ABSTRACT
BACKGROUND: The recent development of deep neural network models for the analysis of breast images has been a breakthrough in computer-aided diagnostics (CAD). Contrast-enhanced mammography (CEM) is a recent mammography modality providing anatomical and functional imaging of the breast. Despite the clinical benefits it could bring, only a few research studies have been conducted around deep-learning (DL) based CAD for CEM, especially because the access to large databases is still limited. This study presents the development and evaluation of a CEM-CAD for enhancing lesion detection and breast classification. MATERIALS & METHODS: A deep learning enhanced cancer detection model based on a YOLO architecture has been optimized and trained on a large CEM dataset of 1673 patients (7443 images) with biopsy-proven lesions from various hospitals and acquisition systems. The evaluation was conducted using metrics derived from the free receiver operating characteristic (FROC) for the lesion detection and the receiver operating characteristic (ROC) to evaluate the overall breast classification performance. The performances were evaluated for different types of image input and for each patient background parenchymal enhancement (BPE) level. RESULTS: The optimized model achieved an area under the curve (AUROC) of 0.964 for breast classification. Using both low-energy and recombined image as inputs for the DL model shows greater performance than using only the recombined image. For the lesion detection, the model was able to detect 90% of all cancers with a false positive (non-cancer) rate of 0.128 per image. This study demonstrates a high impact of BPE on classification and detection performance. CONCLUSION: The developed CEM CAD outperforms previously published papers and its performance is comparable to radiologist-reported classification and detection capability.
ABSTRACT
PURPOSE: The aim of this study was to evaluate the benefit of using quantitative diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping in the initial diagnosis and post-therapeutic follow-up of extremity soft tissue masses. PATIENTS AND METHODS: This study included 90 patients with extremity soft tissue masses. The DWI was obtained with 3 b values, including 0, 400, and 800 s/mm2. Calculation of the ADC value of the lesion was done by placing the region of interest (ROI) to include the largest area of the lesion. ADC values were compared with the histopathology. Eighteen patients had posttherapeutic magnetic resonance imaging (MRI). RESULTS: Benign masses, fibromatosis, and malignant soft tissue masses had mean ADC values of 1.18 ± 1.0191 × 10-3 mm2/s; 1.31 ± 0.245 × 10-3 mm2/sec; and 1.3 ± 0.7 × 10-3 mm2/s, respectively. Myxomatous malignant masses had an ADC value of 2.6 ± 0.55 × 10-3 mm2/s, while nonmyxomatous malignant masses had an ADC value of 1.1 ± 0.8 × 10-3 mm2/s. ADC cutoff value between benign and non-benign (including malignant and locally aggressive masses) was 0.6 × 10-3 mm2/sec with 98.3% sensitivity and 50% specificity (P = 0.5123). The statistical difference between malignant soft tissue masses (mean ADC 1.309 ± 0.723 × 10-3 mm2/s) and fibromatosis masses (mean ADC value 1.31 ± 0.245 × 10-3 mm2/s) using a comparative T-test proved to be of poor significance level (P value ~ 0.9757). Nine patients with soft tissue sarcomas (STSs) had pre and post-therapeutic MRI examinations showing a mean increase of the recorded ADC values by about 0.28 × 10-3 mm2/s in the post-therapy study as compared with the recorded initial pretreatment values. Analysis of the post-therapy follow-up studies of fibromatosis showed that lesions with favorable response to chemotherapy or radiotherapy (8/12) exhibited significantly lower ADC values than those showing progressive disease course. CONCLUSION: DWI with ADC mapping of extremity soft tissue tumors are so complicated that they alone may not be of much value in differentiating between benign and malignant tumors; however, it can be used as a tool for monitoring response to treatment.
ABSTRACT
PURPOSE: The aim of this study was to evaluate contrast-enhanced magnetic resonance imaging (CE-MRI) and quantitative diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping in the detection of recurrent/residual postoperative soft tissue sarcomas. MATERIALS AND METHODS: This study included 36 patients; 27 patients had postoperative recurrent/residual soft tissue sarcomas and 9 patients had postoperative and treatment-related changes (inflammation/fibrosis). The DWI was obtained with 3 b values including 0, 400, and 800 s/mm2. Calculation of the ADC value of the lesion was done via placing the region of interest (ROI) to include the largest area of the lesion. ADC values were compared to histopathology. RESULTS: Our results showed that including CE-MRI improved the diagnostic accuracy and sensitivity in recurrence detection compared to conventional non-enhanced sequences. However, it showed low specificity (55.56%) with a high false-positive rate that may lead to an unnecessary biopsy of a mass such as region of postoperative scar tissue. CONCLUSION: The joint use of gadolinium-enhanced MRI and quantitative DWI with ADC mapping offer added value in the detection of recurrent/residual postoperative soft tissue sarcoma. This combined use increased both the diagnostic sensitivity and specificity with a cut-off average ADC value for detecting nonmyxoid recurrent/residual lesions ≤1.3 × 10-3 mm2/s (100% specificity and 90.48% sensitivity). Our results showed limited value of DWI with ADC mapping in assessing myxoid sarcomatous tumor recurrences.