Towards dense volumetric pancreas segmentation in CT using 3D fully convolutional networks

The fully convolutional networks (FCNs) have been widely applied in numerous medical image segmentation tasks. However, tissue regions usually have large variations of shape and scale, so the ability of neural networks to learn multi-scale features is important to the segmentation performance. In this paper, we improve the network for multi-scale feature fusion, in the medical image segmentation by introducing two feature fusion modules: i) global attention multi-scale feature fusion module (GMF); ii) local dense multi-scale feature fusion module (LMF). GMF aims to use global context information to guide the recalibration of low-level features from both spatial and channel aspects, so as to enhance the utilization of effective multi-scale features and suppress the noise of low-level features. LMF adopts bottom-up top-down structure to capture context information, to generate semantic features, and to fuse feature information at different scales. LMF can integrate local dense multi-scale context features layer by layer in the network, thus improving the ability of network to encode interdependent relationships among boundary pixels. Based on the above two modules, we propose a novel medical image segmentation framework (GLF-Net). We evaluated the proposed network and modules on challenging brain tumor segmentation and pancreas segmentation datasets, and very competitive performance has been achieved.

Deep pancreas segmentation with uncertain regions of shadowed sets

MSFCN-multiple supervised fully convolutional networks for the osteosarcoma segmentation of CT images

The shape of pancreas in different patients is very different and the boundary is fuzzy, so the reliable automatic segmentation of pancreas is an important and difficult task. In this paper, we propose a multi-scale hybrid attention cascade network for challenging pancreas segmentation. First, the original CT images are coarsely segmented through the first-level network, then, the coarse segmentation results are clipped and sent to the second-level network for further training to obtain the fine segmentation results. The network adopts full convolutional network (FCN) integrating multi-scale mixed attention. The cut CT image eliminates the irrelevant background interference and reduces the input size of the secondary network, thus improving the segmentation accuracy. An extensive evaluation of 82 open data sets was performed by quadruple cross validation. The experimental results showed that the Dice coefficient was 84.62±4.20% compared with several advanced methods. In addition, this method has excellent performance in three measurement indexes of precision, Jaccard and recall.

To investigate the usefulness of computed tomographic (CT) spectral imaging parameters in differentiating small (≤3 cm) hepatic hemangioma (HH) from small hepatocellular carcinoma (HCC), with or without cirrhosis, during the late arterial phase (AP) and portal venous phase (PVP). This prospective study was institutional review board approved, and written informed consent was obtained from all patients. The authors examined 49 patients (39 men, 10 women; 65 lesions) with CT spectral imaging during the AP and the PVP. Twenty-one patients had HH; nine, HCC with cirrhosis; and 19, HCC without cirrhosis. Iodine concentrations were derived from iodine-based material-decomposition CT images and normalized to the iodine concentration in the aorta. The difference in iodine concentration between the AP and PVP (ie, iodine concentration difference [ICD]) and the lesion-to-normal parenchyma ratio (LNR) were calculated. Two readers qualitatively assessed lesion types on the basis of conventional CT characteristics. Sensitivity and specificity were compared between the qualitative and quantitative studies. The two-sample t test was performed to compare quantitative parameters between HH and HCC. Normalized iodine concentrations (NICs) and LNRs in patients with HH differed significantly from those in patients with HCC and cirrhosis and those in patients with HCC without cirrhosis: Mean NICs were 0.47 mg/mL ± 0.24 (standard deviation) versus 0.23 mg/mL ± 0.10 and 0.23 mg/mL ± 0.08, respectively, during the AP and 0.83 mg/mL ± 0.38 versus 0.47 mg/mL ± 0.86 and 0.52 mg/mL ± 0.11, respectively, during the PVP. Mean LNRs were 5.87 ± 3.36 versus 2.56 ± 1.10 and 2.29 ± 0.87, respectively, during the AP and 2.01 ± 1.33 versus 0.96 ± 0.16 and 0.93 ± 0.26, respectively, during the PVP. The mean ICD for the HH group (1.37 mg/mL ± 0.84) was significantly higher than the mean ICDs for the HCC-cirrhosis (0.33 mg/mL ± 0.29) (P < .001) and HCC-no cirrhosis (0.82 mg/mL ± 0.99) (P = .03) groups. The combination of NIC and LNR had higher sensitivity and specificity compared with those of conventional qualitative CT image analysis during individual and combined phases. Use of spectral CT with fast tube voltage switching may increase the sensitivity for differentiating small hemangiomas from small HCCs in two-phase scanning.

A Clinical-Radiomic Combined Model based on Dual-Layer Spectral CT for Predicting Pathological T4 in Gastric Cancer.

We propose an automated pancreas segmentation algorithm from contrast-enhanced multiphase computed tomography (CT) and verify its effectiveness in segmentation. The algorithm is characterized by three unique ideas. First, a two-stage segmentation strategy with spatial standardization of pancreas was employed to reduce variations in the pancreas shape and location. Second, patient- specific probabilistic atlas guided segmentation was developed to cope with the remaining variability in shape and location. Finally, a classifier ensemble was incorporated to refine the rough segmentation results. The effectiveness of the proposed algorithm was validated with 20 unknown CT volumes, as well as three on-site CT volumes distributed in a competition of pancreas segmentation algorithms. The experimental results indicated that the segmentation performance was enhanced by the proposed algorithm, and the Jaccard index between an extracted pancreas and a true one was 57.9%. This study verified the effectiveness of two-stage segmentation with spatial standardization of pancreas in delineating the pancreas region, patient-specific probabilistic atlas guided segmentation in reducing false negatives, and a classifier ensemble in boosting segmentation performance.

BackgroundTo compare the effects of laparoscopic-assisted gastrectomy (LAG) and open gastrectomy (OG) on serum interleukin-6 (IL-6) levels in gastric cancer (GC) patients from Asia.MethodsThe following scientific literature databases were searched for relevant clinical studies: PubMed, EBSCO, Ovid, Wiley, Web of Science, Cochrane library, EMBASE, WANFANG and VIP databases. The studies retrieved from database searches were screened based on stringent inclusion and exclusion criteria to select high quality cohort studies for the present meta-analysis. The data extracted from final selected studies were analyzed using STATA 12.0 software.ResultsA total of 54 studies were initially retrieved from database searches, and 11 clinical cohort studies were eventually enrolled in this meta-analysis. The 11 selected studies contained a combined total of 767 GC patients (427 patients in LAG group and 340 patients in OG group). Meta-analysis results demonstrated that postoperative serum IL-6 levels in GC patients in LAG group was significantly lower than the OG group (SMD = −2.16, 95% CI = −3.19 ~ −1.14, P < 0.001). The difference in serum IL-6 levels between the preoperative and postoperative GC patients was significantly lower in the LAG group compared to the difference found in the OG group (SMD = −3.44, 95% CI = −4.87 ~ −2.01, P < 0.001). Subgroup analysis based on country showed that, in both Chinese and Japanese GC patients, the postoperative increase in serum IL-6 levels in LAG group were significantly lower than the increase observed in the OG group (all P < 0.05). In Korean GC patients, the postoperative increase in serum IL-6 levels was not significantly different between the LAG group and OG group (all P > 0.05).ConclusionOur results provide strong evidence that LAG is associated with significantly lower serum IL-6 levels, compared to OG. Thus, LAG carries markedly lower risk of adverse inflammatory reactions in GC patients among Asian population.

In this work, we present a memory-efficient fully convolutional network (FCN)\nincorporated with several memory-optimized techniques to reduce the run-time\nGPU memory demand during training phase. In medical image segmentation tasks,\nsubvolume cropping has become a common preprocessing. Subvolumes (or small\npatch volumes) were cropped to reduce GPU memory demand. However, small patch\nvolumes capture less spatial context that leads to lower accuracy. As a pilot\nstudy, the purpose of this work is to propose a memory-efficient FCN which\nenables us to train the model on full size CT image directly without subvolume\ncropping, while maintaining the segmentation accuracy. We optimize our network\nfrom both architecture and implementation. With the development of computing\nhardware, such as graphics processing unit (GPU) and tensor processing unit\n(TPU), now deep learning applications is able to train networks with large\ndatasets within acceptable time. Among these applications, semantic\nsegmentation using fully convolutional network (FCN) also has gained a\nsignificant improvement against traditional image processing approaches in both\ncomputer vision and medical image processing fields. However, unlike general\ncolor images used in computer vision tasks, medical images have larger scales\nthan color images such as 3D computed tomography (CT) images, micro CT images,\nand histopathological images. For training these medical images, the large\ndemand of computing resource become a severe problem. In this paper, we present\na memory-efficient FCN to tackle the high GPU memory demand challenge in organ\nsegmentation problem from clinical CT images. The experimental results\ndemonstrated that our GPU memory demand is about 40% of baseline architecture,\nparameter amount is about 30% of the baseline.\n

Intrathoracic airway segmentation from computed tomography images is a frequent prerequisite for further quantitative lung analyses. Due to low contrast and noise, especially at peripheral branches, it is often challenging for automatic methods to strike a balance between extracting deeper airway branches and avoiding leakage to the surrounding parenchyma. Meanwhile, manual annotations are extremely time consuming for the airway tree, which inhibits automated methods requiring training data. To address this, we introduce a 3D deep learning-based workflow able to produce high-quality airway segmentation from incompletely labeled training data generated without manual intervention. We first train a 3D fully convolutional network (FCN) based on the fact that 3D spatial information is crucial for small highly anisotropic tubular structures such as airways. For training the 3D FCN, we develop a domain-specific sampling scheme that strategically uses incomplete labels from a previous highly specific segmentation method, aiming to retain similar specificity while boosting sensitivity. Finally, to address local discontinuities of the coarse 3D FCN output, we apply a graph-based refinement incorporating fuzzy connectedness segmentation and robust curve skeletonization. Evaluations on the EXACT’09 and LTRC datasets demonstrate considerable improvements in airway extraction while maintaining reasonable leakage compared with a state-of-art method and the dataset reference standard.

Controversy exists in the neuropsychological literature concerning the existence of gender-associated differences in cognitive functioning and in hemispheric lateralization of cognitive functions. A recent study, based on 14 brains obtained at autopsy, reported sex differences in the splenium of the human corpus callosum and suggested that the larger splenium in females reflects less hemispheric lateralization, or "specialization," than the male brain for visuospatial functions. Our measurements of the human corpus callosum using magnetic resonance images of 37 living subjects failed to confirm reported sex differences in the splenium. A marginally significant sex-related difference in minimum body width and an age-related decrease in anteroposterior distance were found. Most striking were the large variations in callosal size and shape among individuals regardless of age or gender. Existing knowledge of the functions of the corpus callosum does not permit correlations between variations in callosal size and shape and variations in cognitive functions.

The Effect and Mechanism of Subcutaneous and Visceral Adipose Tissue Loss on Gastric Cancer Patients With Cachexia

Fully convolutional networks (FCNs) have shown outstanding performance in image semantic segmentation, which is the key work in license plate detection (LPD). An FCN architecture for LPD is presented. First, a multiscale hierarchical network structure is used to combine multiscale and multilevel features produced by FCN. Then, an enhanced loss structure that contains three loss layers is defined to emphasize the license plates in images. Finally, the FCN generates prediction maps that directly show the location of license plates. Experiments show that our approach is more accurate than many state-of-the-art methods.

Diseases such as diabetes and hypertension can lead to change the shape of the retinal blood vessels. Segmentation of fundus images is a key step in the process of quantitative analysis of the disease, which is instructive in the analysis and diagnosis of clinical diseases. In this paper, a method for the segmentation of retinal image vessels based on fully convolutional network (FCN) with depthwise separable convolution and channel weighting is presented. Firstly, CLAHE and Gamma correction of the green channel of the fundus image are used to enhance the contrast. Then, in order to adapt to network training, the enhanced image is divided into patches to expand the data. Finally, the depthwise separable convolution instead of the standard convolution method is used to increase the network width. Meanwhile, the channel weighting module is introduced to explicitly model the relationship between the characteristic channels in order to improve the distinguishability of the features. The combination of them is applied to the FCN and the results of expert manual identification are used to supervise the experiment on the DRIVE database. The results show that the segmentation accuracy of the proposed method in DRIVE database reached 0.963 0 and AUC reached 0.983 1. The segmentation accuracy in STARE database reached 0.962 0 and AUC achieved 0.983 0. To some extent, the proposed method has better feature resolution and better segmentation performance.

Chronic liver disease (CLD) will affect the enhancement of hepatic parenchyma and portal vein on abdominal-enhanced MRI. To investigate the difference in liver parenchyma and portal vein enhancement in patients with CLD of different liver function grades between Gd- EOB-DTPA and Gd-DPTA in the portal venous phase (PVP). This retrospective study included 218 patients with CLD who had undergone abdominal enhanced MRI from January 2019 to June 2020. Patients with various degrees of liver dysfunction were identified with Child-Turcotte-Pugh and albumin-bilirubin grade. Two readers measured the precontrast and PVP signal intensities of liver parenchyma, portal vein, spleen, and psoas muscle. Relative liver enhancement, liver-to-spleen contrast index, portal vein image contrast, and portal vein-to-liver contrast were calculated. The relative enhancement of liver parenchyma was significantly lower for the Gd-EOB-DTPA group in any degree of liver function than the Gd- DTPA group in the PVP. The Gd-EOB-DTPA group showed significantly lower portal vein-to-liver contrast in the overall study population, CTP class B, and ALBI grade 2 patients compared to the group of Gd-DTPA at PVP. No significant difference was noted in the portal vein image contrast between the two contrast agents, regardless of CTP and ALBI grading. In CLD patients, Gd-EOB-DTPA yielded lower liver parenchymal enhancement and similar portal vein image contrast compared to Gd-DTPA in the PVP. Portal vein-to-liver contrast in the Gd-EOB-DTPA group was lower in the CTP class B and ALBI grade 2 subgroups compared to the Gd- DTPA group.