Clinical utility of contrast-enhanced spectral mammography with histopathological correlation

INTRODUCTION

Breast cancer remains one of the most commonly diagnosed cancers among women. Globally, 2.3 million women were diagnosed with breast cancer in 2022, with 670,000 deaths reported from the disease. In 157 out of 185 countries, it was the most frequently diagnosed cancer among women.^[1] In India, breast cancer accounts for 28.2% of all female cancer cases, making it the most prevalent cancer among females.^[2]

Breast cancer has a multifactorial etiology resulting from a combination of genetic, hormonal, and environmental risk factors. Currently, three primary imaging techniques are used to diagnose and evaluate breast cancer: mammography, ultrasound (US), and magnetic resonance imaging (MRI). Among these, mammography is the recommended screening modality for women over 40 years of age, as it can detect cancers long before they become clinically evident. Early detection through screening can significantly reduce morbidity and mortality rates. US is one of the most valuable modalities in dense breasts; however, the sensitivity drops due to dense heterogeneous tissue producing shadowing.^[3]

Contrast-Enhanced Spectral Mammography (CESM) is a relatively recent and sophisticated digital mammography advancement incorporating a contrast agent to enhance imaging. It is based on dual-energy mammography, where two images of the same breast are captured at different X-ray energy levels—below and above the k-edge of an iodinated contrast agent—following the intravenous administration of an iodine contrast. These images are then processed through subtraction techniques to highlight areas of iodine uptake, aiding in the identification of enhancing lesions.^[4] CESM visualizes contrast enhancement associated with the formation of new blood vessels in breast cancer, similar to contrast-enhanced breast MRI. Since angiogenesis is closely linked to tumor growth, this imaging technique enhances the detection and characterization of malignant lesions.^[5]

In this study, we aimed to contribute to the scientific literature by evaluating the utility of CESM as an adjunct to conventional mammography in improving diagnostic accuracy for differentiating between benign and malignant breast lesions. Additionally, we examined how CESM BI-RADS classifications correlated with conventional mammography BI-RADS with respect to histopathological diagnoses.

MATERIALS AND METHODS

This was a retrospective observational study conducted at our institution after obtaining approval from the institutional review board. Overall, 58 female patients aged 30–70 years who had breast lesions on mammograms and had undergone CESM were included in the study [Figure 1].

Close

Figure 1:

Work flow depicting the chronology of the research. CESM BIRADS: Contrast enhanced spectral mammography breast imaging reporting and data system.

Export to PPT

Procedure of Contrast Administration

Via an 18-gauge catheter, 1.5 ml/kg of the contrast agent was delivered at 3–4 ml/sec, followed by a saline flush at the same rate. CESM imaging began 2 minutes after injection and took around 6–8 minutes. Patients were observed for one hour post-procedure for any allergic or hypersensitivity reactions.

Post-contrast images were assessed for background parenchymal enhancement, mass morphology (shape, margins, internal features, conspicuity), non-mass enhancement, and asymmetry. Each lesion was then assigned a CESM BI-RADS category.

The affected breast was imaged first, followed by the normal breast, using standard CC and MLO views. For each view, dual-energy images were acquired. Low energy was equivalent to conventional mammogram and high energy images were used to generate subtracted interpretable images. This combination allows for improved lesion detection while maintaining normal anatomical detail.

As tumors exhibit increased vascular permeability, contrast medium pools around them, producing a high signal on the subtracted image, while normal breast tissue shows minimal uptake [Figure 2].^[6,7]

Close

Figure 2:

Set of images acquired during CESM – (a) High energy image, (b) low energy image and (c) subtracted or recombined image.

Export to PPT

All patients underwent US-guided core needle biopsy, and histopathological results were classified as benign or malignant. These findings were correlated with BI-RADS assessments from mammography and CESM, following the ACR BI-RADS 5th edition (2013) and ACR BI-RADS CEM Supplement (2022), respectively.

RESULTS

A total of 58 patients with 58 breast lesions were included in the study. Of these, 36 lesions were histopathologically confirmed as malignant, while 22 were benign. Among the benign lesions, the most common diagnoses were fibroadenomas (n = 9), fibrocystic changes (n = 6), benign phyllodes tumors (n = 3), and sclerosing adenosis (n = 4). Among the malignant lesions, the majority were invasive ductal carcinoma (IDC) (n = 30), followed by invasive lobular carcinoma (n = 3), ductal carcinoma in situ (DCIS) (n = 2), and one case of mucinous carcinoma (n = 1).

A comparison between mammogram BI-RADS and CESM BI-RADS, based on biopsy results, showed that CESM provided higher diagnostic accuracy with a better correlation to histopathological findings. Table 1 shows a statistically significant association between mammography BI-RADS and CESM BI-RADS. CESM had higher negative predictive value (NPV) and sensitivity than DM.

A statistically significant relationship was observed between enhancement patterns of lesions and their biopsy results. Table 2 shows that heterogeneous enhancement patterns (44.8%) were more frequently associated with malignant results, while homogeneous patterns (50%) were more commonly linked to benign results. Rim enhancement (5.2%) showed a low overall occurrence, seen only in three cases, and was mostly linked to benignity.

Intense (31.1%) and moderate enhancement (34.5%) as compared to background vessels (subjective assessment by experienced radiologist) were more frequently associated with malignant results, while mild enhancement (32.8%) was more commonly linked to benign results. Patchy enhancement was rare, seen only in one case, and was associated with benign results.

Table 3 shows a statistically significant association between CESM BI-RADS categories and histopathology results.

CESM demonstrated a higher malignancy detection rate, particularly in lesions categorized as BI-RADS 4 and 5, reducing ambiguity seen in DM, as seen in Table 4.

CESM BI-RADS categories showed a strong correlation with histopathology findings. Higher CESM BI-RADS scores (4C and 5) demonstrated malignancy rates of 75% and 100%, respectively, while lower categories (3 and 4A) showed predominantly benign outcomes [Table 5].

CESM frequently upgraded lesions classified as lower BI-RADS on DM particularly moving BI-RADS 3 and 4A lesions into higher-risk categories (4B, 4C, or 5), as shown in Figure 3 and 4.

Close

Figure 3:

(a-b) CC & MLO views of left breast mammogram show irregular circumscribed mass in the retro areolar region (white arrows) assigned as BI-RADS 4A. (c-e) CESM performed shows moderate conspicuous homogenous enhancement of mass with wash out in delayed phase – mass upgraded to BI-RADS 4C & biopsied, HPE revealed invasive ductal carcinoma.

Export to PPT

Close

Figure 4:

(a-b) CC & MLO views of right breast mammogram show irregular circumscribed mass in the upper outer quadrant (white arrows) assigned as BI-RADS 4A. (c-d) CESM performed shows no enhancement of mass, downgraded to BI-RADS 3 – for 6 months short term follow up.

Export to PPT

The overall result shows histopathological diagnoses in malignant cases predominantly included IDC, with occasional instances of invasive lobular carcinoma and DCIS. Among benign findings, fibroadenomas, fibrocystic changes, and benign proliferative lesions were commonly observed. This distribution reinforces CESM’s reliability in lesion stratification, further validated by pathology results.

DISCUSSION

Our study aimed to compare CESM findings in patients with suspicious breast lesions to histopathological results, which are considered the gold standard. By assessing the diagnostic accuracy of CESM—a faster, simpler, and more cost-effective modality—this study evaluates its reliability against the conventional digital mammography.

Dual-energy subtraction was introduced in contrast-enhanced mammography (CEM) to address the limitations of temporal subtraction.^[8] This technique facilitates full breast compression and acquisition of multiple images in a single session. Due to these advantages, dual-energy subtraction has become standard for CEM. To date, over 100,000 CEM examinations have been performed in both clinical and research settings, demonstrating its growing acceptance and value in clinical practice.^[9]

The results highlight that higher CESM BI-RADS categories are strongly associated with a greater likelihood of malignancy, while lower BI-RADS categories are more often linked to benign outcomes, specifically among cases categorized as CESM. In BI-RADS 3, 91.6% were benign, and 8.3% were malignant. In the CESM BI-RADS 4A category, 88.29% were benign and 11.7% malignant. As the level of suspicion increased, the proportion of malignancies cases rose markedly; 75% of BI-RADS 4C cases and 100% of BI-RADS 5 cases were malignant. These findings support the role of CESM BI-RADS as a reliable, non-invasive tool for stratifying malignancy risk in clinical breast cancer diagnosis. The ACR BI-RADS lexicon states that category 4 lesions have a 3%–94% risk of malignancy, subdivided into 4A (3%–10%), 4B (11%–50%), and 4C (51%–94%). In this study, 61.5% of category 4 lesions were histologically benign, consistent with 61.2% reported by Elverici et al.^[10] This highlights the need for more accurate diagnostic tools to reduce unnecessary biopsies.

In this study, PPV, NPV, and accuracy of CEM were 68.97%, 89.65%, and 79.31%, respectively, aligning closely with Lobbes et al.,^[11] who reported a PPV of 76.2% and NPV of 100%. Sensitivity (86.9%) and specificity (74.3%) also approximated their findings (100% and 87.7%). Additionally, 73.9% of malignant lesions showed heterogeneous enhancement, while 68.6% of benign lesions displayed homogeneous enhancement, similar to Liu et al.,^[12] who reported 84.1% and 62.2%, respectively.

Malignant lesions tend to grow rapidly, often exceeding their vascular supply, leading to necrosis and uneven microvessel distribution. This explains the 73.9% heterogeneous enhancement seen in malignant lesions in our study, which aligns with the findings of the study by Metz et al.^[13]

A strong correlation between enhancement patterns and biopsy outcomes was observed. Malignancy was more common with intense (52.2%) and moderate (43.5%) enhancements, while mild enhancement (51.4%) was primarily associated with benign lesions. Patchy enhancement, although rare, was exclusively linked to benign findings. These results differ slightly from those reported by Chi et al.,^[14] where 73.9% of malignant lesions exhibited marked enhancement and 22.7% showed moderate enhancement, while benign lesions demonstrated mild (32.1%) and moderate (45.9%) enhancement.

CESM offers a significant advantage over conventional mammography, especially in women with dense breast tissue, where mammography’s sensitivity is limited. While mammography relies on structural changes, CESM provides both anatomical and functional imaging by highlighting areas of contrast uptake, indicating angiogenesis typical of malignancies.^[15] CESM has demonstrated higher sensitivity and specificity, better lesion conspicuity, and improved multifocal and multicentric disease detection compared to mammography.^[16] Thus, CESM serves as an effective problem-solving tool in cases where mammography is inconclusive.

However, CESM also has limitations. While it improves local staging accuracy, false positives and false negatives still occur. Lesions lying outside the mammographic field of view or those with diffuse infiltration, such as lobular carcinomas, are more likely to result in false negatives. The study by Thibault et al. identified six false negatives on CESM, four of which were aggressive lobular carcinomas.^[17] Additionally, certain tumors, including DCISs, DCIS, and mucinous carcinomas, may show minimal or no enhancement, leading to potential misdiagnosis. Benign lesions like fibroadenomas commonly enhance on CESM, and other lesions such as papillomas, hamartomas, intramammary nodes, and fat necrosis may also contribute to false-positive interpretations.^[18,19]

Like traditional mammography, CESM involves radiation exposure and carries a small risk of contrast-related reactions, particularly in patients with renal impairment. Depth perception is also limited compared to MRI.

However, CESM offers key advantages: faster acquisition, lower cost, and greater patient tolerance, especially in those with pacemakers, metallic implants, or claustrophobia. It can also display benign background parenchymal enhancement, similar to breast MRI.^[20] While CESM is a valuable diagnostic tool, its use should be guided by patient-specific risk-benefit assessment. Its high sensitivity and strong NPV make it reliable for ruling out malignancy, helping reduce unnecessary biopsies and patient anxiety.

CONCLUSION

This study highlights CESM as an effective, non-invasive tool for detecting and characterizing breast cancer. It enhances conventional imaging by distinguishing between benign and malignant lesions, with higher CESM BI-RADS categories strongly associated with malignancy. High NPV and PPV confirm its diagnostic reliability, while typical enhancement patterns support its accuracy. CESM’s affordability, efficiency, and patient acceptability make it a practical alternative to MRI, reinforcing its value in routine breast cancer assessment and care.