CESM is often compared to MRI to test its utility in tumor diagnosis, and indeed, enhancement patterns were moderately in agreement between the two techniques (21). CESM may produce an enhancement intensity weaker in the ER-positive group than in the ER-negative group, as well as weaker in the PR-positive group than in the PR-negative group, and stronger in the HER-2-positive group than in the HER-2-negative group (21). Further, when comparing radiological characterization of the lesions obtained by either CESM or contrast-enhanced MRI, they have a similar association with the proliferation of tumoral cells, as expressed by Ki-67 (22). However, the authors do not describe if there are any differences between CESM and MRI in differentiating hormonal receptor status.

In CESM-enhanced lesions, the expression was found to be 100% for ER and 77.4% for PR, while moderate or high HER2 positivity was found in lesions with non-mass enhancement and with mass closely associated with a non-mass enhancement component (23). Further, via CESM enhancement, neoplasms larger than 5 mm, with a high proliferative index and frequently HER2-positive, are recognized (24). Conversely, the non-enhancing breast cancer lesions were not associated with any prognostic factor, such as status of ER and/or PR, HER2 expression and/or amplification, and percentage of Ki67, which might be associated with the probability of showing enhancement (25).

Nowadays, one of the cutting-edge methods for image analysis is based on radiomics. For non-invasively assessing the hormone receptor status, other than tumor invasiveness and grade, radiomic features were derived from the first-order histogram of primary breast cancer lesions contoured on both CESM and MRI images and the two techniques resulted to be alternative in the assessment of hormone receptor status (26). Further, radiomics on CESM images showed the potential for the non-invasive characterization of heterogeneous tumors with different hormone receptor statuses (27). Lastly, radiomic features may predict histological outcomes and molecular subtypes via discriminating lesions with a positive or negative expression of hormonal receptors, and being associated with HER2. In particular, in an immunohistochemical study, the performances for discriminating positive versus negative expressions were 90.87% for HER2 positive versus HER2 negative, 83.79% for ER positive versus ER negative, and 84.80% for Ki67 positive versus Ki67 negative (28). The list of the final articles and their relationship with biologic prognostic factors is summarized in Table 1 .

The Quadas-2 survey showed that the articles considered in the analysis were at risk of bias, especially for what concerns the study tests conducted in each research. Indeed, at the time of the radiological evaluation the investigators, i.e., the radiologists, were aware of the results of the histological test, in all studies with the exception of one study, who performed a blinded histological analysis (22). However, all articles referred to a proper reference test, i.e., definitive histology or diagnostic biopsy. Further, four articles were biased in patient selection, because they removed either patients with a tumor not easily identifiable, such as that with suspicious but not contrast-enhancing lesions (27), or patients with post-histology edema or not willing to undergo CESM (25), or because they did not clarify whether patients have mono- or multifocal diseases (21, 22). Lastly, one article alone was at risk of flow bias, because of using both definitive histology and diagnostic biopsy as standard reference (21). The Quadas-2 survey is shown in Table 2 .

Prognostic factors are correlated with patient prognosis and allow important information about the efficacy of antitumoral treatment. Literature demonstrated that proliferative activity indicator (Ki67), HER-2, and hormonal receptor, such as ER and PR, statuses are important in treatment choice and that they have prognostic value in predicting pathological response and clinical outcome (29). As a matter of fact, HER-2 status represents a solid prognostic factor that predicts the response to trastuzumab alone or associated with pertuzumab treatment in locally advanced or early disease therapy (30). Also, determination of ER and PR status is crucial as their expression on the tumor cellular surface is related to a good response to endocrine therapy in both neoadjuvant and adjuvant therapy (30).

Among the biomarkers used to define tumor aggressiveness, Ki67, HER-2, ER, and PR are quantitative values. On the contrary, grading, which is used as well to define tumor aggressiveness, is a qualitative biomarker; therefore, we rather avoided to include it as an investigated prognostic factor in this study.

CESM is a recent tool for diagnostic imaging that, although it uses ionizing radiations thus presenting some limitations in terms of radioprotection (7), may overtake the use of MRI in breast cancer monitoring, since it is more accessible, cheaper, faster, and more tolerated by patients (31), while maintaining performance equivalent to MRI and improving specificity (7, 17). As a matter of fact, the promising results of diagnostic performance could suggest CESM to be a valid alternative for patients who are not eligible for MRI. As a matter of fact, CESM and breast MRI similarly detect physiological, benign background parenchymal enhancement, which may be significantly associated with menopausal status, radiation therapy, hormonal treatment, and breast density and that rarely causes diagnostic issues if showing a bilateral, symmetrical appearance (32, 33). At the same time, the background parenchymal enhancement on MRI is considered a biomarker for increased risk of breast malignancy, while it is not known if the same holds true for CESM (34, 35).

Indeed, CESM and MRI show similar enhancement patterns (21) and a similar association with the proliferation of tumoral cells (22). The equivalence of CESM and MRI might rise from tumor vascularization, which is a crucial feature observed by both diagnostic modalities and is influenced by Ki67.

A necessary step to include CESM in everyday clinical practice will be the standardization of diagnostic criteria. Given the similarity of the basic principles of lesion blood supply of the two modalities, MRI morphology descriptors have been already investigated and used to characterize lesions on CESM (36); however, more studies are needed to finalize the use of these descriptors in CESM image evaluation.

As in any imaging modality, patient motion may affect image quality. Due to the simultaneous acquisition of low-energy and high-energy images, the length of each exposure with CESM is longer than a standard full-field digital mammography, increasing the possibility of motion. However, the examination time of CESM is still shorter than the second-level examination MRI, reducing the risk of motion artifacts. Moreover, to instruct well the patient to hold as still as possible during the exposure is fundamental to reducing the possibility of motion (37).

CESM combines an iodinated contrast agent with the standard mammographic technique to improve lesion detectability. Since the growth of tumors is accompanied by angiogenesis, CESM permits to assess the enhancement related to the neovascularity of breast cancers, allowing a functional characterization in addition to the morphological features provided by structural images (16).

In literature, CESM-enhancing lesions have been associated with higher levels of prognostic factors, such as ER, PR, and HER2 (23, 24). On the other hand, non-enhancing lesions have been found not to relate to prognostic factors (25). Indeed, tumors have a higher enhancement compared to normal tissue due to the increase in vascularization, which in turn is associated with different tumor characteristics and therefore different expressions of prognostic factors.

Feature extraction in radiomics is typically realized by means of pattern recognition algorithms and provides, as a result, a set of numbers, each one representing a quantitative description of a specific either geometric or physical property of the image portion under consideration. In oncological applications, examples of features are tumor size, shape, intensity, and texture, collectively providing a comprehensive tumor characterization, called the radiomics signature of the tumor (38). From an epistemological perspective, radiomics is based on the hypothesis that the extracted features reflect mechanisms occurring at genetic and molecular levels (39) and may reveal the relationship of tumor lesion surfaces with prognostic factor expression. The potential of radiomics applied on breast imaging has been investigated recently, and studies have already demonstrated the additive value of radiomics on MRI in breast cancer evaluation and prognosis (40, 41). Indeed, radiomics on CESM images might assess hormone receptor status (26) and characterize the related heterogeneous tumors (27), as well as predict histological outcomes and molecular subtypes associated with hormone receptors’ expression (28). As a matter of fact, radiomics arises from the analysis of cell morphology, which may be influenced by the expression of the different receptors on the cell surface of the different tumors, thus permitting to differentiate the receptor status starting from imaging.

Radiomics could also contribute to differentiating benign from malignant enhancement in complicated cases, as in patients with high background parenchymal enhancement or low vascularized lesions, that may have a high risk for underestimation or even overestimation of the lesion (42), and to predicting response to NAC (43).

The articles included are all published in the last 2 years, and none of them was blinded, except that of Petrillo et al. (22). Nevertheless, this bias did not invalidate the articles, as the goal of these studies was to find a relationship between imaging features and prognostic factors, not just detecting a tumor. Indeed, knowing the histological subtypes was part of patients’ preliminary information needed to obtain a sample of patients with heterogeneous radiologic patterns. Conversely, homogeneous histologic analysis was crucial in order to obtain consistent results among patients, and in one article the authors did not grant this consistency.

Only eight articles investigated the association between CESM imaging and prognostic factors, suggesting that the use of this technique in cancer prognosis and monitoring is still to be deeply investigated. Indeed, the modality is relatively young and large data pools are required to get strong results on this topic.