Breast-lesion assessment using amide proton transfer-weighted imaging and dynamic contrast-enhanced MR imaging

Breast cancer is the most prevalent malignant neoplasm all over the world.¹ Magnetic resonance imaging (MRI), a non-invasive technique with exceptional soft tissue resolution, plays a significant role in diagnosis, treatment, and prognosis assessment of breast diseases.^2,3,4 In an effort to standardize the imaging strategy for breast lesions, the American College of Radiology created the Breast Imaging Reporting Data System (BI-RADS) in 1992.⁵ However, some imaging features of benign and malignant lesions still overlap. False-positive results from conventional breast MRI could result in unnecessary invasive biopsies being performed.⁶

Dynamic contrast-enhanced MRI (DCE-MRI) is used to obtain focal information by injecting a contrast agent, and it has been widely used for determining prognosis, monitoring therapy, and diagnosing many diseases.⁷ The study of Alkhunizi SM et al. provided insights into the consequences of gadolinium-based contrast agent (GBCA) delivery by showing considerable retention of gadolinium in the spinal cord and peripheral nerves 1 day after dose.⁸ Meanwhile, after multiple GBCA injection, the dentate nucleus and pallidum showed abnormally high signals under T1-weighted imaging sequence.⁹ Therefore, the search for a safe tool for evaluating breast lesions is of great importance.

In 2000, the first MR contrast images of several small molecules were acquired by Wolff et al., who called this novel molecular imaging technology chemical exchange saturation transfer (CEST).¹⁰ Amide proton transfer-weighted imaging (APTWI) technique is a new molecular MRI sequence that is based on CEST, focusing on the exchange between amide protons and bulk water and thus generating image contrast at 3.5 parts per million (ppm) away from water frequency.¹¹ Various preclinical and clinical research investigations have been conducted on this imaging technology¹², making it a potential molecular imaging tool that is now used in clinics.¹³ Meanwhile, in recent years, the differentiation of tumor subtypes and grades and the assessment of therapy efficacy are two applications of APTWI in breast cancer that have drawn increasing attention.^14,15,16,17 However, no research has compared the value of APTWI in diagnosing benign and malignant breast diseases with it in DCE.

The purpose of this study was to investigate the potential of APTWI in breast-lesion diagnosis and compare the performance of DCE and APTWI in identifying benign from malignant breast lesions.

In this prospective study, a pilot research of the viability of using APT value in conjunction with DWI and DCE sequences for the distinction between benign and malignant mammary lesions was successfully conducted. The correlation between the APT value and the pathological factors of breast cancer was explored. The feasibility of 3D APTWI MRI for the distinguishing between benign and malignant mammary lesions was demonstrated. With the clinical feasibility of APTWI in breast a concern among researchers, this study showed an excellent agreement in APT value measurement and the high diagnostic efficiency of APT values, similar to the efficiency of TIC, indicating that valuable diagnostic information could be obtained without using GBCA. Thus, APT value could be regarded as a non-invasive biomarker for differentiating mammary lesions.

In this study, the malignant lesions typically had higher APT values than the benign lesions, consistent with what is generally known about malignant tumors in other diseases.^22,23,24,25 APTWI is a molecular MRI technique that is based on chemical exchange saturation transfer that could detect endogenous mobile proteins and peptides at low molecular concentrations. The high intensity in APTWI was made possible by the fact that malignant lesions were highly cellular and that several proteins were overexpressed in comparison to benign lesions. Due to the high levels of hemoglobin and albumin in blood, angiogenesis is another component that may contribute to enhanced protein signaling in malignancies.¹³ However, this discovery differs from the results of earlier research by Meng et al.^14,16 They suggested that one reason could be that some benign lesions’ secretory abilities were intact, whereas the secretory abilities of some malignant lesions were damaged, leading to lower protein and polypeptide concentration. This difference could result from variations in study participants. Loi L et al. demonstrated that breast cancer displayed a substantially higher APT value than typical fibro-glandular tissue²⁶, which fully confirms the results of the present study.

The findings showed that APT value has a strong diagnostic performance similar to TIC, and it could be employed for the differential diagnosis of mammary lesions. The combination of DWI and APT may increase the specificity of the diagnosis. DCE examination was conducted by injecting GBCA. As is well known, the injection of GBCA has many side effects, such as allergic as well as allergoid reactions, including anaphylactic reactions²⁷, contrast-induced nephropathy²⁸, nephrogenic systemic fibrosis²⁹, and gadolinium retention/deposition.³⁰ On the contrary, APTWI does not require the use of GBCA, thus saving costs and avoiding side effects. For patients with a history of allergies or those refusing to undergo DCE, APTWI could be a great substitute for DCE examination.

In this study, the APT value showed no correlations with ER, Ki-67 expression and histological grades, inconsistent with the results of previous studies. Liu Z et al. found a weakly positive connection between APT value and Ki-67 expression.³¹ A notable detail that they used a 20% threshold to report comparable APT values between groups with varying Ki-67 proliferation levels. This conflict could result from the differing Ki-67 proliferation index levels employed for group classification. Zhang N et al. reported that the APT value and ER expression had a negative connection, may be because ER inhibits lesion angiogenesis by regulating the production of the vascular endothelial growth factor.³² On the contrary, the results of syudy by Meng N et al. showed that there was no correlation between the APT value and ER expression.¹⁴ Consequently, a larger sample size is required for further research. Meng N et al. showed that only a weakly positive correlation existed between the pathogenic grade and APT value; they hypothesized that the cause is that high-level tumor cells have a high density, evident nuclear atypia, as well as increased tissue necrosis, all of which alters the proteins and peptides in the nearby microenvironment, in addition to the rate and distribution of internal water molecule diffusion and movement.¹⁶ Nevertheless, our grouping of histological grades was I, II, and III while their grouping of histological grades was low and high, which could explain the differences in the results. In accordance with previous researches, the APT value showed no correlations with PR and Her-2.^32,33

This study has several limitations. First, breast lesions and some subtypes of breast cancer had relatively small sample sizes, which could cause the AUC of APTWI to be overestimated. Future research may also be warranted on the genomic subtyping capabilities of APTWI for breast cancers. Therefore, more research involving a bigger sample size is required. Second, the diagnostic efficacy of DCE assessed using only TIC curves may be underestimated. Liang X et al. recently revealed that the quantitative and semi-quantitative parameters of DCE provided great diagnostic performance. Third, breast tissue has a substantial amount of fat, which could affect the accuracy of APT values.³⁴ Obviously, in the present study, patients with fatty breast tended to have poorer image quality. All CEST signals become essentially dependent on the amount of fat per voxel due to the ubiquitous fat signal's erroneous normalization of the Z-spectrum. Finally, bias may have existed in the selection of patients. Breast MRI is usually conducted in high-risk population.³⁵ Patients receiving MRI in clinical practice had a considerably higher chance of malignant tumor than those receiving mammography or ultrasound, which could induce background bias (more malignant cases than benign cases).

The capability of identifying benign from malignant breast lesions could be achieved using APTWI and DCE, and APTWI may be a great supplement or even replacement for DCE sequence. The findings still need to be confirmed by future investigations on patient cohorts with larger breast lesions.