Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

Nuclear

Rachel F. Brem, MD Occult : Jocelyn A. Rapelyea, MD Gilat Zisman, BS Kevin Mohtashemi, MD Joyce Raub, MS, RT with High-Resolution Christine B. Teal, MD Stan Majewski, PhD Breast-specific Benjamin L. Welch, PhD Published online before print in Women at High Risk for 10.1148/radiol.2371040758 2005; 237:274–280 Breast Cancer1 Abbreviation: BI-RADS ϭ Reporting and Data System PURPOSE: To prospectively evaluate a high-resolution breast-specific gamma cam- era for depicting occult in women at high risk for breast cancer but 1 From Department of Radiology, with normal mammographic and physical examination findings. Breast Imaging and Intervention Cen- ter, (R.F.B., J.A.R., G.Z., K.M., J.R.) and MATERIALS AND METHODS: Institutional Review Board approval and informed Department of , Breast Care Center (C.B.T.), George Washington consent were obtained. The study was HIPAA compliant. Ninety-four high-risk University Medical Center, 2150 Penn- women (age range, 36–78 years; mean, 55 years) with normal mammographic sylvania Ave NW, Washington, DC (Breast Imaging Reporting and Data System [BI-RADS] 1 or 2) and physical exam- 20037; and the Thomas Jefferson Na- ination findings were evaluated with scintimammography. After injection with tional Accelerator Facility, Newport News, Va (S.M., B.L.W.). Received 25–30 mCi (925–1110 MBq) of technetium 99m sestamibi, patients were imaged May 4, 2004; revision requested July with a high-resolution small-field-of-view breast-specific gamma camera in cranio- 14; revision received October 1; ac- caudal and mediolateral oblique projections. Scintimammograms were prospec- cepted November 15. Supported by a tively classified according to focal radiotracer uptake as normal (score of 1), with no grant from Bristol Myers Squibb, Bil- lerica, Mass. Address correspon- focal or diffuse uptake; benign (score of 2), with minimal patchy uptake; probably dence to R.F.B. (e-mail: [email protected] benign (score of 3), with scattered patchy uptake; probably abnormal (score of 4), .edu). with mild focal radiotracer uptake; and abnormal (score of 5), with marked focal See Materials and Methods for pertinent radiotracer uptake. Mammographic breast density was categorized according to disclosures. BI-RADS criteria. Patients with normal scintimammograms (scores of 1, 2, or 3) were followed up for 1 year with an annual mammogram, physical examination, and repeat scintimammography. Patients with abnormal scintimammograms (scores of 4 or 5) underwent ultrasonography (US), and those with focal hypoechoic lesions underwent . If no lesion was found during US, patients were followed up with scintimammography. Specific pathologic findings were compared with scintimam- mographic findings. RESULTS: Of 94 women, 78 (83%) had normal scintimammograms (score of 1, 2, or 3) at initial examination and 16 (17%) had abnormal scintimammograms (score of 4 or 5). Fourteen (88%) of the 16 patients had either benign findings at biopsy or no focal abnormality at US; in two (12%) patients, invasive carcinoma was diagnosed at US-guided biopsy (9 mm each at pathologic examination).

Author contributions: CONCLUSION: High-resolution breast-specific scintimammography can depict Guarantor of integrity of entire study, small (Ͻ1-cm), mammographically occult, nonpalpable lesions in women at in- R.F.B.; study concepts and design, creased risk for breast cancer not otherwise identified at or physical R.F.B.; literature research, all authors; examination. clinical studies, all authors; data acqui- © sition and analysis/interpretation, all RSNA, 2005 authors; statistical analysis, R.F.B., G.Z., J.A.R.; manuscript revision/re- view, R.F.B., G.Z.; manuscript prepa- ration, definition of intellectual con- tent, editing, and final version ap- Mammography remains the mainstay for breast cancer detection, with a sensitivity of proval, all authors 85%–90% but decreasing to 65% in women with dense (1,2). As a result of the © RSNA, 2005 limitations of mammography, adjunct imaging modalities are being investigated to im- prove breast cancer diagnosis. The most commonly used adjunct imaging modality is

274 ultrasonography (US), which has been fit from an adjunct to mammography, ical examination was performed by a shown to improve breast cancer detec- particularly if additional imaging is not health care provider within 6 months tion by 12% when used as a affected by breast density. Thus, the pur- (mean, 2.4 months; median, 1.7 months) modality in women with dense breasts pose of our study was to prospectively of scintimammography. The institu- (3). evaluate a high-resolution breast-specific tional review board approved our study, Mammography and US are anatomic gamma camera for depicting occult which was Health Insurance Portability approaches to breast cancer diagnosis. breast cancer in women who are at high and Accountability Act compliant, and techniques, which rely risk for breast cancer and who have nor- informed consent was obtained from all on biochemical and physiologic charac- mal mammographic and physical exam- participants. teristics of tumors, are currently being ination results. investigated and are increasingly being 99m Imaging used. Technetium 99m ( Tc) sestamibi MATERIALS AND METHODS was approved in the United States in Patients received an injection of 25–30 99m 1990 for clinical use as a cardiac perfu- This study was supported by a grant from mCi (925–1110 MBq) of Tc sestamibi sion agent for the detection of coronary Bristol Myers Squibb (Billerica, Mass). into the dorsalis pedis vein of the foot or artery disease. In 1994, Khalkhali et al (4) the antecubital vein, depending on clin- 99m The radiotracer used was an in-kind do- investigated the use of Tc sestamibi nation by Bristol Myers Squibb. The ical considerations. All injections were for the detection of breast cancer. Since Dilon camera was provided by an in-kind preferentially administered into the dor- that time, authors investigating scinti- donation from Dilon (New- salis pedis vein, unless the patient re- mammography for the diagnosis of port News, Va). No authors are employed fused or it was not clinically possible. breast cancer have reported sensitivities by Bristol Myers Squibb or Dilon Tech- Approximately 10 minutes after the in- ranging from 62% to 96% and specifici- nologies. Dr Brem has stock options in jection of the radiotracer, images were ties ranging from 69% to 100% (5–17). Of Dilon Technologies and has been on the obtained in the craniocaudal and medio- particular note is that the sensitivity of speakers bureau for Bristol Myers Squibb lateral oblique projections with a high- scintimammography for the detection of (previously DuPont Pharmaceuticals). Dr resolution small-field-of-view breast-spe- breast cancer is not adversely affected by Majewski has stock options and had been cific gamma camera (model 6800; Dilon increased breast density (6), which is a a consultant for Dilon Technologies. Dr Technologies, Newport News, Va). The notable advantage. Welch has stock options and stock and is acquisition time for each image was ap- Authors investigating scintimammog- a consultant to Dilon Technologies. No proximately 10 minutes, with a total im- raphy have used a general-purpose gamma other authors have any other relation- aging time of approximately 40 minutes camera, which is not optimally designed ship with Bristol Myers Squibb or Dilon per scintimammographic study. for breast imaging. Specifically, a general- Technologies. The data reported in this purpose gamma camera has limited in- study, as well as the submission of this Image Evaluation trinsic resolution for smaller than manuscript, have always been in the sole 1 cm. Most cancers diagnosed with scin- control of the authors. Scintimammograms were prospec- timammography are larger than 1 cm, tively classified as normal (score of 1), and the mean size of cancers included in with no focal or diffuse uptake (Fig 1); Patients one large multi-institutional trial was 2.2 benign (score of 2), with minimal patchy cm (7). The reported sensitivity of scinti- From September 2001 to March 2004, uptake (Fig 2); probably benign (score of mammography for cancers 1 cm or 94 women with a mean age of 55 years 3), with scattered patchy uptake (Fig 3); smaller is 35%–64% (17–22). In addition, (range, 36–78 years) who had a calcu- probably abnormal (score of 4), with the design of the gamma camera results lated 5-year risk for the development of mild focal radiotracer uptake (Fig 4); and in limited depiction of lesions in the me- breast cancer of 1.66% or higher, as de- abnormal (score of 5), with marked focal dial portion of the breast. Finally, it is termined with the Gail Risk Model, were radiotracer uptake (Fig 5) (2). In addition, difficult to correlate mammographic eligible (1). The calculated risk of 1.66% all mammograms were evaluated for findings with scintimammographic find- (range, 1.7%–7.6%; mean, 3.3%; median, breast density by using the BI-RADS cri- ings, because it is not possible to image 2.9%) was used since this was the crite- teria for density and were assigned a BI- the breasts in positions comparable to rion used for the inclusion in the BCPT1 RADS breast density category of 1–5. those used in mammography by using a Breast Cancer Prevention Trial 1 (24). Of Mammograms and scintimammograms general-purpose gamma camera. the 94 women in this study, 27 had his- were reviewed and classified by two radi- To optimize scintimammography, a tory of invasive carcinoma, three had his- ologists, one with 15 years of experience high-resolution small-field-of-view gamma tory of , three (R.F.B.) and one with 5 years of experi- camera specific to breast imaging was de- had atypical ductal hyperplasia, and two ence (J.A.R.). Any discrepancies in inter- veloped to improve resolution and opti- had lobular neoplasia. All women seen at pretation between the radiologists was mize breast imaging (23). Preliminary our institution with normal mammo- resolved by consensus. Scintimammo- findings with use of the breast-specific graphic and physical examination find- grams were interpreted without knowl- gamma camera for evaluating women ings and with a calculated breast cancer edge of the patient characteristics and with breast lesions prior to biopsy dem- risk of 1.66% or higher were offered par- mammographic reports. Scintimammo- onstrated improved resolution and sensi- ticipation. In addition, women were re- grams assigned a score of 1, 2, or 3 were tivity in the detection of breast cancer, cruited from advertisements placed in considered to be without evidence of dis- with the greatest improvement demon- the health section of a local newspaper. ease, while those with scores of 4 and 5 strated in nonpalpable lesions and le- All patients had a normal mammogram resulted in directed US to the region of sions smaller than 1 cm (6). Women at (Breast Imaging Reporting and Data Sys- the breast in which there was focal radio- increased risk of breast cancer may bene- tem [BI-RADS] category 1 or 2), and phys- tracer uptake. Patients with abnormal

Volume 237 ⅐ Number 1 Occult Breast Cancer at Scintimammography ⅐ 275 Figure 2. Benign scintimammograms in (a) left craniocaudal, (b) right craniocaudal, (c) left medio- lateral oblique, and (d) right mediolateral oblique projections demonstrate minimal patchy radio- tracer uptake compatible with physiologic uptake.

annual mammography, physical exami- nation, and follow-up scintimammogra- phy by using high-resolution breast-spe- cific gamma camera. All patients with new suspicious findings on a 2nd-year mammogram or with physical findings underwent biopsy as clinically indicated, regardless of scintimammographic find- ings. Figure 1. High-resolution breast-specific scintimammogram in a patient with breast implants. (a) Left craniocaudal, (b) right craniocaudal, (c) left mediolateral oblique, and (d) right medio- Statistical Analysis lateral oblique displacement views. Normal scintimammograms in (e) left craniocaudal, (f) right craniocaudal, (g) left mediolateral oblique, and (h) right mediolateral oblique projections The data were analyzed for sensitivity do not demonstrate focal or diffuse radiotracer uptake. The rim of increased radiotracer uptake is and specificity of cancer detection, as normal parenchyma adjacent to the . well as for positive and negative predic- tive values. Sensitivity is the probability that results at imaging are positive in those patients who have the disease. scintimammograms (score of 4 or 5), in terpreted the mammograms and scinti- Specificity is the probability that results which the subsequent directed US dem- mammograms for this study. at imaging are negative in patients who onstrated a focal hypoechoic area, under- do not have the disease. Statistical anal- Follow-up went US-guided core-needle biopsy. US- ysis was performed with Microsoft Excel guided core-needle were per- Those with abnormal scintimammo- 2000 (Microsoft, Redmond, Wash) soft- formed as previously described (25). Both grams and negative US findings returned ware. For the reference standard, we used the US examinations and subsequent US- at 6 months for reimaging with scinti- either pathologic results of biopsy or fol- guided core-needle biopsies were per- mammography. All patients were fol- low-up imaging that did not demonstrate formed by the radiologists who had in- lowed up in the subsequent year with evidence of malignancy.

276 ⅐ Radiology ⅐ October 2005 Brem et al RESULTS bed (Fig 4); the other, a con- tralateral malignancy (Fig 5). Both of Ninety-four women initially underwent these cancers were diagnosed histopatho- mammography, physical examination, logically as infiltrating and intraductal and scintimammography, with 1-year carcinoma. The largest diameter of the follow-up that included mammography, two cancers measured at US was 6 and 8 physical examination, and repeat scinti- mm, and both measured 9 mm in the mammography. Seventy-eight (83%) of greatest diameter at pathologic examina- the 94 patients had normal scintimam- tion. The cancers were located in the up- mograms (scores of 1, 2, or 3) at initial per outer quadrant and upper inner examination and were followed up for 1 quadrant and occurred in breasts with year. Findings of follow-up mammogra- mammographic BI-RADS density catego- phy, clinical examination, and scinti- ries 2 and 3, respectively. mammography at 1 year were normal in all 78 patients, confirming the absence of False-Positive Findings disease (true-negative findings). Ten of Fourteen (88%) of the 16 patients with the 78 patients who had a normal scinti- positive scintimammograms had either mammogram at initial examination and benign findings at biopsy (7) or no focal at 1-year follow-up underwent biopsy as abnormality at US (false-positive find- a result of suspicious findings at mam- ings) (3). Biopsy was not performed in mography, US, or physical examination. five patients in whom there were no focal Histopathologic findings in all 10 pa- US findings. Among the nine patients tients were benign and confirmed the with false-positive findings, pathologic normal scintimammographic findings. examination demonstrated fibrocystic These patients have been followed up change (one with concomitant sclerosing mammographically for a minimum of 1 adenosis) in seven patients, fibroadenoma year subsequent to the biopsy, with no in one, and in one. The five evidence of malignancy. patients with no focal US findings were Sixteen (17%) of the 94 patients had reimaged at 6 months, at which time positive scintimammograms at initial ex- scintimammographic findings were nor- amination. All 16 of these patients un- mal. derwent US to the region of the breast with focal radiotracer uptake, and 11 Figure 3. Probably benign scintimammo- True-Negative Findings grams in (a) left craniocaudal, (b) right cranio- (69%) of these patients went on to un- caudal, (c) left mediolateral oblique, and dergo biopsy owing to a focal hypoechoic Seventy-eight (83%) of the 94 patients (d) right mediolateral oblique projections dem- finding. Two (18%) of the 11 patients had true-negative findings, which were onstrate scattered patchy radiotracer uptake. who underwent US-guided core-needle defined as a normal mammogram, phys- biopsy were found to have invasive car- ical examination findings, and scinti- cinoma (true-positive findings). The mammogram at 1 year and in the subse- other nine (82%) of 11 patients had be- quent year. There were no false-negative raphy at 6 months to ensure that the nign pathologic findings at US-guided findings in this study (ie, a patient with a targeted lesion was not missed. Mammo- core-needle biopsy, with pathologic find- normal scintimammogram who went on graphic findings for all of these patients ings demonstrating fibrocystic change to biopsy because of suspicious mammo- were stable at 6-month follow-up. (n ϭ 6), fat necrosis (n ϭ 1), fibroade- graphic, US, or physical examination noma (n ϭ 1), and benign cyst content findings and was found to have cancer). ϭ (n 1). The remaining five (31%) of 16 Overall Results patients did not have any focal US find- Additional Findings ings and, therefore, did not undergo bi- In 94 patients with normal (BI-RADS Three patients with four lesions (BI- opsy. The five patients had a follow-up category 1 or 2) mammograms, scinti- RADS category 4 or 5) detected on a 2nd- scintimammogram at 6 months, all of mammography performed with the year mammogram went on to undergo whom did not have any abnormal foci of high-resolution breast-specific gamma biopsy. Of these three patients, all had radiotracer uptake. Furthermore, 6 camera depicted two occult and other- normal scintimammograms. Two lesions months later (ie, 1 year from the initial wise undetected breast cancers. It had an were proved at histopathologic examina- abnormal scintimammogram), these five overall sensitivity of 100% (95% confi- tion to be benign tissue, and two lesions patients had normal scintimammograms demonstrated fibrocystic change. The dence limit, 0.22, 1.0), specificity of 85%, as well. pathologic findings were determined to positive predictive value of 12.5%, nega- be concordant with the mammographic tive predictive value of 100%, and diag- manifestation of the lesions. As per our nostic accuracy of 85%. However, it is True-Positive Findings protocol, patients with benign findings essential to emphasize that the 100% Both cancers detected only with high- at , which was performed sensitivity is based on only two patients resolution scintimammography were in because of suspicious mammographic in this study in whom breast cancer was women with a history of breast carci- findings, and normal scintimammo- diagnosed with the high-resolution noma; one was a local recurrence in a grams were followed up with mammog- breast-specific gamma imaging.

Volume 237 ⅐ Number 1 Occult Breast Cancer at Scintimammography ⅐ 277 gamma camera to 66.6% when the high- resolution breast-specific gamma camera was used (22). In our study, the high-resolution breast-specific gamma camera demon- strated a 100% sensitivity by depicting the two histologically proved malignan- cies that were not identified with either mammography or physical examination. Both of these patients had prior breast cancers. Mammography and physical ex- amination failed to depict cancer that was identified only with the high-resolu- tion breast-specific gamma camera (27). Interestingly, both patients had minimal to moderate breast density (BI-RADS den- sity categories 2 and 3), which suggests that the use of high-resolution scinti- mammography may be beneficial even in women who do not have dense breasts. Figure 4. Probably abnormal scintimammograms in (a) left craniocaudal, (b) right craniocau- In the patient with prior lumpectomy in dal, (c) left mediolateral oblique, and (d) right mediolateral oblique projections demonstrate mild focal radiotracer uptake (arrow). Pathologic examination demonstrated a 9-mm infiltrating ductal whom a local recurrence was diagnosed carcinoma in the lumpectomy bed of a patient with prior left lumpectomy. with high-resolution scintimammogra- phy, the posttherapeutic changes, both mammographic and US, resulted in the inability to detect local recurrence. This physiologic approach of high-resolution scintimammography may be particularly beneficial in this group of patients. Fur- ther studies are needed to assess the role of high-resolution scintimammography in women who have undergone lumpec- tomy. The limited number of patients in- cluded in the study is likely the cause for the 100% sensitivity. Larger studies are needed to better define the true sensitiv- ity for identification of occult breast can- cer in high-risk women by using high- resolution scintimammography. The occult cancers detected measured 6 and 8 mm at US and 9 mm each at pathologic examination, which demon- strates the ability of the high-resolution gamma camera to depict subcentimeter occult cancers. Prior studies with use of Figure 5. Abnormal scintimammograms in (a) left craniocaudal, (b) right craniocaudal, (c) left high-resolution breast-specific gamma mediolateral oblique, and (d) right mediolateral oblique projections demonstrate marked focal camera have similarly demonstrated the radiotracer uptake (arrow). Pathologic examination demonstrated 9-mm infiltrating ductal car- cinoma. detection of subcentimeter cancers in both phantoms (23) and clinically (6). Pathologic findings of the false-posi- tive lesions in this study included a fibro- DISCUSSION biopsies (26,27). The sensitivity of scinti- cystic change (with and without scleros- mammography with a general-purpose ing adenosis), fibroadenoma, and fat ne- Recent studies in which nuclear medi- gamma camera is decreased in cancers crosis. The increase in 99mTc sestamibi cine techniques that rely on biochemical smaller than 1 cm, especially those lo- activity often seen in patients with pro- and physiologic characteristics of breast cated in the medial breast, with sensitiv- liferative breast lesions likely reflects the tumors were evaluated have shown ity decreasing to 35%–64% (17–22). increased mitochondrial activity, as well promising results as a potential adjunct Preliminary results of scintimammog- as mitochondrial density (26). Gupta et to mammography (4–17). Scintimam- raphy with the breast-specific gamma al (28) demonstrated the 99mTc sestamibi mography has demonstrated improved camera demonstrated an increase in sen- uptake in benign to be sensitivity and specificity in breast cancer sitivity for breast cancer detection, rang- highly associated with the presence of detection, particularly in women with ing from 64.3% to 78.6%. In lesions proliferative changes. In addition, radio- dense breasts and in women with archi- smaller than 1 cm in size, the sensitivity tracer uptake increases in direct propor- tectural distortion or scarring from prior increased from 46.7% with a traditional tion to the degree of regional blood flow

278 ⅐ Radiology ⅐ October 2005 Brem et al (27). This may be the cause of the false- small cohort of patients, as well as the been developed. Future studies are be- positive findings in the nonmalignant small number of patients with cancer. ing initiated to evaluate time-intensity proliferative lesion, which results in in- Multi-institutional trials with a larger curves obtained by using scintimammog- creased vascularity and/or mitochondrial sample size are needed to more accu- raphy with the high-resolution breast- activity (29). rately evaluate scintimammography per- specific gamma camera and their effect Currently, only 15%–30% of breast bi- formed with high-resolution breast-spe- on the ability to differentiate benign opsies result in a diagnosis of cancer (25). cific gamma camera. from malignant breast lesions. Clearly it would be advantageous to im- The locations of the cancers detected The decreased detector size and added prove the specificity of mammography only with the high-resolution breast-spe- maneuverability of the high-resolution and thereby decrease the number of bi- cific gamma camera were in the upper breast-specific gamma camera increase opsies performed for benign lesions outer quadrant and upper inner quad- the number of available views, reduce the identified at mammography. A high rant. Although this study demonstrated a amount of tissue and the distance be- specificity, 90% or above, of breast-spe- small number of cancers, one occurred in tween the lesion and the collimator, and cific scintimammography may allow for the upper inner quadrant, a region of permit acquisition geometries that mini- improved differentiation of benign from the breast suboptimally visualized with mize scatter radiation from nearby or- malignant lesions and thereby allow for a mammography, and the other occurred gans that exhibit high uptake, such as reduction in the number of breast biop- in a prior lumpectomy site. Both cancers myocardium and liver (33). These refine- sies that are performed because of inde- occurred in clinical situations in which ments in design serve to eliminate some terminate findings, which are found to mammography is suboptimal. of the intrinsic limitations of scintimam- be benign. In our study, all 10 patients There has been dramatic progress in mography with a general-purpose gamma with normal scintimammograms at 1 the field of breast magnetic resonance camera. year who underwent biopsy owing to (MR) imaging since its initial introduc- Optimally, patients with abnormal mammographic or clinical findings had tion. Contrast material–enhanced MR scintimammograms should undergo bi- benign pathologic findings at biopsy. In imaging has demonstrated effectiveness opsy of the area of increased radiotracer addition, of the three patients with four in imaging and helping diagnose breast uptake by using a directed approach to lesions who had abnormal 2nd-year cancer, as well as in helping evaluate the the region of interest. However, currently mammograms (BI-RADS category 4 or 5) extent of disease (33). In a study by Hel- there is no minimally invasive method to that warranted biopsy, all had normal bich et al (9), which compared MR imag- directly target and perform biopsy on ar- scintimammograms and benign patho- ing with planar and single photon emis- eas of focal radiotracer uptake. Therefore, logic findings. These 13 patients with 14 sion computed tomographic (SPECT) the region of the breast with increased lesions warranting biopsy had normal scintimammography, MR imaging had a uptake was evaluated with US. If an ab- scintimammograms and benign findings sensitivity and specificity for breast can- normality was identified, the patient un- at pathologic examination. Scintimam- cer detection of 96% and 82%, respec- derwent US-guided core-needle biopsy. It mography may potentially reduce the tively. The respective sensitivities and is possible that the region with the in- number of breast biopsies that result in a specificities of planar and SPECT scinti- creased radiotracer uptake could not be pathologic finding that is benign by im- mammography were 62% and 88% and identified with US. It would certainly be proving specificity. This decrease in 83% and 80% (9). However, that study optimal to perform a gamma camera– breast biopsies could result in substantial used a general-purpose gamma camera. It guided stereotactic biopsy of the region improvements in patient care, as well as is likely that the sensitivity of scintimam- of increased radiotracer uptake. Although substantial cost savings (30). mography would increase with the use of not yet available, we have begun prelim- Indeterminate mammographic find- a dedicated high-resolution gamma inary studies on phantoms to develop ings are commonly encountered in high- camera and may well be comparable to this direct approach for the biopsy of the risk women, particularly those with that of MR imaging. Further studies are region of interest. dense breasts, prior lumpectomy, or radi- needed to compare MR imaging and As reported earlier, additional patients ation therapy (26,31). Therefore, the role high-resolution scintimammography in underwent biopsy at 2 years following of scintimammography as an adjunct to the diagnosis of breast cancer. examinations performed because of mammography in the screening of high- Dynamic MR imaging of the breast cur- mammographic or clinical findings. All risk women should be considered, both rently allows evaluation of contrast ma- of these patients had normal scintimam- for its ability to detect occult cancers and terial uptake and washout, resulting in mograms, and all had benign findings at its potential to reduce the number biop- additional analyses of breast lesions. biopsy. Although this suggests a possible sies that result in benign pathologic find- Kuhl et al (34) reported a correlation be- role for scintimammography in the re- ings. tween the contrast material washout duction of benign breast biopsies, in this Our study with the high-resolution curves in areas of contrast enhancement study we did not address that question. breast-specific gamma camera resulted in during breast MR imaging by using this Additional studies to specifically investi- a negative predictive value of 100%, in method, with a reported a sensitivity of gate the effect of scintimammography on contrast to the previously published 91%, specificity of 83%, and accuracy of the reduction of breast biopsies for be- studies of scintimammography with gen- 86% in helping distinguish benign from nign lesions is needed. eral-purpose gamma cameras, which re- malignant breast lesions. There are simi- Patients in this study were evaluated ported a negative predictive value of larities between the principals of radio- initially and at 1 year. We chose 1 year to 82%–96% (4,8,9,32). This difference may tracer uptake in scintimammography evaluate for false-negative scintimammo- well be as a result of the ability of the and contrast material uptake in MR grams at the initial study. Although 1 high-resolution breast-specific gamma imaging. Recently, the ability to obtain year is appropriate to exclude a negative camera to depict subcentimeter tumors. time-intensity curves with the high-reso- study finding in a patient with breast However, our study was limited by the lution breast-specific gamma camera has cancer, it might be optimal to evaluate

Volume 237 ⅐ Number 1 Occult Breast Cancer at Scintimammography ⅐ 279 the patient at 2 years following initial for the diagnosis of breast carcinoma. Ra- 20. Scopinaro F, Schillaci O, Ussof W, et al. A examination. Finally, a multi-institu- diology 1995;196:421–426. three center trial on the diagnostic accu- 6. Brem RF, Schoonjans JM, Kieper DA, Ma- racy of Tc-99m MIBI scintimammography. tional study that would include a larger jewski S. High-resolution scintimammog- Anticancer Res 1997;17:1631–1634. number of patients would allow for fur- raphy: a pilot study. J Nucl Med 2002;43: 21. Tolmos J, Cutrone JA, Wang B. Scintimam- ther and more definitive evaluation of 909–915. mographic analysis of nonpalpable breast high-resolution breast-specific scinti- 7. Khalkhali I, Villanueva-Meyer J, Edell S. lesions previously identified by conven- tional mammography. J Natl Cancer Inst mammography. Diagnostic accuracy of Tc-99m sestamibi breast imaging: multi-center trial results. 1998;90:846–849. This study demonstrates that a high- J Nucl Med 2000;41(12):1973–1979. 22. Palmedo H, Grunwald F, Bender H. Scinti- resolution breast-specific gamma camera 8. Burak Z, Argon M, Memis A. Evaluation of mammography with technetium-99m me- can depict small (Ͻ1-cm), mammo- palpable breast masses with Tc-99m-MIBI: thoxyisobutylionitrile: comparison with graphically occult, nonpalpable lesions a comparative study with mammography mammography and magnetic resonance imaging. Eur J Nucl Med 1996;23:940– in women at increased risk for breast can- and ultrasonography. Nucl Med Commun 1994;15:604–612. 946. cer. The high specificity (84.8%–85.0%) 9. Helbich TH, Becherer A, Trattnig S. Differ- 23. Majewski S, Kieper D, Keppel C. Optimiza- and high negative predictive value entiation of benign and malignant breast tion of dedicated scintimammography (100%) indicate the potential of the lesions: MR Imaging versus Tc-99m sesta- procedure using small detector prototypes and compressible breast phantoms (abstr). high-resolution breast-specific gamma mibiscintimammography.Radiology1997; 202:421–429. Presented at the IEEE camera to distinguish benign from malig- 10. Mekhmandarov S, Sandbank J, Cohen M, Conference, Lyon, France, October 2000. nant lesions, thus potentially aiding in Lelcuk S, Lubin E. Technetium 99-m MIBI 24. Fisher B, Dignam J, Wolmark N, et al. Ta- decreasing the number of benign-breast scintimammography in palpable and non- moxifen in treatment of intraductal breast palpable breast lesions. J Nucl Med 1998; cancer: National Surgical Adjuvant Breast biopsies. We conclude that women at and Bowel Project B-24 randomised con- high risk for the development of breast 39:86–91. 11. Taillefer R, Robidoux A, Lambert R. Tech- trolled trial. Lancet 1999;353(9169):1993– cancer may benefit from the high-resolu- netium 99-m sestamibi prone scintimam- 2000. tion breast-specific scintimammography mography to detect primary breast cancer 25. Schoonjans JM, Brem RF. Fourteen-gauge ultrasonographically guided large-core as an adjunct to mammography. How- and axillary involvement. J Nucl Med 1995;36:1758–1765. needle biopsy of breast masses. J Ultra- ever, additional trials with inclusion of a sound Med 2001;20(9):967–972. 12. Clifford EJ, Lugo-Zamudio C. Scintimam- 26. Waxman AD. The role of (99m)Tc me- greater number of patients are needed to mography in the diagnosis of breast can- thoxyisobutylisonitrile in imaging breast further define the sensitivity and specific- cer. Am J Surg 1996;172:483–486. cancer. Semin Nucl Med 1997;27:40–54. ity of this approach to breast cancer di- 13. Kao CH, Wang SJ, Liu TJ. The use of tech- 27. Cutrone JA, Khalkhali I, Yosper LS. Tc-99m netium-99m methoxyisobutylisonitrile agnosis. sestamibi scintimammography for the breast to evaluate palpable evaluation of breast masses in patients breast masses. Eur J Nucl Med 1994;21: with radiographically dense breasts. Breast Acknowledgments: The authors acknowl- 432–436. J 1999;5:383–388. edge Bristol Myers Squibb for support of this 14. Palmedo H, Biersak HJ, Lastoria S. Scinti- project and especially Beverley Lewis, BA, for 28. Gupta P, Waxman A, Nguyen K. Correla- mammography with technitium-99m me- tion of Tc-99m sestamibi uptake with his- her advice and support. We thank Douglas thoxyisobutylisonitrile: results of a pro- Keipfer, BA, for technical support, Heather topathologic characteristics in patients spective European multi-center trial. Eur with benign breast disease (abstr). J Nucl Young, PhD, for assistance with statistical J Nucl Med 1998;25:375–385. guidance, and Katherine Michener, BA, for as- Med 1996;37:250. 15. Tiling R, Khalkhali I, Sommer H. The role sistance with manuscript preparation. We 29. Cotran S, Lester S. The female breast. of Tc-99m sestamibi scintimammography thank Dilon Technologies for in-kind contri- Pathologic basis of disease. 6th ed. Phila- and contrast enhanced magnetic reso- bution of the breast-specific gamma camera. delphia, Pa: Saunders, 1999; 1100–1102. nance imaging for the evaluation of inde- 30. Rice SJ, Khalkhali I, Diggles L. The role of terminate mammograms. Eur J Nucl Med radionuclide imaging in the evaluation of References 1997;24:1221–1229. suspected breast abnormalities. Breast Dis 1. Sickles EA. Mammographic features of 16. Horne T, Pappo I, Cohenpour M. Tc-99m 1998;10:177–187. early breast cancer. AJR Am J Roentgenol MIBI scintimammography for the detec- 31. Taillefer R. Clinical applications of 99mTc- 1984;143:461–464. tion of breast malignancies: the contribu- sestamibi scintimammography. Semin 2. Niloff PH, Sheiner NM. False-negative tion of the count ratio to specificity. Nucl Nucl Med 2005;35:100–115. mammograms in patients with breast can- Med Commun 1999;20:511–516. 32. Klaus AJ, Klingensmith WC 3rd, Parker SH, cer. Can J Surg 1981;24:50–52. 17. Arslan N, Ozturk E, Ilgan S. Tc-99m MIBI Stavros AT, Sutherland JD, Aldrete KD. 3. Kolb TM, Lichy J, Newhouse JH. Compar- scintimammography in the evaluation of Comparative value of Tc-99m sestamibi ison of the performance of screening breast lesions and axillary involvement: a scintimammography and sonography in mammography, physical examination, comparison with mammography and his- the diagnostic workup of breast masses. and breast US and evaluation of factors topathological diagnosis. Nucl Med Com- AJR Am J Roentgenol 2000;174:1779– that influence them: an analysis of 27,825 mun 1999;20:317–325. 1783. patient evaluations. Radiology 2002; 18. Maffioli L, Agresti R, Chiti A, Crippa F. 33. Williams MB, Pisano ED, Schnall MD, Fa- 225(1):165–175. Prone scintimammography in patients jardo LL. Future directions in imaging of 4. Khalkhali I, Mena I, Jouanne E. Prone scin- with nonpalpable breast lesions. Antican- breast disease. Radiology 1998;206:297– timammography in patients with suspi- cer Res 1996;16:1269–1274. 300. cion of carcinoma of the breast. J Am Coll 19. Scopinaro F, Ierardi M, Porfiri LM, et al. 34. Kuhl CK, Mielcarek P, Klaschik S. Dynamic Surg 1994;178:491–497. Tc-99m MIBI prone scintimammography breast MR imaging: are the signal intensity 5. Khalkhali I, Cutrone JA, Mena I. Scinti- in patients with high and intermediate time course data useful for differential di- mammography: the complementary role risk mammography. Anticancer Res 1997; agnosis of enhancing lesions? Radiology of Tc-99m sestamibi prone breast imaging 17:1635–1638. 1999;211:101–110.

280 ⅐ Radiology ⅐ October 2005 Brem et al