Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Erratum (v213,p930)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hrung, J. M.
Right arrow Articles by Schwartz, J. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hrung, J. M.
Right arrow Articles by Schwartz, J. S.
(Radiology. 1999;213:39-49.)
© RSNA, 1999

Breast Imaging

Cost-effectiveness of MR Imaging and Core-Needle Biopsy in the Preoperative Work-up of Suspicious Breast Lesions1

Janie M. Hrung, MD, Curtis P. Langlotz, MD, PhD, Susan G. Orel, MD, Kevin R. Fox, MD, Mitchell D. Schnall, MD, PhD and J. Sanford Schwartz, MD

1 From the School of Medicine (J.M.H.), the Departments of Radiology (J.M.H., C.P.L., S.G.O., M.D.S.) and Medicine (K.R.F., J.S.S.), the Center for Clinical Epidemiology and Biostatistics (C.P.L.), and the Leonard Davis Institute of Health Economics (C.P.L., J.S.S.), University of Pennsylvania Medical Center, Philadelphia. From the 1997 RSNA scientific assembly. Received September 29, 1997; revision requested November 26; final revision received January 27, 1999; accepted April 30. Supported in part by National Institutes of Health grants R01-CA58358 and R29-LM06238, U.S. Army grant RP950855, U.S. Department of Health and Human Services grant RFP #282-96-0026, and an American Heart Association Student Research Fellowship. Address reprint requests to C.P.L., 5 Biddle Way, Mount Laurel, NJ 08054 (e-mail: langlotz@erols.com).


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To assess the clinical and economic consequences of the use of preoperative breast magnetic resonance (MR) imaging and core-needle biopsy (CNB) to avert excisional biopsy (EXB).

MATERIALS AND METHODS: A decision-analytic Markov model was constructed to compare MR imaging, CNB, and EXB without preoperative testing in a woman with a suspicious breast lesion. Stage-specific cancer prevalence, tumor recurrence, progression rates, and MR imaging and CNB sensitivity and specificity were obtained from the literature. Cost estimates were obtained from the literature and from the Medicare fee schedule.

RESULTS: EXB without preoperative testing was associated with the greatest quality-adjusted life expectancy, followed by MR imaging and CNB; life expectancies were 17.409, 17.405, and 17.398 years, respectively. EXB resulted in the greatest lifetime treatment cost ($31,438), followed by MR imaging ($29,072) and CNB ($28,573). Results were robust over a wide range of cancer prevalence, stage distribution, tumor progression rates, and procedure and treatment costs. Incremental cost-effectiveness ratios showed that preoperative testing was cost-effective, but the choice between MR imaging and CNB was highly dependent on the accuracy of each test and to patient preferences.

CONCLUSION: Preoperative testing of most suspicious breast lesions was cost-effective. More precise estimates of MR imaging and CNB test performance characteristics are needed. Until those are available, patient preferences should inform individual decisions regarding preoperative testing.

Index terms: Breast, biopsy, 00.1261, 00.1267 • Breast neoplasms, diagnosis, 00.30 • Breast neoplasms, MR, 00.1214 • Cost-effectiveness, 00.1214, 00.1261, 00.1267 • Economics, medical


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Breast cancer is the most common cancer and is the second leading cause of cancer-related mortality in women (1). When breast cancer is suspected, excisional biopsy (EXB) is the current standard management option, because biopsy provides a sample for histopathologic diagnosis and is often therapeutic in patients with early-stage cancer. Unfortunately, because of the inherent limitations of mammography, 55%–85% of women who undergo EXB are found to have benign breast lesions (2–9). Additional disadvantages of EXB include its expense and the morbidity that arises from the surgical procedure.

Two less invasive diagnostic options are now becoming available and have the potential to decrease the number of biopsies performed for benign disease. Magnetic resonance (MR) imaging of the breast and core-needle biopsy (CNB) are two preoperative tests that may improve discrimination between malignant and benign disease. MR imaging requires a woman to lie prone in the bore of an MR imaging magnet for about 40 minutes and to receive an intravenous injection of a gadolinium-based contrast agent. It is, therefore, a minimally invasive test, but tissue is not obtained for histopathologic examination. CNB requires a woman to lie prone on a biopsy table for about 40 minutes and to receive a subcutaneous injection of a local anesthetic, with approximately five throws of a 14-gauge needle (6).

A suspicious breast lesion can be localized manually at ultrasonography (US) or stereotactically, usually at the discretion of the radiologist. CNB is moderately invasive and yields tissue for histopathologic diagnosis. However, a negative test result could be due to sampling error and could result in a delayed diagnosis of cancer.

Our objective was to assess the potential clinical role for preoperative CNB and MR imaging in the diagnostic evaluation of mammographically or clinically suspicious breast lesions. Specifically, we sought to assess whether preoperative testing with MR imaging or CNB can cost-effectively decrease the high rate of EXB for benign disease, thereby sparing some women the risk, morbidity, and cost of an unnecessary EXB, without substantially increasing the risk of cancer progression during short-interval follow-up.


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Structure of the Decision Model
A decision-analytic Markov model was constructed by using the computer program Decision Analysis by TreeAge (DATA; TreeAge Software, Williamstown, Mass) to examine and compare alternative diagnostic options for a woman with a suspicious breast lesion. A suspicious breast lesion was defined as a palpable abnormality at physical examination or as a category 4 lesion in the American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) that was detected at mammography as a mass with or without microcalcifications or distortion. Figure 1a shows a simplified representation of the decision tree. The base-case patient was a 55-year-old woman without coexistent disease. We compared the following three management options: (a) preoperative MR imaging of the breast, (b) preoperative CNB with or without stereotactic localization, and (c) EXB without preoperative testing.



View larger version (26K):
[in this window]
[in a new window]
  		Figure 1a. Diagram depicts decision tree. (a) A woman with a suspicious breast lesion can choose breast MR imaging, CNB, or EXB as a diagnostic work-up option. CA = cancer, DCIS = ductal carcinoma in situ, FN = false-negative, FP = false-positive, f/u = follow-up, MRI = MR imaging, Node-neg = node-negative, Node-pos = node-positive, TN = true-negative, TP = true-positive, {circ} = transition state, M = Markov node. (b) Typical Markov subtree for a woman who had node-negative breast cancer depicts all possible transition states ({circ}), which appear immediately to the right of the Markov node (M). In b, numbers represent the probability of moving from one state to another for a patient who survives EXB. Chemo = chemotherapy, s/e = side effects.

 


View larger version (22K):
[in this window]
[in a new window]
  		Figure 1b. Diagram depicts decision tree. (a) A woman with a suspicious breast lesion can choose breast MR imaging, CNB, or EXB as a diagnostic work-up option. CA = cancer, DCIS = ductal carcinoma in situ, FN = false-negative, FP = false-positive, f/u = follow-up, MRI = MR imaging, Node-neg = node-negative, Node-pos = node-positive, TN = true-negative, TP = true-positive, {circ} = transition state, M = Markov node. (b) Typical Markov subtree for a woman who had node-negative breast cancer depicts all possible transition states ({circ}), which appear immediately to the right of the Markov node (M). In b, numbers represent the probability of moving from one state to another for a patient who survives EXB. Chemo = chemotherapy, s/e = side effects.

 
Preoperative test results could be either positive or negative. Women (hypothetical patients) with positive preoperative test results underwent EXB. Patients with negative preoperative test results were assigned to short-interval follow-up of 6 months duration. We intentionally assumed that EXB exhibited perfect sensitivity and specificity to bias our results in favor of EXB.

Patients with true-positive test results underwent EXB, which showed cancer; the cancer would be staged and treated in standard fashion (described later in Model Assumptions). Patients with false-positive test results underwent EXB, which showed benign disease, and subsequently were assigned to routine annual mammography and follow-up. Patients with true-negative test results and false-negative test results were initially found to have benign disease and were assigned to short-interval follow-up. Women with true-negative test results had unchanged or normal mammograms at 6 months with return to routine annual screening cycles.

We initially assumed that patients with false-negative preoperative test results had interval changes on their repeat mammograms at 6-month follow-up, with the diagnosis of breast cancer confirmed at EXB (ie, diagnosis of breast cancer was delayed 6 months). For some patients who underwent CNB, an immediate false-negative result could be obtained (ie, when no calcifications were obtained in the core specimens or when no histopathologic diagnosis could have explained an abnormality detected at mammography). These patients were referred for EXB, without a delay in diagnosis. Treatment and prognosis were stratified by stage. Life expectancy for patients with or without cancer was modeled by using Markov subtrees.

Model Assumptions
We made the following additional assumptions in the construction of the model. Mammograms that were interpreted as abnormal were placed into three categories according to the BI-RADS, as follows: probably benign (category 3); suspicious for malignancy, for which biopsy was recommended (category 4); and malignant, according to radiologic criteria (category 5).

For patients with lesions assessed as BI-RADS category 3 lesions, 6-month follow-up was the recommended treatment strategy. BI-RADS category 3 lesions were considered probably benign, with a less than 2% likelihood of being malignant (10). For patients with lesions assessed as BI-RADS category 5 lesions, immediate EXB was the recommended treatment strategy because of the high positive predictive value of this interpretation (11). In the model, we included patients whose lesions were assessed as BI-RADS category 4 lesions, whose pretest likelihood of breast cancer ranged widely from approximately 2% to 80%.

Because late-stage cancers manifest with more definitive findings of malignancy at mammographic and physical examination, thereby obviating preoperative testing, we assumed that no patient had distant metastatic disease and that the diagnosis of cancer was distributed among ductal carcinoma in situ, node-negative breast cancer, and node-positive breast cancer.

Model parameters, including prevalence and stage distribution of cancer, were based on values obtained after a computer search and a comprehensive review of the medical literature. Table 1 shows the parameters for the base case and the range of sensitivity analysis. The doubling time of a typical breast cancer was estimated to be approximately 157 days (19,20). As the volume of a spherical tumor doubles, its diameter increases by the cube root of 4, or 1.59. From this value and from data from Hellman (21), which indicate the proportion of patients with eventual metastases as a function of tumor diameter, we estimated that 15% of tumors progressed in stage during the 6-month follow-up.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Model Parameters and Range of Sensitivity Analysis
 
Characteristics of diagnostic test performance, which included sensitivity and specificity of MR imaging and CNB, were obtained from recent studies in patients who underwent each test (12,13). We revised the test performance of CNB because the investigators considered atypical ductal hyperplasia at CNB as a benign lesion. Atypia diagnosed at CNB is currently considered to be an indication for EXB because the high proportion of lesions that show atypia also contain either ductal carcinoma in situ or invasive carcinoma (51). Although Gisvold et al (13) considered atypia to be a benign condition, they provided information regarding the diagnosis at CNB and at EXB for each lesion. From those data, we were able to recreate a 2 x 2 table, which we used to calculate the sensitivity and specificity with a new threshold for a positive result of a test—one that included a diagnosis of atypia at CNB. The revised test performance is shown in Table 1. We defined a positive test result at CNB as one that included a diagnosis of atypia, after which a patient would undergo immediate EXB.

Estimates of Effectiveness
Life expectancy for patients with cancer was modeled by using Markov subtrees, which were adapted from the model by Hillner and Smith (22) (Fig 1b). The main predictors of survival were modeled as nodal status and the length of the disease-free interval after either breast-conserving therapy or mastectomy (35) (see Table 1 for base-case values).

All patients entered the subtrees in the "well" state, which signified the condition of a patient who had cancer and who had undergone local treatment. All patients with node-positive cancer underwent chemotherapy with either major or minor side effects (23,24). Patients with a positive estrogen-receptor status (25) also underwent treatment with tamoxifen citrate for 5 years (23,24,30). We made a conservative assumption that 30% of the patients with node-negative cancer underwent chemotherapy, on the basis of their negative estrogen-receptor status (23,24). We assumed that the remaining 70% of patients with node-negative results had a positive estrogen-receptor status and that they were treated for 5 years with tamoxifen (23,24,31).

Patients entered a "well (after systemic treatment)" state from either "chemotherapy" or "tamoxifen" states. Adjuvant systemic therapy was not indicated in patients with ductal carcinoma in situ. Thus, these patients went straight to the "well (after systemic treatment)" state (14). We assumed that patients in the "well (after systemic treatment)" state returned to their baseline state of full health and were asymptomatic.

Each Markov cycle represented 1 year. In each subsequent year, a patient in the "well (after systemic treatment)" or "tamoxifen" state remained in the same state, had a recurrence of cancer, or died of causes other than breast cancer. After a disease recurrence, the patient survived, died of breast cancer, or died of causes other than breast cancer. We assumed that all patients who had three recurrences of breast cancer subsequently died of cancer-related causes.

The stage-specific annual probability of the first cancer recurrence was calculated by using the declining exponential approximation of life expectancy (52) to convert stage-specific disease-free survival rates (22–34,36–38) to annual recurrence rates. We assumed that cancer progression rates were independent of initial stage after the first recurrence. The probability of dying of causes other than breast cancer was obtained from U.S. vital statistics data (41). Other transition probabilities, such as the annual probability of second and third recurrences and annual mortality after recurrence, were obtained from Hillner and Smith's (22) previous model of early-stage breast cancer.

For patients without cancer, a simple three-state Markov model was constructed. A patient entered the subtree through the "well" state. In each subsequent year, the patient could become sick (of any cause) or die. If she became sick, she could remain sick or die. We used age-adjusted mortality rates for causes other than breast cancer to approximate the incidence of sickness in a previously well cohort. We assumed a 70% annual mortality for patients after they became sick.

Life expectancy was converted to quality-adjusted life-years (QALYs) for patients once cancer was diagnosed and staged or once benign breast disease was diagnosed. The utilities for various health states and the transition probabilities between health states were obtained from the model by Hillner and Smith (22) and from the literature (Table 1). For the patient without cancer, the "sick (of any cause)" state was assigned a utility of 0.5. In the base case, no quality adjustment for short-term morbidity from diagnostic testing was made because of a lack of available information. We assumed that any quality adjustments were trivial because of the transient nature of any short-term morbidity (ie, inconvenience, hematoma, etc). In the base-case analysis, life-years were discounted for time effects to the year of diagnosis by using an annual rate of 3% (50).

Estimates of Costs
The model was used to examine the direct medical costs that resulted from the diagnosis of disease and treatment of women with suspicious breast lesions. Costs and references are summarized in Table 1. Cost estimates for MR imaging, CNB, EXB, treatment of the primary breast cancer and subsequent recurrences, follow-up, and terminal costs for deaths from breast cancer and other causes were obtained from the published medical literature between 1988 and 1995 (43–49,53) and from the 1995 Medicare Resource-based Relative Value Scale. All costs were adjusted to 1995 dollars by using the inflation-adjusted Consumer Price Index for medical care services (54).

In some cases, cost estimates for an intervention were available from multiple sources. When this occurred, we used the mean value, except in the following cases. If the intervention was a single procedure (ie, mammography, MR imaging, or CNB), we used the 1995 Medicare reimbursement value. The cost of a radiologic study included the technical and professional fee. In addition, the cost for CNB included the fee for either US or stereotactic guidance and pathologic examination of the specimen that was obtained. We estimated that the costs after breast cancer treatment were those for routine surveillance only (44), since investigators in two randomized trials (55,56) found that intensive testing after breast cancer treatment does not improve survival or influence the health-related quality of life. Therefore, each patient in the model with cancer was followed up routinely after each recurrence. The incremental cost of short-interval follow-up was the cost of an office visit and an additional mammographic examination. We varied this cost estimate during sensitivity analysis to consider the possibility that an additional MR imaging examination or CNB would be included in the short-interval follow-up regimen.

We assumed that the subsequent annual costs for a well patient without cancer were half the annual health costs of an asymptomatic patient after treatment for the primary cancer. We assumed $10,000 was the cost of becoming sick for a patient without cancer (which was slightly less than half the terminal cost of dying of causes other than breast cancer). In the base-case analysis, costs were discounted to the year of diagnosis by using an annual rate of 3% (50).

Sensitivity Analysis
Univariate sensitivity analysis was performed for each parameter of the model over a clinically plausible range of values (Table 1). Multivariate sensitivity analysis was performed, where appropriate. The effect of these changes on outcomes, costs, and cost-effectiveness was identified, and threshold values were calculated for sensitive variables.


   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Base-Case Analysis
The results of the base-case analysis are summarized in Table 2. Differences in predicted life expectancy among patients who underwent EXB, preoperative MR imaging, and preoperative CNB were minor, with life expectancies of 17.409, 17.405, and 17.398 QALYs, respectively (1.5 days for EXB vs MR imaging and 4.0 days for EXB vs CNB). Patients who underwent EXB had the greatest life expectancy and also had the highest lifetime treatment cost per patient, which was $31,438. MR imaging was less costly ($29,072). CNB was least costly ($28,573). Thus, it cost $576,258 to gain 1 QALY by using EXB instead of MR imaging and $253,540 to gain 1 QALY by using EXB instead of CNB. Choosing MR imaging instead of CNB cost an additional $69,446 per QALY gained (Table 3).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Base-Case Results
 

View this table:
[in this window]
[in a new window]
  		TABLE 3. Base-Case Incremental Cost-effectiveness Ratios
 
Sensitivity Analysis
One-way sensitivity analysis was performed for all relevant parameters of the model. The rank order of quality-adjusted life expectancy for the three options remained stable over wide ranges of prevalence and stage distributions of cancer, tumor progression rates, annual probabilities of recurrence, and utilities for the various health states in the model. The cost rankings for the three options remained stable over a range of costs for diagnostic testing, from 50% to 150% of base-case estimates. In addition, we varied the cost of short-interval follow-up for patients with false-negative test results to determine how repeating either MR imaging or CNB with mammography at 6 months affected the optimal strategy. CNB remained the least costly option. Results were also insensitive to the costs of treating both primary and recurrent cancer, the costs of routine follow-up for patients with or without breast cancer, the costs of treatment for sickness other than breast cancer, and the terminal costs for mortality from breast cancer and other causes. Table 4 shows incremental cost-effectiveness ratios at discount rates of 0%, 5%, and 10%.


View this table:
[in this window]
[in a new window]
  		TABLE 4. Sensitivity Analysis of Discount Rate on Incremental Cost-effectiveness Ratios
 
Predicted life expectancy from the EXB or MR imaging strategies became identical when the sensitivity of MR imaging increased to 100% or when the sensitivity of EXB decreased to 95%. Two-way sensitivity analysis across the full range of sensitivity and specificity for both MR imaging and CNB revealed that the maximal gain in life expectancy from choosing EXB over preoperative testing was 0.10 year, or 37 days, which reflected the relatively low rates of tumor progression over 6 months.

The incremental cost-effectiveness ratio of EXB relative to MR imaging or CNB consistently remained above $100,000 per QALY gained for all combinations of sensitivity and specificity for preoperative testing. Incremental cost-effectiveness ratios greater than $100,000 per QALY are "generally agreed to be unattractive" for the allocation of health care resources (57).

Two-way sensitivity analysis revealed that the choice between MR imaging and CNB was exquisitely sensitive to estimates of MR imaging and CNB test performance. Table 5 gives the additional cost to gain 1 QALY by choosing MR imaging over CNB. Five sets of sensitivity and specificity data for CNB were compared with three sets of sensitivity and specificity data for MR imaging. The CNB values (6,7,20,59,60) and base-case MR imaging values (12) were obtained from the literature. The remaining two sets of MR imaging values were obtained from a meta-analysis (58) of data from 11 articles from the literature (4,5,8,9,61–67) by using the methods of Moses et al (68). The incremental cost-effectiveness ratio for MR imaging ranged from being borderline cost-effective in the base case to being dominated by CNB (with higher projected life expectancy and lower cost); CNB was proved to be the more cost-effective test at most estimates of test performance.


View this table:
[in this window]
[in a new window]
  		TABLE 5. Sensitivity Analysis of Test Performance on Cost-effectiveness Ratio (MR Imaging vs CNB)
 
The choice between MR imaging and CNB as the most cost-effective preoperative test also proved to be sensitive to estimates of breast cancer risk. Although the risk of cancer ranges widely for the many lesions that are currently placed in BI-RADS category 4, the category is currently not stratified further (ie, categories 4a, 4b, and 4c for minimally, moderately, and highly suspicious lesions, respectively). However, using the model, we projected costs and life expectancy over time for patients at different risk levels; we estimated the risks of "minimally, moderately, and highly suspicious" lesions to be 5%, 25%–50%, and 75%, respectively. Although preoperative testing proved to be cost-effective relative to EXB at all levels of prevalence, the choice of optimal strategy was sensitive to the prevalence of cancer; MR imaging was preferred as the risk of cancer increased (Table 6).


View this table:
[in this window]
[in a new window]
  		TABLE 6. Sensitivity Analysis of Prevalence of Breast Cancer on Cost-effectiveness Ratio (MR Imaging vs NCB)
 
Efficacy and, therefore, cost-effectiveness were extremely sensitive to patient preferences regarding EXB (Fig 2). The disutility of diagnostic testing represents the fraction of a year of perfect health a woman would be willing to give up to avoid having to undergo a diagnostic test and to avoid its associated short-term morbidity. When a patient's disutility for EXB was greater than or equal to 0.005 year, or 2 days, MR imaging became the option that yielded the greatest QALY. As a patient's disutility for EXB increased above 0.058 year, or 21 days, the optimal strategy changed again, and CNB was preferred.



View larger version (14K):
[in this window]
[in a new window]
  		Figure 2. Diagram depicts sensitivity analysis of the disutility of EXB. The optimal strategy changes from EXB to MR imaging at a threshold of 0.005 year, or 2 days, and from MR imaging to CNB at a threshold of 0.058 year, or 21 days. dEXB = disutility of EXB, MRI = MR imaging, NCB = core-needle biopsy, QALE = quality-adjusted life expectancy.

 
Some authors have suggested that a change in life-expectancy of at least 2 months is clinically meaningful, while gains of days to a few weeks are generally considered clinical "toss-ups" (69). Thus, we performed a one-way threshold analysis of the disutility of EXB to examine the conditions under which an incremental gain of 2 months resulted. With a disutility of EXB (as defined previously) equal to 0.18 year, or 66 days, CNB was found to project QALYs that were approximately 2 months greater than those of EXB (17.393 vs 17.230 QALYs, respectively). With a disutility of EXB equal to 0.20 year, or 73 days, the projected QALYs of MR imaging exceeded those of EXB by approximately 2 months (17.378 vs 17.210 QALYs, respectively).

Not all cancers demonstrate a change in 6 months. To test whether two or three short-interval follow-ups would be necessary to identify all patients with false-negative test results, we doubled and tripled the rate of tumor progression over one short-interval follow-up in the sensitivity analysis. The rankings of the three alternatives remained unchanged.

Although the model did not specify whether a recurrence was local or systemic, we examined the issue by considering a best-case scenario, in which every recurrence in a cohort of patients was local or regional, and a worst-case scenario, in which every recurrence in a cohort of patients was systemic. In the best-case scenario, the annual probability of death after first recurrence was set at 10% (39). In the worst-case scenario, the annual probability of death after first recurrence was set at 50% (40). (Our base-case estimate for this value was 30% [Table 1].) For a cohort of patients with suspicious breast lesions, of whom many did not have cancer, the difference in QALYs that resulted from the best- and worst-case scenarios was small—17.471 versus 17.347 QALYs, or approximately 1.5 months. The rank order of EXB, MR imaging, and CNB as diagnostic options remained unchanged over this range of annual mortality after first recurrence.


   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
The results of our model showed that preoperative testing was a cost-effective alternative to EXB, but the choice between MR imaging and CNB was highly dependent on the accuracy of each test and on patient preferences. Until more precise estimates of MR imaging and CNB test performance characteristics are available, individual patient preferences should be a major factor in making decisions among preoperative testing options.

The detailed results of our model were as follows. For a woman with a suspicious breast lesion, EXB provided the greatest quality-adjusted life expectancy, followed by MR imaging, then CNB. This small difference reflected the relatively high estimated accuracy of preoperative testing and the relatively low estimated rate of tumor progression over 6 months. Because the differences in life expectancy among the three options were very small, the differences were likely offset by the short-term discomforts of undergoing EXB. Preoperative testing allows some women with benign disease to avoid the financial and psychologic costs of undergoing an unnecessary EXB, since women with breast cancer can be observed for 6 months without a substantial compromise in life expectancy.

EXB was also associated with the greatest mean cost per patient, followed by MR imaging, then CNB. The cost savings of preoperative testing can be attributed to a reduction in the number of unneeded EXBs. The higher specificity and lower false-positive test rate of CNB relative to MR imaging allows more women to avoid EXB; thus, CNB is the least expensive diagnostic work-up option.

In addition, incremental cost-effectiveness ratios for choosing EXB versus preoperative testing were higher than those society typically finds acceptable. Thus, in the base case, preoperative testing was cost-effective relative to referring every patient with a suspicious breast lesion for EXB, a finding that was confirmed at sensitivity analysis (incremental cost-effectiveness ratio for EXB vs preoperative testing consistently remained above $100,000 per QALY gained).

On the other hand, the choice of optimal preoperative test was extremely sensitive to estimates of preoperative test performance. The incremental cost-effectiveness ratio for MR imaging ranged from being borderline cost-effective in the base case to being dominated by CNB (CNB with higher projected life expectancy and lower cost); CNB was proved to be the more cost-effective test at most estimates of test performance.

The value of preoperative testing, however, goes beyond a simple reduction in diagnostic costs. The results of both CNB and breast MR imaging can provide both the physician and the patient with knowledge that can inform treatment planning decisions. For example, when a malignant tumor is detected at CNB, complete surgical resection often can be performed in a single operative procedure. Likewise, MR imaging can be used to identify the size and extent of likely tumors and the presence or absence of multifocality to facilitate the decision between breast-conserving therapy or mastectomy. When test results indicate that the most likely diagnosis is benign disease, CNB and MR imaging allow a patient to avoid the cost and morbidity of surgical EXB. Preoperative testing also provides a diagnostic alternative for patients who refuse an open EXB.

Our base-case estimate of CNB test performance was obtained from the study by Gisvold et al (13) in 160 patients who underwent stereotactic core biopsy and subsequent EXB. To our knowledge, this is the largest published study to date in which the presumptive diagnosis of each lesion sampled at CNB was confirmed at histopathologic examination with EXB. We revised the estimates of CNB test performance to reflect that atypical ductal hyperplasia is a precancerous lesion that is best managed with EXB, not with short-interval follow-up. The study results were confirmed by findings from three smaller studies, which were used in our sensitivity analysis (6,7,59), in which the diagnosis of each lesion at CNB was confirmed at EXB. Another study included in the sensitivity analysis was a large multiinstitutional study (60); however, only diagnoses of cancer at CNB were confirmed at EXB. The Radiologic Diagnostic Oncology Group V is conducting an ongoing multicenter study of percutaneous stereotactic biopsy as a minimally invasive alternative to open surgical EXB for impalpable breast lesions. The results of this study should provide a more accurate estimate of the sensitivity and specificity of CNB, at least for nonpalpable lesions (70).

Our base-case estimates of MR imaging test performance were obtained from the study by Nunes et al (12) in 192 patients who were evaluated at MR imaging. To our knowledge, this is one of the largest studies to date with histopathologic confirmation at EXB. A larger multiinstitutional study of the sensitivity and specificity of breast MR imaging is also under way. Until its results are available, we believe meta-analysis of data from the current literature for breast MR imaging provides the best estimate of breast MR imaging test performance. The preliminary values obtained from such a study were used in the sensitivity analysis of our model.

While the patient populations in the study by Gisvold et al (13) of CNB and the study by Nunes et al (12) of MR imaging are somewhat different, we do not believe these differences are substantial. Although the patients in the Nunes et al study (12) had palpable lesions and tended to be younger than the patients in the Gisvold et al study (13), we believe the populations are comparable because the patients were primarily those with suspicious lesions who had been referred for EXB. The fact that the prevalence of breast cancer in the two groups was identical also suggests comparability. The patients in the Gisvold et al study (13) ranged in age from 33 to 88 years (mean age, 58 years), and all had nonpalpable lesions and were referred for EXB. Those in the Nunes et al study (12) ranged in age from 21 to 84 years (mean age, 49 years) and had palpable and/or nonpalpable lesions and were referred for EXB (186 of 192 patients) or cyst aspiration (six of 192 patients). Both groups of patients had a 41% prevalence of breast cancer.

Two studies used in the sensitivity analysis (59,60) (Table 5) included patients with both palpable and nonpalpable lesions, which were localized at US or at stereotactic examination performed at the discretion of the radiologist. The patients in the Pijnappel et al study (59) ranged in age from 26 to 83 years (mean age, 55 years). Parker et al (60) did not give an age range or mean age.

No diagnostic test is perfect, even one such as EXB, which is an accepted standard. However, we made a conservative assumption that biased the results in favor of EXB, which is the current standard of care. Despite this intentional bias, the expected quality-adjusted life expectancy with MR imaging or CNB was comparable to that with EXB. If the test performance parameters for EXB were decreased, MR imaging and CNB would become more preferable.

Estimates of the diagnostic test performance of CNB and MR imaging may change, as both technology and technique continue to be refined and as the results of larger multiinstitutional studies become available. As investigators refine protocols and gain experience with increasing numbers of patients, estimates of test performance may become more accurate. On the other hand, less-experienced investigators who are still learning to optimally interpret studies from these modalities may not achieve the same sensitivity and specificity as the investigators who pioneered them. Newer biopsy methods, such as vacuum-assisted breast biopsy, appear to be promising potential alternatives to EXB for the diagnosis of a suspicious breast lesion. At the time this study was designed, these methods were not widely available. To our knowledge, several reports in the literature describe these new biopsy techniques (71,72), but no data are currently available regarding the sensitivity and specificity of these new biopsy methods. When this information becomes available, it will be valuable to model the cost-effectiveness of these methods alongside those evaluated in this study.

Until more precise estimates of sensitivity and specificity for MR imaging and CNB are available, no generalization can be made about the most cost-effective preoperative test. But because our sensitivity analysis indicated that even large differences in test performance had a very small effect on efficacy, it is likely that patient preferences will play a major role in the choice between these two preoperative tests, as they do in many other important clinical decisions (73,74).

As expected, our formal sensitivity analysis of patient preferences showed that both life expectancy and cost-effectiveness ratios were sensitive to patient preferences. Results were most sensitive to the disutility of EXB. If a patient was willing to trade 0.005 year (2 days) of her future life expectancy to avoid EXB, then MR imaging (which is very sensitive and minimally invasive) became the preferred strategy and remained so over the range of disutility from 2 days to 21 days. However, the relatively low specificity of MR imaging predicted that 21% of patients would still have to undergo an EXB, which would show benign disease. If the patient was willing to trade 0.058 year (21 days) of life expectancy to avoid EXB, then CNB, which had a lower false-positive test rate than MR imaging and therefore a lower EXB rate for benign disease, became the preferred option. While the differences in QALYs among the three options was small enough that the choice among them could be considered a "toss-up" (75), we have identified thresholds for the disutility of EXB above which the incremental gain was very likely to be clinically meaningful (disutility of EXB was 66 and 73 days before the EXB QALYs exceeded CNB and MR imaging QALYs, respectively, by 2 months).

To our knowledge, measurements of preferences for these procedures from actual patients, either from women at screening mammography or from women with a known suspicious lesion, are currently unavailable. Also, little is currently known about patient preferences concerning breast cancer diagnosis and treatment or about patient compliance with short-interval follow-up and/or routine screening after a diagnostic work-up that reveals negative findings. Further research in these areas will aid in the development of optimal diagnostic and treatment strategies. The critical role that preferences play in determining an optimal work-up strategy calls for increased research in both the method of measuring patient preferences and the actual measurement of these preferences to inform future decision making. At present, clinicians must rely on a careful empiric exploration of preferences with each individual patient.

CNB is a diagnostic tool that is currently available; its use continues to increase. Investigators in three studies (43,45,76) have recently examined the cost and effectiveness of CNB and EXB for suspicious breast lesions. Our study results corroborated those of Hillner et al (76), which suggest that choosing EXB over CNB results in an increase in survival of approximately 2 days at an incremental cost of $156,700 per year of life gained (without quality adjustments). Our results were also consistent with those of Doyle et al (45) and Lindfors et al (43), who examined the cost-effectiveness of CNB for women who undergo annual screening mammography versus women who do not undergo screening. They found that CNB decreases the marginal cost per year of life saved by 16.5%, when CNB replaces 80% of the EXB indicated at mammography. The cost analysis performed by Doyle et al (45) reveals a decrease in "cost per lesion needing excision," when CNB was used for "suspicious" lesions and EXB for "probably malignant" lesions, as compared with a strategy of EXB for all mammographic abnormalities.

Because the cost savings of preoperative testing are primarily related to the number of EXBs avoided, it is informative to compare this fraction in the CNB and MR imaging cohorts in our model. Based on the test performance of CNB, 32.5% of patients will have a positive test result and will require EXB (calculated as [prevalence x sensitivity] + [{1 - prevalence} x {1 - specificity}]). An additional 7.5% of patients will have interval changes that will require EXB. Thus, the total fraction of the cohort that could avoid an EXB is 100% minus 40.0%, or 60.0%. This was consistent with results from a recent study (77) of stereotactic CNB, which shows that 41.8% of patients have results that are termed "malignant, unusual, atypical, or inadequate," and all of which lead to EXB. The fraction of their cohort that could avoid an EXB is 58.2%. For the MR imaging cohort in this study, 47.3% of patients will have positive test results, and 2.7% of patients will have an interval change. The fraction avoiding EXB is thus 100% minus 50.0%, or 50.0%.

Although breast MR imaging is currently still in an investigational stage, its likely clinical role will be as a diagnostic adjunct to mammography, rather than as a screening tool. The capability of breast MR imaging to provide fat-suppressed images with high spatial resolution points to a potential role in the staging of breast cancer to facilitate surgical treatment planning (15). Another likely role for breast MR imaging is in the detection of cancer recurrence. Breast MR imaging is also very likely to be useful in situations in which the sensitivity of mammography is known to be decreased, such as the situation of women with silicone implants or that of young women with an increased proportion of dense breast tissue (78–80).

To our knowledge, no other investigators to date have examined MR imaging and CNB in a side-by-side comparison or the cost-effectiveness implications of using MR imaging as a preoperative alternative to EXB in women with documented suspicious lesions. Our results suggest that the current practice of referring all patients with suspicious breast lesions for EXB may not be optimal. CNB and MR imaging are diagnostic alternatives that are cost-effective; projections of future quality-adjusted life expectancy for CNB and MR are essentially equivalent to those of EXB.

In summary, our decision-analytic study findings demonstrate that preoperative testing was a cost-effective strategy for a woman with a suspicious breast lesion, subject to the constraints of patient preference, and that more precise estimates of diagnostic test performance are needed to determine the optimal work-up strategy. We underscore the importance of exploring individual preferences when making current decisions regarding EXB, MR imaging, or CNB as diagnostic work-up alternatives.


   Footnotes
 
Dr Schnall's laboratory may benefit from a licensing agreement concerning a phased-array breast MR imaging coil that was invented in his laboratory.

Abbreviations: BI-RADS = Breast Imaging Reporting and Data System EXB = excisional biopsy CNB = core-needle biopsy QALY = quality-adjusted life-years

Author contributions: Guarantors of integrity of entire study, J.M.H., C.P.L.; study concepts, all authors; study design, J.M.H., C.P.L., J.S.S.; definition of intellectual content, J.M.H., C.P.L.; literature research, J.M.H., C.P.L., S.G.O., K.R.F.; data acquisition, J.M.H., C.P.L., S.G.O., K.R.F., M.D.S.; data analysis, J.M.H., C.P.L., J.S.S.; manuscript preparation, J.M.H.; manuscript editing, J.M.H., C.P.L.; manuscript review, all authors.


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Centers for Disease Control. Breast cancer incidence and mortality: United States, 1992. MMWR CDC Surveill Summ 1996; 45:833-837.
  2. Meyer J, Eberlein T, Stomper P, Sonnenfeld M. Biopsy of occult breast lesions: analysis of 1261 abnormalities. JAMA 1990; 263:2341-2343.[Abstract]
  3. Opie H, , Estes N, Jewell W, Chang C, Thomas J, Estes M. Breast biopsy for nonpalpable lesions: a worthwhile endeavor?. Am Surg 1993; 59:490-494.[Medline]
  4. Gilles R, Guinebretiere JM, Lucidarme O, et al. Nonpalpable breast tumors: diagnosis with contrast-enhanced subtraction dynamic MR imaging. Radiology 1994; 191:625-631.[Abstract/Free Full Text]
  5. Gilles R, Meunier M, Lucidarme O, et al. Clustered breast microcalcifications: evaluation by dynamic contrast-enhanced subtraction MRI. J Comput Assist Tomogr 1996; 20:9-14.[Medline]
  6. Elvecrog E, Lechner M, Nelson M. Nonpalpable breast lesions: correlation of stereotaxic large-core needle biopsy and surgical biopsy results. Radiology 1993; 188:453-455.[Abstract/Free Full Text]
  7. Parker S, Lovin J, Jobe W, Burke B, Hopper K, Yakes W. Nonpalpable breast lesions: stereotactic automated large-core biopsies. Radiology 1991; 180:403-407.[Abstract/Free Full Text]
  8. Fobben E, Rubin C, Kalisher L, Dembner A, Seltzer M, Santoro E. Breast MR imaging with commercially available techniques: radiologic-pathologic correlation. Radiology 1995; 196:143-152.[Abstract/Free Full Text]
  9. Turkat T, Klein B, Polan R, Richman R. Dynamic MR mammography: a technique for potentially reducing the biopsy rate for benign breast disease. JMRI 1994; 4:563-568.
  10. Sickles EA. Periodic mammographic follow-up of probably benign lesions: results in 3,184 consecutive cases. Radiology 1991; 179:463-468.[Abstract/Free Full Text]
  11. Kerlikowske K, Grady D, Barclay J, Sickles EA, Eaton A, Ernster V. Positive predictive value of screening mammography by age and family history of breast cancer. JAMA 1993; 270:2444-2450.[Abstract]
  12. Nunes L, Schnall M, Orel S, et al. Breast MR imaging: interpretation model. Radiology 1997; 202:833-841.[Abstract/Free Full Text]
  13. Gisvold J, Goellner J, Grant C, et al. Breast biopsy: a comparative study of stereotaxically guided core and excisional techniques. AJR 1994; 162:815-820.[Abstract/Free Full Text]
  14. Ernster V, Barclay J, Kerlikowske K, Grady D, Henderson C. Incidence of and treatment for ductal carcinoma in situ of the breast. JAMA 1996; 275:913-918.[Abstract]
  15. Orel S, Schnall M, Powell C, et al. Staging of suspected breast cancer: effect of MR imaging and MR-guided biopsy. Radiology 1995; 196:115-122.[Abstract/Free Full Text]
  16. Zavertnik J, McCoy C, Robinson D, Love N. Cost-effective management of breast cancer. Cancer 1992; 69:1979-1984.[Medline]
  17. Swanson G, Lin C. Survival patterns among younger women with breast cancer: the effects of age, race, stage, and treatment. J Natl Cancer Inst Monogr 1994; 16:69-77.
  18. Porta M, Gallen M, Malats N, Planas J. Influence of "diagnostic delay" upon cancer survival: an analysis of five tumour sites. J Epidemiol Community Health 1991; 45:225-230.[Abstract]
  19. Peer P, van Dijck J, Hendriks J, Holland R, Verbeek A. Age-dependent growth rate of primary breast cancer. 1993; 71:3547-3551.
  20. Tubiana M, Koscielny S. The natural history of breast cancer: implications for a screening strategy. Int J Radiat Oncol Biol Phys 1990; 19:1117-1120.[Medline]
  21. Hellman S. Natural history of small breast cancers. J Clin Oncol 1994; 12:2229-2234.[Free Full Text]
  22. Hillner B, Smith T. Efficacy and cost-effectiveness of adjuvant chemotherapy in women with node-negative breast cancer. N Engl J Med 1991; 324:160-168.[Abstract]
  23. Early Breast Cancer Trialists' Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women (review). Lancet 1992; 339:1-15.[Medline]
  24. Early Breast Cancer Trialists' Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 1992; 339:71-85.[Medline]
  25. Tormey D, Gray R, Falkson H. Postchemotherapy adjuvant tamoxifen therapy beyond five years in patients with lymph node–positive breast cancer. J Natl Cancer Inst 1996; 88:1828-1833.[Abstract/Free Full Text]
  26. Hillner B, Desch C, Carlson R, Smith T, Esserman L, Bear H. Trade-offs between survival and breast preservation for three initial treatments of ductal carcinoma-in-situ of the breast. J Clin Oncol 1996; 14:70-77.[Abstract]
  27. Solin L, Kurtz J, Fourquet A, et al. Fifteen-year results of breast-conserving surgery and definitive breast irradiation for the treatment of ductal carcinoma in situ of the breast. J Clin Oncol 1996; 14:754-763.[Abstract/Free Full Text]
  28. Silverstein M, Barth A, Poller C, et al. Ten-year results comparing mastectomy to excision and radiation therapy for ductal carcinoma in situ of the breast. Eur J Cancer 1995; 31A:1425-1427.
  29. Warneke J, Grossklaus D, Davis J, et al. Influence of local treatment on the recurrence rate of ductal carcinoma in situ. J Am Coll Surg 1995; 180:683-688.[Medline]
  30. Swedish Breast Cancer Cooperative Group. Randomized trial of two versus five years of adjuvant tamoxifen for postmenopausal early stage breast cancer. J Natl Cancer Inst 1996; 88:1543-1549.[Abstract/Free Full Text]
  31. Fisher B, Dignam J, Bryant J, et al. Five versus more than five years of tamoxifen therapy for breast cancer patients with negative lymph nodes and estrogen receptor-positive tumors. J Natl Cancer Inst 1996; 88:1529-1542.[Abstract/Free Full Text]
  32. Leitner S, Swern A, Weinberger D, Duncan L, Hutter R. Predictors of recurrence for patients with small (one centimeter or less) localized breast cancer (T1a,b N0 M0). Cancer 1995; 76:2266-2274.[Medline]
  33. Quiet C, Ferguson D, Weichselbaum R, Helllman S. Natural history of node-negative breast cancer: a study of 826 patients with long-term follow-up. J Clin Oncol 1995; 13:1144-1151.[Abstract]
  34. Verhoef L, Stalpers L, Verbeek A, Wobbes T, van Daal W. Breast-conserving treatment or mastectomy in early breast cancer: a clinical decision analysis with special reference to the risk of local recurrence. Eur J Cancer 1991; 27:1132-1137.
  35. Recht A, Hayer DF, Eberlein TJ, Sadowsky NL. Local-regional recurrence after mastectomy or breast conserving therapy. In: Harris JR, Lippman ME, Morrow M, Hellman S, eds. Diseases of the breast. Philadelphia, Pa: Lippincott-Raven, 1996; 649-667.
  36. van Dongen J, Bartelink H, Fentiman I, et al. Factors influencing local relapse and survival and results of salvage treatment after breast-conserving therapy in operable breast cancer: EORTC trial 10801, breast conservation compared with mastectomy in TNM stage I and II breast cancer. Eur J Cancer 1992; 28A:801-805.
  37. Gardin G, Rosso R, Campora E, et al. Locally advanced non-metastatic breast cancer: analysis of prognostic factors in 125 patients homogeneously treated with a combined modality approach. Eur J Cancer 1995; 31A:1428-1433.
  38. Goldhirsch A, Gelber R, Price K, et al. Effect of systemic adjuvant treatment on first sites of breast cancer relapse. Lancet 1994; 343:377-381.[Medline]
  39. Schwaibold F, Fowble BL, Solin LJ, Schultz DJ, Goodman RL. The results of radiation therapy for isolated local regional recurrence after mastectomy. Int J Radiat Oncol Biol Phys 1991; 21:299-310.[Medline]
  40. Sherry MM, Greco FA, Johnson DH, Hainsworth JD. Metastatic breast cancer confined to the skeletal system: an indolent disease. Am J Med 1986; 81:381-386.[Medline]
  41. Economic and Statistics Administration and Bureau of the Census. Vital statistics: statistical abstract of the United States Washington, DC: U.S. Department of Commerce, 1991; 84.
  42. Smith T, Hillner B. The efficacy and cost-effectiveness of adjuvant therapy of early breast cancer in premenopausal women. J Clin Oncol 1993; 11:771-776.[Abstract]
  43. Lindfors K, Rosenquist C. Needle core biopsy guided with mammography: a study of cost-effectiveness. Radiology 1994; 190:217-222.[Abstract/Free Full Text]
  44. Kattlove H, Liberati A, Keeler E, Brook R. Benefits and costs of screening and treatment for early breast cancer: development of a basic benefit package. JAMA 1995; 273:142-148.[Abstract]
  45. Doyle A, Murray K, Nelson E, Bragg D. Selective use of image-guided large-core needle biopsy of the breast: accuracy and cost-effectiveness. AJR 1995; 165:281-284.[Abstract/Free Full Text]
  46. Baker M, Kessler L, Urban N, Smucker R. Estimating the treatment costs of breast and lung cancer. Med Care 1991; 29:40-49.[Medline]
  47. Eddy D, Hasselblad V, McGivney W, Hendee W. The value of mammography screening in women under age 50 years. JAMA 1988; 259:1512-1519.[Abstract]
  48. Taplin S, Barlow W, Urban N, et al. Stage, age, comorbidity, and direct costs of colon, prostate, and breast cancer care. J Natl Cancer Inst 1995; 87:417-426.[Abstract/Free Full Text]
  49. Hurley S, Huggins R, Snyder R, Bishop J. The cost of breast cancer recurrences. Br J Cancer 1992; 65:449-455.[Medline]
  50. Gold M, Siegel J, Russell L, Weinstein M. Cost-effectiveness in health and medicine New York, NY: Oxford University Press, 1996.
  51. Liberman L, Cohen MA, Dershaw DD, Abramson AF, Hann LE, Rosen PP. Atypical ductal hyperplasia diagnosed at stereotaxic core biopsy of breast lesions: an indication for surgical biopsy. AJR 1995; 164:1111-1113.[Abstract/Free Full Text]
  52. Beck J, Kassirer J, Pauker S. A convenient approximation of life expectancy (the "DEALE"). I. Validation of the method. Am J Med 1982; 73:883-888.[Medline]
  53. Lindfors K, Rosenquist J. The cost-effectiveness of mammographic screening strategies. JAMA 1995; 274:881-884.[Abstract]
  54. Bureau of Labor Statistics, U.S. Department of Labor. Annual change in the consumer price index for all urban consumers, medical care and all items 1990–98; Available at: http://stats.bls.gov/opub/ted/1999/Jun/wk2/art01.txt. Accessed August 3, 1999..
  55. GIVIO investigators. Impact of follow-up testing on survival and health-related quality of life in breast cancer patients: a multicenter randomized controlled trial. JAMA 1994; 271:1587-1592.[Abstract]
  56. Del Turco M, Palli D, Cariddi A, Ciatto S, Pacini P. Intensive diagnostic follow-up after treatment of primary breast cancer: a randomized trial. JAMA 1994; 271:1593-1597.[Abstract]
  57. Goldman L, Gordon DJ, Rifkind BM, et al. Cost and health implications of cholesterol lowering. Circulation 1992; 85:1960-1968.[Abstract]
  58. Hrung JM, Sonnad SS, Schwartz JS, Langlotz CP. Accuracy of MR imaging in the work-up of suspicious breast lesions: a diagnostic meta-analysis. Acad Radiol 1999; 6:387-397.[Medline]
  59. Pijnappel RM, van Dalen A, Boren Rinkes IHM, van den Tweel JG, Mali WPTM. The diagnostic accuracy of core biopsy in palpable and non-palpable breast lesions. Eur J Radiol 1997; 24:120-123.[Medline]
  60. Parker SH, Burbank F, Jackman RJ, et al. Percutaneous large-core breast biopsy: a multi-institutional study. Radiology 1994; 193:359-364.[Abstract/Free Full Text]
  61. Boetes C, Barentsz J, Mus R, et al. MR characterization of suspicious breast lesions with a gadolinium-enhanced TurboFLASH subtraction technique. Radiology 1994; 193:777-781.[Abstract/Free Full Text]
  62. Bone B, Aspelin P, Bronge L, Isberg B, Perbeck L, Veress B. Sensitivity and specificity of MR mammography with histopathological correlation in 250 breasts. Acta Radiol 1996; 37:208-213.[Medline]
  63. Flanagan F, Murray J, Gilligan P, Stack J, Ennis J. Digital subtraction in Gd-DPTA enhanced imaging of the breast. Clin Radiol 1995; 50:848-854.[Medline]
  64. Kelcz F, Santyr G, Cron G, Mongin S. Application of a quantitative model to differentiate benign from malignant breast lesions detected by dynamic, gadolinium-enhanced MRI. JMRI 1996; 6:743-752.
  65. Palmedo H, Grunwald F, Bender H, et al. Scintimammography with technetium-99m methoxyisobutylisonitrile: comparison with mammography and magnetic resonance imaging. Eur J Nucl Med 1996; 23:940-946.[Medline]
  66. Perman W, Heiberg E, Herrmann V. Half-Fourier, three-dimensional technique for dynamic contrast-enhanced MR imaging of both breasts and axillae: initial characterization of breast lesions. Radiology 1996; 200:263-269.[Abstract/Free Full Text]
  67. Stomper P, Herman S, Klippenstein D, et al. Suspect breast lesions: findings at dynamic gadolinium-enhanced MR imaging correlated with mammographic and pathologic features. Radiology 1995; 197:387-395.[Abstract/Free Full Text]
  68. Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data analytic approaches and some additional considerations. Stat Med 1993; 12:1293-1316.[Medline]
  69. Krahn M, Naglie G, Naimark D, Redelmeier D, Detsky A. Primer on medical decision analysis. IV. Analyzing the model and interpreting the results. Med Decis Making 1997; 17:142-151.
  70. Shtern F. Technology assessment: a perspective from the National Cancer Institute. Acad Radiol 1996; 3(suppl 1):S84-S87.
  71. Parker SH, Burbank F. A practical approach to minimally invasive breast biopsy. Radiology 1996; 200:11-20.[Abstract/Free Full Text]
  72. D'Angelo PC, Galliano DE, Rosemurgy AS. Stereotaxic excisional breast biopsies utilizing the advanced breast biopsy instrumentation system. Am J Surg 1997; 174:297-302.[Medline]
  73. Barry M, Mulley A, Fowler F, Wennberg J. Watchful waiting vs immediate transurethral resection from symptomatic prostatism: the importance of the patient's preferences. JAMA 1988; 259:3010-3017.[Abstract]
  74. Yin D, Forman HP, Langlotz CP. Evaluating health services: the importance of patients' preferences and quality of life. AJR 1995; 165:1323-1328.[Abstract/Free Full Text]
  75. Kassirer JP, Pauker SG. The toss-up. N Engl J Med 1981; 305:1467-1469.[Medline]
  76. Hillner B, Bear H, Fajardo L. Estimating the cost-effectiveness of stereotaxic biopsy for nonpalpable breast abnormalities: a decision analysis model. Acad Radiol 1996; 3:351-360.[Medline]
  77. Liberman L, Fahs MC, Dershaw DD, et al. Impact of stereotaxic core breast biopsy on cost of diagnosis. Radiology 1995; 195:633-637.[Abstract/Free Full Text]
  78. Harms SE, Flamig DP, Evans WP, Harries SA, Brown S. MR Imaging of the breast: current status and future potential. AJR 1994; 163:1039-1047.[Abstract/Free Full Text]
  79. Adler DD, Wahl RL. New methods for imaging the breast: techniques, findings and potential. AJR 1995; 164:19-30.[Abstract/Free Full Text]
  80. Stelling CB. MR imaging of the breast for cancer evaluation: current status and future directions. Radiol Clin North Am 1995; 33:1187-1204.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Erratum (v213,p930)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hrung, J. M.
Right arrow Articles by Schwartz, J. S.
Right arrow Search for Related Content