta E12, identical to that of bile ducts. In neonates, children, and even in adults, residual MoAbs CAM 5.2-,Ae1, and 34 beta E 12-positive cells were present around the branches of portal veins. These findings suggested that the CK profile of liver parenchymal cells changes during their differentiation into hepatocytes, whereas that of ductal plate cells and bile ducts remains unaltered with respect to the polypeptides tested here. Some ductal plate cells may persist in neonates, children, and even in adults.     

Wide spectrum screening keratin as a marker of metaplastic spindle cell carcinoma of the breast: an immunohistochemical study of 24 patients

Wide spectrum screening keratin as a marker of metaplastic spindle cell carcinoma of the breast: an immunohistochemical study of 24 patients
Aims : Metaplastic spindle cell carcinomas may be difficult to distinguish histologically from other spindle cell lesions in the breast. Variable staining with cytokeratin immunomarkers has been reported for metaplastic carcinomas. We evaluated the diagnostic utility of anti-cytokeratin polyclonal antibody, wide spectrum screening keratin, to assess spindle cell breast lesions.
Methods and results : Twenty-four patients with spindle cell breast carcinoma and 31 patients with benign or malignant spindle cell tumours were studied using a panel of antibodies directed against multiple cytokeratins (AE1/AE3, CAM5.2, wide spectrum screening keratin), epithelial membrane antigen (EMA), and vimentin. Sites of origin for the 31 controls included breast, bone, and soft tissue. All but one (95.8%) metaplastic carcinomas stained positively with wide spectrum screening keratin. Only rare or focal immunoreactivity was observed with AE1/AE3 in four cases; however, sensitivity of AE1/AE3 was improved in 13 cases using steam EDTA as an antigen retrieval technique. Three cases were immunoreactive with CAM5.2 and eight cases were immunoreactive with EMA. All control cases lacked immunoreactivity with the cytokeratin panel and EMA. The spindle cells in the metaplastic breast tumours (88%) and in the controls (97%) stained with vimentin.
Conclusions : Wide spectrum screening keratin may be the most useful and convenient antibody in differentiating metaplastic spindle cell carcinoma from other spindle cell lesions in the breast.     

Diagnostic Value of Silver-Stained Interphasic Nucleolar Organizer Regions in Breast Tumors

Seventy-six specimens of normal breast tissue and benign and malignant breast lesions were studied to assess the mean area occupied by silver-stained proteins of the nucleolar organizer regions (MNORA) of the nucleolus. The assessment was performed with a computer-assisted image analyzer. The results indicate that only 30% of malignant lesions have a MNORA value greater than that of normal breast tissue or benign lesions. On the other hand, MNORA values of ductal carcinoma in situ were significantly greater than those of epitheliosis (papillomatosis). MNORA values were also significantly different in grade I and grade III invasive ductal carcinomas, the latter exhibiting the highest MNORA values of all the cases observed. Evaluation of MNORA values may therefore help in differentiating benign epithelial proliferations from ductal carcinomas in situ. Furthermore, because there is evidence that MNORA values are indicative of the cell duplication rate, MNORA values may ultimately be considered an objective prognostic parameter in addition to grading for invasive ductal carcinomas.     

The histogenesis of mammary and extramammary Paget''s disease

The histogenesis of mammary and extramammary Paget's disease has been studied by immunohistochemical staining of paraffin-embedded tissue using a panel of epithelial cell markers, which react with secretory or ductal epithelium, but not stratified epithelium. These markers included a monoclonal antibody E29 to epithelial membrane antigen EMA, the cytokeratin marker CAM 5.2 and three new monoclonal antibodies raised to human milk fat globule membrane (LICR-LON-TW19 and H.10.A) and a human bladder cell cancer line (3.77). The findings demonstrate that both mammary and extramammary Paget's disease are of epithelial cell origin and share antigens expressed by simple epithelia. Some antigens, such as EMA and low molecular weight cytokeratins are consistently present in both diseases, whereas other antigens, identified by H.10.A and TW19 are found more frequently in cases of extramammary Paget's disease. This panel of monoclonal antibodies also proved useful in distinguishing Paget's disease from pagetoid melanoma and clear cell Bowen's disease.     

Pleomorphic carcinoma of the breast with expression of macrophage markers: report of two cases

Pleomorphic ductal carcinoma of the breast is a rare variant included in the morphological group of infiltrating ductal carcinoma. The pleomorphic carcinoma is composed predominantly of epithelial and multinucleated tumor giant cells. We report here two cases presenting a lesion composed microscopically of a proliferation of large pleomorphic cells with a predominance of multinucleated giant cells. These lesions were negative for estrogen receptor, progesterone receptor and Her2-neu (triple-negative phenotype). Basal markers (cytokeratin 5/6, cytokeratin 17 and epidermal growth factor receptor [EGFR]) were present, accompanied by the presence of histiocyte marker CD163 in most neoplastic giant cells. High-grade pleomorphic breast carcinomas with the triple-negative phenotype and expression of basal markers might be included in the basal subtype. This is the first report about the co-expression of macrophage marker CD163, with tumor (P53) or epithelial markers (CAM5.2), as indicated by double immunohistochemistry in pleomorphic ductal carcinoma of the breast.     

Peritoneal cystic mesothelioma: an electron microscopic and immunohistochemical study of two male patients.

The clinical, pathological, and ultrastructural features of two cases of peritoneal cystic mesothelioma occurring in men were studied. The results of immunohistochemical staining for CAM 5.2, epithelial membrane antigen, carcinoembryonic antigen, and Factor VIII related antigen are reported for the first time and compared with the staining results of two peritoneal cystic lymphangiomas. Although resembling cystic lymphangioma by light microscopy, cystic mesothelioma may have a greater tendency for local recurrence. Staining for CAM 5.2 or epithelial membrane antigen may facilitate the differentiation of these two entities.     

基于卷积神经网络和图像显著性的心脏CT图像分割

目的心脏医学影像中,感兴趣部分的提取与分割是诊断心脏病变部位的关键。由于心脏舒张、收缩以及血液的流动,心脏CT图像易出现弱边界、伪影,传统分割算法易产生过度分割的情况。为此,提出一种基于卷积神经网络和图像显著性的心脏CT图像分割方法。方法采用卷积神经网络对目标区域进行定位,滤除肋骨、肌肉等造影对比不明显部分,截取出感兴趣区域,结合感兴趣区域的对比度计算并提高感兴趣区域的心脏组织的显著值。通过获得的显著值图像截取心脏图像,并与区域生长算法的分割结果进行对比。最后使用泰州人民医院11例患者的影像数据对算法模型进行训练和测试,随机选择9例用于训练,剩余2例用于测试。结果所提算法模型在心底、心中、心尖3个心脏分段的分割正确率分别达到了92.79%、92.79%、94.11%,均优于基于区域生长的分割方法。结论基于卷积神经网络和图像显著性的分割方法能够准确获取心脏的外围轮廓,轮廓边缘更加平滑,完全能够满足CT图像序列的心脏全自动分割任务需求,分割后的图像更有利于医生对患者心脏健康状况和病变部位的观察。     

Sarcomatoid carcinoma of the breast: An immunohistochemical study of six cases

Summary Six cases of sarcomatoid carcinoma of the breast (SCB) were studied with a panel of anti-bodies directed against epithelial and sarcomatoid components. The monoclonal antibodies (MoAb) AE-1/3, CAM 5.2, and CEA were used to detect epithelial differentiation; polyclonal antibodies against S-100 protein and MoAb against the intermediate filaments desmin and vimentin were used to detect mesenchymal differentiation in the sarcomatoid component. Six cases of invasive duct carcinoma (IDC) and two cases of cystosarcoma phyllodes (CP) were compared to SCB using the same panel of antibodies. In all three groups studied, the epithelial component in the majority of cases stained with anti-cytokeratin antibodies. S-100 protein antibodies stained the epithelial and sarcomatoid components in four cases of SCB; vimentin MoAb stained the epithelium in two cases and the sarcomatoid component in four cases of SCB, while MoAb CEA failed to stain any component of SCB. In contrast, the epithelium in five of six cases of IDC stained with CEA MoAb and only one of six stained for S-100 protein. Possible reasons for the discrepant immunohistochemical staining patterns among SCB, IDC and CP are discussed, in addition to the limitations and pitfalls of immunohistochemistry in diagnostic surgical pathology.