Validation of mitosis counting by automated phosphohistone H3 (PHH3) digital image analysis in a breast carcinoma tissue microarray.

Mitosis counting in H&E stained sections is the most informative constituent of the Nottingham histological grade in breast carcinoma prognosis. Phosphohistone H3 (PHH3) immunohistochemistry (IHC) is a highly specific marker of mitoses, with practical application in identifying mitoses in poorly fixed or distorted tissue and is of prognostic significance in breast carcinoma. Our aim was to assess methods of PHH3 IHC mitosis counting in a tissue microarray (TMA) of 2 mm cores from 36 resected breast carcinomas. Mitoses in H&E and PHH3 stained slides were manually scored by pathologist consensus and expressed as counts/2 mm. PHH3 stained cores were also evaluated by automated digital image analysis (DIA). Results were compared using Spearman correlation. A strong and significant correlation was observed between manual PHH3 and manual H&E mitotic counts (correlation = 0.81; p < 0.0001) and between automated PHH3 DIA and manual H&E mitotic counts (correlation = 0.79; p < 0.0001). More mitoses were identified with PHH3 IHC than with H&E. Manual and DIA PHH3 counts were strongly and significantly correlated (correlation = 0.83; p < 0.0001) and of similar absolute values. PHH3 DIA is a valid alternative to manual counting with potential application in breast cancer reporting and prognostication.

Handling of radioactive seed localisation breast specimens in the histopathology laboratory: the Western Australian experience.

Radio-guided occult lesion localisation using iodine-125 seeds (ROLLIS) is a novel method of localisation for impalpable in situ and invasive carcinomas that has been the subject of a recent pilot study and pilot study extension in Western Australia. Robust protocols for radiation safety, specimen labelling, specimen tracking, seed retrieval and seed disposal were developed at two Western Australian laboratories to minimise the risk of seed loss. The processes are safe and effective with no significant radiation exposure to pathologists and with acquisition of all seeds intact and undamaged. The success can be attributed to developing specific seed retrieval techniques, suited to local preferences at each institution, with input from surgeons, radiologists and medical physics personnel. These techniques are now routine and will continue in the randomised control phase of the ROLLIS study.

Pathology reporting of breast cancer: trends in 1989-1999, following the introduction of mammographic screening in Western Australia.

BACKGROUND: A survey of pathology reporting of breast cancer in Western Australia in 1989 highlighted the need for improvement. The current study documents (1) changes in pathology reporting from 1989 to 1999 and (2) changes in patterns of histopathological prognostic indicators for breast cancer following introduction of mammographic screening in 1989. METHODS: Data concerning all breast cancer cases reported in Western Australia in 1989, 1994 and 1999 were retrieved using the State Cancer Registry, Hospital Morbidity data system, and pathology laboratory records. RESULTS: Pathology reports improved in quality during the decade surveyed. For invasive carcinoma, tumour size was not recorded in 1.2% of pathology reports in 1999 compared with 16.1% in 1989 (p<0.001). Corresponding figures for other prognostic factors were: tumour grade 3.3% and 51.6% (p<0.001), tumour type 0.2% and 4.1% (p<0.001), vascular invasion 3.7% and 70.9% (p<0.001), and lymph node status 1.9% and 4.5% (p = 0.023). In 1999, 5.9% of reports were not in a synoptic/checklist format, whereas all reports were descriptive in 1989 (p<0.001). For the population as a whole, the proportion of invasive carcinomas <1 cm was 20.9% in 1999 compared with 14.5% in 1989 (p<0.001); for tumours <2 cm the corresponding figures were 65.4% and 59.7% (p = 0.013). In 1999, 30.5% of tumours were histologically well-differentiated compared with 10.6% in 1989 (p<0.001), and 61.7% were lymph node negative in 1999 compared with 57.1% in 1989 (p = 0.006). Pure ductal carcinoma in situ (DCIS) constituted 10.9% and 7.9% of total cases of breast carcinoma in 1999 and 1989, respectively (p = 0.01). CONCLUSIONS: Quality of pathology reporting improved markedly over the period, in parallel with adoption of standardised synoptic pathology reports. By 1999, recording of important prognostic information was almost complete. Frequency of favourable prognostic factors generally increased over time, reflecting expected effects of mammographic screening.

Diagnosis of breast microcalcifications: a comparison of stereotactic FNA and core imprint cytology as adjuncts to core biopsy.

Stereotactic core biopsy (CB) using 14-gauge needles was adopted as the standard method of diagnosis of screen-detected breast microcalcifications (MC) at Sir Charles Gairdner Hospital in 1996. Fine needle aspiration (SFNA) was included as an adjunct, to optimise sensitivity and to provide immediate reporting. Recently, core imprint cytology (CI) has been shown to have a high sensitivity in diagnosing malignancy. The aims of this paper were to evaluate the accuracy of SFNA as an adjunct to CB, and whether CI could replace SFNA for immediate reporting in MC. Part A is a retrospective review of CB/SFNA of screen-detected MC from May 1998 to February 2000. A minimum of five cores was performed. SFNA samples were restricted to a maximum of three needle passes. Part B is a prospective study of CI from May to November 2000. In Part A, there were 406 MC in 353 women and 81 carcinomas were proven on excision. The complete sensitivity of CB for a diagnosis of malignancy was 97.5% and of SFNA was 65%. No false-positive diagnoses were made by either method. No extra carcinomas were detected using SFNA. In Part B, CB/CI were performed on 203 MC from 165 women. There were 38 carcinomas and 30 of these (79%) were diagnosed as malignant on CI. No false-positive diagnoses were made. The predictive value of a benign diagnosis was 95%. SFNA had little value as an adjunct to core biopsy in MC. CI promises to be useful in providing same day diagnosis for counselling purposes and for planning future surgery.

Can CD34 discriminate between benign and malignant hepatocytic lesions in fine-needle aspirates and thin core biopsies?

BACKGROUND: Distinguishing well differentiated hepatocellular carcinoma (HCC) from benign hepatocellular lesions is a well recognized problem in fine-needle aspiration (FNA) cytology. The endothelial cell marker CD34 is negative in normal hepatic sinusoids and stains vessels diffusely in HCC. This feature is useful in distinguishing benign from malignant hepatocytic lesions in histological specimens, although benign lesions may show focal positivity for CD34 confined to periportal and periseptal areas. In this study, we assess the role of CD34 in cell block and thin core biopsy material from benign and malignant hepatocellular lesions, and compare it with reticulin staining. METHODS: Cell blocks and thin core biopsies were assessed from 40 cases of HCC and 25 benign hepatocytic lesions. HCCs were scored for nuclear grade. Sections were stained for CD34 antigen and scored semi-quantitatively. Previously performed reticulin stains were reviewed. RESULTS: Thirty three of 40 HCCs (82.5%) showed diffuse positivity for CD34. The other seven cases showed either focal positivity (four cases), minimal positivity (two cases) or negative staining (one case). These results did not correlate with the nuclear grade of the tumor. Two of 25 benign cases (8%) showed diffuse positivity for CD34, 8 showed focal positivity, 11 showed minimal positivity, and 4 showed negative staining. All HCCs showed an abnormal reticulin pattern characterized by expanded trabeculae and islands, or sheets, with decreased or absent reticulin. All of the benign hepatocellular lesions showed a normal trabecular reticulin pattern. CONCLUSIONS: Diffusely positive CD34 staining is useful to support a diagnosis of well differentiated HCC, but in our study the reticulin stain distinguished more consistently between benign and malignant lesions.

Trends in the frequency and predictive value of reporting high grade abnormalities in cervical smears.

BACKGROUND: The "organized approach" to cervical screening in Australia includes standardized quality assurance measures for laboratories. This study examines changes in the frequency and the positive predictive value of reporting severe abnormalities in cervical smears over a 3-year period as a guide to the effects of implementing these measures. METHODS: The results of screening in 6-month periods from January 1995 to December 1997 were determined. Biopsy follow-up for results in the high grade epithelial abnormality ("HGEA") and "inconclusive: possible HGEA" categories was obtained from the Western Australian Cervical Cytology Registry (CCR). RESULTS: Approximately 40,000 smears were examined in each 6-month period. The frequencies of reporting HGEA were 0.47%, 0.59%, 0.79%, 0.85%, and 0.84%, and 0.91% for the study periods (P < 0.001). For the inconclusive category, they were 0.24%, 0.18%, 0.24%, 0.31%, 0.38%, and 0.35% (P < 0.001). Biopsy follow-up was available for 83. 9%, 80.5%, 89.9%, 92.4%, 93.1%, and 90.3% of the HGEA results and for 78.6%, 71.7%, 80.5%, 75.0%, 87.1%, and 85.9% of the inconclusive results over the study periods. The yield of high grade lesions for the biopsied cases was 82.6%, 82.3%, 83.1%, 79.5%, 80.9%, and 79% for HGEA cases and 58.2%, 41.9%, 60.6%, 52.8%, 47.5%, and 54.1% for inconclusive cases. CONCLUSIONS: There was a doubling in the reporting of HGEA results, whereas the positive predictive value for biopsied cases remained at about 80%. Reporting rates for inconclusive: possible HGEA cases also doubled, but the yield of biopsy-proven, high grade lesions remained at about 50%. These changes occurred in the absence of ancillary testing and with targeted rescreening methods. A high rate of reporting HGEA, in combination with a high positive predictive value, is among the most important indicators of cervical cytology laboratory performance. Large improvements in results may occur using conventional methods of quality assurance. Cancer (Cancer Cytopathol)

Cytodiagnosis of well differentiated hepatocellular carcinoma: can indeterminate diagnoses be reduced?

BACKGROUND: Distinction of well differentiated hepatocellular carcinoma (HCC) from benign hepatocellular lesions is a well recognized problem in fine-needle aspiration (FNA) cytology, sometimes leading to indeterminate reports. The aim of this study was to critically examine criteria that might allow definitive diagnosis in these cases. METHODS: FNA smears and cell blocks from 65 patients with primary hepatocellular lesions were reviewed. Seventy separate samples had been obtained. The initial reports in these samples were: HCC in 34, benign findings in 27, and indeterminate findings in 9. We defined architectural and cytological features seen in the malignant cases but not seen in the benign cases, including an assessment of reticulin in cell blocks. These criteria were then applied to the indeterminate cases. RESULTS: The most specific cytologic criteria of malignancy in well differentiated HCC were (i) numerous stripped atypical nuclei, (ii) macronucleoli, (iii) increased mitoses, and (iv) multinucleation. The most specific architectural criteria in smears were (i) widened trabeculae, (ii) well defined capillaries traversing tissue fragments, and (iii) solid islands of hepatocytes rimmed by endothelial cells. The most valuable architectural criteria in cell blocks were (i) trabeculae greater than two cells thick and (ii) reduced or absent reticulin framework. Using the above criteria a retrospective diagnosis of HCC was possible in eight of the nine indeterminate cases, all but one of which have subsequently been confirmed as malignant. CONCLUSIONS: Close attention to architectural features in both smears and cell blocks should allow most well differentiated HCCs to be diagnosed by FNA cytology. A reticulin stain should be part of the routine assessment of cell blocks. Cancer (Cancer Cytopathol)

The "Inconclusive--possible high grade epithelial abnormality" category in Papanicolaou smear reporting.

BACKGROUND: The Australian Terminology for Cervical Cytology Reporting includes the category "Inconclusive-Possible high grade epithelial abnormality." METHODS: The frequency of use of this category, the types of associated cell patterns, and the yield of high grade lesions at biopsy were studied. RESULTS: One hundred and two cases categorized as "Inconclusive" were reported between January and June 1995, representing 0.24% of 41,712 Papanicolaou (Pap) smears screened. The abnormal cells were reported as squamous in 74.5% of cases, endocervical in 4.9% of cases, endometrial in 3.9% of cases, and indeterminate in 16.7% of cases. The main cellular patterns included disorganized groups of hyperchromatic squamous, glandular, or indeterminate cells (64.2% of cases) and atypical metaplastic squamous cells (28.4% of cases). Cell preservation was suboptimal. In 25.3% of cases the cells were highly degenerate or air-dried. Follow-up included biopsy (84.3% of cases), colposcopy alone (7.8% of cases), and repeat Pap smears without any detected abnormality (3.9% of cases). No follow-up was available in 3.9% of cases. High grade abnormalities were found in 66.3% of the biopsied cases and 55.9% of the total cases (48 cervical intraepithelial neoplasia [CIN] of Grade 2 or 3; 2 squamous cell carcinomas; 3 endocervical adenocarcinoma in situ [ACIS]; 3 adenocarcinomas of endocervical, ovarian, and endometrial origin; and 1 endometrial stromal sarcoma). In 16.2% of cases a low grade squamous lesion was present on biopsy (CIN, Grade 1 or human papillomavirus effect); and no lesion was found in 17.4% of cases. CONCLUSIONS: The "Inconclusive" category was not overused, and gave a high yield of biopsy abnormalities. Accepting uncertainty in the diagnosis of some high grade lesions reduces their likelihood of being classified incorrectly as reactive changes, ignored because of poor cell preservation, or lost in the larger group of classifications such as atypical cells of undetermined significance, borderline nuclear abnormality, or non-specific minor changes.

Detection of p53 gene mutation by rapid PCR-SSCP and its association with poor survival in breast cancer.

We examined the association between mutation of the p53 gene and survival in a large cohort of breast cancer patients. Using a rapid, non-isotopic single-strand conformation polymorphism (SSCP) method we screened for mutations in exons 4-10 of the p53 gene in 375 primary breast cancers from patients with a median follow-up of 57 months. Mutations were found in 19% of tumours. Statistically significant associations were found between p53 mutation and histological grade, hormone receptor status, ploidy and S-phase fraction. No association was found between p53 mutation and axillary lymph node involvement, histological type, tumour size, vascular invasion or patient age. In univariate survival analysis, p53 mutation was strongly associated with poor prognosis. This was maintained in the lymph node-negative and hormone receptor-positive patient subgroups. In multivariate analysis, p53 mutation was associated with poor survival independent of lymph node status, estrogen receptor status and S-phase fraction. Our results demonstrate the feasibility of using a rapid and simple polymerase chain reaction-SSCP screening procedure to detect p53 gene mutation in breast cancer for the provision of prognostic information.

Stereotactic 14 gauge core-biopsy of the breast: results from 101 patients.

BACKGROUND: Along with fine needle aspiration (FNA) cytology, core-biopsy has become an integral part of the assessment of mammographically detected breast lesions. METHODS: A series of stereotactic large-core-biopsies of mammographically detected breast lesions was studied to assess the accuracy and limitations of the technique in diagnosing malignancy and in giving specific benign diagnoses, and its use in determining surgical management. RESULTS: Eighty per cent of carcinomas were diagnosed as malignant (absolute sensitivity). In 88.8% of the cancers, the core-biopsy was classified as malignant, suspicious or atypical/indeterminate (complete sensitivity), and in 72% of the invasive carcinomas, invasive tumour was present in the core. The technique was more successful for invasive carcinomas than for ductal carcinoma in situ (DCIS) (absolute sensitivity 86.1 and 55.5, respectively; P = 0.28) and for malignant mass lesions than for a mass with associated microcalcifications or for pure microcalcifications (absolute sensitivity 91, 71 and 66.6%, respectively; P = 0.19). In five of the 45 cancers (11.1%), no tumour tissue was present in the core, but all were excised after mammographic review and no delays in diagnosis have been experienced to date. The benign to malignant ratio for excised lesions was 0.11:1. Of the benign lesions, a specific diagnosis was given in 49% (calcifications in the core in a background of fibrocystic change, or postoperative scarring, or fibro-adenoma); the remainder showed non-specific benign findings. All patients where invasive carcinoma was diagnosed in the core underwent axillary clearance and wide local excision or mastectomy at their first operation. CONCLUSIONS: This technique can markedly reduce the number of benign lesions needing open biopsy, and provide information allowing definitive management of most carcinomas at the first operation. The accuracy of core-biopsy was lower in DCIS/microcalcification lesions; extra core samples or a combination of FNA and core-biopsy may be of value in these cases.

Predicting axillary lymph node metastases in breast carcinoma patients.

Routine axillary dissection is primarily used as a means of assessing prognosis to establish appropriate treatment plans for patients with primary breast carcinoma. However, axillary dissection offers no therapeutic benefit to node negative patients and patients may incur unnecessary morbidity, including mild to severe impairment of arm motion and lymphedema, as a result. This paper outlines a method of evaluating the probability of harbouring lymph node metastases at the time of initial surgery by assessment of tumour based parameters, in order to provide an objective basis for further selection of patients for treatment or investigation. The novel aspect of this study is the use of Maximum Entropy Estimation (MEE) to construct probabilistic models of the relationship between the risk factors and the outcome. Two hundred and seventeen patients with invasive breast carcinoma were studied. Surgical treatment included axillary clearance in all cases, so that the pathologic status of the nodes was known. Tumour size was found to be significantly correlated (P < 0.001) to the axillary lymph node status in the multivariate anlaysis with age (P = 0.089) and vascular invasion (P = 0.08) marginally correlated. Using the multivariate model constructed, 38 patients were predicted to have risk of nodal metastases lower than 20%, of these only 4 (10%) patients had lymph node metastases. A comparison with the Multivariate Logistic Regression (MLR) was carried out. It was found that the predictive quality of the MEE model was better than that of the MLR model. In view of the small sample size, further verification of this model is required in assessing its practical application to a larger population.

Entropy maximization networks: an application to breast cancer prognosis.

Describes two artificial neural network architectures for constructing maximum entropy models using multinomial distributions. The architectures presented maximize entropy in two ways: by the use of the partition function (which involves the solution of simultaneous polynomial equations), and by constrained gradient ascent. Results comparing the convergence properties of these two architectures are presented. The practical use of these two architectures as a method of inference is illustrated by an application to the prediction of metastases in early breast cancer patients. To assess the predictive accuracy of the maximum entropy models, we compared the results with those obtained by the use of the multilayer perceptron and the probabilistic neural network.

Breast cancer in Western Australia in 1989: IV. Summary of histopathological assessment in 655 cases.

This study was part of a population-based survey of all cases of breast cancer diagnosed in Western Australia in 1989. The paper concerns histopathology reporting by pathologists in 655 cases of carcinoma of the breast in that year, before the introduction of mammographic screening programmes. Pathological features of the neoplasms are documented, and the extent to which information known to be of clinical or prognostic importance was included in the reports is analysed. 96.5% of all pathology reports included information on breast cancer subtype and, in 98.6% of cases with axillary dissection, the number of lymph nodes dissected, and the number containing metastatic tumor was stated. In 83.7% of cases of invasive carcinoma exact tumor dimensions were recorded. In 44.9% of cases histological grade was recorded, and information about excision margins was present in 60% of reports overall. The reporting of pathological features in many instances was limited by the way in which the specimen was handled prior to reception. At the time of the study, views about the importance of many aspects of histological assessment were still evolving. Even now, for example, consensus is still being reached on the value of histological grading in predicting prognosis and whether reliable histological assessment of such factors as extent of DCIS and completeness of excision of DCIS is possible. The introduction of mammographic screening since 1989 has provided a focus for wider discussion about the value of histological information in prognostication and patient management. A case is made to support the use of "check lists" for surgical pathology reports in cases of breast cancer.

Evaluation of the expression levels of nm23-H1 mRNA in primary breast cancer, benign breast disease, axillary lymph nodes and normal breast tissue.

Expression levels of nm23-H1 were evaluated in a variety of normal benign and malignant breast tissues by Northern and slot blot. Tissues from 153 patients presenting with palpable breast lesions were studied: 132 primary infiltrating breast cancers, 9 pure duct carcinoma in situ lesions, a phyllodes tumor, 9 benign lesions and 2 local recurrences of carcinoma. In addition to lesional tissue, 49 samples of macroscopically normal breast tissue, 37 axillary lymph nodes and 9 samples from patients undergoing cosmetic reduction mammoplasty were studied. Sets of normal breast tissue, primary tumor and lymph node tissue from individual patients were available for comparison in 37 cases. A wide range of gene expression was detected in the various tissue types. The highest levels of expression were detected in malignant samples with in situ carcinomas being associated with the highest levels of gene expression. The expression levels of nm23-H1 in normal breast tissue were lower than the corresponding tumors from the same patients (p < 0.0005). Benign breast lesions (including 6 fibroadenomas) had levels of gene expression approximating those of the normal tissue samples. Normal axillary lymph nodes had significantly lower levels of nm23-H1 expression than nodes with metastatic deposits (p < 0.03). No significant association was observed between nm23-H1 expression levels and axillary node status in patients with infiltrating carcinoma, although there was a slight trend toward lower nm23-H1 mRNA levels in the node negative group.(ABSTRACT TRUNCATED AT 250 WORDS)

c-erbB-2 amplification in breast cancer: detection in formalin-fixed, paraffin-embedded tissue by in situ hybridization.

We describe a sensitive and practical in situ hybridization method, using a digoxigenin-labeled probe, for the detection of c-erbB-2 amplification in breast cancer in formalin-fixed, paraffin-embedded tissue sections. Forty-six primary breast carcinomas were studied. Nuclear hybridization signal was observed in 36 of 46 carcinomas. Signal was confined to malignant cells. Normal breast epithelium and stromal and inflammatory cells were uniformally negative. DNase predigestion, no-probe preparations, and competitive hybridization confirmed the specificity of the reaction. The hybridization reaction was localized to multiple discrete foci in tumor cell nuclei, suggesting multiple sites of gene copy and transcriptional activity in the nucleus. Considerable cell-to-cell variation in hybridization signal was evident within individual tumors and positive reactions were observed in several cases in which amplification could not be detected by either Southern or slot blot analysis. The high sensitivity and specificity of the reaction and its use in a tissue-based system will allow the study of a range of possible precursor lesions of breast cancer for evidence of c-erbB-2 amplification.

What's new in breast cancer? Molecular perspectives of cancer development and the role of the oncogene c-erbB-2 in prognosis and disease.

The oncogene c-erbB-2 is frequently amplified in human breast carcinoma. The c-erbB-2 gene is present as a single copy in normal cells, and has been mapped to chromosome 17 in the region 17q 12-21.32. c-erbB-2 encodes a transmembrane glycoprotein known as p185. The intracellular component of p185 has tyrosine kinase activity; the extracellular domain has a structure resembling a growth factor receptor. c-erbB-2 amplification, p185 overexpression and levels of transcribed c-erbB-2 specific messenger RNA have been studied in a large number of breast carcinomas using a variety of techniques. In general, overexpression of p185 oncoprotein reflects various levels of DNA amplification, though in some cases amplification can be detected in the absence of overexpression of p185 and similarly overexpression of p185 can be present without detectable levels of c-erbB-2 amplification. This findings suggests that multiple mechanisms may be responsible for overexpression. c-erbB-2 amplification and/or overexpression occurs in almost all cases of high grade duct carcinoma in-situ, but has been reported in only 10%-40% of infiltrating duct carcinoma. c-erbB-2 amplification or overexpression occurs rarely in invasive lobular carcinoma, and has not been detected in ductal or lobular epithelial hyperplasia, or in atypical ductal or atypical lobular hyperplasia. It is generally believed that c-erbB-2 amplification/overexpression is an important independent prognostic indicator in breast carcinoma, identifying a subset of patients with poor prognosis tumours, particularly if axillary node metasases are present. However, many unanswered questions remain regarding c-erbB-2 and its role in breast cancer development and progression. The causes of c-erbB-2 amplification are unknown. There is no evidence of mutations in the human gene which might cause amplification or overexpression. The significance of the differences in levels of c-erbB-2 amplification/overexpression in in-situ duct carcinoma and associated invasive duct carcinoma has not been established. Amplification or overexpression have not been reported in atypical duct hyperplasia, a proposed precursor of duct carcinoma in-situ, yet overexpression occurs almost always in high grade duct carcinoma in-situ. c-erbB-2 may play a critical role in the development of a clonal in-situ, proliferation of high histological grade, yet does not obviously influence the acquisition of an invasive phenotype. We would postulated that this instability in amplification/overexpression is of biological significance, and if better understood may aid in the study of progression of human breast carcinoma.

Stromelysin-3 expression in breast cancer biopsies: clinico-pathological correlations.

Stromelysin-3 expression was studied by Northern blotting in 222 tissue samples including primary and metastatic breast carcinoma and normal breast tissue. Uninvolved breast tissue from mastectomy specimens, normal breast tissue from reduction mammoplasties and normal lymph nodes did not contain stromelysin-3 mRNA. About 62% of primary and metastatic breast carcinomas, but only 1 of 10 in situ ductal carcinomas, expressed stromelysin-3. Stromelysin-3 mRNA was found more often in estrogen-receptor-positive carcinomas and in histological grade-1 carcinomas. There was no significant correlation between stromelysin-3 expression and other prognostic factors, including tumor size, lymph-node involvement, age of patient, vascular invasion and cathepsin-D.

Breast cancer in Western Australia in 1989. I. Presentation.

This study was designed as a population-based study of all cases of breast cancer diagnosed in Western Australia (WA) in 1989. Cases were identified from the State Cancer Registry and from computerized hospital inpatient records. Data were obtained from the records of surgeons and oncologists managing the patients, hospital medical records, and pathology and cytology reports. A total of 701 histologically proven tumours were documented in 692 women. Of these 6.8% were not known to the State Cancer Registry. Two-thirds (68%) of tumours were first detected by the woman herself, 11% were found by a doctor and 11% were detected by mammographic screening. Stage I tumours accounted for 40% of tumours and Stage II 39%. The estimated lifetime risk of a WA woman developing at least one malignant breast tumour is 10%. Passive surveillance based upon a legal obligation on doctors to notify cases of cancer may be resulting in a significant under-estimation of the incidence of cancer in WA. Mammographic screening played only a small role in the detection of breast cancer in WA in 1989, but its contribution and the proportion of stage I tumours should both increase as a population-based mammographic screening programme is established. This survey will provide a yardstick against which changes can be measured. Eighty-four per cent of tumours presently occur in women who would have access to mammographic screening although only 44% occur in the 50-69 age bracket that is to be actively recruited. The lifetime risk of breast cancer in WA women is greater than has been appreciated previously.