Detecting Occult Malignancy in Prophylactic Mastectomy: Preoperative ...

2 downloads 0 Views 274KB Size Report
Jun 22, 2007 - Background: High-risk patients undergoing prophylactic mastectomy (PM) may have unsuspected cancers identified on pathology.
Annals of Surgical Oncology 14(9):2477–2484

DOI: 10.1245/s10434-007-9356-1

Breast Oncology

Detecting Occult Malignancy in Prophylactic Mastectomy: Preoperative MRI Versus Sentinel Lymph Node Biopsy Dalliah Black, MD,1 Michelle Specht, MD,1 Janie M. Lee, MD,2 Francisco Dominguez, MD,1 Michele Gadd, MD,1 Kevin Hughes, MD,1 Elizabeth Rafferty, MD,2 and Barbara Smith, MD, PhD1

1

Department of Surgical Oncology, Massachusetts General Hospital, Yawkey Building, 7th Floor, 55 Fruit Street, Boston, Massachusetts 02114, USA 2 Department of Breast Imaging, Massachusetts General Hospital, Yawkey Building, 7th Floor, 55 Fruit Street, Boston, Massachusetts 02114, USA

Background: High-risk patients undergoing prophylactic mastectomy (PM) may have unsuspected cancers identified on pathology. The optimum way to identify and manage them is controversial. Magnetic resonance imaging (MRI) may identify occult cancer preoperatively. Sentinel lymph node biopsy (SLNB) allows intraoperative staging and axillary dissection during the same operation. We determined the efficacy and cost of MRI and/or SLNB in managing high-risk PM patients. Methods: We reviewed 192 PMs in 173 patients from 1999 to 2005. Costs were estimated for MRI and SLNB during PM by the 2005 Medicare Resource-Based Relative Value Scale. We also estimated costs and procedures for the four strategies in a larger hypothetical cohort. Results: A total of 19 (10%) of 192 PMs contained occult cancers, 14 ductal carcinoma-insitu (DCIS) and 5 invasive ductal carcinoma (IDC). In 59 patients, MRI detected an IDC but missed two DCIS and an IDC. Positive MRIs generated an additional average cost of $1207 per patient. In 56 PMs with SLNB, 6 occult cancers were found, 5 DCIS and 1 IDC, all with negative SLNBs. Adding a SLNB costs an additional average of $644. A theoretical analysis demonstrated that PM alone costs $808 per patient, PM with SLNB costs $1420, PM with MRI and selective SLNB costs $1774, and PM with routine MRI and SLNB costs $2379. Conclusions: MRI adds great cost and misses most occult cancers in PMs. SLNB allows the rare patient with occult IDC to avoid axillary dissection but adds cost. Given the low rate of unsuspected invasive cancers and the costs of MRI and SLNB, neither is recommended as standard practice for PM patients. Key Words: Prophylactic mastectomy—Breast MRI—Sentinel lymph node biopsy—Occult cancer—Cost analysis.

Breast surgeons are increasingly presented with patients at high risk for developing breast cancer. Genetic testing and improvements in risk assessment

have led to the ability to better quantify magnitude of risk. Several studies have now confirmed the ability of prophylactic mastectomy (PM) to reduce risk of breast cancer by 90%–95%.1–3 This more clearly defined risk and benefit data, together with improvements in breast reconstruction options, have led more high-risk patients to choose PM to reduce breast cancer risk.4,5 Five to 15% of high-risk patients undergoing PM will have unsuspected cancer in the mastectomy

Received April 2, 2006; accepted January 5, 2007; published online: June 22, 2007. Address correspondence and reprint requests to: Michelle Specht, MD; E-mail: [email protected] Published by Springer Science+Business Media, Inc.  2007 The Society of Surgical Oncology, Inc.

2477

2478

D. BLACK ET AL.

specimen.6–9 In a recent study of 409 patients with 436 prophylactic mastectomies, cancer was identified in 22 (5%) of 436 specimens. Of these patients, 64% had ductal carcinoma-in-situ (DCIS).10 If unsuspected invasive breast cancer was found on pathologic testing of the PM specimen, standard practice would be for that patient to return to the operating room for an axillary lymph node dissection (ALND). A second surgery is both physically and psychologically difficult for the patient, and it can potentially add risk of surgical complications, particularly if a breast reconstruction has been performed. Breast magnetic resonance imaging (MRI) is a preoperative screening tool that could potentially reduce the number of unsuspected malignancies found at PM. MRI has been shown to increase detection of small cancers in high-risk patients compared with mammography.11,12 In addition, preoperative breast MRI screening of women with a known breast cancer led to identification of previously unsuspected cancer in the contralateral breast in 5% of patients.13 However, false-positive findings on screening MRI are common14,15 and may necessitate additional invasive and/or costly diagnostic procedures. It is unclear whether the added workup associated with these false-positive MRIs outweighs the benefit of finding unsuspected cancers in patients already planning PM. Sentinel lymph node biopsy (SLNB) is a technique that accurately stages the axilla in patients with invasive breast cancer.16 It has been shown that SLNB carries a lower morbidity than ALND.17,18 Performing SLNB at the time of PM may potentially avoid the need for reoperation for axillary dissection if an unexpected malignancy is found in the PM specimen. It is unclear whether the costs associated with performing routine SLNB in all women undergoing PM outweigh the benefits of avoiding reoperation for axillary dissection in the few women found to have unsuspected invasive breast cancers. To address these issues, we studied a series of highrisk patients who underwent PM at our institution. We reviewed the findings on preoperative MRIs and on SLNBs that had been performed in these patients (1) to determine the efficacy and cost of preoperative MRI in detecting occult cancer in high-risk PM patients, and (2) to determine the ability of intraoperative SLNB to accurately stage breast cancer in PM cases and avoid the need for reoperation for axillary dissection. We wished to determine which strategy would minimize both the number of additional procedures to which patients are submitted and the overall cost. Ann. Surg. Oncol. Vol. 14, No. 9, 2007

MATERIALS AND METHODS An institutional review board–approved retrospective chart review was performed of all patients who underwent PM at the Massachusetts General Hospital from January 1999 to July 2005. A total of 173 patients had 192 unilateral or bilateral prophylactic mastectomies identified by the simple mastectomy current procedural terminology (CPT) code and preoperative diagnosis. Preoperative breast imaging results, including MRI, computed tomographic scan, and ultrasound results, were reviewed. Pathology results of SLNBs and PM specimens were reviewed. Preoperative Breast MRI MRI was performed with a 1.5-T scanner with a dedicated breast coil (Signa; GE Medical Systems, Milwaukee, WI). Localization with a gradient-echo sequence was performed, followed by a T2-weighted transverse fast spin-echo sequence (repetition time, 3500 ms; echo time, 17 ms). Three-dimensional fatsaturated spoiled gradient-echo sequences (23/6; flip angle, 30) were used to acquire high-spatial-resolution magnetic resonance images (matrix size 512 · 256, field of view 28 to 35 cm), before and for two sequences immediately after the intravenous administration of gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ). Breast MRI was ordered at the discretion of the treating surgeon. Intraoperative SLNB Lymphatic mapping was performed with radioactive 99mTc sulfur colloid alone, isosulfan blue dye alone, or dual labeling via injection into the subareolar region. All blue, hot, or lymph nodes suspicious on palpation were removed and submitted to pathology. Sentinel lymph nodes were sectioned at 2to 3-mm intervals and analyzed by both hematoxylin and eosin and cytokeratin immunohistochemistry staining. Intraoperative frozen section analysis of the sentinel lymph node was not performed at the time of PM. Cost Analysis The costs of procedures performed in our retrospective PM cohort were estimated from the 2005 Medicare Resource-Based Relative Value Scale (RVU) by using the CPT codes listed in Table 1. A theoretical cost analysis was constructed to determine the cost and number of procedures performed when

2479

OCCULT MALIGNANCY IN PROPHYLACTIC MASTECTOMY TABLE 1. 2005 Medicare values and current procedural terminology (CPT) codesa Procedure

Cost

CPT code

Unilateral breast MRI Simple mastectomy SLNB ALND MRI-guided biopsy Ultrasound-guided biopsy Diagnostic ultrasound Weighted average biopsy

$861 $783 $634 $1008 $1373 $493 $75 $832

76093 19180, 88307 38525, 38792, 88307 38745, 88307 19103, 19295, 88305, 76093, 76090 19102, 19295, 88305, 76645, 76090 76645 (ultrasound + 70% ultrasound biopsies + 30% MRI biopsies)

MRI, magnetic resonance imaging; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection. a CPT (c) 2006 American Medical Association. All rights reserved.19

several prospective management algorithms were applied to a hypothetical cohort of 1000 patients. This analysis examined patients undergoing contralateral PM because this was the largest subset in our retrospective review. The cost-analysis model was constructed with decision analysis software (TreeAge Pro 2005) to compare the following strategies: (1) PM only, (2) PM with MRI and selective SLNB if the MRI was abnormal, (3) PM with routine SLNB, and (4) PM with routine MRI and routine SLNB. Figure 1 presents the hypothetical decision analysis tree. The theoretical cost analysis was performed with the following assumptions: (1) a 5% occult cancer rate, with 2.5% invasive tumors, in prophylactic mastectomies,13 (2) patients with a positive sentinel lymph node proceed to completion ALND, (3) patients with invasive cancer in a PM specimen undergo an ALND if they did not have a SLNB, and (4) noninvasive cancers found in PM do not go on to ALND. MRI sensitivity and specificity is 71% and 90%, respectively.11 The theoretical analysis assumes that a positive MRI test result will prompt two additional procedures. A diagnostic breast ultrasound examination will be performed to evaluate whether the lesion seen on MRI is visible by ultrasound. If the lesion can be visualized by ultrasound, the patient will undergo ultrasound-guided core biopsy. If not, she will undergo MRI-guided core biopsy. The model assumes that in the base case, 70% of lesions identified on MRI are visible by ultrasound. SLNB sensitivity and specificity is 93% and 100%.16 The probability of a positive SLNB result in a high-risk patient undergoing PM who is found to have an invasive breast cancer was 15%, on the basis of the report of King et al.,6 in which 2 of 13 occult cancers had positive SLNB. The following costs were included in the model: unilateral PM, unilateral breast MRI, unilateral SLNB, unilateral ALND, diagnostic breast ultrasound, MRI-guided biopsy, and ultrasound-guided core biopsy. The cost of a percutaneous core biopsy

resulting from a positive MRI was calculated as the cost of a diagnostic breast ultrasound added to the weighted average of the costs of MRI- or ultrasoundguided biopsy on the basis of the probability that a lesion seen on MRI was visible by ultrasound. To examine the effect of variation in model parameters on model results, sensitivity analysis was performed. Univariate sensitivity analysis over a clinically plausible range was performed on all diagnostic parameters and costs. Two-way sensitivity analysis was performed for the sensitivity and specificity of both MRI and SLNB to evaluate the effect of varying diagnostic test performance on model results.

RESULTS Patient Characteristics One hundred seventy-three patients had 192 prophylactic mastectomies performed between January 1999 and July 2005. The average patient was 43.7 years old. All patients were women. Most of our patient population (89%) had a personal history of either a previous contralateral breast cancer or a concurrent contralateral breast cancer. Seventeen percent of the patients were known BRCA mutation carriers, and 52.6% had a positive family history. A positive family history was defined as any first-degree relative with breast cancer, any relative with a breast cancer diagnosed before 50 years of age, multiple family members with breast cancer, or a family history of ovarian cancer. Three patients had prior mantle radiation for Hodgkin disease, and one patient had a PTEN mutation. Occult carcinoma was detected in 19 (10%) of 192 prophylactic mastectomies, 5 invasive ductal carcinoma (IDC) (2.6%), and 14 DCIS (7.3%). Treating clinicians applied different combinations of preoperative MRI and SLNB to the 173 patients at his or her discretion. Table 2 lists the patient characteristics. Subgroups were analyzed Ann. Surg. Oncol. Vol. 14, No. 9, 2007

2480

D. BLACK ET AL.

FIG. 1. Management options in a theoretical cohort. PM, prophylactic mastectomy; SNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection; DCIS, ductal carcinoma-in-situ; TP, true positive; TN, true negative; FP, false positive; FN, false negative; Bx, biopsy.

TABLE 2. Characteristics of four retrospective treatment groups Characteristic No. PMs Bilateral Unilateral Concurrent contralateral breast cancer History of breast cancer Family history of breast or ovarian cancer BRCA mutation Occult cancers Invasive Noninvasive

PM alone (n = 92)

PM + MRI (n = 29)

PM + SLNB (n = 22)

PM + MRI + SLNB (n = 30)

104 24 (23%) 80 (77%) 45 (50%) 29 (32%) 42 (46%) 13 (14%)

32 6 (19%) 26 (81%) 18 (62%) 8 (28%) 20 (69%) 4 (14%)

23 2 (9%) 21 (91%) 14 (64%) 7 (32%) 13 (59%) 4 (18%)

33 6 (18%) 27 (82%) 14 (47%) 13 (43%) 16 (53%) 7 (23%)

9 (8.7%) 3 (3%)

0 1 (3%)

3 (14%) 0

2 (6%) 1 (3%)

PM, prophylactic mastectomy; MRI, magnetic resonance imaging; SLNB, sentinel lymph node biopsy.

separately to determine the impact of preoperative MRI and SLNB on overall management and cost. PM Without Preoperative MRI or Intraoperative SLNB Ninety-two patients (53.2%) had 104 prophylactic mastectomies (54.2%) without preoperative MRI or

Ann. Surg. Oncol. Vol. 14, No. 9, 2007

intraoperative SLNB. Twelve occult cancers (11.6%) were found on pathologic examination of the mastectomy specimens, 3 IDC (2.9%) and 9 DCIS (8.7%). The three IDC were all T1 tumors, measuring 0.15 cm, 0.6 cm, and 0.8 cm, respectively. Only the patient with the 0.8-cm IDC had a postmastectomy ALND, which showed no evidence of nodal metastases.

2481

OCCULT MALIGNANCY IN PROPHYLACTIC MASTECTOMY

Preoperative MRI and PM Fifty-nine patients (34%) who underwent 65 prophylactic mastectomies (33.9%) underwent a preoperative MRI. Thirteen (22%) of 59 patients who had a MRI were BRCA mutation carriers, whereas 17 (15%) of 114 patients who did not have a MRI were BRCA mutation carriers (Table 2). Thirty-three of the 65 prophylactic mastectomies also had SLNB at the time of surgery. There were four occult cancers (6.2%) detected at pathologic examination, two IDCs (0.6 cm and 1.0 cm) and two DCIS. There were four false-positive MRI findings and one true positive MRI finding. Suspicious MRI findings led to three breast ultrasounds, one mammogram, one chest computed tomographic scan, and five SLNBs. The true positive MRI was in a patient with a 1.0-cm IDC and a negative SLNB. Three of the four occult cancers found on pathologic review of the PM specimen were not detected by preoperative MRI, two of which were DCIS and one of which was a 0.6-cm IDC. The patient with the 0.6-cm IDC had a concurrent contralateral T1bN0 invasive lobular cancer and did not undergo completion ALND (Fig. 2a).

PM With Intraoperative SLNB Fifty-two patients (30.1%) had 56 prophylactic mastectomies (29.2%) with intraoperative SLNB. Thirty-three of these 56 prophylactic mastectomies also had a preoperative breast MRI. Two of the 33 patients had a suspicious MRI. One patientÕs MRI showed a 5-mm enhancing mass and linear enhancement suggestive of DCIS. A correlative breast ultrasound also confirmed a 5-mm mass. The second patientÕs MRI showed a 1.2-cm enhancing mass, and a breast ultrasound was negative. Because these two patients had already elected to have a mastectomy, diagnostic biopsies were not performed. SLNBs were performed and were negative. There were six occult cancers found in the mastectomy specimens, 5 DCIS (8.9%) and a 1.0-cm IDC (1.8%). The patient with the 1.0-cm IDC was spared an ALND because of a negative SLNB. Two patients were found to have positive SLNB. One patient with DCIS in the PM specimen had a positive SLNB. Tumor characteristics and receptor status of the positive SLNBs were consistent with the patientÕs contralateral multifocal T2 IDC. A second patient had a positive SLNB and did not have cancer detected in the PM despite extensive sectioning of the specimen. The metastasis in the SLNB was consistent with a contralateral inflammatory IDC. A completion

ALND revealed six additional positive lymph nodes (Fig. 2b). Cost Comparison of the Retrospective Cohort Unilateral PM without MRI or SLNB cost $783 per patient. Unilateral MRI costs $861, and the costs of the extra procedures generated from a suspicious MRI increased the average expense per patient to $1990 (Table 3). Performing intraoperative SLNB with PM had an additional average cost per patient of $644, for a total cost of $1417. Cost Analysis in a Theoretical Cohort Four strategies currently used by breast surgeons to detect occult malignancies and stage the axilla in women undergoing PM were modeled. In the first strategy, patients undergo PM without a preoperative MRI or intraoperative SLNB. In the second strategy, patients undergo routine SLNB without preoperative MRI. In the third strategy, patients have a preoperative MRI with selective SLNB for those patients found to have occult breast cancer detected by MRI. Finally, in the fourth strategy, patients undergo both preoperative MRI and intraoperative SLNB. In 1000 high-risk patients undergoing PM without preoperative MRI or intraoperative SLNB, there would be 50 occult cancers (5%) detected on pathologic examination of the mastectomy specimen. Twenty-five would be noninvasive (2.5%) and 25 would be invasive (2.5%). The average cost for this cohort would be $808 per patient, with 25 additional ALNDs performed as a second operation for patients found to have invasive breast cancer at PM (Table 4). If the 1000 patients all undergo routine intraoperative SLNB during PM, average cost would be $1420 per patient with 1004 additional procedures performed. One thousand SLNBs would be performed, and four ALNDs would be performed for positive SLNBs. There would be 975 negative SLNBs performed. If the 1000 patients all have preoperative MRI with selective SLNB based on suspicious MRI findings, MRI would detect 36 cancers preoperatively (18 noninvasive and 18 invasive) and would miss 14 cancers (7 invasive and 7 noninvasive). The 18 invasive cancers detected by MRI would undergo intraoperative SLNB during PM. MRI would not detect seven invasive cancers found on pathology and these patients would undergo ALND on the basis of the assumptions of our model. No patient would undergo Ann. Surg. Oncol. Vol. 14, No. 9, 2007

2482

D. BLACK ET AL.

a

1 TP 5 Suspicious MRI

4 FP

Pre-op MRI 59 patients 51 TN

54 Negative MRI

3 FN 173 patients 192 PMs

15 occult cancers

No MRI 114 patients

b 6 occult cancers 52 patients 56 PMs with SLNB

5 negative SLNBs

1 positive SLNB (contralateral metastases) 49 negative SLNBs

50 no cancer detected

1 positive SLNB (contralateral metastases)

173 patients 192 PMs 121 patients 136 PMs without SLNB

1 IDC – pt spared an ALND

13 occult cancers

TABLE 3. Cost of adding additional procedures to prophylactic mastectomy in the retrospective cohort Strategy

Cost per patient

PM PM + SLNB PM + MRI

$783 $1417 $1990

PM, prophylactic mastectomy; SLNB, sentinel lymph node biopsy; MRI, magnetic resonance imaging.

an unnecessary SLNB. Costs associated with routine MRI, work-up generated from a suspicious MRI, selective SLNB for suspicious MRI findings, and ALNDs performed for invasive cancers missed by MRI were included in the algorithm and resulted in an average cost of $1774 per patient. The total number of additional procedures for this cohort would be 1290. If the 1000 patients had preoperative MRI and intraoperative SLNB, average cost would be $2379 per patient. Preoperative MRI would detect 36 cancers, 18 of which are invasive. It would miss 14 cancers, 7 invasive. Routine SLNB preformed for the 18 invasive cancers would be negative in 15 cases and therefore spare 15 patients ALND. The three positive SLNBs would result in an ALND. A total of 975 patients without invasive cancer would undergo a negative SLNB. The total number of additional procedures for this cohort would be 2266. Ann. Surg. Oncol. Vol. 14, No. 9, 2007

FIG. 2. (a) Use of preoperative breast magnetic resonance imaging in retrospective cohort. PM, prophylactic mastectomy; TP, true positive; FP, false positive; TN, true negative; FN, false negative. (b) Use of SLNB in retrospective cohort. PM, prophylactic mastectomy; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection.

The literature reports a range of unsuspected cancer detected at PM of 5%–15%. On the basis of this information, we performed a sensitivity analysis varying the incidence of occult cancer from 2.5% to 25%. The rank order of the strategies did not change when we varied the occult cancer rate. The strategy of PM alone had a cost of $796–$909, which was based on a corresponding rate of 2.5%–25%. The strategy of PM with routine SLNB ranged $1419–$1435. PM with routine MRI and selective SLNB if a cancer was detected by MRI ranged $1750–$1959. Last, PM with routine MRI and SNB ranged $2386–$2610. Additionally, the results of the theoretical model did not greatly change over a range of sensitivities for MRI and SLNB or for a 0%–30 % probability range for a positive SLNB in a PM with invasive breast cancer.

DISCUSSION In our series, high-risk patients undergoing PM had a 10% incidence of unsuspected malignancy, most often noninvasive and node negative. In patients undergoing preoperative MRI, there was a 7.7% rate of suspicious MRIs, most of which were falsely positive, and MRI missed three of four total occult cancers. As a result, MRI in its present state cannot be recommended as a useful addition to PM.

2483

OCCULT MALIGNANCY IN PROPHYLACTIC MASTECTOMY

TABLE 4. Theoretical cost analysis in a hypothetical cohort undergoing prophylactic mastectomy Strategy

SLNB

ALND

Average cost per patient

Additional no. procedures

PM PM + SLNB PM + MRI + selective SLNB PM + MRI + routine SLNB

0 1000 18 1000

25 4 10 4

$808 $1420 $1774 $2379

25 1004 1290 2266

PM, prophylactic mastectomy; SLNB, sentinel lymph node biopsy; MRI, magnetic resonance imaging.

The use of SLNB to avoid second operative procedures also had many limitations: among 56 patients who underwent SLNB, only one patient was spared a postmastectomy ALND. The cost comparison of our retrospective cohort and the cost analysis of the hypothetical cohort demonstrated that PM alone with delayed ALND if indicated leads to the smallest number of procedures for patients and is the least costly strategy. Although SLNB may potentially avoid an ALND in the rare patient found to have an unsuspected invasive breast cancer, the prevalence of occult invasive cancer is low enough that most routine intraoperative SLNBs are unnecessary. However, there are patients for whom a second operative procedure for ALND would be especially difficult—for example, in those requiring anticoagulation, those with other operative risks, or those having breast reconstructions at risk for damage by a delayed ALND. In these patients, performing a SLNB during the simple mastectomy is the most cost-effective strategy to reduce the risk of a second surgical procedure. Patients electing PM may present to surgeons after having had a breast MRI. If the MRI shows no abnormalities, it would be reasonable to omit a SLNB. MRI will miss some occult cancers, but in our study, most were DCIS, for which simple mastectomy alone would be adequate treatment. It should be noted that the additional procedures and costs associated with preoperative MRI are primarily related to false-positive and false-negative MRIs. A limitation of the retrospective review was the use of a 1.5-T MRI, which may have contributed to a lower sensitivity. However, in the theoretical cohort, the use of sensitivity analysis to change the MRI sensitivity from 30% to 90% did not alter the rank order of cost per strategy. It is possible that ongoing improvements in MRI technology will make this strategy a more effective in way to detect occult breast cancer with less financial expense in the future. Breast MRI and SLNB should not be routinely performed together in the high-risk patient undergoing PM. This represents the most costly strategy

with the highest number of additional procedures, and no improvement in overall management. Our retrospective study is limited by the fact that our cohort includes patients with variable degrees of risk for breast cancer, and as a result varying prevalence of occult cancer. However, previous studies have demonstrated incidence of occult cancer similar to the rate of 10% observed in our cohort.6 Our theoretical cost analysis was designed to help evaluate how management strategies compare when we vary the prevalence rates of occult cancers. For instance, although it is clinically plausible that the probability of positive lymph nodes for an occult cancer found in a PM are overestimated by the report of King et al.,6 varying this parameter over the range of 0%–30% did not change the rank order of the four strategies with regard to the number of additional procedures or cost. In fact, performing a sensitivity analysis on all diagnostic parameters did not change the rank order of the four strategies with regard to the number of additional procedures or the cost. The results for both the cost comparison in the retrospective cohort and the cost analysis in the hypothetical cohort were based on Medicare RVU reimbursements, which does not take into account hospital procedure charges incurred by the patient or operating room and anesthesia charges. The Medicare RVU costs are likely low estimates because costs may vary with private sector health insurance groups. Additionally, most of our high-risk patients were younger than 65 years old and likely had higher costs based on private insurance rates. We have demonstrated that preoperative MRI and intraoperative SLNB in women undergoing PM increases cost and procedures performed without improving overall patient treatment. Because our model did not factor in morbidity related to SLNB, ALND, and additional workup related to a positive MRI, strategies that include MRI and/or SLNB may have even more negative impact. In the future, improvements in understanding breast cancer risk factors will allow clinicians to better quantify the risk for individual women, and improvements in preoperative imaging may result in fewer unsuspected Ann. Surg. Oncol. Vol. 14, No. 9, 2007

2484

D. BLACK ET AL.

malignancies. Future studies can evaluate these evolving options to best individualize management for high-risk breast cancer patients.

REFERENCES 1. Hartmann LC, Schaid DJ, Woods JE, et al. Efficacy of bilateral prophylactic mastectomy in women with a family history of breast cancer. N Engl J Med 1999; 340:77–84. 2. Rebbeck T, Friebel T, Lynch H, et al. Bilateral prophylactic mastectomy reduces breast cancer risk in BRCA1 and BRCA2 mutation carriers: the PROSE Study Group. J Clin Oncol 2004; 22:1055–62. 3. Rebbeck T, Lynch H, Neuhausen S, et al. Prophylactic oophorectomy in carriers of BRCA1 or BRCA2 mutations. N Engl J Med 2002; 346:1616–22. 4. Lostumbo L, Carbine N, Wallace J, Ezzo J. Prophylactic mastectomy for the prevention of breast cancer. Cochrane Database Syst Rev 2004; 18:CD002748. 5. McDonnell S, Schaid D, Myers J, et al. Efficacy of contralateral prophylactic mastectomy in women with a personal and family history of breast cancer. J Clin Oncol 2001; 19:3938–43. 6. King T, Ganaraj A, Fey J, et al. Cytokeratin-positive cells in sentinel lymph nodes in breast cancer are not random events: experience in patients undergoing prophylactic mastectomy. Cancer 2004; 10:926–33. 7. Goldflam K, Hunt K, Gershenwald J, et al. Contralateral prophylactic mastectomy. Predictors of significant histologic findings. Cancer 2004; 101:1977–86. 8. Hoogerbrugge N, Bult P, de Widt-Levert LM, et al. High prevalence of premalignant lesions in prophylactically removed breasts from women at hereditary risk for breast cancer. J Clin Oncol 2003; 21:41–5.

Ann. Surg. Oncol. Vol. 14, No. 9, 2007

9. Kroll SS, Miller MJ, Schusterman MA, et al. Rationale for elective contralateral mastectomy with immediate bilateral reconstruction. Ann Surg Oncol 1994; 1:457–61. 10. Boughey J, Khakpour N, Meric-Bernstam F, et al. Selective use of sentinel lymph node surgery during prophylactic mastectomy. Cancer 2006; 107:1440–7. 11. Kriege M, Brekelmans C, Boetes C, et al. Efficacy of MRI and mammography for breast-cancer screening in women with a familial or genetic predisposition. N Engl J Med 2004; 351:427– 37. 12. Stoutjesdijk M, Boetes C, Jager G, et al. Magnetic resonance imaging and mammography in women with a hereditary risk of breast cancer. J Natl Cancer Inst 2001; 93:1095–102. 13. Liberman L, Morris E, Kim C, et al. MR imaging findings in the contralateral breast of women with recently diagnosed breast cancer. Am J Roentgenol 2003; 180:333–41. 14. Lee S, Orel S, Woo I, et al. MR imaging screening of the contralateral breast in patients with newly diagnosed breast cancer: preliminary results. Radiology 2003; 226:773–8. 15. Bedrosian I, Schlencker J, Spitz F, et al. Magnetic resonance imaging-guided biopsy of mammographically and clinically occult breast lesions. Ann Surg Oncol 2002; 9:457–61. 16. Kim T, Giuliano A, Lyman G. Lymphatic mapping and sentinel lymph node biopsy in early-stage breast carcinoma: a metaanalysis. Cancer 2006; 106:4–16. 17. Wilke L, McCall L, Posther K, et al. Surgical complications associated with sentinel lymph node biopsy: results from a prospective international cooperative group trial. Ann Surg Oncol 2006; 13:491–500. 18. Temple L, Baron R, Cody H, et al. Sensory morbidity after sentinel lymph node biopsy and axillary dissection: a prospective study of 233 women. Ann Surg Oncol 2002; 9:654–62. 19. American Medical Association. Current procedural terminology code: relative value search. 2005. Available at: http:// www.catalog.ama-assn.org/Catalog/cpt/cpt_search.jsp. Accessed March 20, 2006.