AI-Based Cancer Detection Model for Contrast-Enhanced Mammography

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.

Added value of contrast-enhanced spectral mammogram in assessment of suspicious microcalcification and grading of DCIS

Background

Breast microcalcifications are one of the most difficult mammographic findings to assess. The purpose of this study is to assess the ability of contrast-enhanced spectral mammography in the assessment of suspicious microcalcification and in predicting the grade of DCIS.

Methods

Three hundred and forty cases with suspicious microcalcification were reviewed in this study. We excluded 160 cases associated with masses. We enrolled 180 cases for analysis of suspicious microcalcification on mammograms with no underlying masses. We reviewed the microcalcification for their morphology, distribution, and associated pathological enhancement according to BI-RADS lexicon with pathology results reviewed and classified into benign and malignant which subdivided into low, intermediate, or high-grade DCIS or invasive carcinoma.

Results

Three hundred and forty cases with suspicious microcalcification were reviewed in this study. We excluded 160 cases associated with masses. Forty-five of 180 cases were benign, and 135/180 cases were malignant. Twenty-five of 135 cases were diagnosed as invasive breast carcinomas while 110/135 were ductal carcinoma in situ. From the latter, 110 patients with DCIS, 22/110 cases were low grade, 11/110 cases were intermediate grade, and 77/110 cases were high grade (44 with micro-invasion). A total of 25 invasive carcinomas showed pathological non-mass enhancement, 76/77 cases of high-grade DCIS, and 6/11 cases of intermediate-grade DCIS. No abnormal enhancement appeared with benign entities, low-grade DCIS, and 5/11 cases of intermediate DCIS. The diagnostic performance of CESM in anticipation of high grade in DCIS patients was sensitivity of 98%, specificity of 81.8%, and accuracy of 93.1%. CESM sensitivity, specificity, and accuracy in prediction of invasiveness or high-grade DCIS were 98.5%, 81.8%, and 87.5%, respectively.

Conclusion

CESM can provide a fundamental contribution in the evaluation of suspicious microcalcification as high-grade DCIS or invasive component can present by non-mass enhancement, but enhancement paucity is favorable to diagnose benign lesion or non-invasive/low-grade DCIS.

Characterization of breast masses: a comparative study between automated breast ultrasound (ABUS) and digital breast tomosynthesis (DBT)

Background

Automated three-dimensional (3D) breast ultrasound (US) systems and breast tomosynthesis are promising breast imaging modalities. The study aims to compare the diagnostic indices of the 3D imaging techniques: digital breast tomosynthesis (DBT) and automated breast ultrasound (ABUS) in the characterization of breast masses. This prospective study included 32 women with breast masses either detected by means of clinical examination or with the mammographic exam. All of them have been subjected to tomosynthesis and automatic breast ultrasound examinations. The images from both modalities were then analyzed one at a time by two experienced representative radiologists in consensus. Results were compared to each other and to pathology and follow-up of typically benign findings

Results

The masses statistically evaluated in this study were 37 in number, among which 16 were benign and 21 were malignant. The sensitivity and specificity of tomosynthesis in the characterization of breast masses were 100% and 81.25%, respectively, while automated breast ultrasound were 100% and 75%, respectively.

Conclusion

Tomosynthesis and automated breast ultrasound are two promising modalities in breast imaging. Their diagnostic indices in this study were very close to one another; therefore, they can be used as an adjunct modality to mammography for early diagnosis of breast cancer.

Quantitative mathematical objective evaluation of contrast-enhanced spectral mammogram in the assessment of response to neoadjuvant chemotherapy and prediction of residual disease in breast cancer

Background

The aim of the study is to initiate a new quantitative mathematical objective tool for evaluation of response to neoadjuvant chemotherapy (NAC) and prediction of residual disease in breast cancer using contrast-enhanced spectral mammography (CESM). Forty-two breast cancer patients scheduled for receiving NAC were included. All patients underwent two CESM examinations: pre and post NAC. To assess the response to neoadjuvant chemotherapy, we used a mathematical image analysis software that can calculate the difference in the intensity of enhancement between the pre and post neoadjuvant contrast images (MATLAB and Simulink) (Release 2013b). The proposed technique used the pre and post neoadjuvant contrast images as inputs. The technique consists of three main steps: (1) preprocessing, (2) extracting the region of interest (ROI), and (3) assessment of the response to chemotherapy by measuring the percentage of change in the intensity of enhancement of malignant lesions in the pre and post neoadjuvant CESM studies using a quantitative mathematical technique. This technique depends on the analysis of number of pixels included within the ROI. We compared this technique with the currently used method of evaluation: RECIST 1.1 (response evaluation criteria in solid tumors 1.1) and using another combined response evaluation approach using both RECIST 1.1 in addition to a subjective visual evaluation. Results were then correlated with the postoperative pathology evaluation using Miller–Payne grades. For statistical evaluation, patients were classified into responders and non-responders in all evaluation methods.

Results

According to the Miller–Payne criteria, 39/42 (92.9%) of the participants were responders (Miller–Payne grades III, IV, and IV) and 3/42 (7.1%) were non-responders (Miller–Payne grades I and II). Using the proposed technique, 39/39 (100%) were responders in comparison to 38/39 patients (97.4%) using the combined criteria and 34/39 (87.2%) using the RECIST 1.1 evaluation. The calculated correlation coefficient of the proposed quantitative objective mathematical technique, RECIST 1.1 criteria, and the combined method was 0.89, 0.59, and 0.69 respectively. With classification of patients into responder and non-responders, the objective mathematical evaluation showed higher sensitivity, positive and negative predictive values, and overall accuracy (100%, 97.5%, 100%, and 85.7% respectively) compared to RECIST 1.1 evaluation (87.2%, 97.1%, 28.6%, and 54.8% respectively) and the combined response method (97.4%, 97.4%, 66.7%, and 85.7% respectively).

Conclusion

Quantitative mathematical objective evaluation using CESM images allows objective quantitative and accurate evaluation of the response of breast cancer to chemotherapy and is recommended as an alternative to the subjective techniques as a part of the pre-operative workup.

Diffusion-weighted MR imaging diagnostic merits in the post-therapeutic assessment of musculoskeletal soft tissue sarcoma

Background

The purpose of the study is assessing the diagnostic merits of diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping in evaluating tumor response to chemo-radiotherapy. The study included 36 patients with soft tissue sarcoma, who received chemo/radiotherapy. Tumor longest dimension according to response evaluation criteria in solid tumors 1.1 (RECIST 1.1), the longest dimension of the contrast-enhanced portion of the tumor according to modified response evaluation criteria in solid tumors: (mRECIST), the tumor volume (VOL) (cm3), and DWI with ADC values were recorded.

Results

ADC values in the non-progressive group were higher than those of the progressive group after neoadjuvant treatment (1.63 ± 0.42 vs. 1.24 ± 0.35) with (p < 0.005). ADC variations in the non-progressive group were higher than those of the progressive group (27.09 ± 48.09 vs. − 3.08 ± 23.5)% with (p < 0.05). ADC values after neoadjuvant treatment were negatively related to tumor volume variations (VOL%) after neoadjuvant treatment. ADC variations (ADC%) were inversely correlated with morphologic changes, regardless of the effectiveness of anticancer therapy expressed as changes in tumor size based on (RECIST, mRECIST, and three-dimensional volumetric assessment). An increase in the ADC value was not always associated with a reduction of tumor volume.

Conclusion

Quantitative DW imaging after neoadjuvant therapy provides added value in determining treatment response in soft tissue sarcomas. Therapeutic response to neoadjuvant therapy can be underestimated using RECIST 1.1; therefore, the mRECIST should also be considered.

Adding the merits of contrast to the ease of mammography; can we highlight what’s behind breast asymmetries?

Background

Breast symmetry is one of the major things that radiologists assess when looking at mammograms and is one of the most challenging mammographic findings to evaluate. Contrast-enhanced spectral mammography (CESM) is an emerging mammography technique that has shown comparable sensitivity and specificity to MRI. The purpose of this study is to assess the value of CESM in characterization of breast asymmetries (BAs) and if it should be incorporated in its diagnostic work-up.

Results

Three hundred sixty-five patients with mean age of 47 years were included in the study. CESM was performed aiming for characterization of 380 suspicious or indeterminate breast asymmetries. Assessment of subtracted high-energy images (HEI) markedly improves the overall accuracy reaching 88.4%. Further improvement of the overall accuracy was achieved on combined assessment of the low-energy images (LEI), subtracted high-energy images (HEI), and ultrasound reaching 91.3%.

Conclusion

CESM is considered as a valuable complementary imaging tool considering the evaluation of breast asymmetries and should be incorporated in its diagnostic work-up in cases not resolved on an initial combined mammography and targeted ultrasound study especially in the presence of a heterogeneous dense breast parenchyma. Yet, this may be hindered in the presence of inflammatory signs because of the overlapping imaging criteria.

Is MRI diffusion-weighted imaging a reliable tool for the diagnosis and posttherapeutic follow-up of extremity soft tissue neoplasms?

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.

Can MRI diffusion-weighted imaging identify postoperative residual/recurrent soft-tissue sarcomas?

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.