Comparison of automatic liver volumetry performance using different types of magnetic resonance images.

Measurements of liver volume from MR images can be valuable for both clinical and research applications. Automated methods using convolutional neural networks have been used successfully for this using a variety of different MR image types as input. In this work, we sought to determine which types of magnetic resonance images give the best performance when used to train convolutional neural networks for liver segmentation and volumetry. Abdominal MRI scans were performed at 3 Tesla on 42 adolescents with obesity. Scans included Dixon imaging (giving water, fat, and T2* images) and low-resolution T2-weighted scout images. Multiple convolutional neural network models using a 3D U-Net architecture were trained with different input images. Whole-liver manual segmentations were used for reference. Segmentation performance was measured using the Dice similarity coefficient (DSC) and 95% Hausdorff distance. Liver volume accuracy was evaluated using bias, precision, intraclass correlation coefficient, normalized root mean square error (NRMSE), and Bland-Altman analyses. The models trained using both water and fat images performed best, giving DSC = 0.94 and NRMSE = 4.2%. Models trained without the water image as input all performed worse, including in participants with elevated liver fat. Models using the T2-weighted scout images underperformed the Dixon-based models, but provided acceptable performance (DSC ≥ 0.92, NMRSE ≤6.6%) for use in longitudinal pediatric obesity interventions. The model using Dixon water and fat images as input gave the best performance, with results comparable to inter-reader variability and state-of-the-art methods.

Exenatide for weight-loss maintenance in adolescents with severe obesity: A randomized, placebo-controlled trial.

OBJECTIVE: This study sought to evaluate the effect of 52 weeks of exenatide extended release (XR) on the maintenance of meal replacement therapy (MRT)-induced BMI reduction in adolescents with severe obesity. METHODS: In this randomized, double-blind, placebo-controlled trial, 100 participants aged 12 to 18 years with BMI ≥ 1.2 × 95th percentile were enrolled in a short-term MRT run-in phase. Those who achieved ≥5% BMI reduction during the run-in were then randomized to 52 weeks of exenatide XR 2.0 mg or placebo weekly. Both groups also received lifestyle therapy. The prespecified primary end point was mean percent change in BMI from randomization (post run-in) to 52 weeks in the intention-to-treat population. RESULTS: A total of 100 participants were enrolled, and 66 (mean age 16 = [SD 1.5] years; 47% female) achieved ≥5% BMI reduction with MRT and were randomized (33 to exenatide XR and 33 to placebo). From randomization (post run-in) to 52 weeks, mean BMI increased 4.6% and 10.1% in the exenatide XR and placebo groups, respectively. The placebo-subtracted exenatide XR treatment effect was -4.1% (95% CI: -8.6% to 0.5%, p = 0.078). CONCLUSIONS: Although not achieving statistical significance, exenatide XR, compared with placebo, may partly mitigate the propensity toward BMI rebound in adolescents who achieved initial weight loss with dietary intervention.

Detection of lying down, sitting, standing, and stepping using two activPAL monitors.

UNLABELLED: The activPAL is an accelerometer-based monitor worn on the thigh that classifies daily activities into three categories (sitting/lying down, standing, and stepping). The monitor discriminates between sitting/lying and the upright position by detecting the inclination of the thigh. It detects stepping from the acceleration versus time wave form. However, a current limitation of the activPAL is that it does not discriminate between sitting and lying down. PURPOSE: This study aimed to determine whether placing a second activPAL monitor on the torso would allow the detection of seated versus lying postures. METHODS: Fifteen healthy adults (18-55 yr of age) wore an activPAL on the right thigh and another activPAL over the right rib cage. Both monitors were synchronized and initialized to record data in 15-s epochs. Participants performed a semistructured routine of activities for 3 min each. Activities included lying down (while supine, prone, and on the side), sitting, standing, sweeping, treadmill walking at 3 mph, and treadmill running at 6 mph. The spatial orientation of the thigh and chest monitors was used to determine body posture, and the activPAL on the thigh was used to detect ambulation. RESULTS: The use of two activPAL devices enabled four behaviors to be accurately classified. The percentages of observations that were classified accurately were as follows: lying down (100%), sitting (100%), standing/light activity in the upright position (90.8%), and stepping (100%). CONCLUSIONS: The current method allows researchers to obtain more detailed information on postural allocation compared with that in the use of a single activPAL on the thigh.