- Email: [email protected]
Reporting breast biopsies
Current Diagnostic Pathology (2000) 6, 140–145 © 2000 Harcourt Publishers Ltd doi:10.1054/cdip.2000.0028, available online at http://www.idealibrary.com on
REVIEW
Reporting breast biopsies S. Shousha Imperial College School of Medicine, Department of Histopathology, Charing Cross Hospital, Fulham Palace Road, London, W6 8RF, UK
KEYWORDS breast, benign disease, breast carcinoma, oestrogen receptors
Summary The role of the surgical pathologist in dealing with breast biopsies has changed enormously during the last few years. The introduction of mammographic screening, conservative breast surgery, core biopsies and, more recently, sentinel lymph node techniques, have changed the type of material received by the pathologist for diagnosis, and the type of information required from him/her. Requests for frozen sections have markedly diminished, but new challenges have arisen, particularly in respect of diagnosing borderline lesions which are being increasingly detected by mammography. This article presents a detailed account of the handling and reporting of breast biopsies, based partly on National Guidelines and the practice at Charing Cross Hospital, London. © 2000 Harcourt Publishers Ltd
DEALING WITH THE GROSS SPECIMEN Good fixation is essential for proper diagnosis. In our department we prefer to receive all specimens, except core biopsies, fresh in a plastic bag, immediately after surgical excision whenever this is feasible and in the absence of any suspected infections. The specimen is registered, given a laboratory number, and the pathologist in charge is then called to deal with it. For biopsies coming from other hospitals we ask the surgeons to immerse it in an adequate amount of formalin and send it to us as soon as possible where it is dealt with promptly. Core biopsies are immersed in formalin straight after being removed from the patient, or, in case of stereotactic biopsies, after X-raying the specimen by the radiologist to ensure the presence of calcification.
SPECIMEN TYPES Wire localisation biopsies The surgeons usually send these specimens first to the X-ray department where the whole specimen is X-rayed to ensure that the radiologically abnormal area has been removed. The surgeon is notified of the result, and if the abnormal area has not been removed, a further excision is carried out and dealt with in the same way. We receive these biopsies with their specimen X-rays, and with the wires still attached to them. The biopsy is weighed, measured, described, and the wire is gently
Correspondence to: SS. Tel: +44 20 8846 7144; Fax: +44 20 8846 1364; E-mail: [email protected].
removed. The outer surface of the specimen is painted with India Ink, and a few drops of Bowen’s fixative are added to help adherence of the ink to the specimen surface. Two minutes are usually enough for the ink to dry, after which extra ink is removed by blotting paper. The specimen is then sliced into approximately 5 mm thick slices. The slices are X-rayed in a Faxitron using polaroid X-ray films. This helps to identify the slice(s) containing the calcification or other abnormalities that correspond to those present in the specimen X-ray. All slices containing radiological abnormalities and their immediately adjacent ones are processed together with any other slices that may appear to be abnormal by naked eye. This would include the site of any previous recent needle aspiration or core biopsy which is usually indicated by an area of haemorrhage. The gross appearance of any abnormality is described. The minimum number of slices processed from each specimen is six. The total number will depend on the size of the biopsy and the extent of abnormalities in it. Each slice is processed in a separate cassette marked with the case number and an additional serial number which we also write on the corresponding slice X-ray. The cassettes containing the slices are kept in formalin overnight and are processed the following day, with the stained sections usually available for examination within 48 h. Any unprocesed slices are kept in formalin for 4 weeks, in case any extra tissue is needed for examination.
Wide local excision biopsies The specimen is weighed, measured, described and covered with ink as above. If there is a palpable lesion, the specimen is sliced along the maximum dimension of that lesion, in 5 mm thick slices. The lesion is described
REPORTING BREAST BIOPSIES
and its maximum dimension measured, as well as the distance between the lesion and the nearest excision margin. Samples of the biopsy are then taken for processing. We tend to process all slices containing obvious tumour tissue, as well as the immediately adjacent slices, including the nearest tumour margin, as long as this is feasible. We then sample the other biopsy margins, and process any other slices with gross abnormality. If the biopsy contains no palpable abnormality, the specimen is sliced along its narrowest dimension. All grossly abnormal slices are processed, as well as randomly sampled ones.
Cavity wall biopsies Some surgeons prefer to remove extra strips of breast tissue from the wall of the cavity left in the breast after doing a wide local excision. We usually ask the surgeon to mark the surface away from the cavity with a stitch. This surface is inked, and the whole specimen is processed.
Re-excision biopsies The specimens are dealt with in the same manner as wide local excisions. The specimen is sliced after painting the surface with ink. All slices containing previous biopsy cavity wall are processed as well as any abnormal areas in the rest of the specimen. The likelihood of finding small foci of residual tumour increases with the increase in the number of cavity wall slices examined. Recently it has been recommended to process two slices per centimetre of tissue from grossly benign re-excision biopsies.1 If the re-excision biopsy is received as a flat piece of tissue, like a cavity wall biopsy, it is dealt with as such (see above). The samples are then taken perpendicularly so that the inner surface and outer, painted, surface are included in each section.
Axillary dissection specimens We deal with these specimens while fresh, and we do not use any clearing techniques. All the nodes in the specimen are identified by palpation, dissected and dealt with according to their size. If they are big enough to be bisected, >5 mm, this is done along the presumed site of their hilum, and both halves are processed in one cassette. If the nodes are too small for bisection, they are processed intact, and in this case more than one lymph node may be processed in the same cassette. If a lymph node is too big to be included in a cassette and there is obvious tumour involvement, only a part of the node is processed, after recording the maximum dimension of the tumour deposit. If the enlarged node does not show obvious macroscopic deposits, the whole node is sectioned horizontally and processed in two or more cassettes. The whole idea is to keep track of the number of lymph nodes dissected from the specimen. We usually examine one section from each half of the larger nodes, and two step sections from small lymph
141
nodes. We perform immunoperoxidase stains for cytokeratin only occasionally, when micrometastases are suspected after examination of haemotoxylin and eosin stained sections. This sometimes happens in cases of invasive lobular carcinoma where small metastases may be difficult to detect in routinely stained sections.
Sentinel lymph nodes Each sentinel node is processed separately as described above. As we currently follow a (UK) nationwide protocol for dealing with these nodes, we only examine one section from each half node and do not carry out frozen section or keratin immunohistological examination. Frozen section examination is advocated by some centres where full axillary dissection is carried out, if the sentinel node is positive, before the patient leaves the operating theatre.2,3 However, this is still a controversial point as the false negative results of frozen sections can be too high.4,5 On the other hand, there is accumulating evidence that cytokeratin immunohistochemical examination of a section of the sentinel node increases the yield of positive cases3 which could have therapeutic implications.6
Lumpectomy specimens These are usually done for benign lesions, such as fibroadenomas and fibrocystic change, which are removed with a variable amount of ‘normal’ breast tissue around them. In most cases, gross examination is consistent with a benign lesion. According to its size, the specimen is either processed in its entirety, after being bisected, such as small fibroadenomas, or selectively sampled after being sliced, such as cases of fibrocystic change. Sampling must be directed mainly at the non-fatty areas of the biopsy, as the likelihood of finding any epithelial lesions within pure adipose tissue is extremely low.7 If gross examination is suspicious, the specimen surface is painted with ink and dealt with as for wide local excisions.
Mastectomy specimens We paint the posterior surface of the breast with ink. The specimen is then sliced with each slice being approximately 1 cm thick. The lesion(s) are identified and described. Tumours or the walls of previously excised tumours are processed in their entirety if this is feasible. Sections are also taken from any other abnormal areas as well as from the nipple and the four breast quadrants. If lymph nodes are included in the specimen, these are dealt with as described above. Mastectomies are sometimes carried out for small impalpable tumours, for various clinical reasons. In these cases, it may be difficult to locate the tumour as no preoperative localisation procedures has been carried out. A preoperative mammogram can help the Pathologist to identify the possible site of the lesion. The suspected slice can then be X-rayed to confirm the presence of the mammographic abnormality.
142
CURRENT DIAGNOSTIC PATHOLOGY
Core biopsies
Is the tumour in situ or invasive?
These are received fixed in formalin. Accurate description is important. In particular whether the received biopsy is in the form of intact cores of fragments, the number of cores and their colour should be recorded. We have recently started receiving core biopsies in small cassettes, immersed in formalin, which are used for X-raying the cores at the Radiology department immediately after removal. These small cassettes fit exactly into our histology processing cassettes, thus obviating the need to transfer the cores from the formalin container into our cassettes which can be associated with traumatic artefacts or loss of tissue. All cores are processed, and at least four serial sections or ribbons are examined.
If it is in situ, is microinvasion present? Microinvasion is defined as the presence of one or more foci of clearly invasive carcinoma, not more than 1 mm across, in the immediate vicinity of duct(s) involved by ductal carcinoma in situ.13 In this respect, care should be taken not to include dislodged tumour cells in the stroma, as a result of previous needling procedures, as true invasion.14 In these cases, the effects of the needle, in the form of localised haemorrhage or tract fibrosis, are always evident nearby.
MICROSCOPIC REPORTING OF BREAST BIOPSIES It is important that the same pathologist who has dealt with the gross specimen and sampled it does the microscopic reporting. Gross examination, as well as clinical information, can provide the pathologist with a good idea about what to expect on microscopic examination. Even a good gross description cannot replace seeing the actual lesion by the naked eye. A written report should, ideally, provide clear answers to a number of specific questions:
Is the lesion benign or malignant? This is the most important question that the pathologist is required to answer. In most cases, a straightforward answer can be given easily once the slides have been examined. However, there are certain benign conditions that can be confused with malignancy, for instance microglandular adenosis, extensive sclerosing adenosis and entrapped glands in the centre of a radial scar. It has to be noted here that the latter two lesions may be associated with smaller malignant elements. Caution also has to be exercised when reporting biopsies removed using electrocautery instruments, as cautery artefacts mimicking invasive carcinoma may be produced at the excision margin.8 There are also borderline lesions where a benign or malignant decision cannot be easily agreed upon even by experts in the field. These include the atypical hyperplastic9 and atypical apocrine10 lesions, as well as recently described borderline vascular lesions developing in the skin or breast tissue following radiotherapy.11 However, it seems that strict adherence to the standardised criteria for the diagnosis of these lesions can markedly reduce inter-observer disagreements.12 The histological type of the lesion is given, and if follow up is recommended, as in cases of florid and atypical hyperplasia and borderline phyllodes tumours, a note to that effect is added to the report.
If the lesion is malignant The report has to cover the following points:
Histological type For in situ tumours, differentiation between ductal (DCIS) and lobular (LCIS) lesions is important, and it is useful to mention the sub-type of DCIS, for example comedo, solid, cribriform, etc. although nuclear grading of these lesions (see below) is thought to be of more prognostic significance. For invasive tumours, most of the special histological types, for example tubular, invasive cribriform, mucinous, and possibly medullary, carry a better prognosis than the usual invasive ductal and lobular types. For this reason the strict criteria for the diagnosis of these special types should be followed. For example, for the diagnosis of pure tubular or cribriform carcinoma more than 90% of the tumour should be composed of well-formed tubules or cribriform structures, respectively. Similarly, for the diagnosis of pure mucinous carcinoma, 90% of the tumour must exhibit the typical mucinous appearance. However, if the tumour is composed of a mixture of tubules and cribriform structures, it should be called tubular if tubules compose more than 50% of the lesion, and cribriform otherwise.13 Mixed types do occur, but there are also strict rules for diagnosing mixed tumours. For example, for mixed tubular/ductal or tubular/lobular, the tubular elements should comprise between 75–90% of the surface area of the tumour sections examined. In mixed mucinous/ ductal tumours, the mucinous areas should comprise 10–90% of the tumour. Similarly, in mixed ductal/lobular tumours, the ductal elements should form between 10 and 90% of the tumour.13 For medullary carcinoma, the tumour should consist of markedly pleomorphic cells arranged in solid, syncytial, interconnected sheets with well-defined, pushing margins, minimal intratumoural stroma and heavy infiltration with lymphocytes. DCIS elements are usually absent, and the tumour is usually oestrogen and progesterone receptor negative. The pathologist should also be alert to spot the uncommon non-epithelial malignant tumours such as lymphomas, angiosarcomas and other soft tissue sarcomas and to be able to differentiate the latter from metaplastic carcinomas using keratin immunostains.
Histological grade For DCIS, cases are graded according to the shape and size of their nuclei into three grades: low, intermediate and high. Low grade lesions have small, monomorphic
REPORTING BREAST BIOPSIES
nuclei almost the same size as the nuclei of normal epithelial cells. High grade lesions have markedly enlarged and pleomorphic nuclei which are usually also irregularly spaced and show easily detected mitotic activity. Intermediate grade DCIS has only slightly enlarged nuclei showing a mild degree of pleomorphism.13 For invasive carcinoma, the Elston and Ellis grading method is used which depends on the assessment of three histological features: the extent of glandular differentiation, the degree of nuclear pleomorphism and the number of mitotic figures.13,15 A score of 1–3 is assigned to each feature. The added total score for the three features determines the tumour grade. Thus, for glandular differentiation, a score of 1 is given if >75% of the tumour cells are arranged in glandular structures, a score of 2 if glands form 10–75% of the tumour and a score of 3 is given if the glands are <10%. Nuclear pleomorphism is assessed as for DCIS, with a score of 1 given to low grade nuclei, a score of 2 for intermediate grade, and a score of three for high grade nuclei. For mitotic activity, the score depends on the number of mitotic figures/10 high power microscopic field (HPF), at the periphery of the tumour, and the diameter of this field. For a field diameter of 0.48, a score of 1 is given for 0–6 mitoses/10 HPF, a score of 2 for 7–12 mitoses, and a score of 3 for >12 mitoses. On adding the three scores together, a total score of 3–5 indicates a grade 1 tumour, a total score of 6–7 is a grade 2, and 8–9 indicates a grade 3 invasive tumour. All carcinomas are graded regardless of their histological type.
143
and eosin stained sections, without the need for special immunostaining. Distinction between lymphatics and capillaries can be difficult, hence the use of the term lymphovascular to cover both types of vessels. In many cases the involved vessel is present in the immediate vicinity of a vein and an artery, which helps to confirm its lymphovascular nature. There is a good direct correlation between the presence of lymphovascular invasion and the presence of lymph node metastasis,16 but cases with lymphovascular invasion and no obvious lymph node metastasis do exist; and it has been shown that there is a significant relationship between lymphovascular invasion and both survival and local recurrence, which is independent of the lymph node status.17
Is the lesion completely excised? If the tumour is seen at the marked excision, we can presume that it has not been completely excised. If the tumour does not reach the marked margin, the distances between the margin and the invasive and in situ elements, if present, are measured and given separately. The type of in situ elements seen near or at the margin is important, as some surgeons will not do further reexcision if these elements are of the LCIS type. Reexcision biopsies are reported in the same way. Studies have shown that in cases reported as inadequately excised, the chances of finding residual tumour in the reexcision biopsies is around 55%.18 The rate of residual tumour detection increases with the increase in the number of tissue samples examined.1
Size Accurate measurement is important for staging the disease and planning postoperative treatment. The largest dimension is measured during gross examination. For small tumours, where a cross section of the whole lesion can be included in a single slide, microscopic measurement is usually more accurate than gross measurement. In the presence of abundant DCIS elements adjacent to the invasive lesion, these are characterised and measured separately. Cases of pure DCIS can be difficult to measure if the extent of the lesion is not grossly obvious and is found in more than one microscopic section. In these cases, an assessment may have to be made on the basis of the number of consecutive slices involved by the tumour and the approximate thickness of each slice. Thus, if three consecutive slices, each 5 mm thick, are involved, the estimated maximum dimension would be 15 mm, if that is more than the measurement of the lesion in individual slides.
Presence or absence of lymphovascular invasion This is usually assessed in the immediate peri-tumoural breast tissue. The idea is to avoid confusing shrinkage artefacts, which may occur within the tumour because of poor fixation, with lymphatic invasion. The tumour cells should be present within an endothelium-lined vessel. These can usually be identified in routine haematoxylin
Axillary lymph node status Nodes with grossly obvious metastasis are easily diagnosed. The maximum dimension of the largest node is usually given in the report. Sometimes a clinically markedly enlarged node appears microscopically as if it is the result of several amalgamated, adjacent involved nodes. We mention this possibility in the report although the ‘node’ is still counted as one. The presence or absence of extranodal tumour spread outside the lymph node capsule is mentioned, as it might have an effect on management and prognosis.19 For nodes with no obvious gross involvement, sections are scanned using the low power of the microscope, before going to the high power, looking carefully for evidence of metastatic disease. Although small metastatic deposits are more commonly seen near the capsule, they can be present elsewhere. If any suspicious cells are seen, immunohistochemistry for pan-keratin is useful in confirming the presence of metastasis if the suspicious cells are keratin-positive. This is especially useful in cases of invasive lobular carcinoma, where identification of small metastases can be difficult. Very rarely, keratin-positive benign glandular epithelial inclusions may be found in axillary nodes, but these can be easily distinguished from carcinoma by the lack of malignant cytological and histological features.20 Currently, we do not routinely use keratin staining for the detection of micrometastasis, as
144
this would be impractical when dealing with an average of 20 nodes from each case. However, we might have to change this practice in view of recent information strongly suggesting a relationship between immunohistologically demonstrated lymph node micrometastasis and prognosis, particularly in postmenopausal patients.21 Furthermore, with the introduction of the sentinel node concept, efforts for detecting micrometastases, whether by immunohistochemistry or step sectioning, can be concentrated on a single node or a few nodes, and their significance can be determined by long-term follow up. Polymerase chain reaction (PCR)-based techniques for the detection of micrometastases are available, but still under investigation.22,23
Oestrogen and progesterone receptor status Assessment of the oestrogen (ER) and, to a lesser extent, progesterone (PgR) receptor status of breast carcinoma is essential for the proper planning of the management of the disease. This assessment is now routinely done immunohistochemically using the immunoperoxidase technique on routinely processed paraffin sections of the tumour.24 Highly specific monoclonal antibodies that work in paraffin sections are commercially available. The results are thought to be more accurate than the previously used biochemical assay methods, as the staining products are visualised microscopically, thus eliminating the possibility of false positive and false negative results which may arise from sampling errors. Proper positive and negative controls should be used in each batch of staining. The results are usually scored semiquantitatively, although quantitative methods are also available. There are several published semiquantitative methods but we use what is commonly referred to as the H scoring system, which takes into account the percentage of stained tumour cells as well as the intensity of their staining. Only nuclear staining of tumour cells is taken into consideration. Any cytoplasmic staining, which is occasionally seen, is ignored as it is thought to be non-specific. Tumour cells that stain as intensely as normal cells, seen in the same or control sections, are given a factor of 3. Moderate staining is 2, faint staining is 1, and no staining is 0. The whole stained section is scanned and the percentage of stained cells and the intensity of staining are assessed. The percentage of stained cells is multiplied by the staining factor giving a total score which varies between 0 and 300. Cases with a score of 50 or less are usually considered ER negative, although this cut-off point may vary from one institute to another. A score of 51–100 is weakly positive (+), 101–200 is moderately positive (++), and 201–300 is strongly positive (+++).24 Staining and scoring should be carried out according to a strict protocol, and ideally the laboratory should participate in one of the National or Regional Quality Assurance Schemes to guarantee the quality of the procedure.
CURRENT DIAGNOSTIC PATHOLOGY
Other immunohistological studies There is a large number of commercially available antibodies that can be used on routinely processed paraffin sections of breast carcinoma for the detection of specific antigens that are thought to have useful prognostic or predictive value. The most widely studied and best known members of this group are antibodies directed against the protein product of c-erbB-2 (an oncogene), p53 (a tumour suppressor gene) and Ki-67 (a cell proliferative marker). The clinical usefulness of these antibodies is still under study, but most studies suggest that the immunohistological demonstration of c-erbB-2 and p53 and a high Ki-67 count are all associated with poorer prognosis. The expression of c-erbB-2 is also thought to have predictive value as regards the response of the tumour to specific chemotherapeutic agents. Also, there are currently clinical trials using specific anti cerbB-2 antibodies to treat c-erbB2 positive tumours in conjunction with other appropriate chemotherapeutic agents. Thus, the demand for the demonstration of these, and possibly other, antigens in breast cancer tissue is bound to increase, and strict protocols for their demonstration and scoring are necessary.25
PRACTICE POINTS
• Proper fixation of specimens is essential for accurate microscopic interpretation • Proper gross examination and sampling of the •
• • • • •
biopsy is an essential part of the histopathological examination and is best carried out by the person who is going to do the microscopic reporting Specific protocols should be followed in dealing with wire localisation and wide local excision biopsies to secure identifying the mammographic or palpable lesion, and to accurately assess completeness of excision In cases of ductal carcinoma in situ, all tumour tissue has to be processed and examined in order not to miss small foci of microinvasive or invasive carcinoma All lymph nodes received should be identified and examined Thorough sampling of re-excision biopsies will increase the chance of finding smaller foci of residual tumour In reporting cases of fibrocystic change, the presence or absence of epithelial hyperplasia and its type should be mentioned For in situ tumours, the report should include the histological type, nuclear grade, presence or absence of microinvasion, estimated size, completeness of excision and the presence of any other associated abnormalities
REPORTING BREAST BIOPSIES
PRACTICE POINTS (cont)
• For invasive tumours, the histological type and
•
grade should be mentioned as well as the size, associated in situ elements or other abnormalities, presence or absence of peritumoural lymphovascular invasion, completeness of excision, and prominent additional features like the presence of heavy lymphocytic infiltration or areas of necrosis For lymph nodes, the total number of nodes examined and the number of positive nodes should be mentioned. The size of the largest metastatic deposit and the presence or absence of extracapsular tumour spread help the Oncologist to visualise the extent of metastatic disease which may influence the therapeutic decision
RESEARCH DIRECTIONS
• To assess the usefulness of the sentinel lymph node
• •
technique in determining the axillary lymph node status instead of resorting to the more radical surgical procedures of axillary clearance and dissection which can be associated with a relative high morbidity To subclassify using morphological and molecular techniques the heterogeneous group termed ‘invasive ductal carcinoma’ The relationship between specific genetic changes and the morphology of tumours
REFERENCES 1. Abraham S C, Fox K, Fraker D, Solin L, Reynolds C. Sampling of grossly benign breast reexcisions. A multidisciplinary approach to assessing adequacy. Am J Surg Pathol 1992; 23: 316–322. 2. Giuliano A E. Mapping a pathway for axillary staging. A personal perspective on the current status of sentinel lymph node dissection for breast cancer. Arch Surg 1999; 134: 195–199. 3. Turner R R, Ollila D W, Stern S, Giuliano A E. Optimal histopathologic examination of the sentinel lymph node for breast carcinoma staging. Am J Surg Pathol 1999; 23: 263–267. 4. Anderson T J. The challenge of sentinel lymph node biopsy. Histopathology 1999; 35: 82–84. 5. Dixon J M, Mamman U, Thomas J. Accuracy of intraoperative frozen-section analysis of axillary lymph nodes. Br J Surg 1999; 86: 392–395. 6. McIntosh S A, Going J J, Soukop M, Purushotham A D, Cooke T G. Therapeutic implications of the sentinel lymph node in breast cancer. Lancet 1999; 354: 570.
145
7. Schnitt S J, Wang H H. Histologic sampling of grossly bening breast biopsies. How much is enough? Am J Surg Pathol 1989; 13: 505–512. 8. Rosen P P. Letters to the editor. Ann Surg 1986; 204: 612–613. 9. Rosai J. Borderline epithelial lesions of the breast. Am J Surg Pathol 1991; 15: 209–221. 10. Seidman J D, Ashton M, Lefkowitz M. Atypical apocrine adenosis of the breast. A clinicopathologic study of 37 patients with 8.7year follow-up. Cancer 1996; 77: 2529–2537. 11. Fineberg S, Rosen P P. Cutaneous angiosarcoma and atypical vascular lesions of the skin and breast after radiation therapy for breast carcinoma. Am J Clin Pathol 1994; 102: 757–763. 12. Schnitt S J, Connolly J L, Tavassoli F A, Fechner R E, Kempson R L, Gelman R, Page D L. Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria. Hum Pathol 1992; 16: 1133–1143. 13. National Coordinating Group for Breast Screening Pathology. Pathology reporting in breast cancer screening. Second edition. NHSBSP Publication No. 3, 1995. 14. Youngson B J, Liberman L, Rosen P P. Displacement of carcinomatous epithelium in surgical breast specimens following stereotaxic core biopsy. Am J Clin Pathol 1995; 103: 598–602. 15. Elston C W, Ellis I O. Pathological prognostic factors in breast cancer I. The value of histological grade in breast cancer: experience from a large study with long term follow-up. Histopathology 1991; 19: 403–410. 16. Yiangou C, Shousha S, Sinnett H D. Primary tumour characteristics and axillary lymph node status in breast cancer. Br J Cancer 1999; 80: 1974–1978. 17. Pinder S E, Ellis I O, Galea M, O’Rouke S, Blamey R W, Elston C W. Pathological prognostic factors in breast cancer. III. Vascular invasion: relationship with recurrence and survival in a large study with long-term follow-up. Histopathology 1994; 24: 41–47. 18. Gwin J L, Eisenberg B L, Hoffman J P, Ottery F D, Boraas M, Solin L J. Incidence of gross and microscopic carcinoma in specimens from patients with breast cancer after re-excision lumpectomy. Ann Surg 1993; 218: 729–734. 19. Fisher B J, Perera F E, Cooke A L, Opeitum A, Dar A R, Venkatesan V M, Stitt L, Radwan J S. Extracapsular axillary node extension in patients receiving adjuvant systemic therapy: An indication for radiotherapy? Int J Radiation Oncology Biol Phys 1997; 38: 551–559. 20. Fisher C J, Hill S, Millis R R. Benign lymph node inclusions mimicking metastatic carcinoma. J Clin Pathol 1994; 47: 245–247. 21. Cote R J, Peterson H F, Chaiwun B, Gelber R D, Goldhirsch A, Castiglione-Gertsch M, Guterson B, Neville A M. Role of immunohistochemical detection of lymph-node metastases in management of breast cancer. Lancet 1999; 354: 896–900. 22. Schoenfeld A, Luqmani Y, Smith D, O’Reilly S, Shousha S, Sinnett H D, Coombes R C. Detection of breast cancer micrometastases in axillary lymph nodes using polymerase chain reaction. Cancer Res 1994; 54: 2986–2990. 23. Giuliano A E, Kelemen P R. Sophisticated techniques detect obscure lymph node metastases in carcinoma of the breast. Cancer 1998; 83: 391–393. 24. Shousha S, Peston D. Immunohistochemical demonstration of oestrogen and progesterone receptors in paraffin sections of breast carcinoma. Cur Diag Pathol 1997; 4: 176–180. 25. Allred D C, Harvey J M, Berardo M, Clark G M. Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 1998; 11: 155–168.
REVIEW
Reporting breast biopsies S. Shousha Imperial College School of Medicine, Department of Histopathology, Charing Cross Hospital, Fulham Palace Road, London, W6 8RF, UK
KEYWORDS breast, benign disease, breast carcinoma, oestrogen receptors
Summary The role of the surgical pathologist in dealing with breast biopsies has changed enormously during the last few years. The introduction of mammographic screening, conservative breast surgery, core biopsies and, more recently, sentinel lymph node techniques, have changed the type of material received by the pathologist for diagnosis, and the type of information required from him/her. Requests for frozen sections have markedly diminished, but new challenges have arisen, particularly in respect of diagnosing borderline lesions which are being increasingly detected by mammography. This article presents a detailed account of the handling and reporting of breast biopsies, based partly on National Guidelines and the practice at Charing Cross Hospital, London. © 2000 Harcourt Publishers Ltd
DEALING WITH THE GROSS SPECIMEN Good fixation is essential for proper diagnosis. In our department we prefer to receive all specimens, except core biopsies, fresh in a plastic bag, immediately after surgical excision whenever this is feasible and in the absence of any suspected infections. The specimen is registered, given a laboratory number, and the pathologist in charge is then called to deal with it. For biopsies coming from other hospitals we ask the surgeons to immerse it in an adequate amount of formalin and send it to us as soon as possible where it is dealt with promptly. Core biopsies are immersed in formalin straight after being removed from the patient, or, in case of stereotactic biopsies, after X-raying the specimen by the radiologist to ensure the presence of calcification.
SPECIMEN TYPES Wire localisation biopsies The surgeons usually send these specimens first to the X-ray department where the whole specimen is X-rayed to ensure that the radiologically abnormal area has been removed. The surgeon is notified of the result, and if the abnormal area has not been removed, a further excision is carried out and dealt with in the same way. We receive these biopsies with their specimen X-rays, and with the wires still attached to them. The biopsy is weighed, measured, described, and the wire is gently
Correspondence to: SS. Tel: +44 20 8846 7144; Fax: +44 20 8846 1364; E-mail: [email protected].
removed. The outer surface of the specimen is painted with India Ink, and a few drops of Bowen’s fixative are added to help adherence of the ink to the specimen surface. Two minutes are usually enough for the ink to dry, after which extra ink is removed by blotting paper. The specimen is then sliced into approximately 5 mm thick slices. The slices are X-rayed in a Faxitron using polaroid X-ray films. This helps to identify the slice(s) containing the calcification or other abnormalities that correspond to those present in the specimen X-ray. All slices containing radiological abnormalities and their immediately adjacent ones are processed together with any other slices that may appear to be abnormal by naked eye. This would include the site of any previous recent needle aspiration or core biopsy which is usually indicated by an area of haemorrhage. The gross appearance of any abnormality is described. The minimum number of slices processed from each specimen is six. The total number will depend on the size of the biopsy and the extent of abnormalities in it. Each slice is processed in a separate cassette marked with the case number and an additional serial number which we also write on the corresponding slice X-ray. The cassettes containing the slices are kept in formalin overnight and are processed the following day, with the stained sections usually available for examination within 48 h. Any unprocesed slices are kept in formalin for 4 weeks, in case any extra tissue is needed for examination.
Wide local excision biopsies The specimen is weighed, measured, described and covered with ink as above. If there is a palpable lesion, the specimen is sliced along the maximum dimension of that lesion, in 5 mm thick slices. The lesion is described
REPORTING BREAST BIOPSIES
and its maximum dimension measured, as well as the distance between the lesion and the nearest excision margin. Samples of the biopsy are then taken for processing. We tend to process all slices containing obvious tumour tissue, as well as the immediately adjacent slices, including the nearest tumour margin, as long as this is feasible. We then sample the other biopsy margins, and process any other slices with gross abnormality. If the biopsy contains no palpable abnormality, the specimen is sliced along its narrowest dimension. All grossly abnormal slices are processed, as well as randomly sampled ones.
Cavity wall biopsies Some surgeons prefer to remove extra strips of breast tissue from the wall of the cavity left in the breast after doing a wide local excision. We usually ask the surgeon to mark the surface away from the cavity with a stitch. This surface is inked, and the whole specimen is processed.
Re-excision biopsies The specimens are dealt with in the same manner as wide local excisions. The specimen is sliced after painting the surface with ink. All slices containing previous biopsy cavity wall are processed as well as any abnormal areas in the rest of the specimen. The likelihood of finding small foci of residual tumour increases with the increase in the number of cavity wall slices examined. Recently it has been recommended to process two slices per centimetre of tissue from grossly benign re-excision biopsies.1 If the re-excision biopsy is received as a flat piece of tissue, like a cavity wall biopsy, it is dealt with as such (see above). The samples are then taken perpendicularly so that the inner surface and outer, painted, surface are included in each section.
Axillary dissection specimens We deal with these specimens while fresh, and we do not use any clearing techniques. All the nodes in the specimen are identified by palpation, dissected and dealt with according to their size. If they are big enough to be bisected, >5 mm, this is done along the presumed site of their hilum, and both halves are processed in one cassette. If the nodes are too small for bisection, they are processed intact, and in this case more than one lymph node may be processed in the same cassette. If a lymph node is too big to be included in a cassette and there is obvious tumour involvement, only a part of the node is processed, after recording the maximum dimension of the tumour deposit. If the enlarged node does not show obvious macroscopic deposits, the whole node is sectioned horizontally and processed in two or more cassettes. The whole idea is to keep track of the number of lymph nodes dissected from the specimen. We usually examine one section from each half of the larger nodes, and two step sections from small lymph
141
nodes. We perform immunoperoxidase stains for cytokeratin only occasionally, when micrometastases are suspected after examination of haemotoxylin and eosin stained sections. This sometimes happens in cases of invasive lobular carcinoma where small metastases may be difficult to detect in routinely stained sections.
Sentinel lymph nodes Each sentinel node is processed separately as described above. As we currently follow a (UK) nationwide protocol for dealing with these nodes, we only examine one section from each half node and do not carry out frozen section or keratin immunohistological examination. Frozen section examination is advocated by some centres where full axillary dissection is carried out, if the sentinel node is positive, before the patient leaves the operating theatre.2,3 However, this is still a controversial point as the false negative results of frozen sections can be too high.4,5 On the other hand, there is accumulating evidence that cytokeratin immunohistochemical examination of a section of the sentinel node increases the yield of positive cases3 which could have therapeutic implications.6
Lumpectomy specimens These are usually done for benign lesions, such as fibroadenomas and fibrocystic change, which are removed with a variable amount of ‘normal’ breast tissue around them. In most cases, gross examination is consistent with a benign lesion. According to its size, the specimen is either processed in its entirety, after being bisected, such as small fibroadenomas, or selectively sampled after being sliced, such as cases of fibrocystic change. Sampling must be directed mainly at the non-fatty areas of the biopsy, as the likelihood of finding any epithelial lesions within pure adipose tissue is extremely low.7 If gross examination is suspicious, the specimen surface is painted with ink and dealt with as for wide local excisions.
Mastectomy specimens We paint the posterior surface of the breast with ink. The specimen is then sliced with each slice being approximately 1 cm thick. The lesion(s) are identified and described. Tumours or the walls of previously excised tumours are processed in their entirety if this is feasible. Sections are also taken from any other abnormal areas as well as from the nipple and the four breast quadrants. If lymph nodes are included in the specimen, these are dealt with as described above. Mastectomies are sometimes carried out for small impalpable tumours, for various clinical reasons. In these cases, it may be difficult to locate the tumour as no preoperative localisation procedures has been carried out. A preoperative mammogram can help the Pathologist to identify the possible site of the lesion. The suspected slice can then be X-rayed to confirm the presence of the mammographic abnormality.
142
CURRENT DIAGNOSTIC PATHOLOGY
Core biopsies
Is the tumour in situ or invasive?
These are received fixed in formalin. Accurate description is important. In particular whether the received biopsy is in the form of intact cores of fragments, the number of cores and their colour should be recorded. We have recently started receiving core biopsies in small cassettes, immersed in formalin, which are used for X-raying the cores at the Radiology department immediately after removal. These small cassettes fit exactly into our histology processing cassettes, thus obviating the need to transfer the cores from the formalin container into our cassettes which can be associated with traumatic artefacts or loss of tissue. All cores are processed, and at least four serial sections or ribbons are examined.
If it is in situ, is microinvasion present? Microinvasion is defined as the presence of one or more foci of clearly invasive carcinoma, not more than 1 mm across, in the immediate vicinity of duct(s) involved by ductal carcinoma in situ.13 In this respect, care should be taken not to include dislodged tumour cells in the stroma, as a result of previous needling procedures, as true invasion.14 In these cases, the effects of the needle, in the form of localised haemorrhage or tract fibrosis, are always evident nearby.
MICROSCOPIC REPORTING OF BREAST BIOPSIES It is important that the same pathologist who has dealt with the gross specimen and sampled it does the microscopic reporting. Gross examination, as well as clinical information, can provide the pathologist with a good idea about what to expect on microscopic examination. Even a good gross description cannot replace seeing the actual lesion by the naked eye. A written report should, ideally, provide clear answers to a number of specific questions:
Is the lesion benign or malignant? This is the most important question that the pathologist is required to answer. In most cases, a straightforward answer can be given easily once the slides have been examined. However, there are certain benign conditions that can be confused with malignancy, for instance microglandular adenosis, extensive sclerosing adenosis and entrapped glands in the centre of a radial scar. It has to be noted here that the latter two lesions may be associated with smaller malignant elements. Caution also has to be exercised when reporting biopsies removed using electrocautery instruments, as cautery artefacts mimicking invasive carcinoma may be produced at the excision margin.8 There are also borderline lesions where a benign or malignant decision cannot be easily agreed upon even by experts in the field. These include the atypical hyperplastic9 and atypical apocrine10 lesions, as well as recently described borderline vascular lesions developing in the skin or breast tissue following radiotherapy.11 However, it seems that strict adherence to the standardised criteria for the diagnosis of these lesions can markedly reduce inter-observer disagreements.12 The histological type of the lesion is given, and if follow up is recommended, as in cases of florid and atypical hyperplasia and borderline phyllodes tumours, a note to that effect is added to the report.
If the lesion is malignant The report has to cover the following points:
Histological type For in situ tumours, differentiation between ductal (DCIS) and lobular (LCIS) lesions is important, and it is useful to mention the sub-type of DCIS, for example comedo, solid, cribriform, etc. although nuclear grading of these lesions (see below) is thought to be of more prognostic significance. For invasive tumours, most of the special histological types, for example tubular, invasive cribriform, mucinous, and possibly medullary, carry a better prognosis than the usual invasive ductal and lobular types. For this reason the strict criteria for the diagnosis of these special types should be followed. For example, for the diagnosis of pure tubular or cribriform carcinoma more than 90% of the tumour should be composed of well-formed tubules or cribriform structures, respectively. Similarly, for the diagnosis of pure mucinous carcinoma, 90% of the tumour must exhibit the typical mucinous appearance. However, if the tumour is composed of a mixture of tubules and cribriform structures, it should be called tubular if tubules compose more than 50% of the lesion, and cribriform otherwise.13 Mixed types do occur, but there are also strict rules for diagnosing mixed tumours. For example, for mixed tubular/ductal or tubular/lobular, the tubular elements should comprise between 75–90% of the surface area of the tumour sections examined. In mixed mucinous/ ductal tumours, the mucinous areas should comprise 10–90% of the tumour. Similarly, in mixed ductal/lobular tumours, the ductal elements should form between 10 and 90% of the tumour.13 For medullary carcinoma, the tumour should consist of markedly pleomorphic cells arranged in solid, syncytial, interconnected sheets with well-defined, pushing margins, minimal intratumoural stroma and heavy infiltration with lymphocytes. DCIS elements are usually absent, and the tumour is usually oestrogen and progesterone receptor negative. The pathologist should also be alert to spot the uncommon non-epithelial malignant tumours such as lymphomas, angiosarcomas and other soft tissue sarcomas and to be able to differentiate the latter from metaplastic carcinomas using keratin immunostains.
Histological grade For DCIS, cases are graded according to the shape and size of their nuclei into three grades: low, intermediate and high. Low grade lesions have small, monomorphic
REPORTING BREAST BIOPSIES
nuclei almost the same size as the nuclei of normal epithelial cells. High grade lesions have markedly enlarged and pleomorphic nuclei which are usually also irregularly spaced and show easily detected mitotic activity. Intermediate grade DCIS has only slightly enlarged nuclei showing a mild degree of pleomorphism.13 For invasive carcinoma, the Elston and Ellis grading method is used which depends on the assessment of three histological features: the extent of glandular differentiation, the degree of nuclear pleomorphism and the number of mitotic figures.13,15 A score of 1–3 is assigned to each feature. The added total score for the three features determines the tumour grade. Thus, for glandular differentiation, a score of 1 is given if >75% of the tumour cells are arranged in glandular structures, a score of 2 if glands form 10–75% of the tumour and a score of 3 is given if the glands are <10%. Nuclear pleomorphism is assessed as for DCIS, with a score of 1 given to low grade nuclei, a score of 2 for intermediate grade, and a score of three for high grade nuclei. For mitotic activity, the score depends on the number of mitotic figures/10 high power microscopic field (HPF), at the periphery of the tumour, and the diameter of this field. For a field diameter of 0.48, a score of 1 is given for 0–6 mitoses/10 HPF, a score of 2 for 7–12 mitoses, and a score of 3 for >12 mitoses. On adding the three scores together, a total score of 3–5 indicates a grade 1 tumour, a total score of 6–7 is a grade 2, and 8–9 indicates a grade 3 invasive tumour. All carcinomas are graded regardless of their histological type.
143
and eosin stained sections, without the need for special immunostaining. Distinction between lymphatics and capillaries can be difficult, hence the use of the term lymphovascular to cover both types of vessels. In many cases the involved vessel is present in the immediate vicinity of a vein and an artery, which helps to confirm its lymphovascular nature. There is a good direct correlation between the presence of lymphovascular invasion and the presence of lymph node metastasis,16 but cases with lymphovascular invasion and no obvious lymph node metastasis do exist; and it has been shown that there is a significant relationship between lymphovascular invasion and both survival and local recurrence, which is independent of the lymph node status.17
Is the lesion completely excised? If the tumour is seen at the marked excision, we can presume that it has not been completely excised. If the tumour does not reach the marked margin, the distances between the margin and the invasive and in situ elements, if present, are measured and given separately. The type of in situ elements seen near or at the margin is important, as some surgeons will not do further reexcision if these elements are of the LCIS type. Reexcision biopsies are reported in the same way. Studies have shown that in cases reported as inadequately excised, the chances of finding residual tumour in the reexcision biopsies is around 55%.18 The rate of residual tumour detection increases with the increase in the number of tissue samples examined.1
Size Accurate measurement is important for staging the disease and planning postoperative treatment. The largest dimension is measured during gross examination. For small tumours, where a cross section of the whole lesion can be included in a single slide, microscopic measurement is usually more accurate than gross measurement. In the presence of abundant DCIS elements adjacent to the invasive lesion, these are characterised and measured separately. Cases of pure DCIS can be difficult to measure if the extent of the lesion is not grossly obvious and is found in more than one microscopic section. In these cases, an assessment may have to be made on the basis of the number of consecutive slices involved by the tumour and the approximate thickness of each slice. Thus, if three consecutive slices, each 5 mm thick, are involved, the estimated maximum dimension would be 15 mm, if that is more than the measurement of the lesion in individual slides.
Presence or absence of lymphovascular invasion This is usually assessed in the immediate peri-tumoural breast tissue. The idea is to avoid confusing shrinkage artefacts, which may occur within the tumour because of poor fixation, with lymphatic invasion. The tumour cells should be present within an endothelium-lined vessel. These can usually be identified in routine haematoxylin
Axillary lymph node status Nodes with grossly obvious metastasis are easily diagnosed. The maximum dimension of the largest node is usually given in the report. Sometimes a clinically markedly enlarged node appears microscopically as if it is the result of several amalgamated, adjacent involved nodes. We mention this possibility in the report although the ‘node’ is still counted as one. The presence or absence of extranodal tumour spread outside the lymph node capsule is mentioned, as it might have an effect on management and prognosis.19 For nodes with no obvious gross involvement, sections are scanned using the low power of the microscope, before going to the high power, looking carefully for evidence of metastatic disease. Although small metastatic deposits are more commonly seen near the capsule, they can be present elsewhere. If any suspicious cells are seen, immunohistochemistry for pan-keratin is useful in confirming the presence of metastasis if the suspicious cells are keratin-positive. This is especially useful in cases of invasive lobular carcinoma, where identification of small metastases can be difficult. Very rarely, keratin-positive benign glandular epithelial inclusions may be found in axillary nodes, but these can be easily distinguished from carcinoma by the lack of malignant cytological and histological features.20 Currently, we do not routinely use keratin staining for the detection of micrometastasis, as
144
this would be impractical when dealing with an average of 20 nodes from each case. However, we might have to change this practice in view of recent information strongly suggesting a relationship between immunohistologically demonstrated lymph node micrometastasis and prognosis, particularly in postmenopausal patients.21 Furthermore, with the introduction of the sentinel node concept, efforts for detecting micrometastases, whether by immunohistochemistry or step sectioning, can be concentrated on a single node or a few nodes, and their significance can be determined by long-term follow up. Polymerase chain reaction (PCR)-based techniques for the detection of micrometastases are available, but still under investigation.22,23
Oestrogen and progesterone receptor status Assessment of the oestrogen (ER) and, to a lesser extent, progesterone (PgR) receptor status of breast carcinoma is essential for the proper planning of the management of the disease. This assessment is now routinely done immunohistochemically using the immunoperoxidase technique on routinely processed paraffin sections of the tumour.24 Highly specific monoclonal antibodies that work in paraffin sections are commercially available. The results are thought to be more accurate than the previously used biochemical assay methods, as the staining products are visualised microscopically, thus eliminating the possibility of false positive and false negative results which may arise from sampling errors. Proper positive and negative controls should be used in each batch of staining. The results are usually scored semiquantitatively, although quantitative methods are also available. There are several published semiquantitative methods but we use what is commonly referred to as the H scoring system, which takes into account the percentage of stained tumour cells as well as the intensity of their staining. Only nuclear staining of tumour cells is taken into consideration. Any cytoplasmic staining, which is occasionally seen, is ignored as it is thought to be non-specific. Tumour cells that stain as intensely as normal cells, seen in the same or control sections, are given a factor of 3. Moderate staining is 2, faint staining is 1, and no staining is 0. The whole stained section is scanned and the percentage of stained cells and the intensity of staining are assessed. The percentage of stained cells is multiplied by the staining factor giving a total score which varies between 0 and 300. Cases with a score of 50 or less are usually considered ER negative, although this cut-off point may vary from one institute to another. A score of 51–100 is weakly positive (+), 101–200 is moderately positive (++), and 201–300 is strongly positive (+++).24 Staining and scoring should be carried out according to a strict protocol, and ideally the laboratory should participate in one of the National or Regional Quality Assurance Schemes to guarantee the quality of the procedure.
CURRENT DIAGNOSTIC PATHOLOGY
Other immunohistological studies There is a large number of commercially available antibodies that can be used on routinely processed paraffin sections of breast carcinoma for the detection of specific antigens that are thought to have useful prognostic or predictive value. The most widely studied and best known members of this group are antibodies directed against the protein product of c-erbB-2 (an oncogene), p53 (a tumour suppressor gene) and Ki-67 (a cell proliferative marker). The clinical usefulness of these antibodies is still under study, but most studies suggest that the immunohistological demonstration of c-erbB-2 and p53 and a high Ki-67 count are all associated with poorer prognosis. The expression of c-erbB-2 is also thought to have predictive value as regards the response of the tumour to specific chemotherapeutic agents. Also, there are currently clinical trials using specific anti cerbB-2 antibodies to treat c-erbB2 positive tumours in conjunction with other appropriate chemotherapeutic agents. Thus, the demand for the demonstration of these, and possibly other, antigens in breast cancer tissue is bound to increase, and strict protocols for their demonstration and scoring are necessary.25
PRACTICE POINTS
• Proper fixation of specimens is essential for accurate microscopic interpretation • Proper gross examination and sampling of the •
• • • • •
biopsy is an essential part of the histopathological examination and is best carried out by the person who is going to do the microscopic reporting Specific protocols should be followed in dealing with wire localisation and wide local excision biopsies to secure identifying the mammographic or palpable lesion, and to accurately assess completeness of excision In cases of ductal carcinoma in situ, all tumour tissue has to be processed and examined in order not to miss small foci of microinvasive or invasive carcinoma All lymph nodes received should be identified and examined Thorough sampling of re-excision biopsies will increase the chance of finding smaller foci of residual tumour In reporting cases of fibrocystic change, the presence or absence of epithelial hyperplasia and its type should be mentioned For in situ tumours, the report should include the histological type, nuclear grade, presence or absence of microinvasion, estimated size, completeness of excision and the presence of any other associated abnormalities
REPORTING BREAST BIOPSIES
PRACTICE POINTS (cont)
• For invasive tumours, the histological type and
•
grade should be mentioned as well as the size, associated in situ elements or other abnormalities, presence or absence of peritumoural lymphovascular invasion, completeness of excision, and prominent additional features like the presence of heavy lymphocytic infiltration or areas of necrosis For lymph nodes, the total number of nodes examined and the number of positive nodes should be mentioned. The size of the largest metastatic deposit and the presence or absence of extracapsular tumour spread help the Oncologist to visualise the extent of metastatic disease which may influence the therapeutic decision
RESEARCH DIRECTIONS
• To assess the usefulness of the sentinel lymph node
• •
technique in determining the axillary lymph node status instead of resorting to the more radical surgical procedures of axillary clearance and dissection which can be associated with a relative high morbidity To subclassify using morphological and molecular techniques the heterogeneous group termed ‘invasive ductal carcinoma’ The relationship between specific genetic changes and the morphology of tumours
REFERENCES 1. Abraham S C, Fox K, Fraker D, Solin L, Reynolds C. Sampling of grossly benign breast reexcisions. A multidisciplinary approach to assessing adequacy. Am J Surg Pathol 1992; 23: 316–322. 2. Giuliano A E. Mapping a pathway for axillary staging. A personal perspective on the current status of sentinel lymph node dissection for breast cancer. Arch Surg 1999; 134: 195–199. 3. Turner R R, Ollila D W, Stern S, Giuliano A E. Optimal histopathologic examination of the sentinel lymph node for breast carcinoma staging. Am J Surg Pathol 1999; 23: 263–267. 4. Anderson T J. The challenge of sentinel lymph node biopsy. Histopathology 1999; 35: 82–84. 5. Dixon J M, Mamman U, Thomas J. Accuracy of intraoperative frozen-section analysis of axillary lymph nodes. Br J Surg 1999; 86: 392–395. 6. McIntosh S A, Going J J, Soukop M, Purushotham A D, Cooke T G. Therapeutic implications of the sentinel lymph node in breast cancer. Lancet 1999; 354: 570.
145
7. Schnitt S J, Wang H H. Histologic sampling of grossly bening breast biopsies. How much is enough? Am J Surg Pathol 1989; 13: 505–512. 8. Rosen P P. Letters to the editor. Ann Surg 1986; 204: 612–613. 9. Rosai J. Borderline epithelial lesions of the breast. Am J Surg Pathol 1991; 15: 209–221. 10. Seidman J D, Ashton M, Lefkowitz M. Atypical apocrine adenosis of the breast. A clinicopathologic study of 37 patients with 8.7year follow-up. Cancer 1996; 77: 2529–2537. 11. Fineberg S, Rosen P P. Cutaneous angiosarcoma and atypical vascular lesions of the skin and breast after radiation therapy for breast carcinoma. Am J Clin Pathol 1994; 102: 757–763. 12. Schnitt S J, Connolly J L, Tavassoli F A, Fechner R E, Kempson R L, Gelman R, Page D L. Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria. Hum Pathol 1992; 16: 1133–1143. 13. National Coordinating Group for Breast Screening Pathology. Pathology reporting in breast cancer screening. Second edition. NHSBSP Publication No. 3, 1995. 14. Youngson B J, Liberman L, Rosen P P. Displacement of carcinomatous epithelium in surgical breast specimens following stereotaxic core biopsy. Am J Clin Pathol 1995; 103: 598–602. 15. Elston C W, Ellis I O. Pathological prognostic factors in breast cancer I. The value of histological grade in breast cancer: experience from a large study with long term follow-up. Histopathology 1991; 19: 403–410. 16. Yiangou C, Shousha S, Sinnett H D. Primary tumour characteristics and axillary lymph node status in breast cancer. Br J Cancer 1999; 80: 1974–1978. 17. Pinder S E, Ellis I O, Galea M, O’Rouke S, Blamey R W, Elston C W. Pathological prognostic factors in breast cancer. III. Vascular invasion: relationship with recurrence and survival in a large study with long-term follow-up. Histopathology 1994; 24: 41–47. 18. Gwin J L, Eisenberg B L, Hoffman J P, Ottery F D, Boraas M, Solin L J. Incidence of gross and microscopic carcinoma in specimens from patients with breast cancer after re-excision lumpectomy. Ann Surg 1993; 218: 729–734. 19. Fisher B J, Perera F E, Cooke A L, Opeitum A, Dar A R, Venkatesan V M, Stitt L, Radwan J S. Extracapsular axillary node extension in patients receiving adjuvant systemic therapy: An indication for radiotherapy? Int J Radiation Oncology Biol Phys 1997; 38: 551–559. 20. Fisher C J, Hill S, Millis R R. Benign lymph node inclusions mimicking metastatic carcinoma. J Clin Pathol 1994; 47: 245–247. 21. Cote R J, Peterson H F, Chaiwun B, Gelber R D, Goldhirsch A, Castiglione-Gertsch M, Guterson B, Neville A M. Role of immunohistochemical detection of lymph-node metastases in management of breast cancer. Lancet 1999; 354: 896–900. 22. Schoenfeld A, Luqmani Y, Smith D, O’Reilly S, Shousha S, Sinnett H D, Coombes R C. Detection of breast cancer micrometastases in axillary lymph nodes using polymerase chain reaction. Cancer Res 1994; 54: 2986–2990. 23. Giuliano A E, Kelemen P R. Sophisticated techniques detect obscure lymph node metastases in carcinoma of the breast. Cancer 1998; 83: 391–393. 24. Shousha S, Peston D. Immunohistochemical demonstration of oestrogen and progesterone receptors in paraffin sections of breast carcinoma. Cur Diag Pathol 1997; 4: 176–180. 25. Allred D C, Harvey J M, Berardo M, Clark G M. Prognostic and predictive factors in breast cancer by immunohistochemical analysis. Mod Pathol 1998; 11: 155–168.