This prospective study was performed in accordance with the Declaration of Helsinki and approved by the local ethics committee. All eligible patients were willing to undergo all study procedures and provided written informed consent prior to enrolment.
Patients
From October 2012 to December 2014, 54 consenting women (mean age 54 years, range 33–72) were consecutively enrolled in this prospective study. Inclusion criteria were age 18 years or older and histologically proven breast cancer with indications for NAC. Exclusion criteria were: known BRCA-mutation (to ensure the highest radioprotection); pregnancy; contraindications to MRI or contraindications to CESM, including a history of an anaphylactoid or anaphylactic reaction to any contrast medium or impaired renal function with chronic kidney disease stage 3 or higher (creatinine clearance < 60 ml/min).
The diagnosis of breast carcinoma was based on a percutaneous tru-cut biopsy to determine the histological and molecular characteristics of the lesion detected at the initial multimodality assessment based on physical examination, digital mammography and ultrasound. The therapeutic pathway of each patient was discussed and planned by the Breast Unit Multidisciplinary Group. The NAC protocol included the administration of anthracycline for the first 3 months and then taxanes for the next 3 months; trastuzumab was added in cases of human epidermal growth factor receptor 2 (HER2)+ carcinoma. Before NAC, each patient was tattooed to ensure correct surgical positioning of the region to be excised [17].
Study workflow
The study workflow is shown in Fig. 1.
CESM and MRI were performed thrice: time (t)0, before the beginning of NAC (pre-NAC); t3m, after about 3 months (during-NAC) and t6m, after the end of treatment (post-NAC), just before surgery. In premenopausal women, MRI and CESM were performed between the 5th and 12th day after the end of menstrual flux. MRI was performed first, and then CESM was performed within 3 days. According to the first MRI performed before the beginning of NAC, if the neoplasm involved both breasts, then CESM was performed bilaterally, otherwise it was performed only on the affected breast.
CESM protocol
CESM is based on a dual-energy system developed by GE Healthcare (Chalfont St-Giles, UK): after contrast administration, a set of low-energy and high-energy images is acquired in quick succession while the breast remains compressed, obtaining a low-dose image, comparable to a standard digital mammogram, and a post-processing recombined image, which enhances the distribution of the iodine contrast medium. The three digital mammography units (GE Senographe Essential) used in this study were equipped with molybdenum (Mo) and rhodium (Rh) anode tracks and filters, which can be combined to perform standard mammography. Additional copper (Cu) filters were installed to allow the combinations of Mo/Cu or Rh/Cu and shape the X-ray spectra for CESM dual-energy exposures. The examinations were performed using the CESM automatic exposure option, set according to the thickness of the compressed breast and its density. The average glandular dose (AGD) delivered by the CESM examinations was evaluated.
The CESM protocol was designed according to Dromain et al. [9]. The intravenous contrast agent was administered by a nurse under the supervision of a radiologist, using a hand-held battery-powered injector (Optistat, Covidien). The contrast agent was ioversolo 350 mg/ml (Optiray, Covidien) and the administered dose was 1.5 ml/kg of body weight [7,8,9,10,11,12,13,14,15,16]. In a monolateral CESM, the radiographer compressed the breast for the medio-lateral oblique projection (MLO) 2 minutes after administration of contrast agent and then decompressed the breast, and after a further 2 minutes again compressed the breast for the cranio-caudal projection (CC).
MRI protocol
MRI examinations were performed with three different MRI scanners (Philips-Achieva 1.5 T, Siemens-Avanto 1.5 T and GE-Signa 1.5 T), all equipped with phased-array coils (seven coils on the Philips-Achieva; eight coils on the Siemens-Avanto and GE-Signa). The same MRI protocol was applied in the prone position and with no breast compression. The sequences acquired were: T2-weighted with or without fat suppression; 3D dynamic perfusion T1-weighted with or without fat suppression, with a first acquisition before contrast agent administration (mask) and seven acquisitions after the gadolinium contrast agent was administered, and temporal resolution between 50 and 70 s; diffusion weighted sequence where b values of 0 and 600 mm/s2 were set. The images were then processed using commercial software (Cadstream, Merge Healthcare) [18,19,20]. A detailed assessment of the enhancement kinetics was evaluated by the radiologist especially during diagnosis, but its description is not an objective of the study as the comparison is based on measurements of the size of the enhanced lesions.
Endpoint of the study and image analysis
A total of seven independent radiologists, all experts in breast imaging (from 5 to 18 years’ experience), took part in this study to evaluate and report CESM and MRI examinations. In the clinical workflow, the radiologist evaluated either CESM or MRI of the single patient, blinded to the MRI or CESM performed at the same step of the study, but not blinded to all previous breast examinations.
The endpoint of the study was the comparison of post-NAC measurements on CESM and MRI, with the histopathological response identified in surgical specimens considered as the gold standard. Correlation between the measurements of the lesion using the two diagnostic tools was then evaluated for each step of the study (pre NAC, during NAC and post NAC). And last of all, the response to therapy was assessed for each patient according to the parameters set in the response evaluation criteria in solid tumors (RECIST) 1.1 criteria, considering the sum of the largest dimension of malignancies at baseline and its variation at subsequent measurements [3, 21]. Response was classified as follows: complete response (CR, disappearance of all lesions); partial response (PR, ≥ 30% dimensional reduction), stable disease (SD, < 30% dimensional reduction/< 20% dimensional increase) and progressive disease (PD, ≥ 20% dimensional increase).
In CESM measurements were obtained considering the maximum dimension on recombined images, based on contrast uptake. Since the low-energy image of CESM allows the visualization and characterization of microcalcifications, this information, integrating the extension of the enhanced area [16], contributed to defining the size of the tumor bed, especially in the pre-NAC evaluation.
In MRI, the maximum diameter of the enhancing lesion was measured on post-contrast T1-weighted or subtracted axial images at peak enhancement. For both CESM and MRI, in the evaluations during NAC and post NAC, enhancing foci spread inside the tumor bed were considered as the expression of what remained of the same pathologic area, thus defined with a single overall measure. As this evaluation is already quite common in MRI, we considered it also for CESM, according to its high sensitivity and specificity described in the literature [7,8,9,10,11,12,13,14,15,16, 22,23,24,25].
According to the literature [22, 23] and the Breast Imaging-Reporting and Data System (BI-RADS) lexicon, the lesion location and a detailed descriptive analysis of its enhancement were reported both for MRI and CESM pre NAC and post NAC: non-mass enhancement, evaluating its distribution, or mass enhancement, defining lesion shape, margins and vascularization, considering both qualitative evaluation of contrast uptake and internal enhancement patterns.
Surgical specimens
Tumor size was macroscopically measured in formalin-fixed specimens. Histopathological examination was performed in paraffin-embedded specimens. Tumor size was defined as the largest dimension based on macroscopic and histopathological examination. In the case of multifocal breast cancer, the maximum dimension of the largest invasive tumor was used.
In clinical practice the persistence of residual ductal carcinoma in situ (DCIS) components is considered as pathological complete response (pCR) [26]. However, in comparing CESM and MRI measurements with histopathological response on surgical specimens, we considered the DCIS component as part of the residual neoplasm, because when visible on imaging, it impacted surgical planning.
Statistical analysis
In the absence of an a priori hypothesis, given the exploratory nature of the study, no formal sample size calculation was performed. Primary analysis included assessing agreement between CESM and MRI and histopathological assessment, and between the two diagnostic tools (pre, during and post NAC), measured with Lin’s coefficient [27]. The agreement difference estimate was accompanied by a bootstrap (bias-corrected and accelerated) confidence interval. In the first evaluation the Pearson’s correlation coefficient (PCC) was also calculated to allow easier comparison with similar previous studies in the literature.
Secondary analyses included estimation of the diagnostic performance indexes of each method for complete response with 95% confidence intervals (Clopper-Pearson-method [28]). Statistical analysis was carried out using R 3.2.3 (R Foundation for Statistical Computing, Vienna, Austria).