Added Value of Quantitative Ultrasound and Machine Learning in BI-RADS 4-5 Assessment of Solid Breast Lesions.

The purpose of this study was to evaluate various combinations of 13 features based on shear wave elasticity (SWE), statistical and spectral backscatter properties of tissues, along with the Breast Imaging Reporting and Data System (BI-RADS), for classification of solid breast lesions at ultrasonography by means of random forests. One hundred and three women with 103 suspicious solid breast lesions (BI-RADS categories 4-5) were enrolled. Before biopsy, additional SWE images and a cine sequence of ultrasound images were obtained. The contours of lesions were delineated, and parametric maps of the homodyned-K distribution were computed on three regions: intra-tumoral, supra-tumoral and infra-tumoral zones. Maximum elasticity and total attenuation coefficient were also extracted. Random forests yielded receiver operating characteristic (ROC) curves for various combinations of features. Adding BI-RADS category improved the classification performance of other features. The best result was an area under the ROC curve of 0.97, with 75.9% specificity at 98% sensitivity.

The Added Value of Statistical Modeling of Backscatter Properties in the Management of Breast Lesions at US.

PURPOSE: To develop a classification method based on the statistical backscatter properties of tissues that can be used as an ancillary tool to the usual Breast Imaging Reporting and Data System (BI-RADS) classification for solid breast lesions identified at ultrasonography (US). MATERIALS AND METHODS: This study received institutional review board approval, and all subjects provided informed consent. Eighty-nine women (mean age, 50 years; age range, 22-82 years) with 96 indeterminate solid breast lesions (BI-RADS category 4-5; mean size, 13.2 mm; range, 2.6-44.7 mm) were enrolled. Prior to biopsy, additional radiofrequency US images were obtained, and a 3-second cine sequence was used. The research data were analyzed at a later time and were not used to modify patient management decisions. The lesions were segmented manually, and parameters of the homodyned K distribution (α, k, and µn values) were extracted for three regions: the intratumoral zone, a 3-mm supratumoral zone, and a 5-mm infratumoral zone. The Mann-Whitney rank sum test was used to identify parameters with the best discriminating value, yielding intratumoral α, supratumoral k, and infratumoral µn values. RESULTS: The 96 lesions were classified as follows: 48 BI-RADS category 4A lesions, 16 BI-RADS category 4B lesions, seven BI-RADS category 4C lesions, and 25 BI-RADS category 5 lesions. There were 24 cancers (25%). The area under the receiver operating characteristic curve was 0.76 (95% confidence interval: 0.65, 0.86). Overall, 24% of biopsies (in 17 of 72 lesions) could have been spared. By limiting analysis to lesions with a lower likelihood of malignancy (BI-RADS category 4A-4B), this percentage increased to 26% (16 of 62 lesions). Among benign lesions, the model was used to correctly classify 10 of 38 fibroadenomas (26%) and three of seven stromal fibroses (43%). CONCLUSION: The statistical model performs well in the classification of solid breast lesions at US, with the potential of preventing one in four biopsies without missing any malignancy.

Ultrasound monitoring of RBC aggregation as a real-time marker of the inflammatory response in a cardiopulmonary bypass swine model.

OBJECTIVES: In many pathological conditions, including high-risk surgery, the severity of the inflammatory response is related to the patient outcome. However, determining the patient inflammatory state presents difficulties, as markers are obtained intermittently through blood testing with long delay. RBC aggregation is a surrogate marker of inflammation that can be quantified with the ultrasound Structure Factor Size and Attenuation Estimator. The latter is proposed as a real-time inflammation monitoring technique for patient care. DESIGN: Ten swine underwent a 90-minute cardiopulmonary bypass, and surveillance was maintained during 120 minutes in the postbypass period. To promote the inflammatory reaction, lipopolysaccharide was administrated two times prior to surgery in six of those swine (lipopolysaccharide group). During the whole procedure, the Structure Factor Size and Attenuation Estimator cellular imaging method displayed a RBC aggregation index (W) computed from images acquired within the pump circuit and the femoral vein. Interleukin-6, interleukin-10, C-reactive protein, haptoglobin, immunoglobulin G, and fibrinogen concentrations were measured at specific periods. MAIN RESULTS: Compared with controls, the lipopolysaccharide group exhibited higher W within the pump circuit (p < 0.05). In the femoral vein, W was gradually amplified in the lipopolysaccharide group during cardiopulmonary bypass and the postbypass period (p < 0.05), whereas interleukin levels were higher in the lipopolysaccharide group but only at the end of cardiopulmonary bypass and beginning of postbypass (p < 0.05). CONCLUSIONS: Continuous RBC aggregation monitoring can characterize the evolving inflammatory response during and after cardiopulmonary bypass. The Structure Factor Size and Attenuation Estimator is proposed as a real-time noninvasive monitoring technique to anticipate inflammation-related complications during high-risk surgery or critical care situations. Because RBC aggregation promotes vascular resistance and thrombosis, W could also provide early information on vascular disorders in those clinical situations.