Mammography. Until 1960 diagnosis of mammary carcinoma depended on discovery of a breast mass by physical examination, followed by a biopsy of the lesion. It has been said that, with experience, carcinoma as small as 1 cm may be regularly detected by palpation. After 1960, x-ray study of the breast (mammography) began to receive considerable attention. Several favorable reports have been published, and several mass screening surveys have been attempted. To date, available information on the status of mammography includes the following:

1. Breast carcinoma can be visualized on mammography in some cases (23%–42%) in which it is not palpable. Reports indicate that 9%–42% of visualized nonpalpable lesions are malignant, with an incidence of lymph node metastasis of 0%–38%.
2. Screening surveys utilizing mammography are reporting detection of 2-3 times the expected rate of breast carcinoma.
3. Proper technique is of the utmost importance; this calls for special training and conscientious technicians.
4. Mammography is definitely not infallible. The average good radiologist will probably miss a malignant diagnosis in about 15% of cases (range, 4%–44%) and call a benign lesion malignant in about 5%–10% (range, 4%–15%). Reported incidence of palpable lesions not visualized on mammography ranges from 5%–20%. Biopsy or some other type of tissue diagnosis is still essential for all breast lesions.
5. Mammography is best in the postmenopausal or large breast where fatty tissue predominates. In these circumstances, probably 80%–90% of malignant tumors can be diagnosed correctly, whereas the figure decreases to 55% for women under age 45 years.
6. Mammography is useful for indicating the site for biopsy when several breast masses are present, for demonstrating unexpected additional foci of tumor elsewhere in the breast, and for detecting unsuspected tumor in the opposite (contralateral) breast. Mammography has demonstrated simultaneous tumor in the contralateral breast in 2%–3% of patients and eventual development of contralateral breast carcinoma in 6%–8% of patients. However, not all such tumors are detected by mammography. Pathology studies on mastectomy specimens have disclosed multiple foci of carcinoma in the same breast in 20%–30% of cases (literature range, 13%–75%), and biopsies of the contralateral breast have detected invasive carcinoma in about 1%–2% (0.5%–16%) and lobular carcinoma in situ in about 20%–30% (10%–53%).
7. At present, mammography is not an ideal screening procedure, because only a limited number of satisfactory studies can be performed daily under present conditions in the average radiology office.

Radionuclide bone scan. When the diagnosis of breast cancer is first made or suspected, the question may arise as to which tests provide useful information that might influence type or extent of treatment. Bone scan detects lesions on initial workup in about 5% of clinical stage I lesions (literature range, 0%–30%), about 10% of clinical stage II lesions (0%–43%), about 8% of combined stage I and II lesions (1%–40%), about 25% of clinical stage III lesions (0%–62%), and about 15% of all lesions (4%–48%). Also, about one half of patients with solitary bone scan abnormalities have no demonstrable tumor at the abnormal site, the scan changes being due to benign abnormalities of bone. The incidence of overall false positive scan findings in patients with breast carcinoma is about 10% (3%–57%). About 10% (4%–33%) of breast carcinoma metastases to bone seen on x-ray film will be missed by bone scan, because breast carcinoma produces a relatively high number of osteolytic lesions, which are not as frequently seen by bone scan as are osteoblastic lesions. Bone scan detects about 20% (10%–40%) of metastases not seen on x-ray film. The small amount of information available on results of initial-visit brain scanning suggests that fewer than 5% will be abnormal if there are no neurologic signs or symptoms. Surprisingly few data are available on the contribution of liver scan to initial (pretherapy) workup, but one study found that liver scan yielded only 1% true positive results.

Fine-needle aspiration. Another diagnostic modality is fine-needle (22-gauge) aspiration of breast masses with cytologic examination of the aspirate. Studies have shown about 1% (range, 0%–3%) false positive results and about 5%–10% (range, 0%–24%) false negative results. The procedure gives excellent results when the person doing the aspiration and the cytologist are experienced. Reliable interpretation requires special training. A definite cytologic diagnosis of malignancy is much more helpful than a diagnosis of benignity due to the significant rate of false negative results. Whether this procedure should replace surgical biopsy with frozen section is controversial. In cases where the patient refuses biopsy, when surgery cannot be performed, or when the lesion is cystic, there is not the same controversy.

Prognostic tests in breast carcinoma

The earliest prognostic factors were established from gross and microscopic examination. Presence or absence of palpable axillary nodes (and to a lesser extent, the number of nodes containing metastases), tumor size (cutoff point, 2.0 cm diameter), and tumor nuclear grade (size, shape, and degree of chromatin abnormality determining low or high grade) were (and still are) important independent prognostic indicators. Later, various laboratory tests were tried in hopes of further improving accuracy either alone or in combination.

Estrogen receptor assay. Estrogen receptor assay (ERA) is widely used as an aid in selection of breast cancer therapy. It has been shown that approximately 30%–40% of postmenopausal breast cancer patients respond to oophorectomy, adrenalectomy, hypophysectomy, or estrogen therapy. Premenopausal patients respond less frequently. Certain tissues, such as endometrium, vagina, and breast, have been shown to respond to estrogen stimulation of estrogen receptor sites within their epithelial cells. ERA techniques most frequently used involve tissue slices or cytoplasm (cytosol) extracts from the tumor to be evaluated. Radioactive estradiol and another estrogen-binding substance are added, and the tumor estrogen receptors compete with the estrogen binder for the labeled estradiol. The amount of labeled estradiol bound by the tumor estrogen receptors can then be determined. This type of assay is an estimate of the number of unoccupied (“active”) estrogen-binding sites. Immunoassays of several types are also becoming available. This technique estimates total quantity of estrogen receptors, including both active and inactive receptors. According to current information, about 60%–65% (literature range, 30%–80%) of primary breast carcinomas and about 45%–50% of breast carcinoma metastases have tumor cells with estrogen receptors that bind sufficient estrogen per unit of cancer tissue to be considered estrogen receptive, or “positive.” ERA positivity does not show satisfactory correlation with presence or absence of lymph node involvement or the degree of differentiation of the tumor. Breast carcinomas in postmenopausal women are more likely to have estrogen receptors than those in premenopausal women. About 60%–70% of women whose tumors are estrogen-receptor positive respond to hormonal manipulation (estrogens, antiestrogens, endocrine ablation, or androgens). About 5%–10% of those considered ERA negative will respond (some laboratories report < 5%). Thus a negative ERA result is interpreted by many as an indication that chemotherapy is more likely to be effective than endocrine therapy. In general, metastases tend to share the same type of receptor status as the primary tumor. However, some investigators report that metastases from estrogen-treated primary tumors are frequently receptor negative.

Certain laboratory aspects of the test are important. There are several techniques for preparing tissue for the cytosol method for receptor assay, and some of the techniques differ significantly in the number of patient tumors that demonstrate apparently increased receptors. At present, it is necessary to secure at least 1 gm of tumor after eliminating all fat and normal breast tissue. The tumor must be fresh and must be frozen by dry ice within a short time (preferably within 15 minutes) after excision. The specimen must be kept on dry ice and sent to the reference laboratory on dry ice. The estrogen receptors are very labile, and failure to quick-freeze and provide adequate low temperatures produces false negative results. In addition, quality control surveys have shown considerable variation among laboratories in ability to detect low concentrations of estrogen receptors.

As the description of this test indicates, with current techniques the procedure is partially a bioassay and therefore is available only at larger institutions or reference laboratories.

Immunocytochemical estrogen receptor methods. Several investigators have developed immunologic methods to visually demonstrate presence or absence of estrogen receptors in cells within fixed tissue microscopic sections or cytology smears. Although true receptor quantitation cannot be done, this technique permits visualization of receptor distribution, that is, how many tumor cells are visually positive or negative. Many carcinomas do not have a cell population that is uniformly estrogen receptor positive or negative. Studies to date have reported about 80%–85% (range, 60%–90%) correlation with standard ERA results.

Progesterone receptor assay. Progesterone receptors can be assayed (PRA) on the same tumor tissue in a manner similar to estrogen receptors. In general, demonstration that a breast carcinoma is PRA positive adds about 10%–15% more likelihood to ERA positivity that the tumor will respond to hormonal manipulation. Thus, tumors that are both estrogen- and progesterone-receptor positive have about a 70%–80% chance of responding to hormonal therapy. Those negative for both receptors have less than a 10% chance of responding to hormones or antihormones. Eighteen percent to 49% of patients are ERA positive but PRA negative and have about a 30% chance of responding. Three percent to 13% of patients are ERA negative and PRA positive. PRA can be performed on formalin-fixed paraffin-embedded microscopic slides using immunohistochemical stains similar to those used for ERA. As in ERA, this technique does not provide a quantitative answer.

DNA ploidy. DNA (deoxyribonucleic acid) ploidy is a measure of total nuclear DNA content, usually performed by flow cytometry. About 60% of breast cancer overall are aneuploid and 40% are diploid. The amount of nuclear DNA in the active cell stage of DNA synthesis (S-phase) can be calculated and is reported as the S-phase fraction. This is a parameter of cell proliferative activity. There is now some controversy whether DNA ploidy (aneuploid vs. diploid status) provides significant prognostic information. Whereas many of the earlier studies reported that diploid breast carcinomas had significantly or considerably better prognosis than aneuploid ones, some more recent studies do not confirm this or do not find that ploidy is a significant independent risk factor. Increased S-phase activity is somewhat better accepted as an unfavorable prognostic sign. Unfortunately, SPF is technically more difficult to measure accurately and is less standardized between laboratories.

Cathepsin D. Cathepsin D is a protease enzyme found in lysosomes of cells in many tissues. In breast cancer, it appears to be regulated by estrogen. Studies using cytosol (tissue extracts) from breast cancer found that increased quantity (tumor “overexpression”) of Cathepsin D predicted shorter disease-free survival and worse overall prognosis. However, more recent studies using monoclonal antibody immunohistologic stains on routine formalin-fixed and processed tissue microscopic slides found that presence of Cathepsin D staining of tumor cells, especially with strong staining, predicted longer disease-free survival and better overall prognosis (the opposite from cytosol-based studies).

C-erbB2 oncogene amplification. C-erbB2 (HER-2/neu or HER-2) protooncogene is a gene involved with cell growth. When the gene increases (or causes to increase) production or associated C-erbB-2 protein, the gene is said to be amplified (or overproducing). This amplification has been reported in about 10%–30% of most human adenocarcinomas. The reported frequencies from individual site adenocarcinomas can be influenced by several factors: whether the gene itself is identified (usually by nucleic acid probe techniques) or the gene protein is detected (usually by immunohistochemistry), whether frozen sections (fresh tissue) or formalin-fixed paraffin-embedded sections from routine tissue slide examination are used, whether the antibody recognizes the external or internal part of the molecule, and what criteria are used to define a positive result. In breast cancer, formalin-fixed paraffin-embedded slides generally show fewer positive cells (20%–30%; range, 9%–80%) than fresh-frozen tissue (22%–49%). C-erbB2 amplification, in general, correlates with negative estrogen receptor status, higher tumor grade, and higher probability of aneuploidy, therefore suggesting poorer prognosis. However, the majority of studies found status of axillary lymph nodes to be a more powerful independent risk factor than c-erbB2 amplification.

Cell proliferation markers. In general, the more rapidly growing a tumor is, the more aggressive it is, and the prognosis becomes worse. Therefore, various indices of cell proliferation have been proposed to help estimate prognosis. One of the first was the nuclear mitotic count (or index), based on the number of mitoses per (microscope) high-power field (generally at 400x magnification). In breast cancer, this was found to have some correlation with tumor differentiation and therefore with prognosis, but prognostic usefulness was not as good as could be obtained using other tests. The SPF from cell cycle DNA flow cytometry has been discussed. This has proved (under the right conditions) to be a useful and helpful prognostic indicator. Ki-67 is a monoclonal antibody that detects a protein in cell nuclei that appears only in the growth phase of the cell cycle. In certain tumors, including breast carcinoma, abnormal quantity of Ki-67 (Ki index, by immunostaining of microscopic slides from tumor areas) correlates with less differentiated tumors, larger tumor size, increased p53, and less favorable prognosis, and to a lesser degree with negative estrogen and progesterone receptor status. There is disagreement concerning correlation with axillary node metastases.

The p53 assay. The p53 gene is a tumor suppressor gene similar to the retinoblastoma suppressor gene. Mutation of the p53 gene results in production of an altered protein that cannot function normally and may actually promote cell growth. Normally, p53 cannot be detected in breast tissue using immunohistologic stains. In breast carcinoma, about 25% of patients have detectable p53 nuclear protein. This correlates with increased cell proliferative activity and to some extent with lack of estrogen receptors.

Summary of breast carcinoma prognostic tests

At present, axillary lymph node status, tumor size, and estrogen and progesterone receptor assay are still by far the most widely used and accepted prognostic indicators. Nuclear grade could be more widely used if uniform criteria for interpretation were established. S-phase fraction would probably become more important if uniform methodology and interpretation were agreed upon. Of the other current contenders, Cathepsin-D and c-erbB2 oncogene “expression” seem to be the most likely possibilities for at least limited use. However, the prognostic test area can change rapidly.