Abstract Ultrasound-guided hookwire localization was initially introduced to facilitate the excision of nonpalpable breast lesions by guiding surgical exploration, thereby reducing operative time and morbidity. The same technique has since found utility in a range of other applications outside breast and can be useful within the musculoskeletal system. Despite this, there remains limited literature with respect to its technical aspects and practical utility. We describe our technique and a series of preoperative ultrasound-guided wire localizations in the musculoskeletal system to assist surgical excision of 4 soft tissue masses.
Abstract Ultrasound is an important imaging modality for the detection and characterization of breast cancer. Though consistently shown to detect mammographically occult cancers, especially in women with dense breasts, breast ultrasound has been noted to have high false-positive rates. In this work, we present an artificial intelligence (AI) system that achieves radiologist-level accuracy in identifying breast cancer in ultrasound images. To develop and validate this system, we curated a dataset consisting of 288,767 ultrasound exams from 143,203 patients examined at NYU Langone Health, between 2012 and 2019. On a test set consisting of 44,755 exams, the AI system achieved an area under the receiver operating characteristic curve (AUROC) of 0.976. In a reader study, the AI system achieved a higher AUROC than the average of ten board-certified breast radiologists (AUROC: 0.962 AI, 0.924±0.02 radiologists). With the help of the AI, radiologists decreased their false positive rates by 37.4% and reduced the number of requested biopsies by 27.8%, while maintaining the same level of sensitivity. To confirm its generalizability, we evaluated our system on an independent external test dataset where it achieved an AUROC of 0.911. This highlights the potential of AI in improving the accuracy, consistency, and efficiency of breast ultrasound diagnosis worldwide.
Abstract OBJECTIVES The purpose of this study was to review the outcomes of surgical treatment in patients with aorto-left ventricular tunnel and to investigate what kind of patient cohort is more likely to have adverse events. METHODS Twenty-one patients with a median age of 6.58 [interquartile range (IQR) 4.17–24.50] years who received surgical treatment of aorto-left ventricular tunnel from March 2002 to December 2019 were reviewed. The median follow-up time was 64.50 (IQR 25.15–120.50) months. Clinical characteristics, surgical methods and follow-up outcomes were summarized in separate groups of patients with or without preoperative aortic valve (AoV) issues. Composite adverse events were defined as death or requirement of reoperation. Time-related analysis of freedom from death and requirement of reoperation was performed with the Kaplan–Meier method. RESULTS The average tunnel size was 8.68 (standard deviation: 3.62) mm. The most common and the most important associated lesions were AoV lesions. Tunnels in 20 patients were closed with direct sutures or a patch. For 1 patient with an irreparable AoV, the tunnel was cut open simultaneously with aortic valve replacement and aortic root plasty. In the group of patients with preoperative AoV issues, 4 patients received aortic valve replacement with mechanical prosthetic valves and 6 patients received aortic valvuloplasty repair. The follow-up outcomes differed significantly between groups (the incidence of death was 15.38% and the incidence of requiring a reoperation was 46.15% in patients with preoperative AoV issues). In the group without preoperative AoV issues, there were no deaths and no reoperations (Fisher’s exact test; P = 0.018). The probability of freedom from death and of the requirement for reoperation between the 2 groups was not significantly different (log-rank, P = 0.09). Overall, the estimated probability of freedom from death and requirement of reoperation was 77.30% (standard error: 10.20%) [95% confidence interval (CI): 49.53–91.00] at 5 years, 67.64% (standard error: 12.70%) (95% CI: 36.71–85.84) at 10 years. CONCLUSIONS Patients with aorto-left ventricular tunnel with preoperative AoV issues are more prone to die or to require a reoperation. In contrast, patients without preoperative AoV issues can be free from death or reoperation for a longer period of time. Patients with preoperative AoV issues need much stricter postoperative long-term echocardiographic follow-up.
The purpose of this study was to determine (a) the frequency of apocrine metaplasia (ApoM) found on MR core biopsy of suspicious findings, and (b) to determine if there are specific MR imaging features that might obviate the need for biopsy. This HIPAA-compliant retrospective study was performed under IRB exemption for quality assurance studies. Patient demographics, MR imaging features, and pathology were reviewed. Breast lesions which underwent MR-guided biopsy, yielding ApoM on pathology analysis were included. Retrospective review of MR imaging features of these lesions was performed by two radiologists blinded to pathology results except for the presence of ApoM. Imaging features on MR assessed included location, size, morphology, T1 and T2 signals, and enhancement kinetics. Full pathology results were subsequently reviewed during data analysis. The pathology slides and imaging was subsequently reviewed by two fellowship trained radiologists and a breast pathologist to categorize the finding of ApoM into target lesion (imaging corresponds to size of lesion on pathology) versus incidental lesion. Target lesion characteristics were assessed to determine specific MRI features of ApoM. Between January 2011 to November 2012, 155 distinct breast lesions suspicious for malignancy successfully underwent MR-guided biopsy. Of the 155 lesions biopsied, 123 (79%) were benign and 32 (21%) were malignant. Of the 123 benign biopsies, ApoM was found in 57 (46%), of which 35 (61%) had no associated atypia and 22 (39%) had associated atypia. Of the 32 malignant biopsies, three (9%) had associated ApoM (DCIS in two cases and DCIS/LCIS in one case). Of the 60 cases with ApoM, only 11 (18.3%) were target lesions and 49 were incidental lesions (81.7%). Of the 60 cases with ApoM, 35 (58%) were masses (average size 0.8 cm for both with or without atypia) and 25 (42%) were nonmass enhancement (NME) (average size 2.1 cm with and 1.0 cm without atypia). Only five (14%) of 35 masses demonstrated spiculated margins, of which four were associated with atypia (80%). Of 22 lesions with atypia or other high-risk lesion, 14 (64%) were masses, most commonly with irregular margins (64%). Of the 12 T2 hyperintense lesions, only two (1.7)% had associated atypia or high-risk lesion, and none were associated with malignancy. Of the 11 target lesions, seven were T2 hyperintense. Enhancement kinetics were variable: 30 (50%) showed mixed persistent and plateau kinetics, eight (13%) persistent delayed enhancement, 10 (17%) plateau kinetics, four (7%) washout kinetics, and eight (13%) were below threshold for kinetic analysis. ApoM is a common benign pathologic result at MR-guided core biopsy for both masses and NME accounting for 39% of all biopsy results in this series. Although there is considerable variability in imaging characteristics on MR, our results suggest biopsy may be safely obviated for lesions that are subcentimeter T2 hyperintense areas of NME and short term follow-up imaging may be a reasonable alternative for these lesions.
Purpose To compare a novel multicoil compressed sensing technique with flexible temporal resolution, golden‐angle radial sparse parallel (GRASP), to conventional fat‐suppressed spoiled three‐dimensional (3D) gradient‐echo (volumetric interpolated breath‐hold examination, VIBE) MRI in evaluating the conspicuity of benign and malignant breast lesions. Materials and Methods Between March and August 2015, 121 women (24–84 years; mean, 49.7 years) with 180 biopsy‐proven benign and malignant lesions were imaged consecutively at 3.0 Tesla in a dynamic contrast‐enhanced (DCE) MRI exam using sagittal T1‐weighted fat‐suppressed 3D VIBE in this Health Insurance Portability and Accountability Act‐compliant, retrospective study. Subjects underwent MRI‐guided breast biopsy (mean, 13 days [1–95 days]) using GRASP DCE‐MRI, a fat‐suppressed radial “stack‐of‐stars” 3D FLASH sequence with golden‐angle ordering. Three readers independently evaluated breast lesions on both sequences. Statistical analysis included mixed models with generalized estimating equations, kappa‐weighted coefficients and Fisher's exact test. Results All lesions demonstrated good conspicuity on VIBE and GRASP sequences (4.28 ± 0.81 versus 3.65 ± 1.22), with no significant difference in lesion detection ( P = 0.248). VIBE had slightly higher lesion conspicuity than GRASP for all lesions, with VIBE 12.6% (0.63/5.0) more conspicuous ( P < 0.001). Masses and nonmass enhancement (NME) were more conspicuous on VIBE ( P < 0.001), with a larger difference for NME (14.2% versus 9.4% more conspicuous). Malignant lesions were more conspicuous than benign lesions ( P < 0.001) on both sequences. Conclusion GRASP DCE‐MRI, a multicoil compressed sensing technique with high spatial resolution and flexible temporal resolution, has near‐comparable performance to conventional VIBE imaging for breast lesion evaluation. Level of Evidence : 3 Technical Efficacy : Stage 3 J. MAGN. RESON. IMAGING 2017;45:1746–1752
Objective The study was designed to explore the evolution of non-small cell lung cancer (NSCLC) management in the last 20 years. Methods The top 100 most-cited papers on NSCLC treatment were retrieved from the Web of Science Core Collection database. R and VOSviewer were used to extract bibliographic information, including the year of publication, countries/regions, institutions, authors, journals, keywords, impact factor, and total citations. The topic and type of papers were checked independently by authors. Bibliometric analysis was conducted and visualized with R, CiteSpace, Excel and VOSviewer to identify output dynamics, research forces, topics, hotspots, and frontiers in the field. Results The average citation of each retrieved top 100 most-cited NSCLC management papers was 1,725 (range: 615-7,340). Fifty-seven corresponding authors were from the United States. This country contributed the most papers (n=76), followed by Germany (n=34), France (n=33), and South Korea (n=32). The top contributors were Paz-Ares L. (n=12) and Reck M. (n=12). The Memorial Sloan Kettering Cancer Center published the largest number of papers (n=20). There were two significant citation paths, indicating publications in medicine/medical/clinical journals primarily cited journals in molecular/biology/genetics fields, partly cited health/nursing/medicine fields. Top-cited papers mainly came from the New England Journal of Medicine (n=33, citations=80,427), followed closely by the Journal of Clinical Oncology (n=28, citations=32,408). “Chemotherapy” (n=36) was the keyword with the greatest frequency of co-occurrence. “Open-label” was the keyword with the strongest burst strength (=4.01), followed by “nivolumab” (=3.85), “blockade” (=2.86), and “efficacy” (=2.85). Conclusions The United States as a nation and the Memorial Sloan Kettering Cancer Center as an institute contributed the most to this field. The New England Journal of Medicine is the most eye-catching journal. Hotspots of NSCLC management have almost undergone an evolution from chemotherapy and radiotherapy to targeted therapy to immunotherapy. Molecular/biological/genetic fields become the main research base for NSCLC treatment. Immunotherapy and combination therapy are research frontiers.
Recently, many researchers have made successful progress in building the AI systems for MOBA-game-playing with deep reinforcement learning, such as on Dota 2 and Honor of Kings. Even though these AI systems have achieved or even exceeded human-level performance, they still suffer from the lack of policy diversity. In this paper, we propose a novel Macro-Goals Guided framework, called MGG, to learn diverse policies in MOBA games. MGG abstracts strategies as macro-goals from human demonstrations and trains a Meta-Controller to predict these macro-goals. To enhance policy diversity, MGG samples macro-goals from the Meta-Controller prediction and guides the training process towards these goals. Experimental results on the typical MOBA game Honor of Kings demonstrate that MGG can execute diverse policies in different matches and lineups, and also outperform the state-of-the-art methods over 102 heroes.
3D object detection is vital in the environment perception of autonomous driving. The current monocular 3D object detection technology mainly uses RGB images and pseudo radar point clouds as input. The methods of taking RGB images as input need to learn with geometric constraints and ignore the depth information in the picture, leading to the method being too complicated and inefficient. Although some image-based methods use depth map information for post-calibration and correction, such methods usually require a high-precision depth estimation network. The methods of using the pseudo radar point cloud as input easily introduce noise in the conversion process of depth information to the pseudo radar point cloud, which cause a large deviation in the detection process and ignores semantic information simultaneously. We introduce dynamic convolution guided by the depth map into the feature extraction network, the convolution kernel of dynamic convolution automatically learns from the depth map of the image. It solves the problem that depth information and semantic information cannot be used simultaneously and improves the accuracy of monocular 3D object detection. MonoDCN is able to significantly improve the performance of both monocular 3D object detection and Bird's Eye View tasks within the KITTI urban autonomous driving dataset.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)