The relationship between growth fractions and oestrogen receptors in human breast carcinoma, as determined by immunohistochemical staining

Measurements of the growth fraction (GF) and oestrogen receptor (OR) status were performed by immunohistochemical staining, using the monoclonal antibodies Ki-67 and anti-OR, in consecutive cryostat sections of biopsies from 74 women with primary breast cancer, and 12 women with benign breast disease. A significant inverse correlation was identified between the GF and OR status (r = -0.452, P less than 0.0001), but a group of OR-positive tumours with high GFs was also observed. There was a strong positive correlation between immunocytochemically determined OR percentages and values obtained with a cytosolic OR biochemical assay for the 70 cases in which both techniques were performed (r = 0.801, P less than 0.0001). A significant relationship was also revealed between the histological grade of infiltrating ductal carcinomas and their associated GFs (r = 0.622, P less than 0.0001). No correlation was found between the GF and the number of positive axillary nodes. Our findings suggest that GF and OR status are useful independent prognostic parameters in breast cancer, but the former probably has an overriding influence. Both parameters may be useful guides for selection of appropriate endocrine and/or adjuvant chemotherapy.     

Immunohistochemical analysis of p53 protein expression in benign and malignant skin tumors using a panel of anti-p53 antibodies.

The expression of p53 in a variety of benign and malignant skin lesions has been first assessed in frozen sections and then compared with the results obtained in corresponding paraffin-embedded sections using various immunohistochemical staining methods with a panel of anti-p53 antibodies. Of the 48 benign and malignant skin lesions studied, 46(96%) had corresponding paraffin sections and immunohistochemical results obtained with DO7 on frozen and paraffin sections were concordant in 97%, qualitatively. Using streptavidin-biotin complex method, p53 was identified in 33% of dysplastic squamous lesions, 50% of squamous cell carcinomas (SCCs) and 36% of basal cell carcinomas (BCCs) on frozen section, whereas 25% of dysplastic squamous lesions, 40% of SCCs, and 32% of BCCs showed p53 positivity on paraffin-embedded sections. In frozen sections, the same regions of each specimen exhibited similar topographic patterns of positive immunoreactivity with both monoclonal antibodies, PAb 1801 and DO7. In contrast, immunohistochemical staining with polyclonal antibody, CM-1, gave poor morphologic resolution, although effective in paraffin-embedded sections.     

Diagnostic usefulness of dipeptidyl aminopeptidase IV monoclonal antibody in paraffin-embedded thyroid follicular tumours.

Monoclonal antibodies to dipeptidyl aminopeptidase IV (DAP IV, EC 3.4.14.5) were raised and selectively applied to paraffin-embedded sections of thyroid carcinoma. Five monoclonal antibodies were found to stain paraffin sections of thyroid carcinomas. Using one of these antibodies (44-4), we studied retrospectively aberrant expression of DAP IV in thyroid carcinoma to determine whether immunohistochemical staining with DAP IV antibody is useful in pathological diagnosis. In almost all cases of thyroid follicular and papillary carcinoma, tumour cells were positive (99.0 per cent) with DAP IV, whereas the cases of follicular adenoma showed a low incidence (27.1 per cent) of positive staining. Follicular adenoma with incomplete capsular invasion had a higher positive incidence (50 per cent) than follicular adenoma without incomplete capsular invasion (9.6 per cent). In positive staining cases previously diagnosed as benign tumours, 11 benign cases reacting positively with DAP IV were rediagnosed as carcinoma after re-examination of more thyroid paraffin block sections or serial sections. These findings suggest that DAP IV monoclonal antibody is very useful in distinguishing thyroid follicular carcinoma from follicular adenoma.     

Automated immunohistochemical estrogen receptor in fixed embedded breast carcinomas. Comparison with manual immunohistochemistry on frozen tissues

Estrogen receptor (ER) status of breast carcinomas determines prognosis and treatment. Biochemical ER assays are expensive and time-consuming and require fresh tumor. Immunohistochemical ER was assessed in 68 breast carcinomas, by an automated method using routinely processed formalin-fixed paraffin-embedded tissues, and manually with the use of snap-frozen tissues with a monoclonal anti-ER and peroxidase-antiperoxidase technique. The paraffin sections were digested with DNase to enhance development of signal. Positive nuclear ER was obtained in 9 (13%) fixed tissues and 36 (53%) frozen tissues. The sensitivity, specificity, and predictive value of a positive test result, as compared with the biochemical assay, were 25%, 100%, and 100% for the paraffin section technique, and 89%, 88%, and 89% for the frozen sections. Although it is specific, lack of sensitivity, resulting from loss of ER with fixation and room temperature handling, renders this immunohistochemical technique unacceptable on fixed tissues. However, ER immunostain on frozen tissue is an acceptable alternative to biochemical assay.     

Demonstration of estrogen receptor by immunohistochemical staining in paraffin sections of breast carcinoma

Paraffin embedded sections of 64 breast carcinomas were stained immunohistochemically using a commercially available monoclonal antibody to estrogen receptor. To improve the sensitivity of the staining, the authors used a Pronase enzyme pretreatment, biotinylated antibody to rat IgG as secondary antibody, streptavidin-alkaline phosphatase as tertiary reagent and fast red as chromogen. When compared to the results of estrogen receptor enzyme immunoassay, this method yielded an 85.9% concordance rate, 86.2% specificity and 85.7% sensitivity. When compared to estrogen receptor immunocytochemistry(ER-ICA) in frozen section and considering the inherent advantages of immunohistochemical staining over biochemical assay, the major advantages of this method are good morphology, suitability for retrospective study and reduced cost of staining due to dilution of expensive primary antibody. Thus, this method offers an alternative to ER assay using fresh tissue and should provide additional valuable information about estrogen receptor.     

c-erbB-2 oncogene product staining in gastric adenocarcinoma. An immunohistochemical study

The c-erbB-2 oncogene has been shown to be amplified in a variety of human adenocarcinomas. Antibodies to the protein product, p185, have been used for immunostaining of paraffin-embedded material, and have demonstrated that high levels of c-erbB-2 protein expression correlate with gene amplification under certain conditions. In studies by others, amplification has been demonstrated in 40 per cent of tubular type adenocarcinomas of the stomach, and an immunohistochemical study on frozen tissue has demonstrated staining in 3 out of 10 cases. Our study, using paraffin-embedded material, demonstrates staining in 19 per cent of 126 cases using a polyclonal antibody. Of the positive cases, 75 per cent were tubular or papillary type (P less than 0.025), and prominent staining was restricted to this group. Three cases showed well-defined positive areas in keeping with clonal expression of p185. No specific staining of normal or dysplastic epithelium adjacent to the carcinomas was found.     

Assessment of oestrogen receptor content of breast carcinoma by immunohistochemical techniques on fixed and frozen tissue and by biochemical ligand binding assay.

The oestrogen receptor content of 61 breast carcinomas was assessed by biochemical ligand binding assay and three immunohistochemical techniques--a frozen section method (Abbott ER-ICA) and on paraffin wax sections after fixation by two methods. The two fixatives used were Carson's buffered formalin and methacarn, and a DNAse pretreatment of sections was used. Overall agreement for the immunohistochemical methods with the ligand binding technique were 95%, 85%, and 86% for the frozen, formalin, and methacarn methods, respectively. A semiquantitative staining score was performed and all three methods gave significant correlations of staining scores with biochemical ligand binding values. The frozen section method was best (r = 0.88) with the fixed tissue methods yielding poorer correlation coefficients. Several factors affected staining, including the nature of the fixative and variable activity of DNAse. It is concluded that immunohistochemical assessment of oestrogen receptor content on fixed tissue provides acceptable qualitative information but that standardisation of protocols for tissue processing will be necessary for optimal utility and especially for quantitative assessments.     

DEVELOPMENT OF AN OPTIMAL PROTOCOL FOR ANTIGEN RETRIEVAL: A ‘TEST BATTERY’ APPROACH EXEMPLIFIED WITH REFERENCE TO THE STAINING OF RETINOBLASTOMA PROTEIN (pRB) IN FORMALIN-FIXED PARAFFIN SECTIONS

The retinoblastoma (RB) gene, which encodes the nuclear RB protein (pRB), is believed to be involved in cell cycle control and cell differentiation. Studies have demonstrated that loss of RB function may play a role in tumour formation and progression of a variety of human tumours, such as bladder, lung, breast, and prostate cancers. The immunohistochemical detection of pRB expression in formalin–paraffin sections of human cancer has potential advantages of convenience, economy, and compatibility with routine surgical pathology practice. In practice, however, results using pRB antibodies on routinely processed, paraffin-embedded tissue have been inconsistent. In this study, the antigen retrieval (AR) method has been applied to the immunohistochemical detection of pRB in paraffin-embedded tissues and a ‘test battery’ approach has been developed to identify the principal variables that result in the optimal AR protocol. This approach includes the use of buffered solutions at pH 1, 6, and 10 with three different heating conditions (temperatures 120°C, 100°C, and 90°C). In the example described here with antibody RB-WL-1, the low pH solution with the microwave heating at 100°C proved most effective. Both fresh and routinely processed formalin–paraffin tissues of normal and bladder carcinoma were used for a comparison of the pRB immunostaining. The AR method was evaluated by comparing the immunohistochemical staining result on routinely processed formalin–paraffin sections with frozen sections of the same tumour. A consistent intensity of immunohistochemical staining for pRB was achieved using the identified optimal AR protocol on formalin–paraffin sections. All slides showed positive staining of pRB in normal mesenchymal and epithelial tissues. The pattern of pRB localization and intensity of staining was similar to that obtained in frozen sections, though the intensity obtained by AR treatment on paraffin sections was slightly to moderately stronger than that obtained in frozen sections. Once the protocol was identified, it was tested using routinely processed paraffin tissue sections of 245 cases of bladder carcinoma, with consistent pRB immunostaining results. The protocol described is simple to perform and gives reproducible results for evaluation of pRB expression by immunohistochemistry.     

Immunohistochemical demonstration of keratin 7 in routinely fixed paraffin-embedded human tissues.

The immunoreactivity of OV-TL 12/30, a monoclonal anti-keratin 7 antibody (Mab), was investigated on frozen as well as paraffin-embedded human tissues. Its reactivity patterns were compared with another well-characterized monoclonal antibody to keratin 7 (RCK 105), and with broadly cross-reacting monoclonal (OV-TL 12/5) as well as polyclonal (pKer) keratin antisera. In frozen sections of normal and malignant human tissues both keratin 7 Mabs gave similar staining patterns. The immunoreactivity for OV-TL 12/30 and the polyclonal antibody (pKer) in tissue sections fixed in 4 per cent formalin or Bouin solution, was completely restored when pretreated with 0.1 per cent pronase, 0.1 per cent trypsin in phosphate-buffered saline (PBS) or with 0.5 per cent pepsin in 0.01 N HCl. Except for loss of immunoreactivity on human normal stomach surface epithelium and glandular mucous cells, Mab OV-TL 12/30 reacted strongly positive with essentially all those formalin- or Bouin-fixed paraffin-embedded tissues that had been shown to stain in non-fixed, frozen sections. In addition to the good correlation in human tissues, a complete correlation between the reactivity on frozen and paraffin-embedded human carcinomas (n = 86) was found as well. While both RCK 105 (anti-keratin 7) and OV-TL 12/5 (anti-keratin 5, 7, 14, 19) did not stain on paraffin-embedded sections, the polyclonal control antiserum (pKer) lost immunoreactivity in some cell types (e.g. mucous cells in compound glands, hepatocytes, pancreatic acinar cells, and proximal and distal convoluted tubules of the kidney). Our study shows that the keratin 7 Mab OV-TL 12/30 is an excellent marker for tumour histopathology since it is reactive in paraffin-embedded formalin-fixed human tissues.     

Methodology of immunohistological detection of oestrogen receptor in human breast carcinoma in formalin-fixed, paraffin-embedded tissue: a comparison with frozen section methodology

We describe a method of immunohistochemically assessing oestrogen receptor status on routinely processed formalinfixed tissue, using a commercially available monoclonal antibody (Abbott H222). with pronase predigestion of tissue sections and overnight antibody incubation. The staining was assessed using the H score system. A series of 94 cases of breast cancer were analysed and the results were compared with assessment by oestrogen receptor immunocytochemical assay performed on frozen section. Direct comparison of the paired sets of H scores obtained with frozen tissue and formalin-fixed tissue showed a highly significant correlation of 0.8 ( P < 0.001) between the two methods of oestrogen receptor assessment. Chi-squared analysis using H score cut off points of 50 and 100 also showed a similar significant association ( P < 0.001). We conclude that this oestrogen receptor method, applicable to formalin-fixed, paraffin-embedded tissue, gives accurate results on routinely fixed tissue and could be used as an alternative to other methods.     

Automated immunohistochemical estrogen receptor in fixed embedded breast carcinomas.

The authors immunohistochemically assessed the presence of estrogen receptor (ER) in formalin-fixed, paraffin-embedded tissue sections of 68 breast carcinomas by an automated method using Pronase (CalBiochem, La Jolla, CA) predigestion and alkaline phosphatase detection (Method 1). These results were compared with those obtained by an automated peroxidase-antiperoxidase method with DNAse pretreatment of fixed embedded sections (Method 2), with ER immunostain on frozen sections (Method 3), and with biochemical results (dextran-coated charcoal cytosolic [DCC] assay). Compared with the DCC assay, Methods 1, 2, and 3 gave sensitivities of 54%, 25%, and 89%, respectively. The sensitivity for Method 1 was increased to 74% in those cases with DCC results showing greater than 50 fmol/mg protein. These findings indicate that ER immunohistochemical studies on formalin-fixed paraffin-embedded tissues (as assayed by Method 1) provide useful clinical information when the results are positive. A negative result, especially if surrounding normal elements are not positive, may indicate no receptors, receptor levels less than 50 fmol/mg protein, or improper tissue preservation. In the absence of fresh tissue for ER assay by DCC assay or of frozen sections for immunostaining, and with an understanding of its limitations, this method may be useful.     

Cytokeratin and neurofilament in lung carcinomas.

Three monoclonal antibodies, one directed against cytokeratin (clone 80) and two directed against neurofilament (clones 2F11 and 3G6), were used in the study of a series of 77 lung carcinomas by immunohistochemical staining. The anti-cytokeratin antibody, a very broadly reacting antibody directed against an antigenic determinant common to a great number of cytokeratins, was applicable on frozen sections. The two anti-neurofilament antibodies, directed against the 70 kD protein (clone 2F11) and the 160 kD and 200 kD proteins (clone 3G6) of neurofilament, were applicable on both frozen sections and paraffin sections. The staining results on the lung carcinomas indicate that all types of tumors studied, including small-cell anaplastic carcinoma, are markedly positive for cytokeratin. Frozen sections of five and formalin-fixed and paraffin-embedded sections of six other small-cell anaplastic carcinomas were negative with both anti-neurofilament monoclonal antibodies. One poorly differentiated squamous cell carcinoma positive with anti-neurofilament clone 2F11 but negative with clone 3G6. This distribution of cytoskeletal proteins demonstrates the epithelial differentiation of all types of lung carcinomas. Neuroendocrine differentiation of lung carcinomas as found in the small-cell anaplastic types does not result in expression of neurofilament proteins.     

Immunohistochemistry of estrogen receptor protein in paraffin sections. Effects of enzymatic pretreatment and cobalt chloride intensification

Monoclonal antibodies provide important tools for the demonstration of estrogen receptors (ERs) in cases of breast cancer. This study reports an improved immunohistochemical method for the demonstration of ER in formalin-fixed paraffin-embedded tissue samples using the Abbott monoclonal antibody to ER protein. Tissue sections were pretreated briefly with trypsin, followed by DNase before the performance of the immunohistochemical reaction and cobalt chloride was used to intensify the color of the diaminobenzidine reaction product. In 20 cases, the results in paraffin sections were compared with biochemical assays with the dextran-coated charcoal technique or with immunohistochemistry performed on frozen sections. There was an excellent correlation between the results obtained with all three methods. The introduction of cobalt chloride into the chromogen solution significantly increased the sensitivity of this approach as compared with the use of diaminobenzidine alone.     

Variability of immunohistochemical reactivity on stored paraffin slides.

AIMS: To investigate the effects of slide storage on immunohistochemical staining, since recent reports have indicated that storage of unstained paraffin slides for up to 12 weeks may lead to false negative immunostaining of tumour markers. METHODS: 11 antibodies (anti-cytokeratin, epithelial membrane antigen (EMA), vimentin, smooth muscle actin, PS100, chromogranin, CD45, CD20, CD3, CD30, and oestrogen receptor (OR) were tested on unstained paraffin slides of breast carcinomas, lymphomas, and neuroendocrine tumours that had been stored for three to 10 years. All the paraffin blocks were recut less than one week before immunostaining. Immunostainings of years old slides were compared with those of recent slides in at least five cases for each antibody. For three antibodies (antichromogranin, anti-CD3, and anti-OR) we also tested one year old and three months old slides. RESULTS: Intensity of staining on years old slides was strikingly reduced for chromogranin and CD3 in several cases and was slightly stronger for vimentin. In some cases a significant decrease of OR positivity was observed after three months storage, and a complete loss of OR immunostaining after 12 months. No significant difference was noted with the other antibodies. CONCLUSIONS: Immunohistochemical detection of some antigens located either in the nucleus, in the cytoplasm, or on the cytoplasmic membrane could be impaired by storage of paraffin slides as short a time as three months. One should be cautious of doing retrospective immunohistochemical studies on stored unstained slides.     

Immunocytochemical staining of c-erbB-2 oncogene in fine-needle aspirates of breast carcinoma: A comparison with tissue sections and other breast cancer prognostic factors

The role of fine-needle aspiration (FNA) in breast cancer management is expanding to include prognostic in addition to diagnostic studies. This study was undertaken to compare the immunohistochemical staining of c-erbB-2, a breast cancer prognostic factor, on cytologic FNA smears of breast cancer with that on corresponding formalin-fixed, paraffin-embedded tissue sections of the same tumor and to correlate the c-erbB-2 expression with other known breast cancer prognostic factors. FNA smears (destained, alcoholfixed, Pap-stained direct smears) and corresponding tissue sections (formalin-fixed, paraffin-embedded, unstained sections) from 36 primary breast carcinomas were stained immunohistochemically with c-erbB-2 antibody using an avidin-biotin procedure. Ten tumors (28%) showed strong positive staining for c-erbB-2 on the FNA smear and, of those ten, seven were positive on the corresponding tissue section and three were negative. In several of the cytology and tissue positive cases, staining on the tissue section was significantly weaker than on the FNA smear. Two cases involved treatment with preoperative chemotherapy; in one of those cases, tissue staining was weaker than cytologic staining, and, in the other case, both were negative. Correlation of c-erbB-2 staining with other prognostic factors showed an association between positive c-erbB-2 staining and high nuclear grade. Our results indicate that immunohistochemical staining for c-erbB-2 can be performed on destained FNA smears of breast carcinomas and may be more sensitive on the cytologic specimens than on formalin-fixed, paraffin-embedded tissue sections. Diagn Cytopathol 1994;11:250–254. © 1994 Wiley-Liss, Inc.     

CD69 expression correlates with expression of other markers of Th1 T cell differentiation in peripheral T cell lymphomas

CD69, a marker of early T cell activation, is associated with Th1 T cell differentiation. Previously we found that peripheral T cell lymphomas could be subdivided based on the expression of markers of Th1 versus Th2 differentiation, including CXCR3, CD134/OX40, CCR4, and CD30. Here we report immunohistochemical staining for CD69 in frozen and paraffin sections of peripheral T cell lymphomas that exhibit immunoreactivity for markers of Th1 or Th2 differentiation. CD69 expression correlated with immunoreactivity for other Th1 differentiation markers in 18 of 19 frozen specimens of peripheral T cell lymphomas (P = 0.0005). In 10 of these cases in which paraffin-embedded tissue was available for study, CD69 immunohistochemical staining of paraffin sections correlated with frozen section expression. CD69 immunostaining was performed on paraffin sections from 53 additional cases of peripheral T cell lymphoma and correlated with immunoreactivity for other Th1 differentiation markers (P < 0.0001) and was associated with specific subtypes of peripheral T cell lymphoma, including angioimmunoblastic lymphoma, Lennert's lymphoma, and mycosis fungoides/Sezary syndrome, previously noted to express Th1 differentiation-associated markers. Anaplastic large cell lymphoma, both systemic and cutaneous, which typically exhibits immunoreactivity for markers of Th2 expression, was negative for CD69 immunostaining in 22 of 24 cases. CD69 immunostaining results support previous findings that a subset of T cell lymphomas exhibits immunophenotypic features of either Th1 or Th2 T cell differentiation. In addition, CD69 is a useful immunohistochemical marker for specific T cell lymphomas in frozen and paraffin-embedded tissue.     

Detection of surface antigen 17-1A in breast and colorectal cancer

Substantial progress has been made in detecting cell surface or intracytoplasmatic antigens to identify spread tumor cells with monoclonal antibodies (MAbs). The 17-1A antigen is already used as a target for specific immunotherapy in colorectal cancer. The purpose of this study was to compare the expression of 17-1A antigen in colorectal tumors versus breast cancers. MAb against the epithelial-specific antigen (ESA) and a routine staining technique were used to detect the 17-1A antigen in 100 cases of colorectal and 111 cases of breast cancer. The antigen expression of each tumor entity was examined by light microscopy on paraffin sections. Thirty six of the formalin-fixed paraffin sections of breast cancer were compared with their corresponding frozen sections. Evaluation was realized by a histological score (grade 0-9) considering the distribution and the staining intensity. We found an antigen expression of 17-1A in colorectal cancer quantified at 7.1+/-1.8 and at 4.5+/-2.5 for breast cancer in our score. Comparing paraffin sections and frozen sections in the 36 cases of breast cancer, the score was 5.5+/-2.3 in the paraffin and 8.1+/-1.9 in the frozen section group. Our results confirmed the high expression of 17-1A cases of in colorectal carcinoma. Furthermore, 17-1A is expressed in the majority of breast carcinomas, revealing a high difference between paraffin and frozen sections. As a result, a specific immunotherapy with MAbs against 17-1A antigen in minimal residual stages of breast cancer might be considered.     

Immunocytochemical staining of progesterone receptor in paraffin sections of human breast cancers.

Monoclonal antibodies (mAB) against progesterone receptor (PR) and the peroxidase-antiperoxidase (PAP) method to visualize PR in paraffin sections from 68 human breast cancers were used. Ten mAB, which recognize human PR on frozen sections, were tested. Six could detect PR in paraffin sections, with Li 417, LET 456, and LET 126 giving the best results. LET 126 antibody was used for most further studies. The effects of fixation with picric acid-formaldehyde (PAF), buffered-formalin, or with Bouin solution were investigated; all fixation methods allowed PR immunolabeling, although PAF or buffered formalin usually gave the best results. Positive staining was seen in the nucleus of carcinoma cells. Variations in intensity and extent of immunoreactivity were observed in all sections and among different regions of the same specimen. These were probably related to the heterogeneity of the tumor cell population. Results were compared with the PR content in the respective tumor tissues, determined by steroid-binding assay, and with immunocytochemistry on frozen sections. It was shown that there were correlations between the immunocytochemical staining (positive or negative) and the steroid binding assay (80%) and between the immunocytochemical staining on paraffin sections and on frozen sections (78%). Weaker intensity and fewer number of PR-positive cells were found for paraffin-embedded tumors. Estrogen receptors were also detected on adjacent sections from the same paraffin-embedded tissues by use of monoclonal anti-ER antibodies (ERICA-kit[Abbott Laboratories, Chicago, IL]) and DNase pretreatment. In conclusion, this immunocytochemical method for detection of PR and ER on paraffin sections offers an alternative to frozen tissue. It allows histologic and immunocytochemical studies on the same sample and retrospective studies on stored tissue blocks.     

Vimentin expression in benign and malignant breast epithelium.

AIMS--To determine vimentin expression in epithelial cells in benign breast disease and malignant breast tumours; to assess the value of vimentin expression as a prognostic indicator in breast carcinoma. METHODS--Frozen and formalin fixed, paraffin wax embedded sections from 78 carcinomas, three phyllodes tumours, 19 fibroadenomas and 19 cases of fibrocystic disease were examined with a monoclonal antibody from the V9 clone. A correlation between vimentin expression and known prognostic indicators was sought in ductal carcinomas. The intracellular localisation of vimentin was examined in benign and malignant lesions. RESULTS--Vimentin expression was identified on frozen section in the cells of ductal (53%), lobular (86%), and mucinous (33%) carcinomas and in the luminal epithelium of fibroadenomas (68%), cases of fibrocystic disease (47%), and a malignant phyllodes tumour. Formalin fixation reduced the percentage of carcinomas and cases of benign disease in which vimentin was detected. This reduction was more pronounced in fibroadenoma and fibrocystic disease than in ductal carcinoma. Associations were identified between vimentin expression as detected on frozen section and tumour grade, size, number of lymph nodes affected, oestrogen receptor content and growth fraction. Only the association with grade was significant (p = 0.045). There was no significant correlation between any of these prognostic variables and vimentin expression on paraffin wax sections. There was no difference in the intracellular localisation of vimentin staining between benign and malignant lesions, or between low and high grade ductal carcinomas. CONCLUSION--There is some loss of vimentin immunoreactivity after formalin fixation. Vimentin expression does not assist in differentiating between benign and malignant breast disease, but is correlated with tumour grade in ductal carcinoma.