| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
San Francisco Coordinating Center (J.S.L., S.R.C.), California Pacific Medical Center Research Institute, San Francisco, California 94107; Division of Endocrinology and Metabolism (J.S.L.), San Francisco General Hospital, Department of Epidemiology and Biostatistics and Women’s Health Clinical Research Center (E.V., D.G.), and Division of General Internal Medicine, Department of Medicine (M.S.B.), University of California, San Francisco, California 94143; Division of Research (B.E.), Kaiser Permanente Medical Care Program, Northern California, Oakland, California 94612; Departments of Obstetrics and Gynecology, and Preventive Medicine (F.Z.S.), University of Southern California Keck School of Medicine, Los Angeles, California 90033; Berlex Laboratories, Inc. (V.H.), Montville, New Jersey 07045; Department of Epidemiology (J.A.C.), University of Pittsburgh, Pittsburgh, Pennsylvania 15261; Esoterix Endocrinology Inc. (W.C.), Calabasas, California 91301; SFBC Taylor Technology (J.S.), Inc., Princeton, New Jersey 08540; and Department of Biochemistry (E.F., M.D.), Royal Marsden Hospital, London SW3 6JJ, United Kingdom
Address all correspondence and requests for reprints to: Jennifer S. Lee, M.D., Division of Endocrinology, Clinical Nutrition, and Vascular Medicine, University of California Davis Medical Center, 4150 V Street, Suite G400, Sacramento, California 95817. E-mail: catechols{at}gmail.com.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to determine whether an extraction-based (indirect) assay or a non-extraction-based (direct) assay correlates better with mass spectrometry and body mass index (BMI).
Design/Setting: In a pilot study of 40 postmenopausal women, endogenous E2 measurements from three indirect and four direct assay methods and gas chromatography-tandem mass spectrometry (GC-MS/MS) were compared. A confirmatory study compared an indirect and a direct assay, selected among those in the pilot study, to GC-MS/MS; this study was conducted in 374 postmenopausal women not taking hormone therapy from the Ultra Low-dose TRansdermal estrogen Assessment (ULTRA) trial.
Main Outcomes: Pearson correlation coefficients among E2 measurements by assay methods and BMI, and their confidence intervals, by bias-corrected bootstrap method, were used.
Results: In the pilot study, E2 by three indirect assays correlated better (P < 0.03) with GC-MS/MS and with BMI than measurements by four direct assays. In the confirmatory study, the indirect assay correlated better (P < 0.01) with GC-MS/MS and BMI than the direct assay. Measurements by the indirect and direct assays were overestimated, but deviations in direct assay measurements were less precise. Mean E2 by the indirect and direct assays were higher (by 14 and 68%, respectively) and less reproducible than by GC-MS/MS.
Conclusion: Until mass spectrometry is practical for wide use, extraction-based indirect assays may be preferable for measuring low postmenopausal serum E2.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Two major methods are widely used to measure E2: indirect and direct immunoassays. Several studies have compared performances of E2 assays for postmenopausal levels but typically for levels over 15 pg/ml (62.4 pmol/liter) (2, 13, 14, 16). Substantial disparity existed between measurements by direct and indirect assays in breast cancer patients who received aromatase inhibitor therapy, which lowers E2 levels to nearly zero (17). Indirect assays are typically RIAs that consist of an initial extraction step. In contrast, direct assays do not involve extraction, can be automated, and generally require less labor, specimen volume, and cost. Mass spectrometry is considered a reference standard for measuring serum E2 levels (14, 20).
This study’s objective was to compare E2 assay methods in their ability to measure low endogenous serum concentrations. First, we conducted a pilot study to compare E2 measurements by each of seven E2 assays (three indirect and four direct) with those by gas chromatography-tandem mass spectrometry (GC-MS/MS) and with body mass index (BMI), which shows a strong correlation with serum E2 in postmenopausal women (3, 21, 22, 23). Second, we confirmed findings from the pilot study using baseline sera collected from 374 women enrolled in a randomized clinical trial (18).
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
From the Study of Osteoporotic Fractures (SOF), we randomly selected 40 postmenopausal women who reported never having used hormone therapy. These 40 women comprised the population-based pilot study cohort. SOF is a prospective cohort study of women aged 65 yr or older who were recruited in 1986–1987 from four communities in the United States (24, 25). At the baseline examination, participants were instructed to adhere to a fat-free diet during the night before and the morning of the examination to minimize lipemia that may interfere with hormone assays. After blood was drawn between 0800 and 1400 h, serum immediately was frozen to –20 C. Within 2 wk, all samples were shipped to a central repository and stored in liquid nitrogen at –190 C until assays were performed. At baseline, history of hormone therapy was ascertained, and weight (in lightweight clothing without shoes) was measured using a balance-beam scale. BMI was defined as body weight in kilograms divided by height in meters squared. The appropriate institutional review boards approved the study, and prior written informed consent was obtained from all SOF participants.
The confirmatory study cohort was 374 postmenopausal women from the Ultra Low-dose TRansdermal estrogen Assessment (ULTRA), a randomized clinical trial of the effect of 2 yr of treatment with 0.014 mg/d of transdermal E2 on changes in bone density in 417 postmenopausal women aged 60–80 yr with normal bone density for age. The ULTRA trial design and main findings have been published (18). The confirmatory study participants were not taking hormone therapy, including low-dose estrogen, at baseline through yr 1. Fasting baseline and nonfasting 1-yr follow-up sera were available for hormone testing by three E2 assay methods. Aliquots were frozen at the clinical sites and sent to Biomedical Research Institute (Rockville, MD) for storage at –70 C. At baseline, weight and height were measured and BMI was calculated. ULTRA was funded by Berlex Laboratories (Montville, NJ), the manufacturer of Menostar, the transdermal E2 patch used in the trial. The institutional review board of each clinical center and the coordinating center approved the study protocol, and informed consent was obtained from all participants.
E2 assays
The pilot study measured E2 in baseline sera obtained from each of the 40 participants. Esoterix Endocrinology, Inc. (Calabasas, CA, and San Antonio, TX) performed all measurements using commercially available kits from four direct (non-extraction-based) assays obtained from the following suppliers: Diagnostic Products Corp. (Los Angeles, CA; DPC Double Antibody E2 Assay), Diagnostic Systems Laboratories (Webster, TX; DSL-4800 Ultrasensitive E2 RIA), Roche Diagnostics (Alameda, CA; Elecsys E2 Enzyme Immunoassay II), and Ortho-Clinical Diagnostics (Rochester, NY; Vitros). E2 measurements using three in-house indirect (extraction-based) RIAs were performed at their respective laboratories: Esoterix Endocrinology, Inc. (Esoterix E2 RIA), Royal Marsden Hospital (research laboratory of Mitch Dowsett; London, UK; Royal Marsden E2 RIA), and University of Southern California (research laboratory of Frank Z. Stanczyk; Los Angeles, CA; USC E2 RIA). GC-MS/MS, measured at SFBC Taylor Laboratories (Princeton, NJ), served as the reference assay. Reproducibilities of the DSL, DPC, and Vitros assays were measured on duplicate sera from a subset (of 20 women) due to scarcity of serum volume.
In the confirmatory study, E2 was measured by an indirect and a direct immunoassay method that correlated strongly with GC-MS/MS and BMI in the pilot study. Specifically, the direct (DPC Double Antibody E2 Assay) and the indirect assays (Royal Marsden E2 RIA) were compared with GS-MS/MS. Covance Central Laboratory Services Inc. (Indianapolis, IN) conducted the direct assay, and the research laboratory of Mitch Dowsett conducted the in-house indirect assay.
Each laboratory site performed the assays in a single batch and was blinded to the identification or characteristics of the subjects. For each assay, lower detection limit is defined as two SD values above the mean of replicates of the zero E2 standard. For the pilot study’s research focus on very low postmenopausal E2 levels, laboratories reported the measured values of the E2 concentrations that were below the calculated lower limit of detection (and typically reported as "undetectable"), except for levels measured by the Roche assay, in which automation made reporting at those levels impossible. Five Roche measurements reported as less than 5 pg/ml (<18 pmol/liter) were approximated to 2.5 pg/ml (9 pmol/liter) in our data analyses.
Indirect (extraction-based) assays
Esoterix’s total E2 RIA required 1.0 ml serum. Extraction and LH20 column chromatography preceded RIA (Esoterix). The intraassay CVs were 13.1% at a mean of 6.5 pg/ml (23.9 pmol/liter) and 5.5% at 28.6 pg/ml (105.0 pmol/liter). The interassay CV was 12% at 26 pg/ml (95 pmol/liter). Assay controls were human sera, and blanks were water. Each sample’s recovery was assessed using tritiated E2, and samples were corrected for losses; typical recovery was 85%. The reported functional sensitivity was 5 pg/ml (18 pmol/liter), which is the lowest E2 concentration measured for a spiked sample whose coefficient of variation is (CV) 20%. The lower detection limit was 2.0 pg/ml (7.3 pmol/liter).
Royal Marsden Hospital’s total E2 RIA required 0.5 ml serum. Extraction was performed with diethyl ether. E2 concentration was measured by competitive RIA using a highly specific rabbit antiserum raised against an E2-6-carboxymethyloxime-BSA conjugate (EIR, Wurenlingen, Switzerland) and E2-6-carboxymethyloxime-[2-125I]-iodohistamine (
2000 Ci/mmol; Amersham International, Buckinghamshire, UK). Recovery of more than 90% was determined in a separate set of samples in a previous study (26). The intraassay and interassay CVs at a mean of 7.3 pg/ml (27.0 pmol/liter) were 7.6 and 17%, respectively. The lower detection limit was 0.8 pg/ml (2.9 pmol/liter).
USC’s total E2 RIA used 0.8 ml serum, although the minimum volume required is 0.5 ml. Ethyl acetate/hexane (3:2) extraction and Celite column partition chromatography were performed before RIA quantification. Recovery of tritiated E2, which ranged from 73–86%, was used to correct observed E2 values. Intraassay CV was 7.9% at 34 pg/ml (124 pmol/liter), whereas interassay CVs were 8.0 and 12.0% at 16 pg/ml (58.7 pmol/liter) and 27 pg/ml (99.1 pmol/liter), respectively. The lower limit of detection was 1.8 pg/ml (6.6 pmol/liter).
Direct (non-extraction-based) assays
DPC’s Double Antibody E2 assay required 0.2 ml serum and was a competitive RIA that used rabbit antiserum. The intraassay CV was approximately 12% at 5 pg/ml (18 pmol/liter) and 5% at 20 pg/ml (73 pmol/liter). The interassay CV was not determined within the postmenopausal range. The lower detection limit was 1.4 pg/ml (5.1 pmol/liter).
DSL’s Ultrasensitive E2 RIA (DSL-4800) assay required 0.2 ml serum and was a double antibody competitive RIA. Goat antirabbit 
globulin serum in a buffer of polyethylene glycol was the precipitating reagent. The intraassay and interassay CVs were 8.9 and 7.5%, respectively, at 5.3 pg/ml (19.5 pmol/liter). At 28 pg/ml (102.8 pmol/liter), the intra- and interassay CVs were 6.5 and 9.7%, respectively. The lower detection limit was 2.2 pg/ml (8.1 pmol/liter).
The Roche Elecsys E2 Enzyme Immunoassay (EIA) II, an electrochemiluminescence immunoassay, required 0.3 ml serum. The intraassay CV was 3.3% at 35.4 pg/ml (130 pmol/liter), and the interassay CV was 6.2% at 22.9 pg/ml (84.1 pmol/liter). The lower detection limit was 5 pg/ml (18 pmol/liter).
Ortho-Clinical Diagnostics’ Vitros assay required 0.07 ml of sample and was a competitive enzyme immunoassay, carried out with horseradish peroxidase-labeled E2 conjugate and a mixture of biotinylated antibodies (sheep and rabbit anti-E2). CVs were not available for E2 within the postmenopausal range. The lower detection limit of the assay was 2.7 pg/ml (10 pmol/liter).
GC-MS/MS
GC-MS/MS was performed by SFBC Taylor Technologies Inc. and required 1.0 ml serum (27, 28). Deuterated E2 was added to assess recovery. Bond Elut Certify solid-phase cartridges were used for extraction and ethyl acetate was the eluate. Two derivatizations were performed: 1) reaction with pentfluorobenzoyl chloride; and 2) reaction with N-methyl-N-(trimethylsilyl)-trifluoroacetamide. The derivatized analytes were separated by gas chromatography using a DB-17 fused silica capillary column and were detected by negative ionization tandem mass spectrometry. The instruments included a Finnigan (San Jose, CA) MAT mass spectrometer and a Varian (Palo Alto, CA) 3400 gas chromatography system. The validity of this method was assessed by examining intrarun precision and accuracy for calibration standards and quality control (27). Calibration standards and blanks were prepared in water. Quality-control pools were prepared by spiking with E2 above endogenous levels in human postmenopausal serum, and, for very low levels, postmenopausal serum was diluted with charcoal stripped serum and then spiked to approximate concentrations.
Based on data presented in abstract form and unpublished data provided by Taylor Technologies Inc. (27), control samples with an expected E2 concentration of 0.6 pg/ml (2.2 pmol/liter) showed measured mean concentration of 0.6 pg/ml (2.2 pmol/liter), with a CV of 15.8%. At an expected concentration of 2.1 pg/ml (7.7 pmol/liter), the measured concentration was 2.0 pg/ml and the intraassay CV was 17.8%. At an expected concentration of 3.9 pg/ml (14.3 pmol/liter), the measured mean concentration was 3.9 pg/ml and the intraassay CV was 9.0%. At an expected concentration of 26.1 pg/ml (95.8 pmol/liter), the measured concentration was 26.2 pg/ml and the intraassay CV was 3.1%. Recovery from water and charcoal stripped serum were consistent within a matrix and between two matrices. In control specimens ranging from 1.3 pg/ml (4.8 pmol/liter) to 40 pg/ml (147 pmol/liter), recovery ranged from 81–94%. The limit of detectability was 0.6 pg/ml (2.2 pmol/liter).
Statistical analyses
In the pilot study and confirmatory study, we computed Pearson correlation coefficients to evaluate how well, by each of the assays, E2 correlated with that by the reference standard GS-MS/MS and with BMI. Furthermore, to evaluate assay precision in the pilot study, replicate assays were performed in separate assays on 20 split serum specimens. Precision for the three assays used in the ULTRA study was calculated from paired measurements of E2 (at baseline and yr 1) by each assay method in each of 374 participants not on hormone therapy. We assumed that year-to-year levels of endogenous E2 would remain constant, on average, among these women.
Confidence intervals for correlations were computed using the bias-corrected percentile bootstrap method (29). The statistical significance of the differences between pair-wise correlations was based on approximate two-sided P values for the test of the null hypothesis of no difference, computed as twice the proportion of the 500 bootstrap samples for each pair-wise comparison in which the estimated difference did not exceed the null value of zero. P < 0.05 was considered statistically significant. All analyses were performed using SAS software, version 8.02 (SAS Institute, Inc., Cary, NC), and STATA version 8.0 (Stata Corp., College Station, TX).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Pearson correlation coefficients between GC-MS/MS and each of the seven assays ranged from a low of 0.57 (Roche) to a high of 0.94 (Royal Marsden). The three indirect assays showed higher correlations (Royal Marsden, 0.94; USC, 0.91; Esoterix, 0.88) with GC-MS/MS than the direct assays (DPC, 0.83; Vitros, 0.71; DSL, 0.70; Roche, 0.57). The average correlation coefficient of the three direct assays (0.70) was about 20% lower than the average correlation of the four indirect assays (0.91) (P < 0.03). None of the correlation coefficients were significantly different from one another across all assays in this pilot study. E2 measurements by the indirect assays tended to be higher than those by GC-MS/MS (Fig. 1
). Each of the three indirect assays and GC-MS/MS correlated better with BMI than the direct assays (Fig. 2). Vitros and DSL assays were only weakly correlated with BMI (r = 0.25 and 0.27, respectively). Using body weight instead of BMI did not substantially alter these findings (data not shown). Based on correlations with GC-MS/MS and BMI, the direct assay by DPC appeared to perform better than the other direct assays.
|
|
). According to World Health Organization guidelines, 45% were osteopenic (t-score between –1 and –2.5), and only 15% were osteoporotic (t score < –2.5). On average, mean E2 based on the direct assay was 14% higher and based on the indirect assay was 68% higher than GC-MS/MS. Compared with the direct assay, the indirect assay correlated better with GC-MS/MS (r = 0.65 and 0.89, respectively, P < 0.01 for the difference in correlation coefficients; Fig. 3) and tended to systematically overestimate at higher E2 concentrations. The strength of correlations with BMI was nearly the same for the indirect assay and GC-MS/MS (Table 2). In contrast, the direct assay was less precise in overestimating E2 concentrations, and its correlation with BMI was nearly 50% lower than the GC-MS/MS-BMI correlation (P < 0.01 for the difference between the correlations with BMI for the direct and GC-MS/MS assays).
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In our pilot study, the mean and range of E2 values obtained by DPC’s Double Antibody direct assay were low compared with those obtained by GC-MS/MS, Royal Marsden Hospital’s indirect RIA, and the other assays. In contrast, the confirmatory study found that E2 values obtained by DPC’s assay were higher than those by GC-MS/MS but lower than those by Royal Marsden Hospital’s assay. These differences observed for DPC’s direct assay may be due to the fact that the pilot study consisted of only 40 women, whereas the confirmatory study included over six times more women. Different laboratories performed the DPC assay for the initial pilot and later validation study, DPC calibrators were not subjected to GC-MS/MS; and variations in the assay kits themselves may have occurred.
The present study has several limitations. We did not include all commonly used assay methods, newer chemiluminescence assays, or newer estrogen bioassays that have been reported to be sensitive in measuring postmenopausal E2 levels (19, 28). The reasons for overestimations by the indirect and direct assays are not known and deserve further study. Limited availability of serum specimens prevented us from performing more complete intra- or interassay reproducibility experiments for all assays. Difference in calibration standards across assays was not addressed. We assessed reproducibility by comparing E2 measurements a year apart in ULTRA participants. It cannot be ruled out that the observed differences in these assays’ measurements may be true biological change. The small changes in E2 levels did not correlate with changes in BMI and suggest that E2 concentration in late menopause is constant.
Mass spectrometry has been a reference standard for measuring both male and female sex hormones (14, 21, 30). In addition, Wang et al. (28) recently reported that a recombinant cell ultrasensitive estrogen bioassay correlated well with the same GC-MS/MS assay used in the present study. Even for low postmenopausal E2 concentrations, gas or liquid chromatography-mass spectrometry methods have been shown to be quite accurate (19, 23, 31). However, as with RIAs, many variables factor into the accuracy of mass spectrometry methods including standardizing reagents, procedures, and hardware. In the present study, a physiological correlate of estrogen, BMI, was also used to evaluate assay performance. Again, GC-MS/MS E2 showed the strongest correlation with this measure. Furthermore, findings using both GC-MS/MS and BMI reference measures were quite consistent, providing some reassurance of the conclusions. The present study is the first to directly compare several assay methods with mass spectrometry and a biological correlate within the very low E2 range. This study was blinded and used samples from well-characterized large cohorts. Clinical researchers and epidemiologists must be aware of the lack of standardization of E2 assays at low postmenopausal concentrations and should be wary of the variability in E2 measurements across different assays. Until mass spectrometry or an alternate method is practical for wide use to measure low E2 levels, it is difficult to decide which assay method to use. Whereas direct assays require less specimen volume and lend themselves to automation by excluding an initial extraction and purification step, their accuracy may suffer due to factors including ill-defined matrix effects or nonspecific binding. We conclude that studies of E2 in postmenopausal women should use indirect methods or mass spectrometry. Previous studies that relied on direct assays may have underestimated the relationships between endogenous hormones and other conditions in postmenopausal women.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Current address for J.S.L.: Division of Endocrinology, Clinical Nutrition, and Vascular Medicine, University of California Davis Medical Center, Sacramento, California 95817.
First Published Online August 1, 2006
Abbreviations: BMI, Body mass index; CV, coefficient of variation; DPC, Diagnostic Products Corp.; DSL, Diagnostic Systems Laboratories; E2, estradiol; GC-MS/MS, gas chromatography-tandem mass spectrometry; USC, University of Southern California.
Received October 31, 2005.
Accepted July 25, 2006.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Tivesten, L. Vandenput, F. Labrie, M. K. Karlsson, O. Ljunggren, D. Mellstrom, and C. Ohlsson Low Serum Testosterone and Estradiol Predict Mortality in Elderly Men J. Clin. Endocrinol. Metab., July 1, 2009; 94(7): 2482 - 2488. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Ankarberg-Lindgren, E. Norjavaara, M. M. Kushnir, W. L. Roberts, A. L. Rockwood, J. Bergquist, A. M. Bunker, and A. W. Meikle Are Estradiol Results Determined by the Tandem Mass Spectrometry Assay Clinically Useful for Children? Am J Clin Pathol, May 1, 2009; 131(5): 746 - 750. [Full Text] [PDF]
|
|
|
|
 |
|
 P. Fan, W. Yue, J.-P. Wang, S. Aiyar, Y. Li, T.-H. Kim, and R. J. Santen Mechanisms of Resistance to Structurally Diverse Antiestrogens Differ under Premenopausal and Postmenopausal Conditions: Evidence from in Vitro Breast Cancer Cell Models Endocrinology, May 1, 2009; 150(5): 2036 - 2045. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Cummings, J. A. Tice, S. Bauer, W. S. Browner, J. Cuzick, E. Ziv, V. Vogel, J. Shepherd, C. Vachon, R. Smith-Bindman, et al. Prevention of Breast Cancer in Postmenopausal Women: Approaches to Estimating and Reducing Risk J Natl Cancer Inst, March 18, 2009; 101(6): 384 - 398. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Eriksson, M. Lorentzon, L. Vandenput, F. Labrie, M. Lindersson, A.-C. Syvanen, E. S. Orwoll, S. R. Cummings, J. M. Zmuda, O. Ljunggren, et al. Genetic Variations in Sex Steroid-Related Genes as Predictors of Serum Estrogen Levels in Men J. Clin. Endocrinol. Metab., March 1, 2009; 94(3): 1033 - 1041. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Lee, A. Z. LaCroix, L. Wu, J. A. Cauley, R. D. Jackson, C. Kooperberg, M. S. Leboff, J. Robbins, C. E. Lewis, D. C. Bauer, et al. Associations of Serum Sex Hormone-Binding Globulin and Sex Hormone Concentrations with Hip Fracture Risk in Postmenopausal Women J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1796 - 1803. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Ankarberg-Lindgren and E. Norjavaara A purification step prior to commercial sensitive immunoassay is necessary to achieve clinical usefulness when quantifying serum 17 -estradiol in prepubertal children Eur. J. Endocrinol., January 1, 2008; 158(1): 117 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Z. Stanczyk, J. S. Lee, and R. J. Santen Standardization of Steroid Hormone Assays: Why, How, and When? Cancer Epidemiol. Biomarkers Prev., September 1, 2007; 16(9): 1713 - 1719. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J Santen, N. F Boyd, R. T Chlebowski, S. Cummings, J. Cuzick, M. Dowsett, D. Easton, J. F Forbes, T. Key, S. E Hankinson, et al. Critical assessment of new risk factors for breast cancer: considerations for development of an improved risk prediction model Endocr. Relat. Cancer, June 1, 2007; 14(2): 169 - 187. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
