Differentiation of Malignancy and Idiopathic Granulomatous Mastitis Presenting as Non-mass Lesions on MRI: Radiological, Clinical, Radiomics, and Clinical-Radiomics Models.

RATIONALE AND OBJECTIVES: To investigate the effectiveness of machine learning-based clinical, radiomics, and combined models in differentiating idiopathic granulomatous mastitis (IGM) from malignancy, both presenting as non-mass enhancement (NME) lesions on magnetic resonance imaging (MRI), and to compare these models with radiological evaluation. MATERIAL AND METHODS: A total of 178 patients (69 IGM and 109 breast cancer patients) with NME on breast MRI evaluated between March 2018 and April 2022, were included in this two-center study. Age, skin changes, presence of fistula, and abscess were recorded from hospital records. Two experienced radiologists evaluated MRI images according to the breast imaging reporting and data system 2013 lexicon. Lesions were segmented independently on T2-weighted, apparent diffusion coefficient, and post-contrast-T1-weighted sequences. Data were split into training and external testing sets. Machine learning models were built using Light GBM (light gradient-boosting machine). Radiological, clinical, radiomics, and clinical-radiomics models were created and compared. Decision curve analysis was performed. Quality of reporting and that of methodology were evaluated using CLEAR and METRICS tools. RESULTS: IGM group was younger (p = 0.014). Abscesses (p < 0.001), fistulas (p < 0.001), and skin changes (p < 0.001) were significantly more common in the IGM group. No significant difference was detected in terms of lesion size (p = 0.213). In the evaluation of NME, the lowest performance belonged to the radiologists' evaluation (AUC for training, 0.740; for testing, 0.737), while the highest AUC was achieved by the model developed by combined clinical and radiomics features (AUC for training, 0.979; for testing, 0.942). CONCLUSION: Our study has shown that the machine learning-based clinical-radiomics model might have the potential to accurately discriminate IGM and malignant lesions in evaluating NME areas.

Association between lesion enhancement and breast cancer in contrast-enhanced spectral mammography.

BACKGROUND: Contrast-enhanced spectral mammography (CESM) may help to determine the malignancy potential of lesions according to the degree of enhancement. PURPOSE: To investigate the correlation between the degree of contrast enhancement of the lesions in contrast-enhanced spectral mammography (CESM) and the final histopathological diagnosis in patients with BI-RADS 4 and 5 lesions. MATERIAL AND METHODS: CESM was performed in 128 patients who had BI-RADS 4 and 5 lesions on mammography and underwent histopathological examination. A total of 128 index lesions were scored using a 4-point scale regarding the degree of contrast enhancement (0 = no contrast enhancement, 1 = minimal, 2 = moderate, 3 = marked), a score of 2 and 3 was accepted as suggestive of malignancy. The study was approved in our institutional scientific committee. RESULTS: In total, 76 (59.4%) of the lesions had benign histopathological results, whereas 52 of them had malignant results. Contrast enhancement was not observed in 22.7% of the lesions while 24.2% had minimal enhancement, 18.8% had moderate enhancement, and 34.4% had marked enhancement in CESM. The sensitivity of the degree of contrast enhancement in CESM was 98.1%, when the specificity was 77.6%, positive predictive value was 75%, negative predictive value was 98.3%, and accuracy was 85.9%. CONCLUSION: This study demonstrated that the degree of contrast enhancement of the lesions in CESM may be used in daily practice with easily performing a visual scale in predicting the malignancy potential of the lesions.

Adenoid Cystic Carcinoma of the Breast: Multimodality Imaging Findings and Review of the Literature.

RATIONALE AND OBJECTIVES: Adenoid cystic carcinoma (ACC) of the breast is a rare type of breast cancer with favorable prognosis. There is limited data on the radiological findings of this rare tumor in literature. The aim of this study is to determine the most common imaging features and review the literature. MATERIALS AND METHODS: Pathological databases of seven institutions from 2009 to 2021 were retrospectively reviewed, and patients with a diagnosis of ACC of the breast were determined. Thirteen patients whose imaging studies could be recalled from the picture archiving systems (PACS) were included in the study. Clinical and pathological findings as well as follow-up data were recorded. Radiological findings were analyzed and categorized based on BI-RADS 5th edition. RESULTS: There were 16 mass lesions in 13 patients (two multifocal cases, one case with recurrence). Mammography demonstrated 14 masses, while ultrasound (US) demonstrated all. MRI was available in only seven cases, with eight masses. The most common findings were round or oval shape on all modalities (78.57%-93.75%). Other frequent findings were parallel orientation (81.25%), isoechoic or hyperechoic echogenicity (62.5%), high T2 signal (87.5%), restricted diffusion (71.43%), and homogeneous enhancement (62.5%). Mammography, US and MRI showed circumscribed margins resembling a benign lesion in 35.71%, 37.5% and 50% of the lesions respectively. Three patients had a cyst-like echogenicty on US. Half of the lesions were avascular on Doppler US (6/12) and half were soft (2/4) on strain elastography. Although there were benign features on all imaging modalities seperately, all lesions could be categorized as BI-RADS 4 or 5 when the findings were combined. However 9/16 masses were BI-RADS 4A, emphasizing the subtlety of the malignant features. CONCLUSION: ACC of the breast can present with findings resembling a benign lesion on different imaging modalities. Although combination of all imaging findings correctly indicated the suspicious nature of the lesions in all cases, final classification was BI-RADS 4A in most of them. Radiologists should be aware of the more frequent findings of ACC of the breast for early diagnosis. US findings of isoechoic or hyperechoic appearance, and cyst-like echogenicity have not been reported previously in literature.

Magnetic resonance imaging findings associated with recurrence in idiopathic granulomatous mastitis.

PURPOSE: To evaluate magnetic resonance imaging (MRI) findings related to recurrence of idiopathic granulomatous mastitis (IGM). METHODS: Demographic data [age, number of births, duration of lactation period, body mass index (BMI) and presence of recurrence] of 71 patients who were diagnosed with IGM were analyzed retrospectively. Characteristics of IGM (maximum width, location, involvement of the retroareolar region, deep tissue, skin), fibroglandular density (FGD), background parenchymal enhancement (BPE), distribution and pattern of contrast enhancement, presence of prepectoral edema, abscesses, fistulae, axillary lymphadenopathies on MRI and apparent diffusion coefficient (ADC) values from the pathological area were recorded. RESULTS: The recurrence rate in patients was 59% (42/71). We found a statistically significant relationship between recurrence and BPE (p = 0.028) and mean ADC (p = 0.035) values (for the cut-off of 1.00 × 10-3 mm2/s; sensitivity = 61.9%, specificity = 69%, AUC = 0.648). However, patients' age (p = 0.346), lactation period (p = 0.470), number of births (p = 0.774), BMI (p = 0.630) maximum width of the area of enhancement (p = 0.112), involvement of the retroareolar region (p = 0.290), deep tissue (p = 0.285), skin (p = 0.230), distribution (p = 0.857) and enhancement pattern (p = 0.157), presence of prepectoral edema (p = 0.094), abscesses (p = 0.441), fistulae (p = 0.809), lymphadenopathies (p = 0.571), and FGT (p = 0.098) were not significantly associated with recurrence. CONCLUSION: Our results revealed that recurrent IGM patients showed high BPE and lower mean ADC values. We think that high BPE and low mean ADC (<1.00 × 10-3 mm2/s) on MRI at the diagnosis stage may be a sign of possible future recurrence, and it will be beneficial to follow the patients more closely and arrange the treatment algorithms accordingly.