Abstract
Purpose:
To investigate the utility of contrast-enhanced mammography (CEM) as an alternative to breast MRI for the evaluation of residual disease after neoadjuvant treatment (NAT).
Methods:
This prospective study enrolled consecutive women undergoing NAT for breast cancer from July 2017–July 2019. Breast MRI and CEM exams performed after completion of NAT were read independently by two breast radiologists. Residual disease and lesion size on MRI and CEM recombined (RI) and low-energy images (LEI) were compared. Histopathology was considered the reference standard. Statistical analysis was performed using McNemar’s and Leisenring’s tests. Multiple comparison adjustment was made using Bonferroni procedure. Lesion sizes were correlated using Kendall’s tau coefficient.
Results:
There were 110 participants with 115 breast cancers. Residual disease (invasive cancer or ductal carcinoma in situ) was detected in 83/115 (72%) lesions on pathology, 71/115 (62%) on MRI, 55/115 (48%) on CEM RI, and 75/115 (65%) on CEM LEI. When using multiple comparison adjustment, no significant differences were detected between MRI combined with CEM LEI and CEM RI combined with CEM LEI, in terms of accuracy (MRI: 77%, CEM: 72%; p≥0.99), sensitivity (MRI: 88%, CEM: 81%; p≥0.99), specificity (MRI: 47%, CEM: 50%; p≥0.99), PPV (MRI: 81%, CEM: 81%; p≥0.99), or NPV (MRI: 60%, CEM: 50%; p≥0.99). Size correlation between pathology and both MRI combined with CEM LEI and CEM RI combined with CEM LEI was moderate: τ = 0.36 vs 0.33 (p≥0.99).
Conclusion:
Contrast-enhanced mammography is an acceptable alternative to breast MRI for the detection of residual disease after neoadjuvant treatment.
Keywords: Breast neoplasms, neoadjuvant therapy, mammography, magnetic resonance imaging
INTRODUCTION
The field of breast imaging continues to evolve as new imaging technologies and treatment paradigms are introduced. Contrast-enhanced mammography (CEM) is an imaging modality gaining a strong foothold as a diagnostic tool, with mounting evidence supporting its ability to aid in the detection of breast cancer [1–3]. CEM has demonstrated equal or better sensitivity and specificity in the diagnosis of breast cancer compared to standard full-field digital mammography (FFDM) [4, 5]. CEM has also demonstrated superior cancer detection compared with digital breast tomosynthesis [6]. It has the advantage of acquiring, in a single exam, both low-energy images (LEI) that are equivalent to FFDM as well as recombined images (RI) that can demonstrate enhancing lesions even without abnormalities on the LEI [7]. Multiple studies have suggested that CEM is comparable to breast MRI in evaluating disease extent [5, 8–10]. A logical next step is to utilize CEM to monitor disease response to therapy, particularly within the setting of neoadjuvant treatment (NAT).
NAT is the management of choice for large tumors with biological profiles sensitive to chemotherapy and/or targeted treatments [11–13]. It may allow otherwise nonresectable tumors to be removed surgically and can improve cosmetic outcomes in other large tumors [14, 15]. For patients at higher risk of metastatic disease, NAT also provides systemic treatment earlier than if a surgical procedure would have been performed first. Additionally, prognosis is improved in patients with pathologic complete response with lower local and distant recurrence rates [16].
Currently, breast MRI is considered the gold standard for imaging evaluation before and after NAT and can be used in combination with FFDM to detect residual disease [17, 18]. Upon post-treatment examination, malignant tumors may demonstrate reduction or complete resolution of abnormal enhancement in patients with good response to NAT. MRI findings are known to correlate well with pathologic findings but can sometimes over- or underestimate lesion size and extent [9]. Furthermore, residual pathological disease may not be identified on MRI or residual enhancement may be present in patients with pathologic complete response [19]. CEM is less expensive and less time-consuming than MRI, and has the advantage of providing functional and vascular information within the anatomic framework of FFDM [20, 21].
In this study, we investigate the utility of CEM as an alternative to breast MRI for the evaluation of residual disease after NAT in patients with breast cancer. Presence and size of residual invasive carcinoma and ductal carcinoma in situ (DCIS) on pathology of surgical specimens was compared to residual lesions on both CEM and MRI after NAT.
METHODS
Study sample
This prospective, single-institution study, approved by the institutional review board, registered on the clinicaltrials.gov website (NCT03070340; https://www.clinicaltrials.gov/ct2/show/NCT03070340) and compliant with the Health Insurance Portability and Accountability Act, enrolled consecutive women undergoing NAT for breast cancer between July 2017 and July 2019. Participants included in the study had a full standard protocol breast MRI and a full bilateral CEM at our institution no more than 30 days after completion of NAT and prior to surgery. MRI and CEM were performed within a 7-day period. Written informed consent was obtained from each participant. The exclusion criteria were (1) patients who chose to withdraw from the study; (2) surgery performed at an outside institution; (3) surgery performed more than 2 months after completion of the NAT; (4) CEM or MRI not performed due to a medical reason; and (5) lesion not included in the field-of-view of CEM due to far posterior location. Participant accrual is demonstrated in Fig. 1.
Fig. 1.
Patient accrual.
CEM protocol
Imaging acquisition was performed using dual-energy mammography equipment (Selenia Dimensions, Hologic). Iodine-based contrast media (Iohexol, Omnipaque 350 mg I/ml, GE Healthcare) was administered intravenously at a dose of 1.5 mL/kg to a maximum of 150 mL. The injection influx rate was set to 3 mL/s. Imaging acquisition was initiated approximately 2.0–2.5 min after contrast infusion and ideally finished within 8 minutes. LEI and high-energy images were obtained nearly simultaneously in the mediolateral oblique and craniocaudal projections. The LEI served as the diagnostic equivalent of FFDM while high and low exposures created by a vendor-specific proprietary algorithm were used to create the RI, elucidating areas of contrast concentration. CEM LEI and RI were uploaded to the institutional Picture Archiving and Communication System (PACS) for interpretation.
MRI protocol
Breast MRI examinations were performed using 1.5 or 3.0T equipment (GE Healthcare) with a dedicated 8 or 16-channel phased-array breast coil. Imaging acquisition consisted of a T2-weighted axial sequence with fat suppression, a T1-weighted axial sequence without fat suppression, and pre- and postcontrast T1-weighted sequences with fat suppression. Contrast-enhanced sequences were acquired before and at three 60-s intervals after injection of 0.1 mmol/kg of Gadobutrol (Gadavist, Bayer HealthCare). Subtraction post-processed images were obtained by subtracting each of the 3 post-contrast sequences from the pre-contrast sequence. Three-dimensional maximum projection intensity images were also created and uploaded to the institutional PACS system for interpretation. The breast MRI protocols are demonstrated in Table 1.
Table 1.
Breast MRI protocols
| MRI sequences | 1.5 T | 3.0 T |
|---|---|---|
| Axial non-fat-suppressed T2-weighted | STIR; TR 6,000 ms; TE 70 ms; Flip angle 160°; FOV 36 cm; matrix 256 × 256; slice thickness 4 mm | FSE; TR 4,384 ms; TE 102 ms; Flip angle 111°; FOV 35 cm; matrix 288 × 224; slice thickness 3 mm |
| Axial non-fat-suppressed T1-weighted | 3D VIBRANT; TR Min 1.8 Max 1500 ms; TE min 4.2 max 24 ms; Flip angle 10°; FOV 36 cm; matrix 360 × 360; slice thickness 2 mm | 3D VIBRANT; TR 4.3 ms; TE 2.1 ms; Flip angle 10°; FOV 30 cm; matrix 320 × 192; slice thickness 1 mm |
| Axial contrast-enhanced fat-suppressed T1-weighted | 3D VIBRANT; TR 6.2 ms; TE 3 ms; Flip angle 10°; FOV 36 cm; matrix 360 × 360; slice thickness 2 mm | 3D VIBRANT; TR 4.3 ms; TE 2.1 ms; Flip angle 10°; FOV 32 cm; matrix 300 × 300; slice thickness 1 mm |
Abbreviations: FOV = field of view; MRI = magnetic resonance imaging; STIR = Short Tau Inversion Recovery; TE = echo time; TR = repetition time; VIBRANT = volume imaging for breast assessment
CEM and MRI Evaluation
Both CEM and MRI were interpreted independently by two different dedicated breast imaging radiologists with > 12 years of experience in breast imaging and 1–7 years of experience with CEM who remained blinded to the results of the other modality. Each study was read by a single radiologist from the breast imaging service team. Morphologic and enhancement features were recorded, including lesion type and size. Pre-treatment size was based on measurements on initial breast MRI. In participants with multifocal disease, the largest dimension of the lesions combined was noted. In participants with multicentric disease, different lesions were considered separately. Absence of enhancement at the site of malignancy after NAT was considered an imaging complete response while presence of enhancement was considered compatible with residual disease on both CEM RI and MRI. For breast MRI, the most delayed post-contrast sequence was used to document extent of residual enhancement. For CEM LEI, the presence of residual mass, asymmetry, or microcalcifications at the site of biopsy-proven cancer was considered compatible with residual disease. Posttreatment lesion size was based on the largest dimension seen in each modality. When combining CEM RI and LEI or MRI and CEM LEI, residual lesion size was based on the largest extent of any of the modalities.
Pathological Evaluation
Pathological evaluation of the surgical specimens was performed by a dedicated breast pathologist with 8 years of experience. Tumor histological type was based on pathological evaluation of pre-NAT biopsy specimens. Presence of residual invasive tumor and DCIS in surgical specimens was identified and their largest extent was measured. Only measurements of the tumor bed from the main excision were considered when establishing the extent of residual disease. Separated margins or re-excision specimens were not included in the measurement of residual tumor.
Statistical Analysis
Statistical analysis was performed by an expert (VS) utilizing R software version 3.6.3 (R Core Development Team). The results were expressed as medians and interquartile ranges (IQR) for continuous variables and proportions for categorical variables. The presence of residual invasive carcinoma and/or DCIS in surgical specimens was considered the reference standard for residual disease. Comparisons of sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy using breast MRI, CEM RI, and LEI for detection of residual disease were made. Sensitivity, specificity, and accuracy was compared using McNemar’s test while PPV and NPV was compared using Leisenring’s test. Multiple comparison adjustment was made using the Bonferroni procedure. Type I error rate (α) was set to 0.05. Correlation of lesion sizes between pathology and imaging modalities was also assessed using Kendall’s tau correlation coefficient. Bootstrapping was used to generate p values for the differences in the correlation coefficients.
RESULTS
Clinical Characteristics of the Study Sample
There were 110 participants with 115 malignant lesions included in the study, with a median age of 44 years (IQR, 37–54). There were 108 (94%) invasive ductal carcinomas and 7 (6%) invasive lobular carcinomas. The median pre-treatment tumor size on MRI was 4.0 cm (IQR, 2.7–6.1). Mastectomy was the treatment of choice for 50 lesions (43%) and lumpectomy was performed for 65 lesions (57%).
On MRI, 71/115 (62%) residual lesions were identified, including 28 mass enhancements, 23 non-mass enhancements, 15 mass and non-mass enhancements combined, and 5 enhancing foci. On CEM RI, there were 55/115 (48%) residual lesions, including 28 mass enhancements, 23 non-mass enhancements and 4 mass and non-mass enhancements combined. On CEM LEI, there were 75/115 (65%) residual lesions, including 40 calcified and 35 non-calcified lesions. On pathology, residual invasive disease and/or DCIS was detected in 83/115 (72%) lesions, while residual invasive disease alone was detected in 76/115 (66%) lesions. Participant characteristics are demonstrated in Table 2. No serious adverse events occurred during or after the MRI and CEM exams.
Table 2.
Patient characteristics
| Characteristics of patients | n | % | ||
|---|---|---|---|---|
| Total number of participants (100% women) | 110 | 100 | ||
| Total number of tumors | 115 | 100 | ||
| Patient median age = 44 years (IQR, 37–54) | ||||
| Menopausal status | ||||
| Pre-menopausal | 71 | 65 | ||
| Postmenopausal | 39 | 35 | ||
| Family history of breast cancer | ||||
| Positive | 56 | 51 | ||
| Negative | 54 | 49 | ||
| Tumor size (median cm, IQR) | ||||
| Pre-treatment (MRI) | 4.0 (2.7–6.1) | |||
| Post-treatment (Pathology) | 0.8 (0–2.2) | |||
| Histopathology of breast cancers | ||||
| Invasive ductal carcinoma | 108 | 94 | ||
| Invasive lobular carcinoma | 7 | 6 | ||
| Histopathological grade | ||||
| I | 4 | 3 | ||
| II | 30 | 26 | ||
| II/III | 18 | 16 | ||
| III | 63 | 55 | ||
| Receptor profile | ||||
| ER/PR+ Her2− | 57 | 49 | ||
| Her2+ | 33 | 29 | ||
| Triple negative | 25 | 22 | ||
| Post-neoadjuvant surgery type | ||||
| Lumpectomy | 65 | 57 | ||
| Mastectomy | 50 | 43 | ||
| Residual disease on pathology (invasive or DCIS) | ||||
| Residual disease | 83 | 72 | ||
| No residual disease | 32 | 28 | ||
| Residual invasive disease on pathology | ||||
| Residual invasive disease | 76 | 66 | ||
| No residual invasive disease | 39 | 34 | ||
Abbreviations: DCIS = Ductal carcinoma in situ; IQR = interquartile range
MRI vs. CEM RI
Among 83 participants with pathological residual disease, including invasive carcinoma and/or DCIS, MRI detected residual enhancement in 61/83 (73%, IQR 63–83) while CEM RI detected in 53/83 (64%, IQR 53–74). Among 32 participants with pCR, complete imaging response was identified on MRI in 22/32 (69%, IQR 50–84) and on CEM RI in 30/32 (94%, IQR 79–99). No statistically significant differences were identified in accuracy, sensitivity, specificity, PPV, and NPV between MRI and CEM RI in detecting residual disease after multiple comparison correction.
Among 76 participants with pathological residual invasive disease, MRI detected residual enhancement in 59/76 (78%, IQR 67–87) while CEM RI detected in 50/76 (66%, IQR 54–76). Among 39 participants without pathological residual invasive disease, complete imaging response was identified on MRI in 27/39 (69%, IQR 52–83) and on CEM RI in 34/39 (87%, IQR 73–96). No statistically significant differences were identified in accuracy, sensitivity, specificity, PPV, and NPV between MRI and CEM RI in detecting residual invasive disease. Tables 3 and 4 summarize the results obtained with MRI and CEM RI in detecting residual disease, and residual invasive disease, respectively. Table 5 demonstrates the comparisons of accuracy, sensitivity, specificity, PPV, and NPV between MRI and CEM RI.. Fig. 2 demonstrates a case of residual disease detected on both MRI and CEM. Fig. 3 demonstrates a case of false-positive post-treatment MRI and CEM in a patient with pCR.
Table 3.
Diagnostic performance of each imaging modality in the detection of residual disease (invasive disease and/or DCIS)
| Residual Disease on Imaging Modality | Pathological Residual Disease | No Pathological Residual Disease | Accuracy (95% CI) |
Sensitivity (95% CI) |
Specificity (95% CI) |
PPV (95% CI) |
NPV (95% CI) |
|
|---|---|---|---|---|---|---|---|---|
| MRI | positive | 61 | 10 | 72% (63–80) |
73% (63–83) |
69% (50–84) |
86% (76–93) |
50% (35–65) |
| negative | 22 | 22 | ||||||
| CEM RI | positive | 53 | 2 | 72% (63–80) |
64% (53–74) |
94% (79–99) |
96% (88–100) |
50% (37–63) |
| negative | 30 | 30 | ||||||
| MRI + CEM LEI | positive | 73 | 17 | 77% (68–84) |
88% (79–94) |
47% (29–65) |
81% (72–89) |
60% (39–79) |
| negative | 10 | 15 | ||||||
| CEM RI + LEI | positive | 67 | 16 | 72% (63–80) |
81% (71–89) |
50% (32–68) |
81% (71–89) |
50% (32–68) |
| negative | 16 | 16 | ||||||
Abbreviations: CEM = contrast-enhanced mammography; CI = confidence interval; DCIS = ductal carcinoma in situ; LEI = low-energy images; MRI = magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value; RI = recombined images
Table 4.
Diagnostic performance of each imaging modality in the detection of residual invasive disease
| Residual Invasive Disease on Imaging Modality | Pathological Residual Invasive Disease | No Pathological Residual Invasive Disease | Accuracy (95% CI) |
Sensitivity (95% CI) |
Specificity (95% CI) |
PPV (95% CI) |
NPV (95% CI) |
|
|---|---|---|---|---|---|---|---|---|
| MRI | positive | 59 | 12 | 75% (66–83) |
78% (67–87) |
69% (52–83) |
83% (72–91) |
61% (46–76) |
| negative | 17 | 27 | ||||||
| CEM RI | positive | 50 | 5 | 73% (64–81) |
66% (54–76) |
87% (73–96) |
91% (80–97) |
57% (43–70) |
| negative | 26 | 34 | ||||||
| MRI + CEM LEI | positive | 66 | 24 | 70% (61–79) |
87% (77–94) |
38% (23–55) |
73% (63–82) |
60% (39–79) |
| negative | 10 | 15 | ||||||
| CEM RI + LEI | positive | 61 | 22 | 68% (59–76) |
80% (70–88) |
44% (28–60) |
74% (63–83) |
53% (35–71) |
| negative | 15 | 17 | ||||||
Abbreviations: CEM = contrast-enhanced mammography; CI = confidence interval; LEI = low-energy images; MRI = magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value; RI = recombined images
Table 5.
Performance comparison between imaging modalities (α < 0.001).
| Residual Disease | Residual Invasive Disease | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MRI | CEM RI | p-value | MRI + CEM LEI | CEM RI + LEI | p-value | MRI | CEM RI | p-value | MRI + CEM LEI | CEM RI + LEI | p-value | |
| Accuracy | 72% | 72% | > 0.99 | 77% | 72% | > 0.99 | 75% | 73% | > 0.99 | 70% | 68% | > 0.99 |
| Sensitivity | 73% | 64% | > 0.99 | 88% | 81% | > 0.99 | 78% | 66% | 0.70 | 87% | 80% | > 0.99 |
| Specificity | 69% | 94% | 0.47 | 47% | 50% | > 0.99 | 69% | 87% | > 0.99 | 38% | 44% | > 0.99 |
| PPV | 86% | 96% | 0.28 | 81% | 81% | > 0.99 | 83% | 91% | > 0.99 | 73% | 74% | > 0.99 |
| NPV | 50% | 50% | > 0.99 | 60% | 50% | > 0.99 | 61% | 57% | > 0.99 | 60% | 53% | > 0.99 |
Abbreviations: CEM = contrast-enhanced mammography; LEI = low-energy images; MRI = magnetic resonance imaging; NPV = negative predictive value; PPV = positive predictive value; RI = recombined images
Fig. 2.
33-year-old woman with left breast cancer (arrows) seen on pre-treatment MRI subtraction image (a). After neoadjuvant chemotherapy, MRI subtraction image demonstrates no significant change in tumor size (b). Contrast-enhanced mammography (CEM) low-energy images on mediolateral oblique (MLO) view (c) and craniocaudal (CC) view (d) shows a residual mass. CEM recombined images demonstrate residual enhancement on MLO and CC views (e, f) with good correlation to the post-neoadjuvant MRI.
Fig. 3.
39-year-old woman with right breast cancer (arrows) seen on pre-treatment MRI subtraction image (a). Additional non-mass enhancement was biopsy-proven benign (arrowheads). After neoadjuvant chemotherapy, the MRI subtraction image demonstrates a reduction in the size of the cancer with a residual enhancing mass (b). No evident residual mass in identified on contrast-enhanced mammography (CEM) low-energy images (c, d). A residual enhancing mass is identified on CEM recombined images (e, f). No residual disease was identified on pathology.
Combined MRI and CEM LEI vs. CEM
For the detection of residual disease, combining breast MRI and CEM LEI identified a residual lesion in 73/83 (88%, IQR 79–94) participants while CEM, including both RI and LEI, detected a residual lesion in 67/83 (81%, IQR 71–89). Among 32 participants with pCR, no residual lesion was identified on MRI combined with LEI in 15/32 (47%, IQR 29–65) and on CEM in 16/32 (50%, IQR 32–68). No statistically significant differences were identified in accuracy, sensitivity, specificity, PPV, and NPV between MRI combined with CEM LEI and CEM RI combined with CEM LEI in detecting residual disease after multiple comparison correction.
For the detection of residual invasive disease only, combining MRI with CEM LEI identified a residual lesion in 66/76 (87%, IQR 77–94) participants while on CEM a residual lesion was identified in 61/76 (80%, IQR 70–88). Among participants with no residual invasive disease, no residual lesion was detected on MRI combined with LEI in 15/39 (38%, IQR 23–55) and on CEM in 17/39 (44%, IQR 28–60). No statistically significant differences were identified in accuracy, sensitivity, specificity, PPV, and NPV between MRI combined with CEM LEI and CEM RI combined with CEM LEI in detecting residual invasive disease.
The presence of calcifications on CEM LEI was responsible for the identification of 9/83 (11%) residual lesions that were not detected on MRI and 11/83 (13%) not detected on CEM RI in patients with pathological residual disease and in 5/76 (7%) on MRI and 8/76 (11%) on CEM RI in patients with pathological residual invasive disease. On the other hand, calcifications were responsible for false positive CEM LEI in 7/32 (22%) lesions that were negative on MRI and 12/32 (38%) negative on CEM RI in patients with no pathological residual disease and in 11/39 (28%) lesions negative on MRI and in 15/39 (38%) lesions negative on CEM RI in patients with no pathological residual invasive disease.
Tables 3 and 4 summarize the results obtained with MRI combined with CEM LEI and CEM RI combined with CEM LEI in detecting residual disease and residual invasive disease, respectively. Table 5 demonstrates the comparisons of accuracy, sensitivity, specificity, PPV and NPV between MRI combined with CEM LEI and CEM RI combined with CEM LEI.
Size Correlation
On pathology, median residual invasive tumor size was 1.2 cm (IQR, 0.5–2.1) while median DCIS size was 1.7 cm (IQR, 0.4–2.3). The imaging modalities tended to overestimate residual disease size. Median residual disease size was 1.9 cm on MRI (IQR, 1.2–3.3), 2.0 cm on CEM RI (IQR, 1.2–3.3), and 2.2 cm on CEM LEI (IQR, 1.5–4.1). Overestimation was higher on combination of MRI with CEM LEI, 2.6 cm (IQR, 1.7–4.9) and on combination of CEM LEI with RI, 2.5 cm (IQR, 1.5–4.4).
In patients with residual disease after NAT, residual mass enhancement on MRI had a median size of 1.8 cm (IQR, 1.0–2.1) while other types of enhancement had a median size of 1.9 cm (IQR, 1.3–3.6). On CEM RI, median mass enhancement size was 1.6 cm (IQR, 1.0–2.2) while other enhancements measured 3.2 cm (IQR, 1.7–5.8). On CEM LEI, calcified residual lesions had a median size of 3.6 cm (IQR 2.1–6.2) while non-calcified lesions had a median size of 1.6 cm (IQR, 1.2–2.1).
Size correlation between MRI and CEM RI was moderate for residual disease (τ = 0.41 vs. 0.47) and for residual invasive disease (τ = 0.45 vs. 0.34) without a significant statistical difference (p ≥ 0.99 and p = 0.52, respectively). CEM LEI showed only slight correlation (τ = 0.17) which was statistically significantly different compared with MRI (p = 0.01) but not with CEM RI (0.54). No significant difference was observed comparing MRI combined with CEM LEI with CEM RI and LEI for residual disease (τ = 0.36 vs. τ = 0.33, p ≥ 0.99) or for residual invasive disease (τ = 0.29 vs. τ = 0.22, p ≥ 0.99). A size correlation comparison between CEM and MRI is demonstrated in Table 6. Correlation between residual disease size on pathology and imaging modalities is demonstrated in Figure 4. Correlation between residual invasive disease size on pathology and imaging modalities is demonstrated in Figure 5.
Table 6.
Agreement correlation (tau) between pathological residual disease (invasive disease and/or DCIS) and residual invasive disease size and imaging modalities
| Imaging Modality | Residual Disease | p-value | Residual Invasive Disease | p-value |
|---|---|---|---|---|
| MRI | 0.41 | > 0.99 | 0.45 | 0.52 |
| CEM RI | 0.47 | 0.34 | ||
| MRI + CEM LEI | 0.36 | > 0.99 | 0.29 | > 0.99 |
| CEM RI + LEI | 0.33 | 0.22 |
Abbreviations: CEM = contrast-enhanced mammography; DCIS = ductal carcinoma in situ; LEI = low-energy images; MRI = magnetic resonance imaging; RI = recombined images
Fig. 4.
Bland-Altman plots for size correlations between pathological residual disease and (a) MRI, (b) contrast-enhanced mammography recombined images, (c) contrast-enhanced mammography low-energy images.
Fig. 5.
Bland-Altman plots for size correlations between pathological residual invasive disease and (a) MRI, (b) contrast-enhanced mammography recombined images, (c) contrast-enhanced mammography low-energy images.
DISCUSSION
In this study, we investigated the use of contrast-enhanced mammography (CEM) as an alternative to MRI for the evaluation of residual disease after NAT. We have demonstrated that the presence of residual disease can be adequately assessed with CEM or breast MRI. CEM and breast MRI were similarly accurate in detecting residual disease, which included the combination of invasive disease and DCIS or invasive disease alone. All imaging modalities tended to overestimate residual lesion size. Size correlation with pathological residual disease was moderate for both CEM and breast MRI.
Our study showed identical accuracy (72%) for both CEM RI and breast MRI in the detection of pathological residual disease, invasive disease and/or DCIS; the accuracies were similar when considering only residual invasive disease (73%. vs. 75%, respectively). MRI showed higher sensitivity than CEM RI for both residual disease (73% vs 64%) and for residual invasive disease (78% vs. 66%), while specificity was lower for MRI (69% vs. 94% and 69% vs. 87%); however, multiple comparison analysis showed these differences to be non-significant.
Current standard practice utilizes breast MRI combined with FFDM to evaluate presence and extent of residual disease [17]. When combined with LEI, the respective sensitivities of MRI and CEM improved to 88% and 81% for residual disease and 87% and 80% for residual invasive disease; however, the combination lowered the respective specificities to 47% and 50% for residual disease and 38% and 44% for residual invasive disease. The increase in sensitivity and decrease in specificity with the use of CEM LEI can be attributed to the detection of calcifications, which can be seen in patients with false negative MRI and CEM RI but may be present in false positive CEM LEI. In a previous study with 33 participants by Barra et al., MRI showed higher sensitivity (92% vs 76%) and lower specificity (75% vs. 88%) than CEM for the detection of residual disease [22]. In contrast, in a study by Iotti et al., MRI failed to identify residual disease in 15 out of 38 participants and was outperformed by CEM, which failed to identify residual disease in only 6 patients [23]. Finally, Patel et al. found similar sensitivities and specificities for MRI and CEM in detecting residual disease [24]. The limited population of these studies likely contributed to the mixed results. Our study was the largest prospective sample to date. Although there was a trend favoring higher sensitivity and lower specificity for MRI, no statistically significant difference was identified between the two modalities. Future studies with larger sample sizes are necessary to further evaluate differences in performance between these two modalities.
It has been proposed that inflammatory response at the tumor bed after NAT can cause false-positive residual enhancement, which can lead to overestimation of extent of residual disease. On the other hand, reduced vascular inflow at the tumor bed and disruption of the carcinomatous tissue by necrosis or fibrosis caused by the medications administered during NAT may explain underestimation or false-negative imaging studies [25–27]. Lesion size was overestimated on imaging modalities in our study, in agreement with previous studies [9, 22]. Correlation of lesion size between pathology and MRI or CEM RI was moderate. Residual disease size, invasive disease and/or DCIS, had the highest size correlation with CEM RI (τ = 0.47), while for residual invasive disease alone the highest size correlation was achieved with MRI (τ = 0.45), but no statistically significant difference was observed between MRI and CEM RI. Size correlation was lower with CEM LEI, as previously observed by Steinhof-Radwanska et al [28]. The presence of calcifications is a major factor for overestimation on CEM LEI, as demonstrated in our study.
CEM may be preferred to breast MRI in numerous scenarios. Patients may have contraindications to breast MRI, e.g., presence of ferromagnetic implanted devices that in some cases cannot undergo MRI. Some patients are claustrophobic and cannot tolerate MRI. Post-NAT CEM may be preferred in patients who were initially diagnosed by a screening or diagnostic CEM, facilitating comparison between pre- and post-treatment studies. Furthermore, costs related to CEM can be as low as 10% of those of a breast MRI study and may appeal to health systems with restricted budgets [29, 30]. Additionally, patients who undergo breast MRI still need a mammogram for adequate post-NAT evaluation and preoperative planning, whereas CEM offers RI and LEI in a single study.
On the other hand, CEM has limitations that should be considered in the post-NAT setting. The limited field-of-view and lack of 3-dimensional capability can jeopardize the detection of infiltration of surrounding structures by breast tumors such as the pectoralis muscle and the chest wall. Moreover, allergy to iodine-based contrast agents is not rare and patients with a known history of allergy should not undergo CEM [31].
Criteria for pCR vary between institutions. In our institution, pCR is considered as absence of invasive carcinoma and DCIS. Although patients with residual DCIS only have better prognosis than patients with residual invasive disease, absence of both DCIS and invasive disease has superior disease-free survival than residual DCIS [32, 33]. No significant differences in performance of MRI and CEM were identified regardless of which criteria for pCR is used.
There are some limitations to our study. First, pre-treatment CEM was not performed; therefore, LEIs were compared to pre-treatment mammogram while RIs were correlated with enhancement on pre-treatment MRI. Second, because each imaging study was read by only a single breast imaging radiologist, we could not evaluate inter-reader agreement. Last, only the largest dimensions of each lesion on both imaging and pathological evaluation were considered. Breast cancer may present as multiple and scattered foci in a single tumor bed, which can make size comparison difficult, likely reflected in the only moderate agreement in size between imaging modalities and pathology.
In conclusion, contrast-enhanced mammography is an acceptable alternative to breast MRI for the detection of residual disease after neoadjuvant treatment. Contrast-enhanced mammography has comparable accuracy to breast MRI combined with low-energy images and may be preferred in a number of scenarios. Although there were no significant statistical differences in accuracy, sensitivity, specificity, positive predictive value, or negative predictive value between the two modalities, future studies with larger sample sizes are necessary to validate our findings.
ACKNOWLEDGEMENTS
The authors thank Garon Scott, BA, and Joanne Chin, MFA, ELS, for their help in editing this manuscript.
Funding
This study was funded by Hologic, Inc., and partially by the National Institutes of Health/National Cancer Institute Cancer Center Support Grant P30 CA008748.
ABBREVIATIONS
- CEM
contrast-enhanced mammography
- DCIS
ductal carcinoma in situ
- FFDM
full-field digital mammography
- IQR
interquartile range
- LEI
low-energy images
- NAT
neoadjuvant treatment
- NPV
negative predictive value
- PACS
Picture Archiving and Communication System
- pCR
pathologic complete response
- PPV
positive predictive value
- RI
recombined images
Footnotes
Conflict of Interest
Maxine S Jochelson has received an honorarium from GE for speaking, and an honorarium for speaking at the Lynn Sage Breast Cancer Symposium and at MD Anderson. Janice S Sung has received research grants from Hologic and GE. The remaining authors have no relevant financial or non-financial interests to disclose.
Ethical approval
This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Memorial Sloan Kettering Cancer Center (IRB# 17–045). Study start date: February 28, 2017. ClinicalTrials.gov Identifier: NCT03070340.
Consent to participate
Written informed consent was obtained from all individual participants included in the study.
Data availability
Data generated or analyzed during the study are available from the corresponding author by request.
REFERENCES
- 1.Gluskin J, Rossi Saccarelli C, Avendano D, Marino MA, Bitencourt AGV, Pilewskie M, et al. Contrast-Enhanced Mammography for Screening Women after Breast Conserving Surgery. Cancers (Basel). 2020;12(12). doi: 10.3390/cancers12123495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sung JS, Lebron L, Keating D, D’Alessio D, Comstock CE, Lee CH, et al. Performance of Dual-Energy Contrast-enhanced Digital Mammography for Screening Women at Increased Risk of Breast Cancer. Radiology. 2019;293(1):81–8. doi: 10.1148/radiol.2019182660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cozzi A, Magni V, Zanardo M, Schiaffino S, Sardanelli F. Contrast-enhanced Mammography: A Systematic Review and Meta-Analysis of Diagnostic Performance. Radiology. 2022;302(3):568–81. doi: 10.1148/radiol.211412. [DOI] [PubMed] [Google Scholar]
- 4.Fallenberg EM, Schmitzberger FF, Amer H, Ingold-Heppner B, Balleyguier C, Diekmann F, et al. Contrast-enhanced spectral mammography vs. mammography and MRI - clinical performance in a multi-reader evaluation. Eur Radiol. 2017;27(7):2752–64. doi: 10.1007/s00330-016-4650-6. [DOI] [PubMed] [Google Scholar]
- 5.Jochelson MS, Dershaw DD, Sung JS, Heerdt AS, Thornton C, Moskowitz CS, et al. Bilateral contrast-enhanced dual-energy digital mammography: feasibility and comparison with conventional digital mammography and MR imaging in women with known breast carcinoma. Radiology. 2013;266(3):743–51. doi: 10.1148/radiol.12121084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sudhir R, Sannapareddy K, Potlapalli A, Krishnamurthy PB, Buddha S, Koppula V. Diagnostic accuracy of contrast-enhanced digital mammography in breast cancer detection in comparison to tomosynthesis, synthetic 2D mammography and tomosynthesis combined with ultrasound in women with dense breast. Br J Radiol. 2021;94(1118):20201046. doi: 10.1259/bjr.20201046. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Francescone MA, Jochelson MS, Dershaw DD, Sung JS, Hughes MC, Zheng J, et al. Low energy mammogram obtained in contrast-enhanced digital mammography (CEDM) is comparable to routine full-field digital mammography (FFDM). Eur J Radiol. 2014;83(8):1350–5. doi: 10.1016/j.ejrad.2014.05.015. [DOI] [PubMed] [Google Scholar]
- 8.Fallenberg EM, Dromain C, Diekmann F, Engelken F, Krohn M, Singh JM, et al. Contrast-enhanced spectral mammography versus MRI: Initial results in the detection of breast cancer and assessment of tumour size. Eur Radiol. 2014;24(1):256–64. doi: 10.1007/s00330-013-3007-7. [DOI] [PubMed] [Google Scholar]
- 9.Sumkin JH, Berg WA, Carter GJ, Bandos AI, Chough DM, Ganott MA, et al. Diagnostic Performance of MRI, Molecular Breast Imaging, and Contrast-enhanced Mammography in Women with Newly Diagnosed Breast Cancer. Radiology. 2019;293(3):531–40. doi: 10.1148/radiol.2019190887. [DOI] [PubMed] [Google Scholar]
- 10.Lee-Felker SA, Tekchandani L, Thomas M, Gupta E, Andrews-Tang D, Roth A, et al. Newly Diagnosed Breast Cancer: Comparison of Contrast-enhanced Spectral Mammography and Breast MR Imaging in the Evaluation of Extent of Disease. Radiology. 2017;285(2):389–400. doi: 10.1148/radiol.2017161592. [DOI] [PubMed] [Google Scholar]
- 11.Korde LA, Somerfield MR, Carey LA, Crews JR, Denduluri N, Hwang ES, et al. Neoadjuvant Chemotherapy, Endocrine Therapy, and Targeted Therapy for Breast Cancer: ASCO Guideline. J Clin Oncol. 2021;39(13):1485–505. doi: 10.1200/JCO.20.03399. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Leon-Ferre RA, Hieken TJ, Boughey JC. The Landmark Series: Neoadjuvant Chemotherapy for Triple-Negative and HER2-Positive Breast Cancer. Ann Surg Oncol. 2021;28(4):2111–9. doi: 10.1245/s10434-020-09480-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pease AM, Riba LA, Gruner RA, Tung NM, James TA. Oncotype DX((R)) Recurrence Score as a Predictor of Response to Neoadjuvant Chemotherapy. Ann Surg Oncol. 2019;26(2):366–71. doi: 10.1245/s10434-018-07107-8. [DOI] [PubMed] [Google Scholar]
- 14.Mamounas EP. Impact of neoadjuvant chemotherapy on locoregional surgical treatment of breast cancer. Ann Surg Oncol. 2015;22(5):1425–33. doi: 10.1245/s10434-015-4406-6. [DOI] [PubMed] [Google Scholar]
- 15.(EBCTCG) EBCTCG. Long-term outcomes for neoadjuvant versus adjuvant chemotherapy in early breast cancer: meta-analysis of individual patient data from ten randomised trials. Lancet Oncol. 2018;19(1):27–39. doi: 10.1016/S1470-2045(17)30777-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Spring LM, Fell G, Arfe A, Sharma C, Greenup R, Reynolds KL, et al. Pathologic Complete Response after Neoadjuvant Chemotherapy and Impact on Breast Cancer Recurrence and Survival: A Comprehensive Meta-analysis. Clin Cancer Res. 2020;26(12):2838–48. doi: 10.1158/1078-0432.CCR-19-3492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Jochelson MS, Lampen-Sachar K, Gibbons G, Dang C, Lake D, Morris EA, et al. Do MRI and mammography reliably identify candidates for breast conservation after neoadjuvant chemotherapy? Ann Surg Oncol. 2015;22(5):1490–5. doi: 10.1245/s10434-015-4502-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Feliciano Y, Mamtani A, Morrow M, Stempel MM, Patil S, Jochelson MS. Do Calcifications Seen on Mammography After Neoadjuvant Chemotherapy for Breast Cancer Always Need to Be Excised? Ann Surg Oncol. 2017;24(6):1492–8. doi: 10.1245/s10434-016-5741-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Marinovich ML, Houssami N, Macaskill P, Sardanelli F, Irwig L, Mamounas EP, et al. Meta-analysis of magnetic resonance imaging in detecting residual breast cancer after neoadjuvant therapy. J Natl Cancer Inst. 2013;105(5):321–33. doi: 10.1093/jnci/djs528. [DOI] [PubMed] [Google Scholar]
- 20.Patel BK, Garza SA, Eversman S, Lopez-Alvarez Y, Kosiorek H, Pockaj BA. Assessing tumor extent on contrast-enhanced spectral mammography versus full-field digital mammography and ultrasound. Clin Imaging. 2017;46:78–84. doi: 10.1016/j.clinimag.2017.07.001. [DOI] [PubMed] [Google Scholar]
- 21.Bicchierai G, Tonelli P, Piacenti A, De Benedetto D, Boeri C, Vanzi E, et al. Evaluation of contrast-enhanced digital mammography (CEDM) in the preoperative staging of breast cancer: Large-scale single-center experience. Breast J. 2020;26(7):1276–83. doi: 10.1111/tbj.13766. [DOI] [PubMed] [Google Scholar]
- 22.Barra FR, Sobrinho AB, Barra RR, Magalhães MT, Aguiar LR, de Albuquerque GFL, et al. Contrast-Enhanced Mammography (CEM) for Detecting Residual Disease after Neoadjuvant Chemotherapy: A Comparison with Breast Magnetic Resonance Imaging (MRI). Biomed Res Int. 2018;2018:8531916. doi: 10.1155/2018/8531916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Iotti V, Ravaioli S, Vacondio R, Coriani C, Caffarri S, Sghedoni R, et al. Contrast-enhanced spectral mammography in neoadjuvant chemotherapy monitoring: a comparison with breast magnetic resonance imaging. Breast Cancer Res. 2017;19(1):106. doi: 10.1186/s13058-017-0899-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Patel BK, Hilal T, Covington M, Zhang N, Kosiorek HE, Lobbes M, et al. Contrast-Enhanced Spectral Mammography is Comparable to MRI in the Assessment of Residual Breast Cancer Following Neoadjuvant Systemic Therapy. Ann Surg Oncol. 2018;25(5):1350–6. doi: 10.1245/s10434-018-6413-x. [DOI] [PubMed] [Google Scholar]
- 25.Wasser K, Sinn HP, Fink C, Klein SK, Junkermann H, Lüdemann HP, et al. Accuracy of tumor size measurement in breast cancer using MRI is influenced by histological regression induced by neoadjuvant chemotherapy. Eur Radiol. 2003;13(6):1213–23. doi: 10.1007/s00330-002-1730-6. [DOI] [PubMed] [Google Scholar]
- 26.Denis F, Desbiez-Bourcier AV, Chapiron C, Arbion F, Body G, Brunereau L. Contrast enhanced magnetic resonance imaging underestimates residual disease following neoadjuvant docetaxel based chemotherapy for breast cancer. Eur J Surg Oncol. 2004;30(10):1069–76. doi: 10.1016/j.ejso.2004.07.024. [DOI] [PubMed] [Google Scholar]
- 27.Marinovich ML, Macaskill P, Irwig L, Sardanelli F, von Minckwitz G, Mamounas E, et al. Meta-analysis of agreement between MRI and pathologic breast tumour size after neoadjuvant chemotherapy. Br J Cancer. 2013;109(6):1528–36. doi: 10.1038/bjc.2013.473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Steinhof-Radwańska K, Grażyńska A, Lorek A, Gisterek I, Barczyk-Gutowska A, Bobola A, et al. Contrast-Enhanced Spectral Mammography Assessment of Patients Treated with Neoadjuvant Chemotherapy for Breast Cancer. Curr Oncol. 2021;28(5):3448–62. doi: 10.3390/curroncol28050298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Son D, Phillips J, Mehta TS, Mehta R, Brook A, Dialani VM. Patient preferences regarding use of contrast-enhanced imaging for breast cancer screening. Acad Radiol. 2021. doi: 10.1016/j.acra.2021.03.003. [DOI] [PubMed] [Google Scholar]
- 30.Sogani J, Mango VL, Keating D, Sung JS, Jochelson MS. Contrast-enhanced mammography: past, present, and future. Clin Imaging. 2021;69:269–79. doi: 10.1016/j.clinimag.2020.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Covington MF, Pizzitola VJ, Lorans R, Pockaj BA, Northfelt DW, Appleton CM, et al. The Future of Contrast-Enhanced Mammography. AJR Am J Roentgenol. 2018;210(2):292–300. doi: 10.2214/AJR.17.18749. [DOI] [PubMed] [Google Scholar]
- 32.von Minckwitz G, Untch M, Blohmer JU, Costa SD, Eidtmann H, Fasching PA, et al. Definition and impact of pathologic complete response on prognosis after neoadjuvant chemotherapy in various intrinsic breast cancer subtypes. J Clin Oncol. 2012;30(15):1796–804. doi: 10.1200/JCO.2011.38.8595. [DOI] [PubMed] [Google Scholar]
- 33.Provenzano E, Bossuyt V, Viale G, Cameron D, Badve S, Denkert C, et al. Standardization of pathologic evaluation and reporting of postneoadjuvant specimens in clinical trials of breast cancer: recommendations from an international working group. Mod Pathol. 2015;28(9):1185–201. doi: 10.1038/modpathol.2015.74. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data generated or analyzed during the study are available from the corresponding author by request.