Medical Device Link.

Originally Published IVD Technology March 2003

Assay Development

A new assay may offer a universally-applicable method of assessing breast cancer risk. 

Jeanne Ohrnberger

The mammastatin serum assay (MSA) may improve current methods of breast cancer risk assessment and screening.

Breast cancer is one of the most common cancers among women, second only to skin cancer. In 2001, an estimated 192,200 new cases of invasive breast cancer were diagnosed in the United States, and 40,200 women died from this disease. One in 8 women in this country will develop breast cancer in her lifetime. Only lung cancer claims more women’s lives each year.¹

Mammography and breast examination are commonly used to detect an abnormal breast mass. Once detected, a mass is biopsied to rule out or confirm a breast cancer diagnosis. Though it is the most powerful tool for breast cancer detection, mammography is not entirely reliable, as it yields a false-negative rate of 20% and a false-positive rate of 25–30%, depending upon the age of the patient.² 

The BRAC1/2 markers have enabled women with a family history of cancer to analyze their risk of developing breast cancer. However, this risk-analysis tool addresses the needs of only part of the population that will develop breast cancer. Traditional tumor markers like CA-15.3 and CA-27.29 do not fill this void either, as they are used primarily for monitoring cancer progression and recurrence. 

A tool that could assess the risk of spontaneously developing breast cancer for the majority of potential patients would be of great value to the medical field. It would alert these patients to their risk of developing the disease, influence women at risk to receive an annual mammogram, and aid physicians in interpreting mammograms. 

The serum assay to mammastatin, a protein indicating breast health, is such a tool and can be used to provide information about breast cancer risk. This article examines the development of this assay and its diagnostic applications.
Cancer Cell Growth Inhibition 

Mammastatin was first described as a tissue-specific growth-inhibitory protein in 1989.³ It was first isolated from normal human mammary cells (NHMCs) grown in a laboratory tissue culture. The protein was detected inside the cells, and had secreted into the culture media. 
Mammastatin has since been identified in tissue sections within the epithelial cells lining the lumen of the mammary duct. It has not been found in human breast milk, but has been detected in human serum, suggesting basolateral secretion. Three species of the protein, with molecular weights of 44, 49, and 53 kD, have been identified by polyacrylamide gel electrophoresis. 

The growth of breast cancer cells is inhibited in culture when a medium containing mammastatin is applied to them. This inhibition is specific to breast cancer cell lines. The 53-kD species of mammastatin is necessary for maximal growth inhibition. 

Figure 1. The process of analyzing a sample with the mammastatin serum assay allows for an efficient and minimally invasive method of breast cancer risk assessment. (Click to enlarge)

The 53-kD species has never been detected in any breast cancer cell line or tumor tissue lysate. However, the protein has been identified by Western blot in all isolates of normal cells. When examining breast tissue from breast cancer patients, neither tumor cells nor normal cells contain the 53-kD mammastatin protein, indicating a down-regulation of the protein throughout the breast when cancer is present.⁴

Technology Behind the Test

The mammastatin serum assay (MSA) detects the level of mammastatin in serum. A colorimetric indirect immunoassay, the MSA was developed in a dot blot format. Critical components, including the mammastatin calibration curve and the primary antibody, are produced and purified in-house.

The calibration curve is generated by in vitro–grown NHMCs. A medium conditioned with the 53-kD secreted protein is used as the calibration standard. Following processing, each lot of NHMCs is assayed for mammastatin content and tested for performance. 
The first mammastatin monoclonal antibody was identified by its ability to block mammastatin activity. Subsequent monoclonal antibodies were made by immunizing Balb C mice with protein purified by antibody-affinity chromatography. The antibodies were then identified by their ability to recognize the same proteins as the original antibody, as determined by Western blot. 

The 7G6 monoclonal antibody, which recognizes mammastatin with the greatest specificity, is the primary antibody used in the MSA. The capabilities of this antibody were identified by screening immunoassays with mammastatin, comparing recognition patterns, and examining immunoprecipitating activity. Antibody 7G6 is produced by in vitro tissue culture and is then purified. Commercial lots of human serum are screened for acceptance as internal controls of high mammastatin (>7.0 relative density units [RDU]) and low mammastatin (<4.0 RDU) for the assay.

The MSA analyzes serum by housing the sample on a nitrocellulose membrane and then probing it with antimammastatin antibody (see Figure 1). Horseradish peroxidase–labeled antimouse antibody is used in conjunction with the colorimetric substrate 3,3'-Diaminobenzidine to visualize the reaction on the membrane.
 

Table I. Description of current methods of risk assessment, detection, and diagnosis. (Click to enlarge)

A calibrated imaging densitometer scans the membrane and quantifies the RDU, a measure of the color intensity of the dots on the membrane. The data from the densitometer is used to generate a standard curve from 0–25 RDU. The mammastatin concentration is then determined by comparing the scanned serum samples with the calibrated standard curve. 

Assessing Breast Cancer Risk

The MSA is not meant to be a stand-alone cancer diagnostic. It may be used in conjunction with standard mammography and patient history to help physicians make decisions about how to manage the care of their patients. The presence of mammastatin may indicate, with high probability, that a patient does not have an elevated risk of developing breast cancer. 

Mammograms, ultrasound, and related advancements in imaging and analysis are excellent tools for identifying an abnormal growth, but before such a growth exists, they offer no value. Although most women have access to these tools, they are underutilized by many, including older and economically challenged women, who truly need them.¹

Mammograms are useful for detecting masses, but without any complementary source of information, they often put unnecessary burdens on patients. Approximately 5% of screening mammograms are positive or suspicious. Of these positive or suspicious reports of a malignant tumor, 80–93% are false-positives.⁶ The high rates of both false-positive and false-negative mammogram results lead to extra costs for follow-up mammograms, biopsies, and the intensive treatment required for late-stage cancers resulting from missed diagnoses.

Table II. Initially, the difference between the mean mammastatin levels of participants with a positive family history of breast cancer and those with no family history of the disease decreases as age increases. However, this trend reverses for women 60 and older. (Click to enlarge)

Generally, the only way to determine whether a mass is malignant is through a relatively expensive biopsy. Up to 10% of women receiving follow-up mammograms require biopsies, and 80% of those biopsies will indicate that the tumor is benign.1 Based on these data, the estimated number of biopsies revealing a false-positive mammogram ranges from 240,0001 to 1.4 million per year.⁶ 

Using the more conservative estimate of 240,000 to determine the annual cost of these biopsies yields a figure of nearly $320 million. Applying MSA technology at the risk-analysis stage could save a good deal of this money by alerting practitioners to the reduced risk of breast cancer in patients with high mammastatin levels. 

The MSA may offer the advantages of early detection, a low error rate, and low costs when combined with current detection methods. Additionally, developments in imaging and analysis are improving detection and monitoring technology (see Table I), and new biological cancer markers are helping to determine prognosis after the disease is diagnosed.  Yet, a more accurate risk assessment tool is needed. 

This need is borne out when considering that family history, personal history, and Gail model calculations are the only tools available for risk assessment. Further, women at a high risk for developing breast cancer, due to family and personal history, are generally aware of being at increased risk.⁵ Thus, tools such as the Gail model do not generally add value to the risk information that is already known. Since the MSA technology uses an entirely new and biologically based parameter to assess risk, it may contribute valuable and complementary information about both patients with a known risk for breast cancer as well as those with no known risk. 

Assessing the Effectiveness of the MSA 

Current screening efforts suffer from a lack of compliance, with an estimated 25-40% of women in the United States participating in recommended screening programs. Since the MSA is a simple blood test, and is less embarrassing, awkward, and costly than a mammogram, it should be more easily tolerated by most women, and it could encourage higher compliance with screening efforts. 
As an initial screening tool, the test could funnel women at high risk who truly need a mammogram into programs of more-intensive screening. The MSA has a negative predictive value of 97–98%, suggesting that no further diagnostic testing beyond the recommended standard screening is necessary for women with no known risk for breast cancer and normal serum mammastatin levels. Low serum mammastatin may suggest a risk of breast cancer, even in women with no other known risk factor.

Two clinical trials of the MSA have been conducted. The goals of these trials were to evaluate the assay and determine its diagnostic parameters. The first study was a double-blind, single-site, nonrandomized trial comparing serum mammastatin concentrations in women without breast cancer with those of breast cancer patients. Additional data were collected to analyze associations between potential baseline prognostic factors and mammastatin concentrations. 

The cohort of women recruited for this study was divided into two groups. One group consisted of patients who had been diagnosed with stage I, II, or III breast cancer. The other group consisted of “healthy” patients, who had no family history of breast cancer, or had experienced only benign breast disease. Family history was defined as having a first-degree relative with breast cancer. Benign breast disease patients were placed in the healthy group, because their MSA levels (mean MSA = 17.0 RDU) were comparable with those of women with no history of breast cancer (mean MSA = 15.4 RDU).

Several factors precluded women from inclusion in the study. Several women did not submit data detailing their family history of breast cancer, and were excluded from the study (N = 17). Women with breast cancer who were in remission when the serum samples were collected (N = 16, mean MSA = 6.3 RDU), were also excluded because their MSA levels were expected to reflect a different disease process than those of women with active cancer.

Two breast cancer patients with advanced (stage IV) disease had a mean MSA of 19.2 RDU, but were excluded from analysis due to concerns of tumor lysis and release of cellular material. Patients whose disease status was unknown were also excluded from analyses (N = 5). The group of women whose data were used in the final analyses consisted of 38 breast cancer patients and 210 healthy patients, resulting in a total sample size of 248 women.

The ability of the MSA to distinguish serum samples given by healthy patients from those of breast cancer patients was evaluated using receiver operating characteristic (ROC) curve analysis. Sensitivity was calculated as the percentage of women with breast cancer that tested low in mammastatin. Specificity was calculated as the percentage of women without breast cancer that tested high for mammastatin.
By ROC curve analysis, the highest sum of sensitivity and specificity was at an MSA value of 4.0 RDU. Other performance parameters at the 4.0 cutoff level included a sensitivity of 77% (95% confidence interval 0.5418, 0.8964), and a specificity of 85% (95% confidence interval 0.7959, 0.8918).

Most breast cancer patients had a significantly lower mean serum mammastatin level than healthy participants. In the subgroup of participants younger than 40 years of age, those with a positive family history of breast cancer had a mean MSA value (14.2 RDU) lower than women without a family history (24.6 RDU) (p = 0.033). The difference between these means was –10.4 RDU. As the age of the participants in the subgroup increased, the magnitude of the difference between the mean MSA values of those with a family history of breast cancer and the values of those without such history decreased.

The decreasing gap between mean MSA levels becomes evident as the MSA levels for older age groups are examined. For the subgroup of patients aged 40–49 years, the difference of the means was –6.2 RDU. For subgroup 50–59 years, the difference of the means was –2.5 RDU.

However, for patients aged 60–69, the trend appeared to reverse, revealing a difference of the means of 2.5 RDU, whereby mean MSA levels were smaller for patients lacking a family history of breast cancer relative to those with a positive history. In the subgroup of patients aged 70 or older, the difference of the means was 1.6. 

Since the size of the difference is smaller for the subgroup of participants aged 70+ than for the subgroup of participants aged 60–69, the difference of the means cannot be deemed a linear or systematic trend (see Table II). 

A one-degree-of-freedom statistical test revealed that the observed trend in mean MSA level differences between those with a positive family history of breast cancer versus those with no family history of the disease across the age categories was significant (p = 0.039). 
Thus, women who have a positive family history of breast cancer and low mammastatin levels may be at increased risk for the disease, relative to women lacking a history of breast cancer and to women with higher values of mammastatin. 

Moreover, it is suggested that this relationship among risk factors diminishes for older women. The presence of mammastatin decreases with age and as the rate of spontaneous breast cancer occurrence increases.

The second study used a double-blind, multicenter, nonrandomized trial to evaluate serum mammastatin levels in healthy, high-risk, and breast cancer patients and to confirm the diagnostic parameters established in the first study. A total of 316 women with validated MSA levels and known disease status were included in the analyses. The participants were divided into subgroups using the same parameters as in study one, resulting in a sample of 13 breast cancer patients and 296 “healthy” patients. 

A review of the ROC curve was performed again for the MSA test. At the mammastatin cutoff point of 4.0 RDU, sensitivity was 62% (95% confidence interval 0.3888, 0.8964) and specificity was 87% (95% confidence interval 0.7678, 0.8589).

Table II. Initially, the difference between the mean mammastatin levels of participants with a positive family history of breast cancer and those with no family history of the disease decreases as age increases. However, this trend reverses for women 60 and older. (Click to enlarge)

The findings indicate that the best MSA cutoff point for determining high or low mammastatin was 4.0 RDU for both trials one and two, with a sensitivity of 77% and 62% and specificity of 85% and 87%, respectively. 

Most women in the healthy group yielded mammastatin levels higher than 4.0, while women testing below 4.0 had breast cancer. Eighteen women from the “healthy” group in the first study tested below 4.0 RDU, but did not have breast cancer. However, two of these women with low mammastatin have subsequently developed breast cancer. 

The negative predictive value of the assay in studies one and two was 97% and 98%, respectively. This value indicates that women testing high for mammastatin do not have breast cancer and appear not to be at immediate risk of developing breast cancer. 
Furthermore, women under age 41 and with a positive family history of breast cancer appear to have lower mean levels of MSA than do women without a family history of the disease. This relationship diminished as the age of the women increased. 

Conclusion

The correlation between healthy breast tissue and high levels of mammastatin indicates that the MSA could be used to screen young women for increased risk of developing breast cancer (see Figure 2). Currently, the assay is offered for informational purposes to assist established diagnostic procedures. 

A longitudinal study demonstrating a decrease in mammastatin prior to the onset of cancer is necessary to document the risk assessment value of the test. Illuminating the correlations between mammastatin levels and hormone replacement therapy is also a high priority because of the elevated cancer risk associated with hormone replacement therapy. These trials have been planned and will commence in the near future. 

References

1. Professionals Statistics: Breast Cancer Facts and Figures 2001-2002. [on-line] (Atlanta, GA: American Cancer Society Inc., 2001 [cited April, 2002]); available from Internet. http://www.cancer.org/docroot/stt/content/stt_ 1x_breast_cancer_facts_and_figures_2001-2002.asp. 
2. Cancer Facts: Screening Mammograms: Questions and Answers. [on-line] (Bethesda, MD: National Cancer Institute, 2002 [cited April, 2002]); available from Internet: http://cis.nci.nih.gov/fact/5_28.htm.
3. PR Ervin et al., “Production of Mammastatin, a Tissue-Specific Growth Inhibitor, by Normal Human Mammary Cells,” Science 244 (1989): 1585–1587. 
4. L Liberman et al., “US Guided Core Breast Biopsy: Use and Cost-Effectiveness,”Radiology 208 (1998): 718–723.
5. CJ Wright, C Mueller, and C Barber, “Screening Mammography and Public Health Policy: The Need for Perspective,” The Lancet 346 (1995) 29–32.
6. MN Prout, “Breast Cancer Risk Reduction: What Do We Know and Where Should We Go?” Medscape Women’s Health eJournal [on-line] 5, no. 5 (2000) (cited February, 2003); available from Internet: http://www.Medscape.com/viewarticle/408929.  

Jeanne Ohrnberger, PhD, is laboratory director for Biomedical Diagnostics LLC (Ann Arbor, MI), a wholly owned subsidiary of Genesis Bioventures Inc. (Surrey, BC, Canada). She can be contacted at johrnberger@bio-diagnostics.com. 

Copyright ©2003 IVD Technology