| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Editorial |
From the Department of Surgical Oncology, University of Texas M.D. Anderson Cancer Center, Houston, Texas.
Carriers of deleterious BRCA-1 mutations have a 60% to 85% lifetime risk of breast cancer. We, however, do not yet have the ability to predict which carriers will develop breast cancer nor in whom risk-reducing procedures or chemopreventive strategies can be avoided. Thus, the idea of developing a serum test that can help further stratify risk is appealing. Furthermore, studies have shown that even in patients with BRCA mutations who are managed with close observation, about half of the breast tumors are diagnosed when they have already metastasized to the lymph nodes and some of these patients die from their disease.1,2 Clearly, current imaging modalities do not always allow us to diagnose breast cancer early, especially in young women with dense breasts. An urgent need exists for better breast cancer screening tools and, especially, for a blood test for breast cancer.
Proteomic profiling is a rapidly evolving field that has attracted a lot of attention. Surface-enhanced laser desorption ionization-time of flight (SELDI-TOF) mass spectroscopy analysis of serum proteomic profiles is especially being actively pursued as a tool for early diagnosis of many cancers. Early results with ovarian cancer, reported by Petricoin et al.,3 were especially promising, with the identified algorithm discriminating cases of ovarian cancer from nonmalignant cases with a sensitivity of 100% and a specificity of 95%. Zhu et al.4 later reported that they have developed a comprehensive pattern recognition procedure that can achieve perfect discrimination for ovarian cancer (100% sensitivity, 100% specificity). Others5,6 have reported promising results with the use of SELDI-TOF to detect proteomic profiles for prostate cancer and breast cancer as well. Despite these promising reports, the role of proteomic profiling for the diagnosis of cancer has remained controversial.7 The reproducibility of serum proteomic patterns obtained by SELDI-TOF has been questionable. It has been proposed that proteomics data can be biased by the nature of the clinical samples used, artifacts of sample collection, storage and processing, mass spectometry instruments, and bioinformatics analysis.8
In this issue of Annals of Surgical Oncology, Becker et al.9 present a study of SELDI-TOF mass spectometry as a tool to distinguish women with germ-line BRCA-1 mutations who will subsequently develop breast cancer from those who will not. The sera from 15 women with BRCA-1 mutations who developed breast cancer within 3 years of the collection of baseline serum specimens (BRCA-1 cancer) were compared with sera from 15 women with BRCA-1 mutations who were cancer-free at a 7-year follow-up (carriers). SELDI-TOF analysis revealed 107 differentially expressed proteins between these two groups; a decision tree generated from these protein peaks correctly identified the patient’s status with 87% sensitivity and 87% specificity on cross-validation analysis. Patients with BRCA-1 cancer were then compared with patients with sporadic breast cancer (SBC), identifying 35 proteins overexpressed in the BRCA-1 cancer group. The decision tree generated from these protein peaks differentiated the BRCA-1 cancer group from the SBC group with a sensitivity of 94% and specificity of 100%. The proteomic profile of carriers, in contrast, resembled that of normal volunteers.
The study by Becker et al.9 is a small pilot study and can be considered only exploratory in nature. The results are exciting, however, because the authors demonstrate that it may be feasible to differentiate protein profiles for occult cancer detection or prediction in BRCA-1 patients. Whether the proteomic pattern identified in this study represents earlier detection of occult cancer or a predictive pattern for risk is unclear. The mean time from the baseline serum sample to diagnosis of breast cancer in the patients with BRCA-1 cancer was 1.3 years, with a range of 1 to 36 months. Thus, these patients most likely already had an occult tumor at baseline serum evaluation and, therefore, the proteomic profile identified may represent early diagnosis. It has been argued, however, that mass spectrometry is incapable to distinguish early cancer, because it does not have the adequate sensitivity to pick up small changes.7 If the proteomic profile identified represents epiphenomena that can be associated with a higher risk of developing breast cancer in women with BRCA-1 mutations, rather than being a diagnostic for occult disease, that would still be useful in these patients who are uniformly treated as being at very high risk at the present time. A serum test with a specificity of 87% would be unacceptable in the general population, because it would lead to millions of women having additional imaging studies (bilateral ultrasound and bilateral magnetic resonance imaging) and potentially may lead to many anxious patients seeking unnecessary procedures in the setting of normal imaging studies but abnormal serum tests. In contrast, a test with a specificity of 87% for cancer detection or prediction can be of clinical utility for the high-risk, BRCA-1 mutation carriers.
The Becker et al.9 results suggest that the profiles in patients with BRCA-1 cancer are different from those with SBC. This would not be surprising, because the gene expression profile of tumors with BRCA1 mutations have been shown to differ significantly from those of sporadic tumors.10 This result should be approached with caution, however, both because the patients with BRCA-1 cancer had clinically occult disease, compared with SBC patients with clinically apparent disease, and because the serum samples for both groups of patients were obtained at two geographically separate locations, Creighton University and Eastern Virginia Medical School, respectively, and may have been subject to differences in sample preparation, number of freeze-thaw cycles, and sample storage conditions, all potentially affecting the proteomic profiles.
Although the serum proteomics decision tree generated was tested with a cross-validation analysis in this study, the true potential of this proteomics tool can be determined only after testing its prediction rate in a larger, independent serum set. In the interim, it is worth discussing with patients who present for high-risk counseling and consideration for risk-reducing mastectomies, that a serum test may be available that can assist in risk-stratification and early detection in the near future.
FOOTNOTES
Received August 3, 2004; accepted August 23, 2004.
Address correspondence to: Funda Meric-Bernstam, MD, Department of Surgical Oncology, Unit 444, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston TX 77030-4009; Fax: 713-745-4926; E-mail: fmeric{at}mdanderson.org.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
