This prospective study involved 50 patients with early-stage TNBC according to TNM-staging 8thedition [23].All patients were evaluated after the end of treatment, which consisted of 3D conformal radiotherapy (whole breast radiotherapy 4240 cGy /16 fx -265 cGy /fx with electron boost 1000 cGy/ 5 fx- 200cGy/fx), and in case of chemotherapy, it consisted of 4 cycles AC, followed by 12 weeks of paclitaxel. During the follow-up period, when a suspicious mass was detected by ultrasonography or mammography, multiparmetric MRI was done. Patients were referred from the Oncology Department to the Radiodiagnosis Department, Tanta University hospital during the period of April 2017 to April 2020.
Breast lesions detected during sonomammographic examinations that scored from BIRADS 3 to 5 according to the BIRADS classification were included in the analysis [24].
Pregnant women, patients with chronic renal impairment, previous allergies to contrast, implantable devices that are not MRI-compatible, and those with particularly large breasts were excluded from the study. Patients were aware of the examination, and informed consent was obtained prior to the analysis. Ethics approval was acquired from institutional ethical committee.
Complete history of patients was obtained, including personal details, menstrual cycle, and complete clinicopathological data, i.e., details of oncological treatment (e.g., radiotherapy, chemotherapy). TNBC was defined as the lack of estrogen, progesterone, and her2 receptor staining by immunohistochemistry or gene amplification using fluorescent in situ hybridization according to the ASCO/CAP guidelines [25]. The conducted metastatic workup included chest X-ray, abdomenopelvic ultrasonography, bone scan, and computed tomography (CT) scan of the chest or abdomenopelvis with contrast, if indicated. Sonomammography was performed for all patients. Suspicious lesions on the surgical bed were scored from BIRADS 3 to 5. Subsequently, MRI was conducted to examine the suspicious lesions.
MRI technique
A closed high-speed MRI machine (General electric SIGNA 1.5 T) equipped with bilateral breast coils was used for 50 patients. For premenopausal patients, examination was conducted on day 6–13 of the menstrual cycle. Transverse, sagittal, and coronal plane localization scans were subsequently done. Patients were examined by fast spin echo (FSE) T1WI (TR 8.6ms, TE 4.7 ms), T2WI with short tau inversion recovery (STIR) (TR 5600ms, TE 59 ms), and DWI in transverse plan using a single excitation echo planar imaging sequence (TR 8400ms, TE 98 ms).
Dynamic contrast MRI was performed by two-dimensional fast spoiled gradient recalled echo with fat suppression in T1WI (TR 4.3ms, TE 1.3 ms). Five phase-dynamic images were acquired at 1,2, 3, 4, and 5 min. Dynamic analysis with generation of percent of enhancement vs. time curves was performed by positioning the area of interest for all identified lesions with a diameter greater than 5mm.
Lesions showing enhancement were assessed for pattern of enhancement, i.e., rim-, mass-, and non-mass-like enhancement. Diameter of the region of interest (ROI) was 5–25 mm2. A small ROI was allowed near the tumor edge to achieve greatest accuracy.ROI was placed over the enhancing lesion to obtain a dynamic curve pattern. Initial phase of enhancement occurred within 10 min of the contrast injection. Delayed phase was described as persistent, plateau, or washout phase.
Subsequently, a visual analysis of the diffusion-weighted images was conducted, and the apparent diffusion coefficient (ADC) measurement was performed. Images were classified based on the acquired diffusion images and ADC value. High-diffusion images and low ADC value were conducive to restricted diffusion. Moderate signal intensity in both diffusion images and ADC map were conducive to non-restricted diffusion.
ADC values were identified from b-800 DWI-MRI. If identification based on b-800 DWI-MRI was not possible, the lesion was evaluated by b-50 or b-400 images.
The enhancement percentages as well as the shapes of the curves (type I, II, and III curve) were examined. Breast imaging reporting and data system (BIRADS) was used in the study [26].
Heterogeneous mass or non-mass enhancement with ill-defined or irregular margins and observation of a type 3 curve in dynamic analysis were consistent with a malignant lesion diagnosis (BIRADS 4 and 5). Observation of a well-defined regular non-enhancing mass and type 1 curve in dynamic analysis were conducive to a benign lesion diagnosis (BIRADS 2).
No definite diagnosis was achieved for cases in-between these two types and those exhibiting type 2 curves. Hence, further assessment by biopsy and follow-up was recommended (BIRADS 3).
Pathological assessment
Patients were assessed using trucut biopsy to define the nature of suspicious lesion.
Statistical analysis
Data was analyzed in terms of range and mean + standard deviation. Student’s t-test was used for comparison of the data. Further, p-values of < 0.05 were considered statistically significant. Accuracy, sensitivity, specificity, positive predictive value, and negative predictive value were determined.
A true positive was defined as BI-RADS >4and proven as recurrence on pathological evaluation, while a false positive was defined as the same BI-RADS but proven to be a benign lesion on pathological evaluation
A false negative was defined as BI-RADS < 3 and proven as recurrence on histopathology, while a true negative was defined as the same BI-RADS but proven as a benign lesion on histopathology.