AI-Based Automated Lipomatous Tumor Segmentation in MR Images: Ensemble Solution to Heterogeneous Data.

Deep learning (DL) has been proposed to automate image segmentation and provide accuracy, consistency, and efficiency. Accurate segmentation of lipomatous tumors (LTs) is critical for correct tumor radiomics analysis and localization. The major challenge of this task is data heterogeneity, including tumor morphological characteristics and multicenter scanning protocols. To mitigate the issue, we aimed to develop a DL-based Super Learner (SL) ensemble framework with different data correction and normalization methods. Pathologically proven LTs on pre-operative T1-weighted/proton-density MR images of 185 patients were manually segmented. The LTs were categorized by tumor locations as distal upper limb (DUL), distal lower limb (DLL), proximal upper limb (PUL), proximal lower limb (PLL), or Trunk (T) and grouped by 80%/9%/11% for training, validation and testing. Six configurations of correction/normalization were applied to data for fivefold-cross-validation trainings, resulting in 30 base learners (BLs). A SL was obtained from the BLs by optimizing SL weights. The performance was evaluated by dice-similarity-coefficient (DSC), sensitivity, specificity, and Hausdorff distance (HD95). For predictions of the BLs, the average DSC, sensitivity, and specificity from the testing data were 0.72 [Formula: see text] 0.16, 0.73 [Formula: see text] 0.168, and 0.99 [Formula: see text] 0.012, respectively, while for SL predictions were 0.80 [Formula: see text] 0.184, 0.78 [Formula: see text] 0.193, and 1.00 [Formula: see text] 0.010. The average HD95 of the BLs were 11.5 (DUL), 23.2 (DLL), 25.9 (PUL), 32.1 (PLL), and 47.9 (T) mm, whereas of SL were 1.7, 8.4, 15.9, 2.2, and 36.6 mm, respectively. The proposed method could improve the segmentation accuracy and mitigate the performance instability and data heterogeneity aiding the differential diagnosis of LTs in real clinical situations.

Alternate Cervical Venous Access Sites for Implantable Port Catheters: Experience at a Single Quaternary Care Institution.

INTRODUCTION: Clinical outcomes of implantable port catheters (IPCs) placed via alternative veins such as the external jugular and cervical collaterals have not been well established. This investigation evaluates the short- and long-term outcomes of IPCs inserted via alternate cervical veins (ACV) compared to traditionally inserted IPCs via the internal jugular vein (IJV). MATERIALS AND METHODS: A total of 24 patients who received an IPC between 2010 and 2020 via an ACV-defined as the external jugular vein, superficial cervical vein, or unnamed collateral veins-were identified. Based on power analysis, a matched control group of 72 patients who received IPCs via the IJV was identified. Non-inferiority analysis for port complications was performed between the two groups based on the selected non-inferiority margin of 20%. Secondary end points included complication-free survival and comparison of complications by the time at which they occurred. RESULTS: ACV access was non-inferior to traditional access for overall complications. Alternate access resulted in fewer complications than traditional access with an estimated reduction of - 7.0% [95% CI - 23.6%, 39.7%]. There was no significant difference in peri-procedural and post-procedural complications between the two groups. Complication-free survival was also equivalent between the two groups. CONCLUSION: IPC placement via ACVs was non-inferior to IPCs placed via traditional access through the IJV. When abnormal pathology obviates the use of IJV access, other cervical veins may be considered prior to seeking alternate locations such as femoral, translumbar, inferior vena cava, and hepatic veins.

Characterization of the regional distribution of skeletal muscle adipose tissue in type 2 diabetes using chemical shift-based water/fat separation.

PURPOSE: To show the feasibility of assessing the spatial distribution of skeletal muscle adipose tissue using chemical shift-based water/fat separation and to characterize differences in calf intermuscular adipose tissue (IMAT) compartmentalization in patients with type 2 diabetes mellitus (T2DM) compared to healthy age-matched controls. MATERIALS AND METHODS: A chemical shift-based water/fat separation approach using a multiecho 3D spoiled gradient echo sequence was applied in a study of 64 patients, including 35 healthy controls and 29 subjects with T2DM. Masks were defined based on manual segmentations to compute fat volume within different compartments, including regions of subcutaneous adipose tissue (SAT) and six muscular regions. IMAT was divided into two compartments representing fat within the muscular regions (intraMF) and fat between the muscular regions (interMF). Two-sample Student's t-tests were used to compare fat volumes between the two groups. RESULTS: The subjects with T2DM had a lower volume of SAT compared to the healthy controls (P = 4 × 10(-5) ). There was no statistically significant difference in the IMAT volume between the two groups. However, the intraMF volume normalized by the IMAT volume was higher in the diabetics compared to the controls (P = 0.006). CONCLUSION: Chemical shift-based water/fat separation enables the quantification of fat volume within localized muscle regions, showing that the IMAT regional distribution is significantly different in T2DM compared to normal controls.