To assess feasibility of training a convolutional neural network (CNN) to automate liver segmentation across different imaging modalities and techniques used in clinical practice and to apply this technique to enable automation of liver biometry. Materials and Methods: A two-dimensional U-Net CNN was trained for liver segmentation in two stages by using 330 abdominal MRI and CT examinations. First, the neural network was trained with unenhanced multiecho spoiled gradient-echo images from 300 MRI examinations to yield multiple signal weightings. Then, transfer learning was used to generalize the CNN with additional images from 30 contrast material-enhanced MRI and CT examinations. Performance of the CNN was assessed by using a distinct multiinstitutional dataset curated from multiple sources (498 subjects). Segmentation accuracy was evaluated by computing Dice scores. These segmentations were used to compute liver volume from CT and T1-weighted MRI examinations and to estimate hepatic proton density fat fraction (PDFF) from multiecho T2*-weighted MRI examinations. Quantitative volumetry and PDFF estimates were compared between automated and manual segmentation by using Pearson correlation and Bland-Altman statistics. Results: Dice scores were 0.94 ± 0.06 for CT (n = 230), 0.95 ± 0.03 (n = 100) for T1-weighted MRI, and 0.92 ± 0.05 for T2*weighted MRI (n = 168). Liver volume measured with manual and automated segmentation agreed closely for CT (95% limits of agreement: −298 mL, 180 mL) and T1-weighted MRI (95% limits of agreement: −358 mL, 180 mL). Hepatic PDFF measured by the two segmentations also agreed closely (95% limits of agreement: −0.62%, 0.80%). Conclusion: By using a transfer-learning strategy, this study has demonstrated the feasibility of a CNN to be generalized to perform liver segmentation across different imaging techniques and modalities. With further refinement and validation, CNNs may have broad applicability for multimodal liver volumetry and hepatic tissue characterization.

No abstract

Electronic skins, often with recognition and sensing capabilities that are beyond those associated with biological skin, provide important medical information for human health. However, how to make electronic skins with both tactile and touchless perceptions for applications in real-time health monitoring is a challenge due to biological complexity. Herein, flexible bimodal e-skins are demonstrated using a lamellated MXene/chitosan film as the kernel sensing layer. These biomimetic hybrid films show excellent biocompatibility in a cytotoxicity test, indicating a reduced risk of application in the human body. The flexible devices display two modes that can detect pressure (with a limit of detection (LoD) of 3 Pa, stability >3500 times, and response time of 143 ms) and humidity (stability >20 days). The bifunctional sensor can also be used in detecting and discriminating electrophysiological signals (including voice recognition, human pulses, and finger bending) and biochemical molecules (respiratory rate). This work may lead to the application of biocomposite materials in multifunctional flexible sensing technology.

Purpose To evaluate the repeatability and reproducibility of 2D and 3D hepatic MRE with rigid and flexible drivers at end expiration and inspiration in healthy volunteers. Material and Methods Nine healthy volunteers underwent two same-day MRE exams separated by a 5–10 minute break. In each exam, 2D and 3D MRE scans were performed, each under four conditions (2 driver types [rigid, flexible] × 2 breath-hold phases [end-expiration, end-inspiration]). Repeatability (measurements under identical conditions) and reproducibility (measurements under different conditions) were analyzed by calculating bias, limit of agreement, repeatability coefficient (RC), reproducibility coefficient (RDC), intraclass correlation coefficient (ICC), and concordance correlation coefficient (CCC), as appropriate. Results For 2D MRE, RCs and ICCs range between 0.29 – 0.49 and 0.71 – 0.91, respectively. For 3D MRE, RCs and ICCs range between 0.16 – 0.26 and 0.84 – 0.96, respectively. Stiffness values were biased by breath-hold phase, being higher at end-inspiration than end-expiration, and the differences were significant for 3D MRE (p < 0.01). No bias was found between driver types. Inspiration-vs.- expiration RDCs and CCCs ranged between 0.30–0.54 and 0.61–0.72, respectively. Rigid-vs.-flexible driver RDCs and CCCs ranged between 0.10–0.44 and 0.79–0.94, respectively. Conclusion This preliminary study suggests that 2D MRE and 3D MRE under most conditions potentially have good repeatability. Our result also points to the possibility that stiffness measured with the rigid and flexible drivers is reproducible. Reproducibility between breath-hold phases was modest, suggesting breath-hold phase might be a confounding factor in MRE-based stiffness measurement. However, larger studies are required to validate these preliminary results.

No abstract

Purpose To demonstrate the feasibility of performing single breath-hold, non-cardiac gated, ultrafast, high spatial-temporal resolution whole chest MR pulmonary perfusion imaging in humans. Materials and Methods Eight (8) subjects (5 male, 3 female) were scanned with the proposed method on a 3T clinical scanner using a 32-channel phased-array coil. Seven (88%) were healthy volunteers, and one was a patient volunteer with sarcoidosis. The peak lung enhancement phase for each subject was scored for gravitational effect, peak parenchymal enhancement and severity of artifacts by 3 cardiothoracic radiologists independently. Results All studies were successfully performed by MR technologists without any additional training. Mean parenchymal signal was very good, measuring 0.78 ± 0.13 (continuous scale, 0 = “none” → 1 = “excellent”). Mean level of motion artifacts was low, measuring 0.13 ± 0.08 (continuous scale, 0 = “none” → 1 = “severe”). Conclusion It is feasible to perform single breath-hold, non-cardiac gated, ultrafast, high spatial-temporal resolution whole chest MR pulmonary perfusion imaging in humans.

Post-contrast liver magnetic resonance imaging is typically performed with breath-hold 3D gradient echo sequences. However, breath-holding for >10 s is difficult for some patients. In this study, we compared the quality of hepatobiliary phase (HBP) imaging without breath-holding using the prototype pulse sequences stack-of-stars liver acquisition with volume acceleration (LAVA) (LAVA Star) with or without navigator echoes (LAVA Star navi+ and LAVA Star navi−) and Cartesian LAVA with navigator echoes (Cartesian LAVA navi+). Methods: Seventy-two patients were included in this single-center, retrospective, cross-sectional study. HBP imaging using the three LAVA sequences (Cartesian LAVA navi+ , LAVA Star navi− , and LAVA Star navi+) without breath-holding was performed for all patients using a 3T magnetic resonance system. Two independent radiologists qualitatively analyzed (overall image quality, liver edge sharpness, hepatic vein clarity, streak artifacts, and respiratory motion/pulsation artifacts) HBP images taken by the three sequences using a five-point scale. Quantitative evaluations were also performed by calculating the liver-to-spleen,-lesion, and-portal vein (PV) signal intensity ratios. The results were compared between the three sequences using the Friedman test. Results: LAVA Star navi+ showed the best image quality and hepatic vein clarity (P < 0.0001). LAVA Star navi− showed the lowest image quality (P < 0.0001-0.0106). LAVA Star navi+ images showed fewer streak artifacts than LAVA Star navi− images (P < 0.0001), while Cartesian LAVA navi+ images showed no streak artifacts. Cartesian LAVA navi+ images showed stronger respiratory motion/pulsation artifacts than the others (P < 0.0001). LAVA Star navi− images showed the highest liver-to-spleen ratios (P < 0.0001-0.0005). Cartesian LAVA navi+ images showed the lowest liver-to-lesion and-PV ratios (P < 0.0001-0.0108). Conclusion: In terms of image quality, the combination of stack-of-stars acquisition and navigator echoes is the best for HBP imaging without breath-holding.

Purpose To develop and evaluate a method for volumetric contrast-enhanced MR imaging of the liver, with high spatial and temporal resolutions, for combined dynamic imaging and MR angiography using a single injection of contrast. Methods An interleaved variable density (IVD) undersampling pattern was implemented in combination with a real-time-triggered, time-resolved, dual-echo 3D spoiled gradient echo sequence. Parallel imaging autocalibration lines were acquired only once during the first time-frame. Imaging was performed in ten subjects with focal nodular hyperplasia (FNH) and compared with their clinical MRI. The angiographic phase of the proposed method was compared to a dedicated MR angiogram acquired during a second injection of contrast. Results A total of 21 FNH, 3 cavernous hemangiomas, and 109 arterial segments were visualized in 10 subjects. The temporally-resolved images depicted the characteristic arterial enhancement pattern of the lesions with a 4 s update rate. Images were graded as having significantly higher quality compared to the clinical MRI. Angiograms produced from the IVD method provided non-inferior diagnostic assessment compared to the dedicated MRA. Conclusion Using an undersampled IVD imaging method, we have demonstrated the feasibility of obtaining high spatial and temporal resolution dynamic contrast-enhanced imaging and simultaneous MRA of the liver.