Deep learning for automatic segmentation of thigh and leg muscles

Objective

In this study we address the automatic segmentation of selected muscles of the thigh and leg through a supervised deep learning approach.

Material and methods

The application of quantitative imaging in neuromuscular diseases requires the availability of regions of interest (ROI) drawn on muscles to extract quantitative parameters. Up to now, manual drawing of ROIs has been considered the gold standard in clinical studies, with no clear and universally accepted standardized procedure for segmentation. Several automatic methods, based mainly on machine learning and deep learning algorithms, have recently been proposed to discriminate between skeletal muscle, bone, subcutaneous and intermuscular adipose tissue. We develop a supervised deep learning approach based on a unified framework for ROI segmentation.

Results

The proposed network generates segmentation maps with high accuracy, consisting in Dice Scores ranging from 0.89 to 0.95, with respect to “ground truth” manually segmented labelled images, also showing high average performance in both mild and severe cases of disease involvement (i.e. entity of fatty replacement).

Discussion

The presented results are promising and potentially translatable to different skeletal muscle groups and other MRI sequences with different contrast and resolution.

     

Variable echo time imaging for detecting the short T2* components of the sciatic nerve: a validation study

Objective

The aim of this study was to develop and validate an MRI protocol based on a variable echo time (vTE) sensitive to the short T2* components of the sciatic nerve.

Materials and methods

15 healthy subjects (M/F: 9/6; age: 21–62) were scanned at 3T targeting the sciatic nerve at the thigh bilaterally, using a dual echo variable echo time (vTE) sequence (based on a spoiled gradient echo acquisition) with echo times of 0.98/5.37 ms. Apparent T2* (aT2*) values of the sciatic nerves were calculated with a mono-exponential fit and used for data comparison.

Results

There were no significant differences in aT2* related to side, sex, age, and BMI, even though small differences for side were reported. Good-to-excellent repeatability and reproducibility were found for geometry of ROIs (Dice indices: intra-rater 0.68–0.7; inter-rater 0.70–0.72) and the related aT2* measures (intra-inter reader ICC 0.95–0.97; 0.66–0.85) from two different operators. Side-related signal-to-noise-ratio non-significant differences were reported, while contrast-to-noise-ratio measures were excellent both for side and echo.

Discussion

Our study introduces a novel MR sequence sensitive to the short T2* components of the sciatic nerve and may be used for the study of peripheral nerve disorders.

     

Elimination of cardiopulmonary bypass: a prime goal in reoperative coronary artery bypass surgery

OBJECTIVE: The purpose of this study was to evaluate morbidity and mortality in reoperative coronary artery bypass surgery using the New York State database. METHODS: Patients undergoing reoperative coronary artery bypass between January-1995 and December 1996 were included. Patients were operated using cardiopulmonary bypass (CPB group, n = 184) or without cardiopulmonary bypass (non-CPB group, n = 105) by surgeon preference. Groups were compared for preoperative risk factors, postoperative mortality and major complications. RESULTS: Crude mortality was lower in the non-CPB group, despite a higher expected mortality, resulting in a risk-adjusted mortality of 1.3% versus 2.7% for the CPB group (NS). Of non-CPB patients, 91.4% were without complications, while only 72.1% of CPB patients (P < 0.0001) were complication-free. Major complications were significantly reduced in non-CPB patients compared to CPB patients: stroke 0% versus 3.8% (P < 0.04), cardiovascular complications 4.8% versus 15.8% (P < 0.005), other major complications 1.9% versus 10.4% (P < 0.007). Postoperative IABP support was needed in 1.9% of the non-CPB group patients and in 14.2% of the CPB group (P < 0.0007). CONCLUSIONS: The main object of reoperative CABG is to relieve symptoms, since the survival benefit of the procedure has not been demonstrated. Performance of reoperative coronary artery bypass surgery without cardiopulmonary bypass significantly reduces morbidity. We conclude that cardiopulmonary bypass should be avoided whenever possible in reoperative coronary bypass surgery.     

Comparison of Standard 1.5 T vs. 3 T Optimized Protocols in Patients Treated with Glatiramer Acetate. A Serial MRI Pilot Study

This study explored the effect of glatiramer acetate (GA, 20 mg) on lesion activity using the 1.5 T standard MRI protocol (single dose gadolinium [Gd] and 5-min delay) or optimized 3 T protocol (triple dose of Gd, 20-min delay and application of an off-resonance saturated magnetization transfer pulse). A 15-month, phase IV, open-label, single-blinded, prospective, observational study included 12 patients with relapsing-remitting multiple sclerosis who underwent serial MRI scans (Days -45, -20, 0; the minus ign indicates the number of days before GA treatment; and on Days 30, 60, 90, 120, 150, 180, 270 and 360 during GA treatment) on 1.5 T and 3 T protocols. Cumulative number and volume of Gd enhancing (Gd-E) and T2 lesions were calculated. At Days -45 and 0, there were higher number (p < 0.01) and volume (p < 0.05) of Gd-E lesions on 3 T optimized compared to 1.5 T standard protocol. However, at 180 and 360 days of the study, no significant differences in total and cumulative number of new Gd-E and T 2 lesions were found between the two protocols. Compared to pre-treatment period, at Days 180 and 360 a significantly greater decrease in the cumulative number of Gd-E lesions (p = 0.03 and 0.021, respectively) was found using the 3 T vs. the 1.5 T protocol (p = NS for both time points). This MRI mechanistic study suggests that GA may exert a greater effect on decreasing lesion activity as measured on 3 T optimized compared to 1.5 T standard protocol.