Dataset Information

Deep Learning-Based Radiomics of B-Mode Ultrasonography and Shear-Wave Elastography: Improved Performance in Breast Mass Classification.

ABSTRACT: Objective:Shear-wave elastography (SWE) can improve the diagnostic specificity of the B-model ultrasonography (US) in breast cancer. However, whether deep learning-based radiomics signatures based on the B-mode US (B-US-RS) or SWE (SWE-RS) could further improve the diagnostic performance remains to be investigated. We aimed to develop the B-US-RS and SWE-RS and determine their performances in classifying breast masses. Materials and Methods:This retrospective study included 291 women (mean age ± standard deviation, 40.9 ± 12.3 years) from two centers who had US-visible solid breast masses and underwent biopsy and/or surgical resection between June 2015 and July 2017. B-mode US and SWE images of the 198 masses in 198 patients (training cohort) from center 1 were segmented, respectively, to construct B-US-RS and SWE-RS using the least absolute shrinkage and selection operator regression and tested in an independent validation cohort of 65 masses in 65 patients from center 1 and in an external validation cohort of 28 masses in 28 patients from center 2. The performances of B-US-RS and SWE-RS were assessed using receiver operating characteristic (ROC) analysis and compared with that of radiologist assessment [Breast Imaging Reporting and Data System (BI-RADS)] and quantitative SWE parameters [maximum elasticity (E max), mean elasticity (E mean), elasticity ratio (E ratio), and elastic modulus standard deviation (E SD)] by using the McNemar test. Results:The single best-performing quantitative SWE parameter, E max, had a higher specificity than BI-RADS assessment in the training and independent validation cohorts (P < 0.001 for both). The areas under the ROC curves (AUCs) of B-US-RS and SWE-RS both were 0.99 (95% CI = 0.99-1.00) in the training cohort, 1.00 (95% CI = 1.00-1.00) in the independent validation cohort, and 1.00 (95% CI = 1.00-1.00) in the external validation cohort. The specificities of B-US-RS and SWE-RS were higher than that of E max in the training (P < 0.001 for both) and independent validation cohorts (P = 0.02 for both). Conclusion:The B-US-RS and SWE-RS outperformed the quantitative SWE parameters and BI-RADS assessment for classifying breast masses. The integration of the deep learning-based radiomics approach would help improve the classification ability of B-mode US and SWE for breast masses.

SUBMITTER: Zhang X

PROVIDER: S-EPMC7485397 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep Learning-Based Radiomics of B-Mode Ultrasonography and Shear-Wave Elastography: Improved Performance in Breast Mass Classification.

Zhang Xiang X Liang Ming M Yang Zehong Z Zheng Chushan C Wu Jiayi J Ou Bing B Li Haojiang H Wu Xiaoyan X Luo Baoming B Shen Jun J

Frontiers in oncology 20200828

<h4>Objective</h4>Shear-wave elastography (SWE) can improve the diagnostic specificity of the B-model ultrasonography (US) in breast cancer. However, whether deep learning-based radiomics signatures based on the B-mode US (B-US-RS) or SWE (SWE-RS) could further improve the diagnostic performance remains to be investigated. We aimed to develop the B-US-RS and SWE-RS and determine their performances in classifying breast masses.<h4>Materials and methods</h4>This retrospective study included 291 wo ...[more]

PMID: 32984032

Similar Datasets

Project description:ObjectivesThe aim of this study was to assess the potential of machine learning based on B-mode, shear-wave elastography (SWE), and dynamic contrast-enhanced ultrasound (DCE-US) radiomics for the localization of prostate cancer (PCa) lesions using transrectal ultrasound.MethodsThis study was approved by the institutional review board and comprised 50 men with biopsy-confirmed PCa that were referred for radical prostatectomy. Prior to surgery, patients received transrectal ultrasound (TRUS), SWE, and DCE-US for three imaging planes. The images were automatically segmented and registered. First, model-based features related to contrast perfusion and dispersion were extracted from the DCE-US videos. Subsequently, radiomics were retrieved from all modalities. Machine learning was applied through a random forest classification algorithm, using the co-registered histopathology from the radical prostatectomy specimens as a reference to draw benign and malignant regions of interest. To avoid overfitting, the performance of the multiparametric classifier was assessed through leave-one-patient-out cross-validation.ResultsThe multiparametric classifier reached a region-wise area under the receiver operating characteristics curve (ROC-AUC) of 0.75 and 0.90 for PCa and Gleason > 3 + 4 significant PCa, respectively, thereby outperforming the best-performing single parameter (i.e., contrast velocity) yielding ROC-AUCs of 0.69 and 0.76, respectively. Machine learning revealed that combinations between perfusion-, dispersion-, and elasticity-related features were favored.ConclusionsIn this paper, technical feasibility of multiparametric machine learning to improve upon single US modalities for the localization of PCa has been demonstrated. Extended datasets for training and testing may establish the clinical value of automatic multiparametric US classification in the early diagnosis of PCa.Key points• Combination of B-mode ultrasound, shear-wave elastography, and contrast ultrasound radiomics through machine learning is technically feasible. • Multiparametric ultrasound demonstrated a higher prostate cancer localization ability than single ultrasound modalities. • Computer-aided multiparametric ultrasound could help clinicians in biopsy targeting.

Project description:BackgroundBefore the role of shear wave elastography (SWE) and B mode ultrasound (US) in the diagnosis of different forms of idiopathic inflammatory myopathies (IIM) can be investigated, normative data is required. This study aimed to describe and then compare normative SWE and B mode ultrasound metrics of muscles in healthy controls and patients with IIM.MethodsTwenty nine healthy adult controls and 10 IIM patients (5 with inclusion body myositis and 5 with necrotising autoimmune myopathy) underwent a full clinical examination, laboratory investigations, SWE and US measurements of selected proximal and distal limb muscles. Shear wave speed (SWS) and multiple US domains [echogenicity, fascial thickness, muscle bulk and power Doppler (PD)] were measured in both groups.ResultsIn healthy controls (n = 29; mean age 46.60 ± 16.10; 44.8 % female), age was inversely correlated with SWS at the deltoid (stretch) (Rs. -0.40, p = 0.030) and PD score at the deltoid (rest) (Rs. -0.40, P = 0.032). Those ≥ 50 years old had a lower SWS at the deltoid (stretch) compared to the < 50 year group (2.92 m/s vs. 2.40 m/s, P = 0.032). Age correlated with increased echogenicity in the flexor digitorum profundus (Rs. 0.38, P = 0.045). Females had a smaller muscle bulk in the deltoid (P = 0.022). Body mass index (BMI) was inversely associated with SWS in the deltoid (stretch) (Rs - 0.45, P = 0.026), and positively correlated with echogenicity in the deltoid (Rs. 0.69, P = 0.026). In patients ≥50 years of age, patients with IIM (mean age 61.00 ± 8.18; females 20.0 %) had a higher proportion of abnormal echogenicity scores at the flexor digitorum profundus (FDP) (40.00 % vs. 14.30 %, P = 0.022) and tibialis anterior (TA) (80.00 % vs. 28.60 %, P = 0.004). Fascial thickness was lower in the FDP (0.63mm vs. 0.50mm, p = 0.012) and TA (0.58mm vs. 0.45mm, P = 0.001).ConclusionsOur findings suggest there is scope for US techniques to be useful for diagnostic screening of affected muscles in patients with IIM, especially in those with suspected inclusion body myositis or necrotising autoimmune myopathy. We provide normative data for future studies into SWE and US techniques in skeletal muscle. The differences between IIM patients and controls warrant further study in a broader IIM patient cohort.

Project description:Objectives:The aim of the study was to compare elastographic means in parathyroid adenomas, using shear wave elastography and strain elastography. Methods:This prospective study examined 20 consecutive patients diagnosed with primary hyperparathyroidism and parathyroid adenoma, confirmed by biochemical assay, technetium-99 sestamibi scintigraphy, and pathology report, after parathyroid surgery. All patients were examined on conventional 2B ultrasound, 2D shear wave elastography, and strain elastography. We determined using 2D shear wave elastography (SWE) the elasticity index (EI) in parathyroid adenoma, thyroid parenchyma, and surrounding muscle and examined using strain elastography the parathyroid adenoma, and determined the strain ratio with the thyroid tissue and muscle tissue. Results:All patients had positive sestamibi scintigraphy and underwent surgery, with confirmation of parathyroid adenoma in all cases. The mean parathormone (PTH) value before surgery was 153.29?pg/ml (36.5, 464.8) and serum calcium concentration was 10.5?mg/dl (9, 11.5). We compared using 2D-SWE and strain elastography parathyroid adenoma with thyroid tissue and with surrounding muscle. The mean EI measured by SWE in parathyroid adenoma was 4.74?±?2.74?kPa and in thyroid parenchyma was 11.718?±?4.206?kPa (mean difference?=?6.978?kPa, p < 0.001), and the mean EI value in muscle tissue was 16.362?±?3.829?kPa (mean difference?=?11.622, p < 0.001). Using ROC analysis, we found that an EI below 7?kPa correctly identifies parathyroid tissue. We evaluated parathyroid adenomas using strain elastography by color mapping and strain ratio as a semiquantitative measurement; however, we could not find any statistical correlation comparing the strain ratio obtained from the parathyroid adenoma with the thyroid tissue (p=0.485). Conclusion:Ultrasound elastography is a helpful tool in identifying parathyroid adenomas. A cutoff value below 7?kPa can be used in 2D-SWE. Color maps in strain elastography without adding strain ratio can be used, parathyroid adenoma being identified as score 1 in the Rago criteria.

Dataset Information

Deep Learning-Based Radiomics of B-Mode Ultrasonography and Shear-Wave Elastography: Improved Performance in Breast Mass Classification.

Publications

Deep Learning-Based Radiomics of B-Mode Ultrasonography and Shear-Wave Elastography: Improved Performance in Breast Mass Classification.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets