Abstract
OBJECTIVE.
The purposes of this study were to compare the tumor appearance of invasive breast cancer on direct-conversion molecular breast imaging using a standardized lexicon and to determine how often direct-conversion molecular breast imaging identifies all known invasive tumor foci in the breast, and whether this differs for invasive ductal versus lobular histologic profiles.
MATERIALS AND METHODS.
Patients with prior invasive breast cancer and concurrent direct-conversion molecular breast imaging examinations were retrospectively reviewed. Blinded review of direct-conversion molecular breast imaging examinations was performed by one of two radiologists, according to a validated lexicon. Direct-conversion molecular breast imaging findings were matched with lesions described on the pathology report to exclude benign reasons for direct-conversion molecular breast imaging findings and to document direct-conversion molecular breast imaging–occult tumor foci. Associations between direct-conversion molecular breast imaging findings and tumor histologic profiles were examined using chi-square tests.
RESULTS.
In 286 patients, 390 invasive tumor foci were present in 294 breasts. A corresponding direct-conversion molecular breast imaging finding was present for 341 of 390 (87%) tumor foci described on the pathology report. Invasive ductal carcinoma (IDC) tumor foci were more likely to be a mass (40% IDC vs 15% invasive lobular carcinoma [ILC]; p < 0.001) and to have marked intensity than were ILC foci (63% IDC vs 32% ILC; p < 0.001). Direct-conversion molecular breast imaging correctly revealed all pathology-proven foci of invasive disease in 79.8% of cases and was more likely to do so for IDC than for ILC (86.1% vs 56.7%; p < 0.0001). Overall, direct-conversion molecular breast imaging showed all known invasive foci in 249 of 286 (87%) patients.
CONCLUSION.
Direct-conversion molecular breast imaging features of invasive cancer, including lesion type and intensity, differ by histologic subtype. Direct-conversion molecular breast imaging is less likely to show all foci of ILC compared with IDC.
Keywords: breast cancer, invasive ductal carcinoma, invasive lobular carcinoma, molecular breast imaging
Gamma camera–based breast imaging is a technique that uses dedicated breast gamma cameras after IV injection of 99mTc-sestamibi to obtain nuclear medicine images of the breast. Because metabolically active cells preferentially accumulate 99mTc-sestamibi, breast malignancy can be distinguished from the surrounding tissue (even if it is mammographically dense) and a physiologic-based image of the breast can be acquired [1–4]. A few different types of gamma camera imaging units exist and have been variably referred to as breast-specific gamma imaging or molecular breast imaging. Breast-specific gamma imaging usually refers to a camera configuration of a single scintillating-crystal detector and compression paddle. Molecular breast imaging usually refers to a configuration of two detectors made of a semiconductor cadmium zinc telluride that directly converts gamma ray energy into electronic signals [5]. The term “direct-conversion molecular breast imaging” has been introduced to specify the dual-head semiconductor configuration [6, 7].
Multiple possible uses for breast gamma imaging are being actively investigated, including screening [8, 9], diagnostic problem solving [1, 5], evaluation of extent of disease and contralateral screening for patients with newly diagnosed breast cancer [10–12], and prediction and evaluation of response to neoadjuvant chemotherapy [6, 13]. In the screening setting, adjunctive molecular breast imaging in combination with screening mammography has been shown to be 91% sensitive for breast cancer in women with dense breast tissue [8]. In the diagnostic setting, prior studies have shown similar sensitivities (> 90%) in patients recommended for biopsy [1, 4, 14–16]. In recent years, several innovations in detector and collimator configuration of the direct-conversion molecular breast imaging system allow the same performance at a substantially lower dose of 99mTc-sestamibi—300 MBq compared with 740–1100 MBq previously used [7, 17–20]. Because of the high sensitivity achieved at lower doses, combined with relatively high specificity, ranging from 86% to 94%, breast gamma imaging may be a valuable tool in the evaluation of breast cancer [8, 18].
Recently, a lexicon specific for breast gamma imaging and based on the existing BI-RADS [21] terms for mammography and MRI has been proposed and validated [22, 23]. To date, the imaging features of breast cancer based on the proposed gamma imaging lexicon are not known. Determining the variety of ways a cancer might appear on a new imaging modality using a standardized reproducible method is an important step in establishing a new modality [24]. In addition, it is not known whether the gamma imaging features of breast cancer might differ by tumor subtype. Because the oncologic approach to breast cancer is changing to reflect an increasing understanding of tumor biology, it is important to determine how tumor biology (including histologic subtype and hormone receptor status) might similarly affect the appearance of cancer on physiology-based imaging techniques [25, 26]. Finally, though breast gamma imaging has been shown to be sensitive for detecting the presence of cancer in the breast, it is not known how well it performs in identifying all tumor foci in a breast with a known cancer, and there may be treatment planning implications for some patients [27]. Given the infiltrative nature of invasive lobular carcinoma (ILC), the full extent of disease for these tumors is more often not depicted on conventional imaging compared with invasive ductal carcinoma (IDC); it is not known how direct-conversion molecular breast imaging performs in identifying all known tumor foci for ILC.
The purposes of this study were to compare the tumor appearance of invasive breast cancer on direct-conversion molecular breast imaging using a standardized lexicon and to determine how often direct-conversion molecular breast imaging identifies all known foci of invasive tumor in the breast and whether this differs for invasive ductal versus lobular histologic profiles.
Materials and Methods
This retrospective study was approved by the Mayo Clinic institutional review board. Images reviewed were acquired as part of previous institutional review board–approved studies or were acquired for clinical use.
Study Population
All patients who underwent direct-conversion molecular breast imaging at Mayo Clinic from August 2005 through June 2011 and who also had an International Classification of Diseases, Ninth Revision, code for breast cancer were retrospectively identified by search of the electronic medical record. Patients whose direct-conversion molecular breast imaging had been performed after definitive surgical treatment or more than 1 year before breast cancer diagnosis were excluded, as were patients whose pathology reports were not available. Patients who had only ductal carcinoma in situ (DCIS) were also excluded.
Direct-Conversion Molecular Breast Images
All direct-conversion molecular breast images were acquired on one of two dual-head dedicated breast gamma imaging systems at our institution (Discovery NM 750b [prototype], GE Healthcare; or LumaGem, Gamma Medica) after IV injection of 99mTc-sestamibi (Cardiolite, DuPont Merck). These systems use two opposing semiconductor detectors with cadmium zinc telluride pixel elements of either 1.6 × 1.6 mm (LumaGem) or 2.5 × 2.5 mm (Discovery NM 750b). The spatial resolution of gamma cameras degrades with distance from the detector and depends on collimator design; at each detector face, the spatial resolution is equivalent to the pixel size but at a distance of 3 cm, which represents the middle of an average 6 cm compressed breast, the spatial resolution was approximately 3–5 mm for images included in this study [20].
Patients in this retrospective study underwent imaging as part of research or clinical protocols with targeted doses of 296, 370, 740, or 999–1221 MBq (8, 10, 20, or 27–33 mCi), or they had received 99mTc-sestamibi for cardiac stress testing and had direct-conversion molecular breast images obtained between stress and rest cardiac imaging. The current standard at our institution is now 300 MBq because dose reduction techniques have allowed preservation of image count density at the lower dose, such that we consistently achieved at least 800 counts/cm2 for included patients regardless of which dose they received. Imaging commenced within 5 minutes of injection and consisted of two 10-minute acquisitions per breast with the breast gently stabilized by the dual detectors in craniocaudal and mediolateral oblique analogous orientations.
Blinded Retrospective Direct-Conversion Molecular Breast Imaging Review
All included direct-conversion molecular breast imaging studies were reviewed by one of two fellowship-trained breast imagers with at least 3 years of experience in interpreting direct-conversion molecular breast imaging. The radiologists were blinded to pathology findings, all prior imaging, and clinical history. Each finding seen on molecular breast imaging was described according to the validated lexicon [22]; the interobserver agreement previously shown using this lexicon was nearly perfect for lesion type (κ = 0.82) and was substantial for lesion intensity (κ = 0.77). For this study, findings were recorded as mass (discrete lesion with convex borders and homogeneous uptake, seen on both views) or nonmass uptake (uptake that differs from the background but does not meet criteria for a mass). The distribution of nonmass uptake was described as focal (less than one fourth of a quadrant and smaller than 2 cm), segmental (triangular and pointed toward the nipple), regional (greater than 2 cm, not segmental), multiple regional (uptake in at least two large areas of tissue), or diffuse (throughout the breast). Finding intensity was determined subjectively and was described as mild (equal to or slightly more than subcutaneous fat), moderate (greater than, but not more than twice as intense as, subcutaneous fat), or marked (more than twice as intense as subcutaneous fat). Subcutaneous fat intensity is the intensity seen peripherally on the image or the same as the background intensity if the background appears uniform.
Pathology
For each tumor focus described on the pathology report, the histologic type (IDC, ILC, invasive mammary cancer, mixed ductal and lobular phenotypes, or other, such as tubular or mucinous), grade (Nottingham 1–3), estrogen and progesterone receptor status, and ERBB2 (formerly HER2-neu) status were recorded. Axillary node status (positive or negative) based on sentinel node biopsy or lymph node dissection was also recorded. The surgical pathology report was referred to whenever available. If only a core biopsy result was present (in cases in which surgery was performed elsewhere), this was accepted as confirmation of the presence and type of invasive cancer in the targeted area only.
Radiologic-Pathologic Correlation
After completion of blinded direct-conversion molecular breast imaging review, pathology reports and direct-conversion molecular breast images were compared by one of the study radiologists to match individual invasive tumor foci described on the pathology report with corresponding direct-conversion molecular breast imaging findings. A direct-conversion molecular breast imaging finding was deemed to correspond with an invasive tumor focus if it was located in the same quadrant of the breast and was at the same depth (anterior, middle, or posterior, if recorded on the pathology report). Any invasive focus found by pathologic analysis that did not have a matching direct-conversion molecular breast imaging finding was considered occult (false-negative). Any direct-conversion molecular breast imaging finding that did not match with an invasive focus on the pathology report was not considered in the analysis (false-positive or DCIS). Two or more invasive foci could be deemed to match a single direct-conversion molecular breast imaging finding if the invasive foci corresponded in location with a larger direct-conversion molecular breast imaging finding (e.g., the pathology report describes several small adjacent masses in a single quadrant, where on direct-conversion molecular breast imaging, a single larger region of uptake in that quadrant is described). Given that 3D measurement of multiple adjacent tumors is limited by the two-view 2D images obtained with direct-conversion molecular breast imaging, the decision of whether several invasive foci matched a single direct-conversion molecular breast imaging finding (assuming the location matched), was validated primarily by visual qualitative or morphologic assessment. For cases in which it was not clear qualitatively whether all foci could be accounted for by the one direct-conversion molecular breast imaging finding, the direct-conversion molecular breast imaging uptake needed to be within 1 cm of the combined size of the invasive foci for them to all be included. If all invasive foci in a single quadrant could not be accounted for by a single direct-conversion molecular breast imaging finding, it was assumed that the larger focus matched the direct-conversion molecular breast imaging finding and the smaller focus or foci were occult. Some patients (n = 50) who had undergone a core biopsy positive for invasive cancer went on to neoadjuvant chemotherapy before definitive surgery. A direct-conversion molecular breast imaging finding in these patients could be deemed true-positive even if surgical pathologic findings were negative, as long as the finding corresponded with the quadrant and depth of the pretherapy positive core biopsy (validated by review of postclip films or conventional imaging of the core biopsy). Invasive tumor foci found on surgical pathologic examination after chemotherapy that did not have a matching direct-conversion molecular breast imaging finding were considered occult.
Statistical Analysis
Associations of direct-conversion molecular breast imaging findings and tumor histologic profile at the lesion level were examined using generalized estimating equations to account for correlation of multiple lesions within individual subjects. Chi-square tests were used at the lesion and subject levels to test for associations of molecular breast imaging findings with tumor characteristics and IDC or ILC status. Exact 95% CIs are also included for tumor subtypes.
Results
Patient Demographics
By medical record search, 589 patients were identified who had undergone direct-conversion molecular breast imaging in the study date range and who had received a diagnosis of breast cancer. Of these, 237 were excluded because their molecular breast imaging was performed after definitive cancer treatment or more than a year before diagnosis, 65 were excluded because they had DCIS only with no invasive disease, and a single patient was excluded because the images were not of diagnostic quality, resulting in an analysis set of 286 patients. Target and administered doses are shown in Table 1. The mean (± SD) patient age was 60.0 ± 12.3 years. Fifty (17%) patients had undergone neoadjuvant chemotherapy before surgery. Ninety-four (33%) patients were node positive at surgery. Eight (2.8%) patients had bilateral disease.
TABLE 1:
Targeted and Administered Doses and Ranges of 99mTc-Sestamibi
| Targeted Dose | No. (%) of Patients | Dose Average, mCi (MBq) | Minimum, Maximum Dose, mCi (MBq) | 10th Percentile, 90th Percentile, mCi (MBq) |
|---|---|---|---|---|
| 8 mCi (296 MBq) | 80 (28) | 8.5 (314.5) | 7.3, 8.8 (270.1, 325.6) | 7.8, 8.8 (288.6, 325.6) |
| 10 mCi (370 MBq) | 2 (1) | 10.5 (388.5) | 10.2, 10.8 (377.4, 399.6) | 10.3, 10.7 (381.1, 395.9) |
| 20 mCi (740 MBq) | 112 (39) | 21.1 (780.7) | 18.0, 22.0 (666, 814) | 19.5, 22.0 (721.5, 814) |
| 27–33 mCi (999–1221 MBq) | 89 (31) | 30.6 (1132.2) | 27.1, 33.0 (1002.7, 1221) | 28.6, 32.4 (1058.2, 1198.8) |
| Cardiac stress test dose | 2 (1) | 51 (1887) | 51, 51 (1887) | 51, 51 (1887, 1887) |
| Dosed during sestamibi shortage | 1 (0.3) | 5.9 (218.3) | 5.9, 5.9 (218.3, 218.3) | 5.9, 5.9 (218.3, 218.3) |
| Total | 286 | 20.6 (762.2) | 5.9, 51 (218.3, 1887) | 8.3, 31.7 (307.1, 1172.9) |
Tumor Pathology Characteristics
In the 286 qualifying patients, there were 390 pathologically identified invasive tumor foci in 294 breasts. The mean invasive tumor size was 1.5 ± 1.4 cm. The mean size of the largest tumor in each breast was 1.7 ± 1.5 cm. The 390 invasive foci included 245 IDCs, 67 ILCs, 60 mixed IDC and ILC subtypes, and 18 foci of other histopathologic subtypes. Patient and tumor characteristics are shown in Table 2, with results of comparative statistical analysis performed between tumor foci of pure IDC versus pure ILC. The 18 other lesions (in 13 patients) included 12 tubular carcinomas (in eight patients); two solid papillary carcinomas (in one patient); one each of medullary, mucinous, tubuloductal cancer; and a phyllodes tumor. Hormone receptor status was available for 298 lesions, 77.5% of which were hormone receptor positive and ERBB2 negative. The number of hormone receptor–negative, ERBB2-positive, and triple-negative lesions was too small to allow subgroup analysis.
TABLE 2:
Tumor and Molecular Breast Imaging Characteristics for Invasive Ductal Carcinoma (IDC) and Invasive Lobular Carcinoma (ILC)
| Characteristic | IDC | ILC | p |
|---|---|---|---|
| Number of lesionsa | 245 | 67 | |
| Number of breasts | 208 | 37 | |
| Number of patients | 202 | 37 | |
| Patient age (y), mean ± SD | 59.6 ± 12.9 | 62.7 ± 10.6 | |
| Cancer histologic type | |||
| Unifocal | 168 | 25 | |
| Multifocal or multicentricb | 26 | 11 | |
| Bilateral | 8 | 0 | |
| Axillary lymph node status | 66 | 9 | |
| Size of largest focus per breast (cm), mean ± SD | 1.42 ± 1.36 | 1.56 ± 1.48 | 0.908 |
| Receptor status | |||
| ER positivec | 168 (78.9) | 37 (100.0) | |
| ER negative | 45 (21.1) | 0 (0.0) | |
| PR positive | 133 (63.0) | 25 (67.6) | |
| PR negative | 78 (37.0) | 12 (32.4) | |
| ERBB2 positive | 29 (13.8) | 2 (5.4) | |
| ERBB2 negative | 181 (86.2) | 35 (94.6) | |
| Triple negative | 29 (13.7) | 0 (0.0) | 0.017 |
| Cancer grade | |||
| Grade 1 | 54 (22.0) | 38 (56.7) | |
| Grade 2 | 123 (50.2) | 29 (43.3) | |
| Grade 3 | 68 (27.8) | 0 (0.0) | |
| Molecular breast imaging intensity | |||
| Marked | 135 (63.4) | 12 (31.6) | < 0.001 |
| Moderate | 54 (25.4) | 16 (42.1) | |
| Mild | 24 (11.3) | 10 (26.3) | |
| Molecular breast imaging lesion type, no. (%) of all lesions | |||
| Mass | 90 (39.6) | 8 (14.6) | < 0.001 |
| Nonmass | 120 (52.9) | 30 (54.6) | |
| Focal | 89 | 19 | |
| Segmental | 10 | 5 | |
| Regional | 12 | 5 | |
| Multiple Regional | 7 | 0 | |
| Diffuse | 2 | 1 | |
| No lesion | 17 (7.5) | 17 (30.9) |
Note—Except where noted otherwise, data are number (%) of patients. ER = estrogen receptor, PR = progesterone receptor.
Patients with other tumors (18 tumors in 13 patients) are not reflected in the tables.
Patients with multifocal or multicentric or bilateral cancer who had more than one tumor histologic type (n = 8) are included under both histologic types.
IDC missing receptor status data: ER positive, n = 32; PR positive, n = 34; ERBB2 positive, n = 35. ILC missing receptor status data: ER positive, n = 30; PR positive, n = 30; ERBB2 positive, n = 30.
Lesion Direct-Conversion Molecular Breast Imaging Characteristics
Imaging characteristics of the IDC and ILC foci are detailed in Table 2, and characteristics of mixed phenotype foci are shown in Table 3. IDC was more likely than ILC to be of marked intensity (63% vs 32%; p = 0.0008). IDC was also more likely than ILC to appear as a mass on direct-conversion molecular breast imaging (odds ratio, 1.23; 95% CI, 1.07–1.42; p = 0.0035). Example cases of IDC and ILC, showing some of the imaging characteristics in the lexicon, are shown in Figures 1 and 2.
TABLE 3:
Tumor and Molecular Breast Imaging Characteristics for Invasive Cancers With Mixed Ductal and Lobular Features
| Characteristic | Value |
|---|---|
| Number of tumor foci | 60 (in 42 breasts) |
| Patient age (y), mean ± SD | 62.1 ± 10.5 |
| Cancer histologic type, no. of lesions | |
| Unifocal | 30 |
| Multifocal or multicentric | 12 |
| Bilateral | 0 |
| Axillary lymph node status | 19 positive (of 42 patients) |
| Size of largest focus per breast (cm), mean ± SD | 1.57 ± 1.35 |
| Receptor status, no. (%) of patients (n = 45) | |
| ER positive | 39 (86.7) |
| ER negative | 6 (13.3) |
| PR positive | 27 (60.0) |
| PR negative | 18 (40.0) |
| ERBB2 positive | 1 (2.4) |
| ERBB2 negative | 41 (97.6) |
| Triple negative | 5 (11.9) |
| Cancer grade, no. (%) of lesions | |
| Grade 1 | 24 (40.0) |
| Grade 2 | 29 (48.3) |
| Grade 3 | 7 (11.7) |
| Molecular breast imaging intensity, no. of lesions | |
| Mild | 5 |
| Moderate | 15 |
| Marked | 22 |
| Molecular breast imaging finding type, no. of lesions | |
| Mass | 20 |
| Nonmass | 21 |
| Focal | 19 |
| Segmental | 1 |
| Regional | 1 |
| Multiple regional | 1 |
| Diffuse | 0 |
| No finding | 9 |
Note—ER = estrogen receptor, PR = progesterone receptor
Fig. 1—
66-year-old woman with grade 3 invasive ductal carcinoma (IDC) in left breast. Craniocaudal (left) and mediolateral oblique (right) analogous direct-conversion molecular breast images show marked intensity mass (arrows) in inferior and slightly medial left breast at posterior depth, corresponding with 1.3-cm grade 3 IDC found at pathologic analysis. No other findings were present in left breast on imaging or pathologic analysis.
Fig. 2—
62-year-old woman with multifocal mixed ductal and lobular invasive cancer in upper outer left breast. Craniocaudal (left) and mediolateral oblique (right) analogous direct-conversion molecular breast images show marked-intensity mass at posterior depth (solid arrows), moderate-intensity mass at middle depth (open arrows), and mild-intensity focal area of nonmass uptake at middle depth (arrowhead) seen on craniocaudal image only, likely obscured on mediolateral oblique image by adjacent mass. On pathologic analysis, these corresponded with grade 2 mixed ductal and lobular cancers measuring 1.9, 1.7, and 1.0 cm, respectively.
At the lesion level, direct-conversion molecular breast imaging showed a corresponding finding for 341 of 390 (87%) pathology-described tumor foci. Excluding lesions in patients who had received neoadjuvant chemotherapy, the average size (measured on surgical pathologic examination) of tumor focus seen on direct-conversion molecular breast imaging is 1.7 cm (range, 0.0–7.2 cm; excluding one patient who had no residual tumor at surgery because the entire lesion had been removed at needle biopsy the lower end of the range is 0.2 cm). ILC was more likely to be occult on direct-conversion molecular breast imaging than IDC (30.9% vs 7.5%; odds ratio, 1.18; 95% CI, 1.05–1.32; p = 0.0062). Of the 49 lesions that were occult on direct-conversion molecular breast imaging, 31 were present in breasts where a separate tumor focus was detected by direct-conversion molecular breast imaging (i.e., were occult satellite lesions; an example case is illustrated in Fig. 3). The median size on pathologic examination of occult satellite lesions was 6 mm (range, 1–21 mm).
Fig. 3—
54-year-old woman with invasive lobular carcinoma (ILC) in left breast.
A, Craniocaudal (left) and mediolateral oblique (right) analogous direct-conversion molecular breast images show moderate-intensity focal area of nonmass uptake (arrows) in upper outer quadrant of left breast at middle depth; on pathologic analysis, matching 1.5-cm grade 2 ILC was identified. No additional molecular breast imaging findings were described. On pathologic analysis, three additional masses were present (grade 2 ILC, 2.1, 1.7, and 0.6 cm).
B, Two larger molecular breast imaging–occult lesions were in far lateral breast and were not included in molecular breast imaging FOV. One of these lesions (arrow) is visible with biopsy clip artifact on axial contrast-enhanced MRI.
When analyzed per breast, direct-conversion molecular breast imaging depicted all known foci of invasive cancer in 257 of 294 breasts (87%). Direct-conversion molecular breast imaging showed all foci in 192 of 202 (95%) breasts with IDC, in 39 of 42 (93%) breasts with invasive tumors that had mixed ductal and lobular features, and in 34 of 38 (89%) breasts with ILC. Fifty-five of the 294 breasts with cancer had multifocal or multicentric disease: direct-conversion molecular breast imaging depicted all known multifocal or multicentric invasive cancer foci in 36 of 55 (65%) breasts. Eighteen lesions (in 18 breasts) were present in breasts in which direct-conversion molecular breast imaging showed no abnormal findings (i.e., false-negative direct-conversion molecular breast imaging). The median size of invasive cancers present in breasts that had a false-negative direct-conversion molecular breast imaging was 6 mm (range, 3–12 mm).
Overall, 49 tumor foci were molecular breast imaging occult. Possible reasons for nonvisualization on molecular breast imaging are detailed in Table 4. A case showing both direct-conversion molecular breast imaging–occult and visible lesions is presented in Figure 4.
TABLE 4:
Molecular Breast Imaging–Occult Findings and Possible Reasons
| Reasons Why a Tumor Could Be Molecular Breast Imaging Occult (n = 49 Findings) | No. (%) of Findings | Comment |
|---|---|---|
| Out of FOV | 8 (16) | |
| Obscured by moderate or marked background uptake | 11 (22) | Three molecular breast imaging–occult lesions in breasts with moderate or marked background uptake were also ≤ 5 mm |
| Size ≤ 5 mm | 18 (37) | |
| No known reason | 17 (35) | |
| Histologic profile | ||
| ILC grade 1 | 3 (18) | Sizes 0.8, 0.8, 0.6 cm |
| ILC grade 2 | 2 (12) | Sizes 0.6, 0.6 cm |
| Tubular | 2 (12) | Sizes 1.4 cm, size not known (no residual carcinoma seen at surgical pathology) |
| Mucinous | 1 (6) | Size 0.7 cm |
| IDC grade 3 | 1 (6) | Size 0.7 cm |
| IDC grade 1 | 2 (12) | Sizes 0.6 and 0.8 cm |
| Mixed grade 1 | 2 (12) | Sizes 0.7 and 0.9 cm |
| Mixed grade 2 | 2 (12) | Sizes 0.6 and 0.7 cm |
Note—IDC = invasive ductal carcinoma, ILC = invasive lobular carcinoma.
Fig. 4—
43-year-old woman with multicentric invasive cancer in left breast.
A, Craniocaudal (left) and mediolateral oblique (right) analogous direct-conversion molecular breast images show marked intensity bilobed mass (arrows) present in inferior left breast. On pathologic analysis, two adjacent masses (grade 3 mixed phenotype, 3.7 and 1.4 cm) were described in same location and were scored as single molecular breast imaging finding, accounting for two pathology-described lesions.
B, Representative microscopy image shows invasive tumor cells with nuclear pleomorphism and no discernable tubule formation. Three additional cancers (grade 1 tubular carcinomas, 1.6, 1.3, and 0.8 cm) were present in central left breast on pathologic analysis and were not described on molecular breast imaging (therefore scored as occult), likely obscured by marked background parenchymal uptake throughout breast.
C, Representative microscopy image shows well-formed tubules infiltrating stroma.
Of the 18 other tumor types (in 13 patients), 12 were tubular carcinomas; six of these appeared as mild intensity focal nonmass uptake, and the other six were occult on direct-conversion molecular breast imaging. Six lesions in four patients were in the same breast as nontubular invasive cancer. All known foci of disease were seen in three of the eight patients. The medullary cancer, phyllodes tumor, and two solid papillary tumors appeared as marked intensity masses. The mucinous and tubuloductal cancers appeared as mild intensity focal nonmass uptake.
Discussion
This study shows a significant difference in direct-conversion molecular breast imaging appearance between IDC and ILC. Lobular cancers are less likely to appear as a mass and are less intense. This study also shows that direct-conversion molecular breast imaging is more likely to depict all known invasive tumor foci for IDC than for ILC, though overall the full extent of invasive disease was accurately depicted by direct-conversion molecular breast imaging in 87% of patients, which compares favorably with MRI (75%) and positron emission mammography (65%) [28].
ILC is known to present diagnostic challenges on all breast imaging modalities [29–32]. Given the single-file pattern of invasion seen with ILC, tumor biology likely explains why the full extent of ILC is less likely to be depicted on imaging than that of IDC.
Few studies have evaluated the relationship between sestamibi uptake and breast tumor histology. These studies looked at a related technology, breast-specific gamma imaging, which also uses injected 99mTc-sestamibi, though it differs in that the gamma camera is composed of a single scintillating-crystal (NaI) detector and a compression paddle and the necessary injected dose is higher [33]. One prior study looked at the use of breast-specific gamma imaging in assessing invasive breast cancer [34]; 149 invasive cancers in 139 patients were retrospectively reviewed and classified as normal or abnormal without lesion classification. Though radiologic-pathologic correlation was not done on the lesion level, there was some uptake noted in the corresponding quadrant 98% of the time. Only nine of the cases in that study were ILCs. An earlier study of breast-specific gamma imaging showed that it identified six of six ILCs, including a case of multifocal disease [35]. Finally, a study of breast-specific gamma imaging in patients with a known or suspected primary tumor showed that breast-specific gamma imaging detected occult cancer in 9% of cases [36]. This differs from our study in that it did not include malignant lesions not seen on breast-specific gamma imaging.
Of interest is why the lesions that were occult on direct-conversion molecular breast imaging were not seen. In our experience, the most common circumstances that lead to a false-negative molecular breast image are small tumor size, lower grade tumors, obscuration by moderate or marked background uptake, or location outside the FOV of direct-conversion molecular breast imaging. (Though the amount of tissue included in the FOV varies by patient, positioning, and compression used, we estimate that compared with MRI, direct-conversion molecular breast imaging shows 1–2 cm less of posterior tissue, much of the axilla, and all of the chest wall.) In this study, 42 of 49 molecular breast imaging–occult lesions were accounted for by one of these reasons. That said, these confounding circumstances were present in many of the cases in which the tumor foci were not occult on molecular breast imaging. The reason for this variability is not known and may reflect the technique or variations in patient and tumor biology. For lesions seen on direct-conversion molecular breast imaging, the lower range of tumor focus size measured at surgery (0.2 cm) is less than would be expected given the detector resolution. This may reflect cases in which a significant amount of a small tumor is removed during needle biopsy performed after direct-conversion molecular breast imaging, leaving minimal residual tumor to measure at surgery. Of note, our study included a wide range of administered doses of 99mTc-sestamibi. This reflects the retrospective design, which included some earlier direct-conversion molecular breast imaging examinations performed during pilot stages and before dose-reduction strategies were developed and used. Currently, all direct-conversion molecular breast imaging examinations are performed on systems that are optimized for lower dose imaging, and doses of 8 mCi (300 MBq) are routine, which results in an approximate 2.4 mSv effective dose.
In this study, we did not address false-positive findings, in part because of our focus on characterizing invasive disease and in part because there were not typically follow-up or corresponding pathologic examination results available for false-positive findings. Previous direct-conversion molecular breast imaging work shows false-positive findings on direct-conversion molecular breast imaging to be due to atypia, fibroadenomas, fat necrosis, papillomas, stromal fibrosis, lymph nodes, sclerosing adenosis, pseudoangiomatous stromal hyperplasia, and benign tissue [1, 8]. In the screening setting, 6.7% of women undergoing direct-conversion molecular breast imaging were called back for additional imaging, with 67% of the findings resolved with immediate diagnostic mammography or ultrasound and single 6-month follow-up molecular breast imaging [37].
This study has several limitations. The determination of whether direct-conversion molecular breast imaging findings and pathology-identified lesions match was made by the study radiologists. Although this was necessary for accurate radiologic-pathologic comparison, it introduced the possibility for bias. Because pathologic results were taken retrospectively from the medical record, there were a variety of interpreting pathologists with a variety of reporting styles, which sometimes made it difficult to consistently match pathologic and radiologic findings. This study does not address the performance of direct-conversion molecular breast imaging with DCIS (and therefore does not study the total extent of disease). Given the retrospective design and the study population limited to patients with known cancer, a true measurement of sensitivity of direct-conversion molecular breast imaging cannot be determined. Finally, we relied on final surgical pathologic examination as a reference standard, though it is possible that additional disease could have been present and undetected in patients who underwent lumpectomy, who had small tumor foci not seen by pathologic examination at mastectomy, or who had initially undetected tumor foci that were eradicated by neoadjuvant chemotherapy before surgery.
Opportunities for future research include prospective study of direct-conversion molecular breast imaging accuracy at predicting the extent of malignancy (both invasive and DCIS). A prospective design would allow radiologic-pathologic correlation at the time of surgery and therefore improved correlation. In addition, differences in imaging appearance on direct-conversion molecular breast imaging between tumor histologic types could be further investigated to determine whether there is prognostic meaning or whether it predicts neoadjuvant chemotherapy treatment response.
In conclusion, direct-conversion molecular breast imaging shows significant differences in imaging appearance of IDC compared with ILC. Direct-conversion molecular breast imaging detected all pathology-proven invasive breast lesions in 87% of breasts.
Acknowledgments
We thank Ricky Carter for statistical advice and Sejal Shah and Beiyun Chen for their pathology expertise and images.
Supported by the Mayo Clinic Center for Translational Science Activities through grant UL1 RR024150 from the National Center for Research Resources, a component of the National Institutes of Health. C. B. Hruska receives royalties for licensed technologies relating to molecular breast imaging per an agreement between Mayo Clinic and Gamma Medica.
References
- 1.Hruska CB, Phillips SW, Whaley DH, Rhodes DJ, O’Connor MK. Molecular breast imaging: use of a dual-head dedicated gamma camera to detect small breast tumors. AJR 2008; 191:1805–1815 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fleming RM. Mitochondrial uptake of sestamibi distinguishes between normal, inflammatory breast changes, pre-cancers, and infiltrating breast cancer. Integr Cancer Ther 2002; 1:229–237 [DOI] [PubMed] [Google Scholar]
- 3.Yang WT. Emerging techniques and molecular imaging in breast cancer. Semin Ultrasound CT MR 2011; 32:288–299 [DOI] [PubMed] [Google Scholar]
- 4.Rechtman LR, Lenihan MJ, Lieberman JH, et al. Breast-specific gamma imaging for the detection of breast cancer in dense versus nondense breasts. AJR 2014; 202:293–298 [DOI] [PubMed] [Google Scholar]
- 5.Hruska CB, O’Connor MK. Nuclear imaging of the breast: translating achievements in instrumentation into clinical use. Med Phys 2013; 40:050901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mitchell D, Hruska CB, Boughey JC, et al. 99mTc-sestamibi using a direct conversion molecular breast imaging system to assess tumor response to neoadjuvant chemotherapy in women with locally advanced breast cancer. Clin Nucl Med 2013; 38:949–956 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hruska CB, Conners AL, Jones KN, et al. Half-time Tc-99m sestamibi imaging with a direct conversion molecular breast imaging system. EJNMMI Res 2014; 4:5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rhodes DJ, Hruska CB, Phillips SW, Whaley DH, O’Connor MK. Dedicated dual-head gamma imaging for breast cancer screening in women with mammographically dense breasts. Radiology 2011; 258:106–118 [DOI] [PubMed] [Google Scholar]
- 9.Hruska CB, Rhodes DJ, Collins DA, Tortorelli CL, Askew JW, O’Connor MK. Evaluation of molecular breast imaging in women undergoing myocardial perfusion imaging with Tc-99m sestamibi. J Womens Health (Larchmt) 2012; 21:730–738 [DOI] [PubMed] [Google Scholar]
- 10.Edwards C, Williams S, McSwain AP, et al. Breast-specific gamma imaging influences surgical management in patients with breast cancer. Breast J 2013; 19:512–519 [DOI] [PubMed] [Google Scholar]
- 11.Zhou M, Johnson N, Gruner S, et al. Clinical utility of breast-specific gamma imaging for evaluating disease extent in the newly diagnosed breast cancer patient. Am J Surg 2009; 197:159–163 [DOI] [PubMed] [Google Scholar]
- 12.Killelea BK, Gillego A, Kirstein LJ, et al. George Peters Award: how does breast-specific gamma imaging affect the management of patients with newly diagnosed breast cancer? Am J Surg 2009; 198:470–474 [DOI] [PubMed] [Google Scholar]
- 13.Wahner-Roedler DL, Boughey JC, Hruska CB, et al. The use of molecular breast imaging to assess response in women undergoing neoadjuvant therapy for breast cancer: a pilot study. Clin Nucl Med 2012; 37:344–350 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Weigert JM, Bertrand ML, Lanzkowsky L, Stern LH, Kieper DA. Results of a multicenter patient registry to determine the clinical impact of breast-specific gamma imaging, a molecular breast imaging technique. AJR 2012; 198:[web]W69–W75 [DOI] [PubMed] [Google Scholar]
- 15.Siegal E, Angelakis E, Morris P, Pinkus E. Breast molecular imaging: a retrospective review of one institution’s experience with this modality and analysis of its potential role in breast imaging decision making. Breast J 2012; 18:111–117 [DOI] [PubMed] [Google Scholar]
- 16.Sun Y, Wei W, Yang HW, Liu JL. Clinical usefulness of breast-specific gamma imaging as an adjunct modality to mammography for diagnosis of breast cancer: a systemic review and meta-analysis. Eur J Nucl Med Mol Imaging 2013; 40:450–463 [DOI] [PubMed] [Google Scholar]
- 17.Weinmann AL, Hruska CB, Conners AL, O’Connor MK. Collimator design for a dedicated molecular breast imaging-guided biopsy system: proof-of-concept. Med Phys 2013; 40:012503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hruska CB, Weinmann AL, Tello Skjerseth CM, et al. Proof of concept for low-dose molecular breast imaging with a dual-head CZT gamma camera. Part II. Evaluation in patients. Med Phys 2012; 39:3476–3483 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hruska CB, Weinmann AL, O’Connor MK. Proof of concept for low-dose molecular breast imaging with a dual-head CZT gamma camera. Part I. Evaluation in phantoms. Med Phys 2012; 39:3466–3475 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Weinmann AL, Hruska CB, O’Connor MK. Design of optimal collimation for dedicated molecular breast imaging systems. Med Phys 2009; 36:845–856 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.D’Orsi CJ, Mendelson EB, Ikeda DM. Breast Imaging Reporting and Data System: ACR BI-RADS—breast imaging atlas. Reston, VA: American College of Radiology, 2003 [Google Scholar]
- 22.Conners AL, Hruska CB, Tortorelli CL, et al. Lexicon for standardized interpretation of gamma camera molecular breast imaging: observer agreement and diagnostic accuracy. Eur J Nucl Med Mol Imaging 2012; 39:971–982 [DOI] [PubMed] [Google Scholar]
- 23.Conners AL, Maxwell RW, Tortorelli CL, et al. Gamma camera breast imaging lexicon. AJR 2012; 199:[web]W767–W774 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Schillaci O The importance of standardized interpretation of molecular breast imaging with dedicated gamma cameras. Eur J Nucl Med Mol Imaging 2012; 39:967–970 [DOI] [PubMed] [Google Scholar]
- 25.Boisserie-Lacroix M, Hurtevent-Labrot G, Ferron S, Lippa N, Bonnefoi H, Mac Grogan G. Correlation between imaging and molecular classification of breast cancers. Diagn Interv Imaging 2013; 94:1069–1080 [DOI] [PubMed] [Google Scholar]
- 26.Birdwell RL, Mountford CE, Iglehart JD. Molecular imaging of the breast. AJR 2009; 193:367–376 [DOI] [PubMed] [Google Scholar]
- 27.Sung JS, Li J, Da Costa G, et al. Preoperative breast MRI for early-stage breast cancer: effect on surgical and long-term outcomes. AJR 2014; 202:1376–1382 [DOI] [PubMed] [Google Scholar]
- 28.Berg WA, Madsen KS, Schilling K, et al. Breast cancer: comparative effectiveness of positron emission mammography and MR imaging in pre-surgical planning for the ipsilateral breast. Radiology 2011; 258:59–72 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lopez JK, Bassett LW. Invasive lobular carcinoma of the breast: spectrum of mammographic, US, and MR imaging findings. RadioGraphics 2009; 29:165–176 [DOI] [PubMed] [Google Scholar]
- 30.Berg WA, Gutierrez L, NessAiver MS, et al. Diagnostic accuracy of mammography, clinical examination, US, and MR imaging in preoperative assessment of breast cancer. Radiology 2004; 233:830–849 [DOI] [PubMed] [Google Scholar]
- 31.Sickles EA. Findings at mammographic screening on only one standard projection: outcomes analysis. Radiology 1998; 208:471–475 [DOI] [PubMed] [Google Scholar]
- 32.Hilleren DJ, Andersson IT, Lindholm K, Linnell FS. Invasive lobular carcinoma: mammographic findings in a 10-year experience. Radiology 1991; 178:149–154 [DOI] [PubMed] [Google Scholar]
- 33.Dickerscheid D, Lavalaye J, Romijn L, Habraken J. Contrast-noise-ratio (CNR) analysis and optimisation of breast-specific gamma imaging (BSGI) acquisition protocols. EJNMMI Res 2013; 3:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Tadwalkar RV, Rapelyea JA, Torrente J, et al. Breast-specific gamma imaging as an adjunct modality for the diagnosis of invasive breast cancer with correlation to tumour size and grade. Br J Radiol 2012; 85:e212–e216 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kieper D, Brem RF, Hoeffer R, Keppel C, Wymer D. Detecting infiltrating lobular carcinoma using scintimammographic breast specific gamma imaging. Phys Med 2006; 21(suppl 1):125–127 [DOI] [PubMed] [Google Scholar]
- 36.Brem RF, Shahan C, Rapleyea JA, et al. Detection of occult foci of breast cancer using breast-specific gamma imaging in women with one mammographic or clinically suspicious breast lesion. Acad Radiol 2010; 17:735–743 [DOI] [PubMed] [Google Scholar]
- 37.Hruska CB, Conners AL, Jones KN, O’Connor MK, Rhodes DJ. False positive findings on adjunct screening molecular breast imaging with Tc-99m sestamibi. (abstract) RSNA 2013 Oak Brook, IL: Radiological Society of North America, 2013. rsna2013.rsna.org/pdf/PDFFiles/13016859.pdf. Accessed May 1, 2015 [Google Scholar]