|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2005.04.0345 on January 8 2007 © 2007 American Society of Clinical Oncology. Ductal Carcinoma In Situ in BRCA Mutation Carriers
From the University of California, San Francisco (UCSF) Cancer Risk Program and the Carol Franc Buck Breast Care Center, UCSF Comprehensive Cancer Center, San Francisco CA Address reprint requests to E. Shelley Hwang, MD, MPH, University of California, San Francisco Cancer Center, 1600 Divisadero Ave, Box 1710, San Francisco, CA 94143-1710; e-mail: shelley.hwang{at}ucsfmedctr.org
PURPOSE: The current literature suggests that ductal carcinoma in situ (DCIS) of the breast is infrequently diagnosed in patients with BRCA germline mutations. We studied women at high risk of hereditary breast cancer syndromes who underwent testing for BRCA1 and BRCA2 to estimate DCIS prevalence and incidence in known BRCA-positive women compared with high-risk women who were mutation negative. METHODS: We analyzed breast event outcomes in a retrospective cohort of 129 BRCA-positive and 269 BRCA-negative women undergoing genetic testing for a BRCA mutation between September 1996 and December 2003 at University of California, San Francisco. We estimated the frequency of DCIS and invasive cancer and time to breast events from birth using a Cox proportional hazard model for competing risks. Histologic grade of DCIS was also compared between groups. RESULTS: Among BRCA carriers, 48 (37%) had DCIS (with or without invasive cancer) compared with 92 noncarriers (34%). Univariate analysis showed that both DCIS and invasive cancer had an earlier onset in mutation carriers than in noncarriers, although on a per-woman basis, this difference was not statistically significant. High-grade DCIS was more common in BRCA1 mutation carriers than in patients without a mutation (P = .02). CONCLUSION: DCIS is equally as prevalent in patients who carry deleterious BRCA mutations as in high familial-risk women who are noncarriers, but occurs at an earlier age. Our results argue for the consideration of DCIS as a criterion for BRCA risk assessments with appropriate weighting in prediction models such as BRCAPRO.
The incidence of ductal carcinoma in situ (DCIS) has increased rapidly since the widespread use of mammographic screening. Nearly 60,000 patients are now diagnosed each year in the United States, representing a significant public health problem. The risk of developing invasive breast cancer after a diagnosis of DCIS ranges from 14% to 60% at 10 years.1 Premenopausal women with DCIS, the age range when familial breast cancer is most common, have particularly adverse outcomes.1,2 Pathologists agree that DCIS is a preneoplastic lesion, and DCIS shares certain features with invasive breast cancer such as genetic signature and epidemiologic risk factors.1-7 However, it remains unclear whether DCIS is part of the BRCA mutation cancer pathway, and it is thought by some to be rare in BRCA1 mutation carriers.8-10 The best data on the relative risk of DCIS in BRCA mutation carriers comes from the Breast Cancer Linkage Consortium.11 This study showed that DCIS associated with invasive breast cancer was significantly under-represented compared with the general population (41% v 56%, respectively). A follow-up study12 also showed a marginally significant lower DCIS prevalence among BRCA mutation carriers. At present, DCIS is not significantly weighted in most of the standard models of BRCA hereditary risk assessment, although it is considered in BRCAPRO (CaGene version 4.0; University of Texas Southwestern Medical Center, Dallas, TX).13,14 Furthermore, large studies of prophylactic oophorectomy in high-risk women do not document DCIS as an outcome.15,16 The recognition of DCIS as part of the spectrum of BRCA-related disease would have important implications for BRCA mutation carriers and would allow an opportunity for risk-reducing interventions before the onset of invasive breast or ovarian cancer. Several groups have reported the frequency of BRCA mutations in women with DCIS. Frank et al17 analyzed 10,000 individuals whose blood was tested for BRCA mutations at Myriad Genetic Laboratories (Salt Lake City, UT). Twenty-six (13%) of 199 women with DCIS diagnosed before age 50 years had deleterious BRCA mutations. This prevalence was significantly lower (P < .001) than BRCA mutation prevalence in women with invasive breast cancer before age 50 years (587 of 2,466 patients; 24%).17 In the only population-based study that has examined family history and diagnosis of breast carcinoma in situ, Claus et al18 found a positive family history to be a significant risk factor for development of DCIS (odds ratio [OR], 1.62; 95% CI, 1.26 to 2.09). In a subset of 369 women from the population sample, Claus et al18 recently found 11 BRCA mutations (three BRCA1, seven BRCA2, and one both) for a prevalence of 3%, which is a BRCA mutation rate similar to that in invasive cancer in other populations. Although the sample size was very small, BRCA mutation carriers with DCIS had a stronger family history of breast/ovarian cancer and were diagnosed at an earlier age than noncarriers. A few studies have observed high-risk women or mutation carriers to discover DCIS as an outcome. Kerlikowske et al19 found that family history was a risk factor (OR, 2.2; 95% CI, 1.0 to 4.2) for DCIS among 39,542 women older than 50 years undergoing screening mammography. In studies of prophylactic mastectomy specimens from high-risk women, DCIS was diagnosed in a range of 4% to 15%.20,21 Kauff et al22 compared 24 prophylactic mastectomy specimens from women with BRCA mutations with 48 autopsy breast specimens from women without a cancer predisposition. They found an elevated risk (OR, 12.7; 95% CI, 3.1 to 52.4; P < .001) of premalignant lesions in the mutation carrier specimens. Kriege et al23 diagnosed six patients with DCIS and one patient with lobular carcinoma in situ using magnetic resonance imaging in 1,909 women at high hereditary risk of breast cancer over a 3-year period. Scheuer et al24 observed 194 women with BRCA mutations over a mean of 2 years. Of 29 women who had risk-reducing mastectomy, two (7%) had DCIS in the surgical specimens. Of 89 women choosing surveillance only, DCIS was diagnosed in three women (3%). Familial breast cancer accounts for up to 20% of breast cancer diagnosed in the United States, but fewer than 5% are attributable to mutations in BRCA1 or BRCA2. Women with known BRCA mutations face up to an 85% lifetime risk of breast cancer and an up to 45% risk of ovarian cancer.9,10 On average, BRCA mutation carriers are diagnosed with invasive breast cancer about 10 years younger than sporadic patients.9 This accelerated time course of tumor development has led some to postulate that BRCA-associated breast cancers bypass the preinvasive stage altogether.25 To further clarify the role of DCIS in the development of BRCA-associated invasive breast cancer spectrum, we analyzed the frequency, histopathology, and time to DCIS diagnoses in BRCA-positive patients compared with high familial risk, BRCA mutation–negative patients in a large cohort of women referred for genetic counseling.
We studied a cohort of women seen at the University of California, San Francisco Cancer Risk Program referred for genetic counseling between September 1996 and December 2003, all of whom were part of a 20-year follow-up protocol approved by the university's Committee on Human Research. All consenting patients in this cohort had a family history of breast and/or ovarian cancer that exceeded a 10% prior probability of carrying a BRCA mutation using the BRCAPRO model.13 All women were either self-referred or physician referred based on personal and/or family history. Both mutation-negative and mutation-positive patients were similar in every respect (ie, referred to the same genetic counseling center for genetic testing because of a family history of breast and/or ovarian cancer) except for the presence or absence of a deleterious BRCA mutation. The main outcome variables of interest were DCIS and invasive breast cancer. For each woman, age at the following nine specific competing events was recorded: diagnosis of DCIS, diagnosis of invasive breast cancer, and mastectomy (three events for each breast), and bilateral salpingo-oophorectomy, death, and initiation of tamoxifen treatment. We used three different statistical models to analyze our data, and each method was based on different a priori assumptions. In the first analysis, each woman comprised the unit of analysis, and we measured time (age) to first cancer event in either breast. Competing risks (ie, those that precluded an observation of a cancer event after their occurrence) included mastectomy, salpingo-oophorectomy, tamoxifen use, and death. We estimated the hazard ratio for first DCIS as a function of BRCA status and other personal characteristics, such as age at first menarche, using a Cox proportional hazards model. We conducted a similar analysis where first invasive cancer, in either breast, was the event of interest. In the subsequent analyses, each woman's breast was the unit of observation. We used a Cox proportional hazards model both with and without shared frailty for the two breasts of each woman.26 In the standard Cox model, all breasts are assumed to have the same risk of an event. In a frailty model, both breasts of each woman are assumed to share the same risk, but the risk can vary from woman to woman. Censoring occurred at the age at mastectomy, development of a breast event (DCIS, invasive breast cancer, or both), or on December 30, 2003 if there was no breast event by the end of the study period. Both breasts were censored at the age of bilateral salpingo-oophorectomy, age of initiation of tamoxifen treatment, or age of death if these events occurred during the period of observation. Women whose breast cancer was diagnosed before 1980 were excluded from the analysis because DCIS was not widely reported before that year. Also excluded were women whose original pathology reports were not available for review and women whose BRCA mutations were variants of unknown significance. We examined univariate and multivariate Cox regression models of competing risk to compare the rates of DCIS and/or invasive breast cancer in carriers and noncarriers.25,27 For the multivariate model, backwards selection was used with P > .05 as the criterion for stepwise elimination of nonsignificant variables. Differences in proportions were tested using an exact contingency table method that computes the probability of obtaining the observed result by chance alone (STATA 9; StataCorp LP, College Station, TX). All significance tests were two sided.
The cohort was comprised of 618 patients referred for genetic counseling between September 1996 and December 2003. More than 95% were white. Four hundred seventy-seven women underwent genetic testing. Eight women were excluded for a breast cancer diagnosis rendered before 1980 (see Methods); in 38 patients, the original pathology reports could not be obtained. One patient was excluded for a diagnosis of lobular carcinoma in situ alone. Thirty-two women had mutation variants of unknown significance on full sequencing of BRCA1 and BRCA2 and were excluded from analysis. The remaining 398 women with 796 breasts at risk comprise the study population. One hundred thirty of the women in the study cohort (33%) registered for genetic testing and counseling with a prior diagnosis of DCIS or invasive breast cancer. Seventy-one of the women in the study cohort (18%) had a first or additional breast cancer event or events (DCIS or invasive breast cancer in either or both breasts) after entering the registry. Table 1 lists the characteristics and family histories of cancer of patients who were included in the final analysis. From a total of 398 eligible women who had BRCA gene testing, 129 were found to be BRCA positive (BRCA1 positive, n = 72; BRCA2 positive, n = 56; and both BRCA1 and BRCA2 positive, n = 1), and 269 were BRCA negative. The proportions of women in each category of reproductive risk factors (ie, age at menarche, age at first live birth) were similar, although mutation carriers tended to be slightly younger (later mean birth year) and were more likely to be nulliparous. This latter effect was partially a result of collinearity between age and parity, in that women in their twenties and early thirties were less likely to have commenced childbearing. Mutation carriers were more likely to be of Ashkenazi Jewish heritage and to have breast cancer among first- and second-degree relatives. As expected, members from BRCA1 families had more ovarian cancer than BRCA2 families and mutation-negative patients.
Table 2 lists the occurrence of breast events. Overall, 66% of breast events occurred before registry for genetic testing. The proportion of DCIS and invasive breast cancers did not differ in BRCA carriers and noncarriers. Thirty-nine percent of carriers and 44% of noncarriers had invasive breast cancer, DCIS, or both (ie, any breast event; Table 2). Carriers and noncarriers did not significantly differ with respect to the frequency of breast cancer events, including DCIS. There were no significant differences between breast cancer event rates for BRCA1 and BRCA2 carriers. Therefore, we combined the results for both mutation groups to evaluate disease-free survival between mutation-positive and mutation-negative patients.
All three analytic models (Table 3) show that mutation carriers had greater hazard for both DCIS and invasive cancer when compared with high-risk mutation-negative patients. The CIs show that, although the point estimates for hazards of invasive cancer are greater than those for DCIS, they do not differ statistically.
Using a process of backwards selection, we constructed a multivariate Cox model for each analysis using reproductive variables (age at menarche, age at first birth, and parity v nulliparity), birth cohort, and Ashkenazi Jewish ethnicity. We also tested for interaction between the variables in the model and found none. The final model included BRCA mutation status, birth cohort, and Ashkenazi ethnicity and did not significantly alter the findings reported earlier using BRCA mutation status as the only predictor variable (data not shown). We compared the histopathologic grade of DCIS in both groups. BRCA1 carriers were more likely to have high-grade DCIS than noncarriers (P = .02; Table 4). However, the grade of DCIS did not differ significantly between BRCA2 carriers and noncarriers (P = .75). There were no low-grade DCIS lesions seen in BRCA1/2 mutation carriers, whereas 7% of DCIS in noncarriers with high familial risk were low grade.
This study found that the prevalence of DCIS in carriers is not different from that in noncarriers. The mean age at diagnosis of DCIS was younger in BRCA mutation carriers than in noncarriers (Table 2), although the hazard ratio when analyzed on a per-women basis was not statistically significant (Table 3). DCIS in BRCA1 mutation–positive patients was more frequently high grade compared with BRCA2-positive and BRCA-negative patients. These findings are consistent with higher tumor grade reported for invasive cancers in mutation carriers.9 The mean age at diagnosis for breast cancer events is younger in this high-risk cohort than for the general population when compared with 1998 SEER data (http://seer.cancer.gov/). On average, BRCA-positive carriers in our cohort were more than 20 years younger than SEER patients at diagnosis of invasive cancer. The same trend was seen for women with DCIS without concurrent invasive cancer. BRCA-positive patients with DCIS were more than 12 years younger at diagnosis than those with DCIS in the general population. The comparison of different analytic models was of interest and relevant to any consideration of disease in paired organs at risk. The standard model that censors the patient at time of first breast cancer event has been well described in the literature. However, this model is limited in that the contralateral breast follow-up time does not contribute to the analysis. In comparison, a model that treats each breast as a separate individual incorporates all follow-up data but does not account for the probable high correlation between breasts in the same individual. One way to address the potential weaknesses of each approach is to treat each breast separately and to use a frailty model that assumes the same risk in paired organs of the same individual. In our study, the three models showed similar point estimates with overlapping CIs. However, comparison of disease-free survival curves of per-patient and per-breast analyses shows a longer disease-free interval in the latter by approximately 5 to 10 years. This may partially be explained by the greater disease-free follow-up time contributed by the contralateral breast. These results implicate DCIS in the breast cancer pathway in BRCA carriers. We projected how DCIS might influence the estimation of BRCA mutation risk in the standard, BRCAPRO model, updated version of CAGene 4b, which was released in March 2006 from the University of Texas Southwestern Medical Center.13 This Bayesian prediction model is based on the cumulative incidence of breast and ovarian cancer in the general population and in families with BRCA1/2 mutations. A recent study found this model to have the highest accuracy for assessment of the probability of a deleterious mutation.27 The penetrance estimates used in BRCAPRO are constantly updated, but the original BRCA-positive and -negative pedigree information entered into the model only included invasive cancers and, therefore, did not include DCIS. Although the current version, BRCAPRO 4b, allows the inclusion of DCIS in the model, DCIS is not a heavily weighted predictor of mutation status. Figure 1 illustrates the hypothetical pedigree of a 38-year-old non–Ashkenazi Jewish proband with invasive breast cancer. Her paternal grandmother had ovarian cancer at age 54 years; her father was diagnosed with prostate cancer at age 58 years. For this proband, the combined prior probability of a positive BRCA1 or BRCA2 test is 21% using BRCAPRO. If this patient had been unaffected, the prior probability would be reduced to 0.008%. Interestingly, with a diagnosis of DCIS instead of invasive cancer, the prior probability of a positive test remains unchanged at 0.008% with this prediction model. In our study cohort of women with pure DCIS, the median predicted probability of carrying a BRCA mutation using the BRCAPRO model was 1.7%. Results of testing identified 20% of patients as actual carriers of a deleterious BRCA mutation. Thus, BRCAPRO clearly underestimated the risk of a mutation in this group.
Our study has some limitations. BRCA carriers and noncarriers are comparable in that all women referred for genetic counseling and testing have a strong family history of breast and/or ovarian cancer. However, BRCA carriers differ in the proportion with Ashkenazi Jewish ancestry, more recent birth cohort, history of ovarian cancer, and parity. Thus, these variables were controlled for in the multivariate analyses. Our control group tested negative for BRCA mutations, but we cannot rule out the possibility that some of these women carried undetected mutations, such as deletions or BRCA mutations that required full sequencing in Ashkenazi Jewish women negative for the three founder mutations. We believe that this type of misclassification would be sufficiently rare as to not affect our findings. Because our cohort was selected for a strong family history of cancer, our data cannot be generalized to all patients. Aside from patient selection, biases might have been introduced because of exclusions from the study, but there is no evidence to suggest that the largest groups excluded (38 patients because of unavailable records and 33 patients because of mutation variants of unknown significance) would differentially bias the outcomes. Our testing protocol requests permission for 20-year follow-up, and no women were lost during the period of study. Furthermore, survivor bias (ie, only affected women who lived were BRCA tested) would likely influence the carrier and noncarrier groups equally because of the strong effect of family history on breast cancer risk. If BRCA carriers died preferentially before testing, again, our estimate of DCIS prevalence would be lower. Volunteer bias would also be expected to affect carriers and noncarriers equally because inclusion in the database occurred before knowledge of BRCA mutation status. Taken together, the potential biases in the study would lead, if anything, to an underestimate of DCIS prevalence in mutation carriers. Our data show that women with BRCA mutations develop earlier onset of both DCIS and invasive breast cancer than comparable BRCA-negative women. The higher than expected prevalence and early onset of DCIS in BRCA carriers suggest that this preinvasive lesion is on the causal pathway to invasive disease. Therefore, premenopausal patients with DCIS and a strong family history of breast and/or ovarian cancer should be considered for BRCA risk assessment. It will be important for others to confirm these findings because the inclusion and appropriate weighting of DCIS in risk models such as BRCAPRO would offer an important opportunity for clinicians to identify BRCA mutation carriers before the onset of invasive breast or ovarian cancer.
Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment: N/A Leadership: N/A Consultant: N/A Stock: N/A Honoraria: N/A Research Funds: E. Shelley Hwang, Novartis Testimony: N/A Other: N/A
Conception and design: E. Shelley Hwang, Jane L. McLennan, Beth B. Crawford, John L. Ziegler Provision of study materials or patients: Jane L. McLennan, Beth B. Crawford Collection and assembly of data: E. Shelley Hwang, Jane L. McLennan, Dan H. Moore, Beth B. Crawford, John L. Ziegler Data analysis and interpretation: E. Shelley Hwang, Jane L. McLennan, Dan H. Moore, Beth B. Crawford, John L. Ziegler Manuscript writing: E. Shelley Hwang, Jane L. McLennan, Dan H. Moore, Beth B. Crawford, John L. Ziegler Final approval of manuscript: E. Shelley Hwang, Jane L. McLennan, Dan H. Moore, Beth B. Crawford, Laura J. Esserman, John L. Ziegler
We thank the genetic counselors and clinicians at the University of California, San Francisco Breast Care Center for their important contributions as well as our patients who willingly and generously provided data and samples for research. This work is dedicated to the memory of Young Hee Hwang.
published online ahead of print at www.jco.org on January 8, 2007. Supported in part by a gift from the Avon Foundation, private donors, and a National Cancer Institute Comprehensive Cancer Center Support Grant (NCI/NIH Grant No. P30 CA82103). Presented in part at the 27th Annual Meeting of Annual Breast Cancer Symposium, December 8-11, 2004, San Antonio, TX. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Burstein HJ, Polyak K, Wong JS, et al: Ductal carcinoma in situ of the breast. N Engl J Med 350:1430-1441, 2004 2. Vicini FA, Recht A: Age at diagnosis and outcome for women with ductal carcinoma-in-situ of the breast: A critical review of the literature. J Clin Oncol 20:2736-2744, 2002 3. Kerlikowske K, Molinaro A, Cha I, et al: Characteristics associated with recurrence among women with ductal carcinoma in situ treated by lumpectomy. J Natl Cancer Inst 95:1692-1702, 2003 4. Radford DM, Phillips NJ, Fair KL, et al: Allelic loss and the progression of breast cancer. Cancer Res 55:5180-5183, 1995 5. O'Connell P, Pekkel V, Fuqua SA, et al: Analysis of loss of heterozygosity in 399 premalignant breast lesions at 15 genetic loci. J Natl Cancer Inst 90:697-703, 1998 6. Fujii H, Szumel R, Marsh C, et al: Genetic progression, histological grade, and allelic loss in ductal carcinoma in situ of the breast. Cancer Res 56:5260-5265, 1996 7. Hwang ES, Devries S, Chew KL, et al: Patterns of chromosomal alterations in breast ductal carcinoma in situ. Clin Cancer Res 10:5160-5167, 2004 8. Sun CC, Lenoir G, Lynch H, et al: In-situ breast cancer and BRCA1. Lancet 348:408, 1996[Medline] 9. Narod SA, Foulkes WD: BRCA1 and BRCA2: 1994 and beyond. Nat Rev Cancer 4:665-676, 2004[CrossRef][Medline] 10. Narod SA, Offit K: Prevention and management of hereditary breast cancer. J Clin Oncol 23:1656-1663, 2005 11. Breast Cancer Linkage Consortium: Pathology of familial breast cancer: Differences between breast cancers in carriers of BRCA1 or BRCA2 mutations and sporadic cases. Lancet 349:1505-1510, 1997[CrossRef][Medline] 12. Lakhani SR, Jacquemier J, Sloane JP, et al: Multifactorial analysis of differences between sporadic breast cancers and cancers involving BRCA1 and BRCA2 mutations. J Natl Cancer Inst 90:1138-1145, 1998 13. Parmigiani G, Berry D, Aguilar O: Determining carrier probabilities in breast cancer susceptibility genes BRCA1 and BRCA2. Am J Hum Genet 62:145-146, 1998[CrossRef][Medline] 14. Antoniou AC, Pharoah PP, Smith P, et al: The BOADICEA model of genetic susceptibility to breast and ovarian cancer. Br J Cancer 91:1580-1590, 2004[Medline] 15. Rebbeck TR, Lynch HT, Neuhausen SL, et al: Prophylactic oophorectomy in carriers of BRCA1 or BRCA2 mutations. N Engl J Med 346:1616-1622, 2002 16. Kauff ND, Satagopan JM, Robson ME, et al: Risk-reducing salpingo-oophorectomy in women with a BRCA1 or BRCA2 mutation. N Engl J Med 346:1609-1615, 2002 17. Frank TS, Deffenbaugh AM, Reid JE, et al: Clinical characteristics of individuals with germline mutations in BRCA1 and BRCA2: Analysis of 10,000 individuals. J Clin Oncol 20:1480-1490, 2002 18. Claus EB, Stowe M, Carter D: Family history of breast and ovarian cancer and the risk of breast carcinoma in situ. Breast Cancer Res Treat 78:7-15, 2003[CrossRef][Medline] 19. Kerlikowske K, Barclay J, Grady D, et al: Comparison of risk factors for ductal carcinoma in situ and invasive breast cancer. J Natl Cancer Inst 89:76-82, 1997[Abstract] 20. Khurana KK, Loosmann A, Numann PJ, et al: Prophylactic mastectomy: Pathologic findings in high-risk patients. Arch Pathol Lab Med 124:378-381, 2000[Medline] 21. Hoogerbrugge N, Bult P, DeWidt-Levert LM, et al: High prevalence of premalignant lesions in prophylactically removed breasts from women at high hereditary risk for breast cancer. J Clin Oncol 21:41-45, 2003 22. Kauff ND, Brogi E, Scheuer L, et al: Epithelial lesions in prophylactic mastectomy specimens from women with BRCA mutations. Cancer 97:1601-1608, 2003[CrossRef][Medline] 23. Kriege M, Brekelmans CT, Boetes C, et al: Efficacy of magnetic resonance imaging and mammography for breast cancer screening in women with a familial or genetic predisposition. N Engl J Med 351:427-437, 2004 24. Scheuer L, Kauff N, Robson M, et al: Outcome of preventive surgery and screening for breast and ovarian cancer in BRCA mutation carriers. J Clin Oncol 20:1260-1268, 2002 25. Adem C, Reynolds C, Soderberg CL, et al: Pathologic characteristics of breast parenchyma in patients with hereditary breast carcinoma, including BRCA1 and BRCA2 mutation carriers. Cancer 97:1-11, 2003[CrossRef][Medline] 26. Aalen OO: Heterogeneity in survival analysis. Stat Med 7:1121-1137, 1988[Medline] 27. James PA, Doherty R, Harris M, et al: Optimal selection of individuals for BRCA mutation testing: A comparison of available methods. J Clin Oncol 24:707-715, 2006 Submitted August 30, 2005; accepted September 7, 2006. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
