Background

Breast cancer is the most common cancer and the leading cause of cancer-related death in women [1]. The prognosis of breast cancer is impacted by several factors including tumor grade, axillary lymph node status and molecular subtypes all of which are vital in treatment planning with molecular subtypes being the most crucial [2,3,4]. Human epidermal growth factor receptor 2 (HER2) overexpression is seen in one out of five invasive breast cancer cases, which is linked to a lower survival rate [5, 6]. Published clinical trial studies reported that HER2-positive (HER2+) are a diverse group of cancers [7, 8]. Among HER2+ patients, two distinct hormone receptor (HR) subtypes can be defined: HR-positive (HR+) and HR-negative (HR-) each of which with unique therapeutic response, recurrence rates, and survival pattern with HR-/HER2+ tumors more likely to cause death within the first five years of diagnosis, a higher tendency to metastasize to brain compared to bone, and a more favorable response to neoadjuvant chemotherapy [9,10,11].

Identification of the imaging features of different molecular types of breast cancer is crucial as it can potentially reflect its biological features [4]. Several studies have described the MRI and mammographic characteristics of HER2+ breast tumors. Fine pleomorphic and linear branching microcalcifications on mammograms while irregular tumor margins, tendency to multifocality and wash-out kinetics on MRI were characteristic of HER2+ tumors [12,13,14,15]. Contrast-enhanced spectral mammography (CESM) is a novel diagnostic imaging tool that utilizes iodinated contrast media using digital mammography to provide low-energy and subtraction images reflecting tumor vascularity, and displaying morphological and anatomical features of neoplastic breast lesions [16, 17]. According to previous reports, CESM offers diagnostic accuracy that is comparable to contrast-enhanced MRI (CE-MRI) [18, 19]. Wide availability, shorter examination time and lower cost are among other CESM advantages [20, 21].

While several studies reported the CE-MRI imaging features of molecular subtypes of breast cancer [22,23,24], studies on CESM focused mainly on differentiating benign from malignant breast cancers, or monitoring treatment response in patients receiving neoadjuvant chemotherapy [25,26,27,28,29]. We believe that CESM can provide morphological and anatomical information that are comparable to CE-MRI, which reflects breast cancer molecular subtypes. Therefore, the purpose of this retrospective study was to describe the CESM features of HER2+ breast cancers according to hormone receptor status and to assess whether specific imaging features can distinguish between HER2+/HR+ and HER2+/HR− breast cancers that can potentially be useful for treatment planning in patients with HER2+ breast cancer.

Methods

Our retrospective single-institution study included all consecutive female patients who had pathologic diagnosis of HER2+ invasive breast cancers. Data were collected from January 2017 to December 2021. The study was approved by our institution Ethics Committee before data collection. A waiver of informed consent was granted due to retrospective study design. The data were collected using our hospital electronic medical records. Images were retrieved from our Radiology Department Picture Archiving and Communication System (PACS). Inclusion criteria included female patient with age more than 18 years, HER2+ breast cancer diagnosis based on core needle biopsy and immunohistochemistry and completion of standard mammogram with CESM examination. Exclusion criteria include contrast media reaction, renal impairment, pregnancy as well as neoadjuvant chemotherapy administration, biopsy, or surgery prior to CESM examination.

CESM examination

All mammographic examinations were performed at the institution’s Radiology Department using GE Healthcare machine (Chalfont St. Giles, UK) allowing dual-energy CESM acquisitions with some specific hardware and software enhancements for image processing and acquisition.

Unlike conventional mammography, the X-ray spectrum was tailored to produce energies just above the K-edge of iodine (33.2 keV) to increase the sensitivity to low concentration of iodine to optimize the visualization of iodine. Typically, for a 50% glandular and 5-cm-thick breast, exposure times were around 1 s and 3 s for low -and high-energy images, respectively.

Standard mammography was done followed by an intravenous administration of 1.5 ml/kg body weighted of a non-ionic contrast material (Omnipaque) manually at one shot. Two sets of low- and high-energy exposures were taken: one 2 min after the contrast medium administration with the breast compressed in a mediolateral oblique (MLO) view and the second 4 min after the contrast medium administration with the breast compressed in the craniocaudal (CC) view. Specific image processing for both exposures was then done with the final subtraction images (one in the MLO and one in the CC view) highlighting the mass and removing the noise created by the fibroglandular tissue. The compression applied after the injection minimized the motion blur keeping the image details seen with the standard mammography. The average examination time was 10 min. After the examination, the patients were monitored for around 30 min for any potential contrast medium adverse reactions.

Image interpretation

One of the participating radiologists interpreted standard mammography and CESM studies in accordance with the Breast Imaging Reporting and Data System (BI-RADS) lexicon designed by the American College of Radiology [30]. In case of disagreement, the case was discussed, and a joint consensus was reached. In patients with multiple masses, the largest lesion was considered for assessment. All radiologists were blinded to the hormonal status of the lesions. On mammograms, imaging presentation was described as mass, architectural changes, or asymmetry and the presence or absence of calcifications was recorded. A mass lesion was defined when a lesion larger than 5 mm in three dimensions was visible in both mammographic views. On CESM examination, the disease was recorded as unifocal or multifocal. The size, shape and margin of the lesion were recorded, and the pattern of enhancement was documented as homogeneous, heterogeneous, ring or non-mass enhancement (NME), which is defined as an area of enhancement not confined to a mass or a three-dimensional volume effect. Table 1 summarizes mammographic and CESM imaging criteria assessed in our study.

Table 1 Mammographic and CESM imaging features characteristics assessed in our study
Full size table

Histopathological assessment

Following surgery or image-guided core biopsy, all specimens were sent to our institution laboratory for histopathological confirmation. Specimens underwent immunohistochemistry to detect the levels of ER, PR and HER2 overexpression. Imaging findings were available during histopathologic assessment. HER2 expression on IHC was categorized according to the cell membrane staining pattern with grades 2+ and 3+ regarded as positive, whereas grades 1+ and 0 considered negative. Fluorescence in situ hybridization (FISH) was used to further analyze all the equivocal samples. In all cases, a FISH ratio larger than 2.2 or a HER2 gene copy greater than 6.0 was considered positive. Stained slides were examined by pathologists for nuclear ER or PR expression using the College of American Pathologists guidelines (≥ 1% cutoff value for positive) [31].

Statistical methods

Data were statistically described in terms of mean ± standard deviation (±SD), median and range, or frequencies (number of cases) and percentages when appropriate. Comparison of numerical variables between the study groups was done using Student t test for independent samples. For comparing categorical data, Chi-square (χ2) test was performed. Exact test was used instead when the expected frequency is less than 5. Accuracy of different findings in differentiating HR+ from HR− cases was represented using the terms sensitivity, specificity, +ve predictive value, −ve predictive value and overall accuracy. Two-sided p values less than 0.05 were considered statistically significant. IBM SPSS (Statistical Package for the Social Science; IBM Corp, Armonk, NY, USA) release 22 for Microsoft Windows was used for all statistical analyses. For accuracy calculations, the following formulas were used:

  • Sensitivity = T(+)ve ÷ [T(+)ve + F(−)ve].

  • Specificity = T(−)ve ÷ [T(−)ve + F(+)ve].

  • Positive predictive value = T(+)ve ÷ [T(+)ve + F(+)ve].

  • Negative predictive value = T(−)ve ÷ [T(−)ve + F(−)ve].

  • Overall accuracy = [T(+)ve + T(−)ve] ÷ all sample.

Results

Following exclusion criteria, 61 patients were included in our study. Twenty-nine cases (47.5%) were HER2+/HR+, and 32 cases (52.5%) were HER2+/HR−. Their ages ranged from 29 to 84 years with a mean of 52.34 ± 11.5 years. Twenty-six patients (42.6%) were < 50 years, while 35 patients (57.4%) were ≥ 50 years. HER2+/HR−. Cases were significantly associated with older age compared to HER2+/HR+ (P value 0.003) with 24 out of 35 (68.6%) patients ≥ 50 years seen in HR- group. Invasive duct carcinoma (IDC) was the most common histopathological type in our series seen in 58 cases (91.1%). Invasive lobular carcinoma (ILC) was seen in two cases (3.3%) and clear cell carcinoma in one case (1.6%). Twenty-eight out of 61 cases (45.9%) were histological grade III, 26 cases (42.6%) were grade II and seven cases (11.5%) were grade I. Compared to HER2+/HR+, HER2+/HR− tumors were strongly associated with higher histologic grade (P value < 0.001) with 22 out of 28 (78.6%) of high histological grade III tumors seen in HR− group in contrast to 23 out of 33 lesions (69.7%) of histological grade I and II in the HR+ group. Table 2 summarizes the demographic, pathologic and imaging features of HER+ tumors according to hormone receptor status.

Table 2 Demographic, pathologic and imaging features of HER + tumors according to hormone receptor status
Full size table

On mammography, 40 out of 61 (65.6%) cases in our series were categorized as mass evenly distributed between the HR− and HR+ groups, 10 cases (16.4%) as architectural changes and 11 cases (18.0%) as asymmetry. We did not find statistically significant difference between mammographic imaging presentations and the HER2+/HR+ and HER2+/HR− groups with P values of 0.788, 0.170 and 0.307 for mass, architectural changes and asymmetry, respectively. Twenty-one out of our 61 HER2+ cases (34.4%) exhibited calcifications on mammography: 11 cases (52.4%) in the HR+ and 10 cases (47.6%) in the HR− group. There was no statistically significant difference between the HR+ and HR− groups regarding the presence or distribution of calcifications with a P value of 0.602.

On CESM, the mean tumor size was 5.4 ± 2.3 cm and 5.9 ± 2.7 cm for HR− and HR+ groups, respectively. We did not find statistically significant difference in tumor size between the two groups with a P value of 0.541. Multifocal tumors were reported in 25 out 61 cases (41.0%). We found that HR− tumors were more likely to be multifocal compared to HR+ group with 18 out of 25 cases (72.0%) of multifocal tumors in our series seen in HR− group (P value 0.018) (Fig. 1). HR-tumors were more likely to be rounded or oval (P value 0.008) compared to HR+ lesions with 18 out of 24 (75%) rounded or oval tumors seen in the HR− group. Additionally, HR-lesions were more likely to have circumscribed margins (P value 0.004) and with associated NME (P value < 0.001) on CESM compared to HR+ group (Fig. 2). Lesions in the HR+ group, on the other hand, had a higher tendency for irregular shape (P value 0.008), spiculated outline (P value v0.004) and heterogeneous (P value 0.021) or ring enhancement (P value 0.046) (Figs. 3 and 4). Table 3 demonstrates CESM imaging features associated with HR negativity within HER2+ tumors.

Fig. 1
figure 1

75-year-old female patient with a pathologically proven left breast IDC (HER2+/HR−) a and b Full-field digital mammography in CC and MLO views. c and d CESM in CC and MLO views showing a partially defined roughly oval-shaped mass lesion with spiculated outline at the retro-aerolar region of the left breast eliciting intense heterogeneous enhancement with smaller mass lesions at the deep retro-aerolar region indicating a multifocal neoplasm

Full size image
Fig. 2
figure 2

64-year-old female patient with a pathologically proven right breast IDC (HER2+/HR−) a and b full-field digital mammography in CC and MLO views showing a large partially circumscribed dense lesion with an adjacent smaller nodule and overlying suspicious amorphous clustered microcalcifications. c and d CESM in CC and MLO views showing an intense heterogeneous enhancement of the large mass lesion at the upper outer quadrant of the right breast associated with non-mass enhancement with segmental distribution

Full size image
Fig. 3
figure 3

68-year-old female patient with a pathologically proven left breast IDC (HER2+/HR+) a full-field digital mammography in CC view and b CESM in CC view showing left breast outer quadrant irregular shaped hyperdense mass lesion with spiculated outline eliciting intense heterogeneous enhancement

Full size image
Fig. 4
figure 4

52-year-old female patient with a pathologically proven right breast IDC (HER2+/HR+: a full-field digital mammography in CC view and b CESM in CC view showing an irregular shaped hyperdense mass lesion with spiculated outline at the retro-areolar region eliciting intense heterogeneous ring enhancement

Full size image
Table 3 CESM imaging features associated with HR negativity within HER2+ tumors
Full size table

Assessment of the sensitivity and accuracy of mammographic and CESM imaging features revealed that the accuracy and sensitivity of NME in differentiating between HR- and HR+HER2+ neoplasms was high at 70.5% and 96.6% respectively. The imaging features with the highest sensitivity for HR+ tumors were irregular mass shape and spiculated margins, with a value of 79.3%. The accuracy of the mammographic imaging features was low at 52.5%, 59.0%, 45.9% and 54.1% for mass, architectural changes, asymmetry, and calcifications respectively.

Discussion

HER2 overexpression accounts for 20 out of 100 breast cancer cases, and it is associated with lower survival rates [5, 6]. In the last two decades, since the molecular subtyping has become the standard of care in breast cancer management, the role of imaging has been to attempt to find specific imaging features that correlate to certain molecular subtypes with the intention to improve patient survival through prompt diagnosis and treatment [4]. CESM is well tolerated by patients, and its use in clinical practice has increased in the last few years [32]. Published studies have concluded that MRI and CESM have similar accuracy for breast cancer detection [33]. In this study, we demonstrated that HER2+ tumors have different demographic, pathologic and imaging features according to hormonal status, and these features were more evident on CESM compared to mammography.

In the current study, HER2+/HR− tumors were significantly associated with older age and higher histological grade compared to HR+ group with P values of 0.003 and < 0.001, respectively, similar to previous investigators [34, 35]. We also reported no statistically significant difference in tumor size between the HR− and HR+ groups as found in our study similar to the study by Song et al. [34].

In this study, the mammographic imaging features were not statistically significant in differentiating between HER2+/HR− and HER2+/HR+ groups similar to previous investigators who did not find significant difference in mammographic imaging presentations as well as the distribution and morphology of calcifications between the HR- and HR group in their series [34].

In the current study, we categorized the pattern of enhancement elicited by CESM into homogeneous, heterogenous, ring and NME based on MRI-BIRADS lexicon morphological characteristics [36]. Heterogeneous pattern of enhancement was the most common enhancement pattern encountered in our series seen in 49.2% of cases. Our findings were consistent with previous investigators who reported that masses eliciting heterogeneous enhancement showed a significantly higher tendency of malignancy compared to masses with homogeneous enhancement [26, 37, 38]. NME was the second most common pattern observed in our series in line with the Kamal et al. [26] and Wilhelm et al. [39], who reported strong correlation between NME and malignant pathology with P values of ≤ 0.001. On CESM, we found that round or oval shape, circumscribed margins and associated NME were strongly associated with HER2+/HR− tumors similarly reported by previous investigators who based their conclusions on mammography and MRI findings [14, 34, 35].

We found that tumors with irregular shape, spiculated outline and heterogeneous or ring enhancement were more consistent with HER2+/HR+ tumors in line with previous investigators who also concluded that HER2+/HR+ tumors were more likely to resemble luminal A (LA) tumors [40, 41]. Our study was contrasted by previous investigators who did not find significant correlation between ring enhancement on CESM and malignant pathology in their series of 109 malignant lesions and came to a conclusion that reviewing T2 MRI images is very reliable in ruling out infected cysts and abscess cavities—an advantage that CESM lacks [26].

The current study had some limitations. Selection bias may have been present due to retrospective study design. The study was carried out at a single institution with a small sample size for HR− and HR+ groups. More multi-institution studies with larger sample sizes are needed in the future to further document and confirm our findings. Additionally, we excluded patients who had neoadjuvant chemotherapy before the CESM examination; therefore, our findings might not be applicable to locally advanced breast cancer. The study did not investigate the interobserver variability in image interpretation. We observed that despite the participation of expert radiologists who were blinded to the IHC status, the evaluation of mass shapes and margins was subjective. The use of automated computer-aided detection (CAD) techniques may be helpful in the future in obtaining reproducible results potentially aiding breast cancer molecular subtype characterization.

Conclusions

CESM is a well-tolerated widely available novel diagnostic tool that offers accuracy comparable to CE-MRI. HER2+ tumors have different demographic, pathologic and imaging features according to the hormone receptor status. On CESM, HER2+/HR− tumors were more likely to be multifocal, rounded or oval with circumscribed margins and associated with NME, while HER2+/HR+ cancers were more likely to be irregular in shape, and spiculated in margin, and elicit heterogeneous or ring enhancement. Mammographic imaging features were not helpful in differentiating HR− and HR+ cancers. Because the two subtypes of HER2 breast cancer have different clinical outcomes, CESM imaging features can potentially enhance patient outcome by accelerating the diagnosis and treatment. More prospective multicenter studies with larger sample sizes and interobserver variability are needed in the future to further validate and document our findings.