The Use of Breast-specific Gamma Imaging as a Low-Cost Problem-Solving Strategy for Avoiding Biopsies in Patients With Inconclusive Imaging Findings on Mammography and Ultrasonography

Abstract

Objective

To evaluate the clinical performance and financial costs of breast-specific gamma imaging (BSGI) as a biopsy-reducing problem-solving strategy in patients with inconclusive diagnostic imaging findings.

Methods

A retrospective analysis of all patients for whom BSGI was utilized for inconclusive imaging findings following complete diagnostic mammographic and sonographic evaluation between January 2013 and December 2018 was performed. Positive BSGI findings were correlated and biopsied with either US or stereotactic technique with confirmation by clip location and pathology. After a negative BSGI result, patients were followed for a minimum of 24 months or considered lost to follow-up and excluded (22 patients). Results of further imaging studies, biopsies, and pathology results were analyzed. Net savings of avoided biopsies were calculated based on average Medicare charges.

Results

Four hundred and forty female patients from 30 to 95 years (mean 55 years) of age were included in our study. BSGI demonstrated a negative predictive value (NPV) of 98.4% (314/319) and a positive predictive value for biopsy of 35.5% (43/121). The overall sensitivity was 89.6% (43/48), and the specificity was 80.1% (314/392). In total, 78 false positive but only 5 false negative BSGI findings were identified. Six hundred and twenty-one inconclusive imaging findings were analyzed with BSGI and a total of 309 biopsies were avoided. Estimated net financial savings from avoided biopsies were $646 897.

Conclusion

In the management of patients with inconclusive imaging findings on mammography or ultrasonography, BSGI is a problem-solving imaging modality with high NPV that helps avoid costs of image-guided biopsies.

Introduction

Breast cancer is the most common cancer diagnosed worldwide and the leading cause of cancer-related death in women.¹ Although screening mammography is the primary imaging modality used for early detection of breast cancer, it is an imperfect screening tool with a reported overall sensitivity of 85%, which drops to 68% in women with dense breasts.² These statistics may be improving with the advent of 3D tomosynthesis techniques; however, false positive results remain common. Approximately 10% to 12% of women screened with modern digital mammography will require further imaging, and a smaller subset will also undergo biopsy.³

Breast US and MRI are 2 commonly available supplemental imaging tests that may be used to further characterize breast findings on other modalities, allow image-guided biopsy and localization, and examine the axilla for metastasis. Breast US was found to have a sensitivity of 100% and specificity of 89.1% when used to work up Breast Imaging Reporting and Data System (BI-RADS) category 0 findings on screening mammography.⁴ However, US does have limitations, which include operator dependence, low soft-tissue contrast, and lower specificity compared with mammography.⁵ A meta-analysis in 2016 found that breast MRI had a high sensitivity of 99% and negative predictive value (NPV) of 99.7% (7 false negative results out of 2316 cases) in the evaluation of equivocal imaging findings without microcalcifications.⁶ While these authors agree that there is disagreement regarding what qualifies as an “equivocal finding,” they state in the meta-analysis that descriptors included “asymmetry without associated microcalcification, architectural distortions and other ambiguous abnormalities such as multiple lesions, discrepancy between clinical symptoms and imaging findings, benign biopsy results with insufficient radiological-pathological concordance, lesions that could not be sufficiently localized during biopsy attempts as well as scars.” However, breast MRI does have disadvantages, including variable specificity and positive predictive value (PPV) depending on the reader, variable and sometimes very complex background parenchymal enhancement, intravenous contrast requirement, limitation in some patients with claustrophobia and obesity, intolerance by many patients, and the highest cost of all breast-specific imaging modalities.⁷ A multicenter study found that 42.1% of women at elevated risk of breast cancer declined screening MRI, with the most common reasons being claustrophobia, time constraints, and financial concerns.⁸ Also, due to high relative cost, insurance providers may not cover breast MRI as a problem-solving tool. Furthermore, if there is no correlate for breast MRI–detected findings on conventional imaging, breast MRI–guided biopsy is often the least comfortable intervention from the patient’s perspective and may also be the most technically challenging for the radiologist to accurately perform.

Recognition of the limitations of mammography and US and challenges inherent to MRI have led to investigation and comparison with other breast-imaging techniques. Breast-specific gamma imaging (BSGI), also known as molecular breast imaging, has been shown to have a very high sensitivity and specificity in breast cancer diagnosis. Breast-specific gamma imaging offers radiotracer imaging that is generally less adversely affected by breast tissue density and is a promising imaging modality for patients who are unable to tolerate MRI due to claustrophobia, obesity, implanted devices, or renal insufficiency. When used in conjunction with mammography as a screening tool, BSGI showed a comparable increase in breast cancer detection rate among high-risk women when compared with mammography plus MRI.⁹ In addition, numerous studies conducted over the past 2 decades have shown that BSGI is more specific than other imaging modalities and has comparable sensitivity to MRI and greater sensitivity than mammography or US.^10-14

Although multiple studies have shown that BSGI has promising prospects as a supplemental imaging tool for improving breast cancer screening, the radiation exposure associated with BSGI limits its utility as an annual screening tool, potentially even for high-risk patients. Therefore, BSGI may be better suited for the diagnostic setting, such as to further evaluate inconclusive findings on mammography with or without US.

At our imaging center, BSGI has been in use as a judiciously applied problem-solving tool to resolve diagnostic uncertainties. Over time, we observed that BSGI was well tolerated by patients, easy to interpret, and spared multiple patients from unnecessary biopsies. This study sought to more formally and retrospectively evaluate the clinical performance and financial cost savings of BSGI when used as a problem-solving strategy to avoid unnecessary biopsies for findings deemed equivocal after conventional diagnostic imaging evaluation. It is important to note that standard-of-care workflows, such as imaging follow-up for clearly probably benign findings and biopsy for clearly suspicious findings, were implemented for the vast majority of cases and that BSGI was only used in equivocal or complicated cases.

Methods

Study population

In this retrospective observational cohort study, all patients who underwent BSGI at our health system for inconclusive or complex diagnostic imaging findings between January 2013 and December 2018 were analyzed (approximate 40 000 yearly mammographic screening volume). All patients had undergone complete diagnostic workup at 1 of the 4 hospitals or 6 imaging centers in our health system with diagnostic mammography (approximate 8% yearly group callback rate), very commonly with both additional mammographic views and diagnostic US, before BSGI evaluation, which was only available at the main breast-imaging diagnostic center. Patients were recommended for BSGI evaluation due to 1 or more of the following: multiple-variable low-suspicion lesions or multifocal but lower-suspicion calcifications on mammography, developing asymmetry or equivocal global asymmetry without comparison, heterogenous to extremely dense tissue with numerous masses, extensive equivocal calcifications or asymmetries adding a high degree of complexity, or inability to undergo an MRI, among others. In most cases, US findings had been similarly heterogeneous or complex, suspicion was toward the lower end, and no clear focal intermediate or higher suspicion target for biopsy was readily available. Patients who did not have a minimum of 24-month follow-up at our health system following initial BSGI testing were labeled as “lost to follow-up” (LTFU), and their data were excluded from the study. This study was approved by our Institutional Review Board and is HIPAA compliant. All results and data were obtained retrospectively from patients’ medical records with waived patient consent.

Breast-specific gamma imaging protocol and interpretation

Patients underwent BSGI with a single-head breast-specific gamma-camera (Dilon 6800 Gamma Camera, Dilon Technologies, Newport News, VA). Patients were imaged in a seated position. Imaging was performed 10 minutes after intravenous injection of an average dose of 20 to 22.5 mCi ^99mTc-sestamibi into the arm contralateral to the breast with the suspected findings or the leg if bilateral breast cancer was suspected. Bilateral craniocaudal (CC) and mediolateral oblique (MLO) images were obtained. Individual image acquisition was conducted for approximately 5 to 10 minutes, with a minimal range of 100 000 counts/image.

All BSGI examinations were reviewed by 1 of 6 board-certified radiologists (1 fellowship-trained in breast imaging). As of 2018, the 6 radiologists had 7 (including fellowship), 15, 22, 27, 31, and 35 years of experience, respectively. Approximately 1 to 2 studies were performed per week within the entire practice. Clinical reports and outcomes were reviewed, and examinations were not reinterpreted for the purpose of this study. Breast-specific gamma imaging examinations with no focally increased radiotracer uptake or with scattered lower-level minimally heterogeneous uptake were classified as normal, whereas those with an area or areas of focally or regionally increased radiotracer uptake were classified as abnormal.

Measures and definitions

True positive results were defined as positive BSGI with breast cancer diagnosed, based on final pathology after surgical excision. False positive results were defined as positive BSGI without a breast cancer diagnosis during the 24-month follow-up period. False positive results included high-risk lesions, such as atypical ductal epithelial hyperplasia or similar lesions for which excision was recommended but without upgrade to breast cancer at surgical excision. Negative results were defined as true negative if no cancer was diagnosed and false negative if cancer was diagnosed during the 24-month follow-up period.

Positive BSGI findings were correlated and biopsied with either US or stereotactic technique with confirmation by clip location and pathology. The vast majority of biopsies done were core needle biopsies performed under US guidance. Correlates for all BSGI findings were able to be biopsied by US or stereotactic guidance. Therefore, BSGI-guided biopsy, which was not available at our institution, was not necessary.

For each BSGI examination, patients were followed up for 24 months afterward by imaging or, in fewer cases, by clinical notes and cancer registry data. Notably, this 2-year follow-up standard chosen for this review is intentionally greater than the 1-year follow-up typically utilized to determine the interval cancer rate following a screening mammogram. We desired a robust follow-up of 2 years to reduce the likelihood that a high NPV was in any way artificially inflated by the lower sensitivity of conventional imaging relative to physiologic imaging done during the follow-up period.¹³ For example, if BSGI failed to detect a lesion, a 2-year follow-up window would increase the chances of the same lesion being detected on subsequent imaging compared with a 1-year follow-up window.

Financial cost savings of BSGI were estimated by calculating total net costs of procedures saved in the diagnostic workup of patients after initial mammography with or without ultrasonography. Costs of all procedures were based on Medicare charges for year 2024. For suspicious sites identified on initial imaging, the number of biopsies saved for each true negative result determined by BSGI were assigned a positive value. It is important to note that a maximum value of 3 per case could be assigned because even if more lesions were present, it was unlikely that >3 biopsies would be needed to determine the presence or absence and extent of disease. On the other hand, each false positive detection by BSGI was assigned a negative number of biopsies saved. In this case, BSGI added additional biopsy targets and additional cost. False negative results and true positive results had no impact on cost analysis and were only used in performance analyses. The net sum of biopsies saved was then used to calculate the net estimated savings from avoiding procedures, including stereotactic-guided, US-guided, or MRI-guided biopsies and laboratory/pathology expenses.

Results

Patient characteristics

A total of 462 female patients had BSGI examinations between January 2013 and December 2018. Of those, 22 were LTFU. Patients ranged in age from 30 years to 95 years, with a mean age of 55 years.

Reported indications for BSGI

Indications reported for the 440 BSGI exams performed included the following: developing asymmetry or global asymmetry without prior comparison, intermediate risk, and inconclusive findings on diagnostic mammography and US (n = 209); multiple low-suspicion masses with variable features, often in the setting of dense breast tissue, intermediate risk, and inconclusive findings on diagnostic mammography and US (n = 105); and multiple bilateral low-suspicion regional calcifications without definite mass or dominant suspicious grouping to optimize targeting for biopsy (n = 40). Less frequent indications reported were either less clear or, in some cases, thought to be less appropriate on review. These included dense breast tissue, high anxiety, intermediate risk, complex patient (n = 25), suspicion for previous nonconcordant biopsy (n = 13), outside referral, high clinical concern despite negative imaging findings (n = 14), concern for radiologic–pathologic discordance and/or possible sampling error (BSGI was performed after negative biopsy results, generally 1 to 2 months later to allow for healing of the biopsy site) (n = 11), and unusually high risk of bleeding from biopsy (n = 3). Finally, additional, even rarer uses were related to risk or in place of MRI. These included extent of disease in the setting of a new cancer, MRI contraindicated (n = 7), family history of breast cancer (n = 6), known breast cancer follow-up to therapy (n = 4), and concerning changes too close to an implant to enable image-guided biopsy (n = 3) (Table 1).

Biopsy results

A total of 440 BSGI exams were performed on 621 suspicious sites. With this, 131/440 (29.8%) BSGI exams demonstrated uptake, many in a multifocal pattern. A total of 312 biopsies were performed: 294 for suspicious BSGI sites and 18 additional biopsies despite a negative BSGI due to persistent clinical concern and shared decision-making. The vast majority of these 312 biopsies were core needle biopsies performed under US guidance, and the remainder were performed using stereotactic guidance. Biopsy results were benign in 250/312 (80.1%) of cases. Detected benign lesions included fibrocystic changes (n = 104), benign breast tissue (n = 52), fibroadenoma (n = 19), cyst contents (n = 5), intraductal papilloma (n = 9), focal apocrine metaplasia (n = 16), usual ductal hyperplasia (n = 13), chronic inflammation (n = 9), adenosis (n = 10), pseudoangiomatous stromal hyperplasia (n = 12), and lobular carcinoma in situ (n = 1). Biopsy results were malignant in 62/312 (19.9%) cases. Five of the 18 biopsies performed despite a negative BSGI but with sufficiently high persistent concern on conventional imaging or a change on follow-up imaging were found to be malignant (false negative). Detected malignancies in the 62 malignant cases included invasive ductal carcinoma (n = 29), ductal carcinoma in situ (DCIS; n = 28), invasive lobular carcinoma (n = 4), and invasive squamous cell carcinoma (n = 1) (Table 2).

Breast-specific gamma imaging diagnostic performance

In this patient cohort, BSGI achieved an overall sensitivity of 89.6% (43/48) and specificity of 80.1% (314/392). In total, 78 false positive and 5 false negative BSGI findings were identified. Positive predictive value was 35.5% (43/121) and NPV was 98.4% (314/319). (Table 3).

Of the 5 false negative BSGI cases, 4 involved small, early calcification clusters or sub–5-mm masses. One case was of a positive lymph node with an occult breast primary. In this case, pathology was also negative after surgery, but the patient had undergone chemotherapy, so it remained uncertain whether a lesion was too small for detection at time of BSGI due to treatment response. In 1 case, the area of concern was also too deep for detection on BSGI, suggesting that BSGI was not the best strategy to have employed. In 1 case, multifocal multicentric malignancy mimicked high background activity on BSGI, leading to an error in interpretation of the case being called negative despite the presence of activity. Taken together, of the 5 false negative results, 2 were detected by biopsy at the time of BSGI due to persistent concern based on initial diagnostic evaluation. Only 2 cases were detected through follow-up imaging, one at 6 and the other at 12 months, respectively. The presence of malignancy in the positive lymph node was already known at the time of BSGI in the fifth case. Notably, despite a study design of a 24-month follow-up period, in an effort to keep the data analysis robust and to maximize confidence that we were not missing any malignancies, for functional purposes, all the malignancies that were detected on follow-up imaging were detected within the first 12 months after BSGI.

A case of true positive BSGI results is shown in Figure 1, a case of true negative BSGI results is shown in Figure 2, and a case of false negative BSGI results is shown in Figure 3.

Estimated financial savings from BSGI

A total of 621 suspicious areas were evaluated with BSGI, and a total of 309 biopsies (49.8% of suspicious areas) were avoided across 6 years, averaging 52 biopsies avoided per year.

The Medicare reimbursement rate for a breast biopsy is $1479 in 2024, whether it is performed under US guidance, stereotactic guidance, or MRI guidance. The Medicare reimbursement rate for BSGI is $355 in 2024. The total cost for biopsies of 621 suspicious sites in this population without use of BSGI would be $918 459. The total cost of utilizing BSGI would be the cost of BSGI examinations plus the cost of biopsies performed in false positive BSGI cases, which would be a total of $271 562. Therefore, the estimated net savings from avoided biopsies were $646 897. Given that other payers typically have higher reimbursement rates than Medicare, it is expected that the actual amount saved will be higher depending on the payer mix.

Discussion

In this retrospective analysis of 440 patients who were evaluated with BSGI for further assessment of inconclusive or complex imaging findings on diagnostic mammography or US, we achieved an overall sensitivity of 89.6% and specificity of 80.1%. These statistics fell within the lower range of recent systematic review reports of BSGI sensitivities from 80.4% to 95.5% and specificities from 78.1% to 90.9%.¹⁴ However, it is important to note that many of the included studies evaluated selected populations including patients who are high risk, have BI-RADS category 4 findings, are of a specific race, or have known DCIS. Overall, we found that BSGI was a useful diagnostic tool for further evaluation of inconclusive findings on mammography or ultrasonography. This is largely attributable to the fact that BSGI is a functional imaging modality that utilizes radiotracers to distinguish physiological differences between benign and malignant breast tissue.

In our study, we evaluated our use of BSGI as a problem-solving tool for persistent but lower-suspicion indeterminate findings in a primarily low to intermediate risk population. This low-suspicion problem-solving application results in a lower prevalence of malignancies compared with applications, such as for suspicious findings or when there is known malignancy. In our practice, cases with a clear suspicious target or targets were biopsied, and then MRI was generally utilized for extent of disease as appropriate. Also, according to the American College of Radiology BI-RADS, the PPV of 35.5% that we found for BSGI is within the normal acceptable range of 20% to 45% for biopsy recommendations following diagnostic mammography for a workup of abnormal screening results or of 30% to 55% for biopsy recommendations for a workup of a palpable lump.¹⁵ The NPV of 98.4% even at 2 years of follow-up in this setting is extremely high and provides a high level of reassurance that no cancer is present, which can spare patients anxiety, cost, and other potential harms from unnecessary biopsies with minimal compromise in disease detection. At a much lower examination cost than MRI, BSGI is also a well-tolerated and easily interpretable option.

In our experience, the vast majority of BSGI false negative cases involved small groups of calcifications or small masses. For example, our false negative case shown in Figure 3 demonstrated no suspicious uptake, but biopsy of mammographic calcifications showed DCIS. This lack of BSGI uptake is consistent with how suspected DCIS does not always have increased blood supply/enhancement on MRI, and lesions may be too small for significant neoangiogenesis. One false negative case involved a patient who had positive lymph nodes and was unable to tolerate MRI. In this patient, BSGI failed to find an occult breast primary. Pathology was negative after surgery, but the patient had undergone chemotherapy, so it remained uncertain if a cancer was present in the breast at the time of BSGI but was too small for detection. We were conservative and considered this case a false negative result, but without definitive pathology, it may have been a true negative finding. Also notable is that occult breast primary is missed in approximately half of breast MRI studies performed on patients with positive lymph nodes but no findings within the breast on conventional imaging.¹⁶ In another case, the area of concern was too deep for confident detection on BSGI, and selection of this modality as a problem-solving tool was likely not the most appropriate choice. It is important to note that the posterior-most 1 cm of breast tissue is slightly more difficult to visualize on BSGI than on mammography, which necessitates careful positioning similar to mammographic imaging angles during location triangulation. This may be a function of the differences in the detection paddle and/or the relative experiences of nuclear technologists compared with mammography technologists with regard to the complexities of breast positioning. In 1 extremely unusual case, rare multifocal multicentric malignancy mimicked rare high background activity on BSGI, leading to incorrect interpretation. Careful review of conventional imaging, representative site sampling (especially if any intermediate to higher suspicion findings are present), and use of follow-up imaging when appropriate were effective strategies for mitigating these issues. Finally, the use of the newer dual-head gamma detector system now available in the market instead of the single-head system we used in our study will likely further improve the sensitivity of BSGI because dual-head systems have been shown to significantly increase sensitivity for breast cancer. A prospective study has demonstrated that the sensitivity of dual-head BSGI was 90% compared with 80% for single-head BSGI, and sensitivity for subcentimeter lesions was 82% for dual-head BSGI compared with 68% for single-head BSGI.¹⁷

According to the results of our study, lesions that show some positive uptake on BSGI include malignant lesions, such as DCIS, invasive ductal carcinoma, and invasive lobular carcinoma. Positive uptake can also occur in some high-risk as well as benign lesions, including fibroadenoma (more commonly premenopausal), intraductal papilloma, focal apocrine metaplasia, usual ductal hyperplasia, and (rarely) fibrocystic changes possibly due to inflamed cysts. To date, there are insufficient data in the literature regarding types of breast lesions that do not demonstrate uptake on BSGI because BSGI examinations with negative uptake typically do not undergo biopsy.

Biopsies were successfully avoided in 309 of the 621 (49.8%) suspicious lesions in our study. At 20% to 25% of the cost of an MRI, BSGI is a relatively inexpensive strategy for avoiding biopsies when clinical concern persists following initial imaging. This is especially valuable given that 71% of breast biopsies performed following an abnormal mammogram result in a benign tissue diagnosis.¹⁸ The decision to biopsy a low-suspicion or indeterminate lesion due to diagnostic uncertainty leads to significant health care costs for the United States, from $1 billion according to conservative estimates up to $2.18 billion.^14,18-19

One primary concern of BSGI is the risk of whole-body radiation exposure. In our study, an average dose of 740 to 832.5 MBq (20–22.5 mCi) of ^99mTc-sestamibi was used. The Food and Drug Administration–recommended activity for ^99mTc-sestamibi is 740 to 1100 MBq (20–30 mCi) for breast imaging, which corresponds to an effective dose of 5.9 to 9.4 mSv.²⁰ In addition, due to the known adherence of ^99mTc-sestamibi to plastic syringes, an average of 20% of activity is retained in the syringe after injection, resulting in an overall decrease in delivered radioactivity.²¹ Of note, there have now been multiple reports of usage of lower dose ^99mTc-sestamibi, indicating that even lower dosages can be adequate and can further decrease radiation exposure in patients, especially with a dual-head detector array.^22,23 It has also been demonstrated that the use of imaging processing algorithms can decrease the dose required to acquire images of sufficient quality.²⁴ In addition, since this is a 1-time problem-solving indication, radiation exposure is substantially less than if BSGI were used as an annual screening study. Overall, the radiation dose from BSGI is within the worldwide annual background radiation level of 2 to 10 mSv, at which current national and international recommendations indicate that concerns about radiation-induced cancer or mortality are unwarranted.²⁰

There are multiple published studies, similar yet different from our study, that have investigated the diagnostic efficacy of BSGI. Hruska et al found that the addition of BSGI to screening mammography in women with dense breasts increased the cancer detection rate and decreased the cost per cancer detected compared with screening mammography alone.²⁵ This study demonstrated the improved diagnostic yield and reduced cost that BSGI can offer when it is used as a supplemental screening tool rather than used as a problem-solving strategy like in our study. A study by Johnson et al found that BSGI is equivalent in sensitivity and superior in specificity for breast cancer detection compared with breast MRI while only costing a quarter as much.¹⁹ This study differed from ours in that many BSGI exams were performed for surgical planning in the setting of newly diagnosed breast cancer as opposed to problem solving in our study. However, it did demonstrate the favorable diagnostic efficacy and low cost of BSGI.

Limitations of this study include the fact that what constitutes an inconclusive or complex imaging finding is subjective and may not be generalizable beyond our practice. In addition, we have a relatively small sample size despite spanning 6 years of BSGI examinations. Due to the intensive and complex nature of data collection, data pertaining to patients who underwent BSGI after 2018 could not be included in this study. Another limitation of this study is that 22 out of 462 patients (4.8%) were LTFU, and their data were not included in the study. It is considered that <5% LTFU leads to little bias and does not pose serious threats to validity.²⁶ Other limitations of our study include a small number of patients with known cancer (11 patients) and lack of including reported specific conventional diagnostic workup BI-RADS assessments.

Conclusion

In summary, the results of our study reveal that BSGI is an effective imaging modality in patients with inconclusive findings despite appropriate diagnostic evaluation with mammography and sonography such that leveraging the strong NPV in inconclusive or complex scenarios may be a useful strategy to avoid unnecessary (often multisite) biopsies. It is our opinion that BSGI is a highly valuable and easily interpreted study that can be used judiciously as a problem-solving tool.

Funding

None declared.

Conflict of interest statement

None declared.

References