Graph cut based automatic aorta segmentation with an adaptive smoothness constraint in 3D abdominal CT images

ObjectiveTo study the feasibility and the potential benefits of defining the internal gross tumor volume (IGTV) of hepatocellular carcinoma (HCC) using contrast-enhanced 4D CT images obtained by combining arterial-phase (AP) contrast-enhanced (CE) 3D CT and non-contrast-enhanced (NCE) 4D CT images using deformable registration (DR).MethodsTen HCC patients who had received radiotherapy beforehand were selected for this study. The following CT simulation images were acquired sequentially: NCE 4D CT in free breathing, NCE 3D CT and APCE 3D CT in end-expiration breath holding. All 4D CT images were sorted into ten phases according to breath cycle (CT00 ~ CT90). Gross tumor volumes (GTVs) were contoured on all CT images and the IGTV-1 was obtained by merging the GTVs in each phase of 4D CT images. The GTV on the APCE 3D CT image was deformably registered to each 4D CT phase image according to liver shape using RayStationTM 3.99.0.7 version treatment planning system. The IGTV-DR was obtained by merging the GTVs after DR on the 4D CT images. Volume differences among the GTVs and between the IGTV-1 and the IGTV-DR were compared.ResultsThe edge of most lesions could be definitively identified using APCE 3D CT images compared to NCE 4D and 3D CT images. The GTV volume on APCE 3D CT images increased by an average of 34.79% (P < 0.05). There was no significant difference among the GTV volumes obtained using NCE 4D and 3D CT images (P > 0.05). The GTV volumes after DR on 4D CT different phase images increased by an average of 36.29% (P < 0.05), as was observed using the APCE 3D CT image (P > 0.05). Lastly, the volume of IGTV-DR increased by an average of 19.91% compared to that of IGTV-1 (P < 0.05).ConclusionNCE 4D CT imaging alone has the potential risk of missing a partial volume of the HCC. The combination of APCE 3D CT and NCE 4D CT images using the DR technique improved the accuracy of the definition of the IGTV in HCC.

This paper presents an automated multi-organ segmentation method for 3D abdominal CT images based on a spatially-divided probabilistic atlases. Most previous abdominal organ segmentation methods are ineffective to deal with the large differences among patients in organ shape and position in local areas. In this paper, we propose an automated multi-organ segmentation method based on a spatially-divided probabilistic atlas, and solve this problem by introducing a scale hierarchical probabilistic atlas. The algorithm consists of image-space division and a multi-scale weighting scheme. The generated spatial-divided probabilistic atlas efficiently reduces the inter-subject variance in organ shape and position either in global or local regions. Our proposed method was evaluated using 100 abdominal CT volumes with manually traced ground truth data. Experimental results showed that it can segment the liver, spleen, pancreas, and kidneys with Dice similarity indices of 95.1%, 91.4%, 69.1%, and 90.1%, respectively.

BackgroundSuprascapular nerve entrapment is a medical condition characterized by clinical symptoms such as neck and shoulder pain, discomfort, upper limb weakness, and heaviness. These symptoms are caused by the compression of the suprascapular nerve at the suprascapular notch. The treatment of this nerve entrapment has been a long-standing challenge. Intelligent-assisted radiofrequency neurolysis is a novel and promising non-pharmacological treatment. Herein, we presented the therapeutic efficacy of this technique in 40 patients.ObjectiveThis study was designed to evaluate the feasibility of three-dimensional (3D) CT imaging integrated with intraoperative intelligence-assisted calibration for radiofrequency ablation of the suprascapular nerve.MethodsA total of 40 patients were enrolled. The control group received the conventional surgery, while the treatment group was given 3D CT imaging plus the intraoperative intelligence-assisted calibration. Intelligence-assisted refers to the use of a smartphone app based on gravity calibration to realize the precise measurement and correction of the puncture needle angle during the operation, combined with 3D CT imaging to achieve the accurate positioning of the suprascapular nerve entrapment site. Data was analyzed on the two groups, including the time taken from starting positioning to successful puncture, Visual Analogue Scale (VAS) scores and ROWE scores before and after treatment, the therapeutic effect, complications, and number of cases with abnormal electromyography (EMG) items before and after treatment.ResultsTime of puncture completion in the treatment group was significantly shorter than that in the control group (8.06 vs 15.98 min, p < 0.05). Postoperative therapeutic effect in the treatment group was also superior to that in the control group. The treatment group had significant differences between pre- and post-treatment VAS scores and had superior ROWE scores compared to the control group.ConclusionThe application of 3D CT imaging combined with intraoperative intelligence-assisted calibration for suprascapular nerve radiofrequency ablation can effectively shorten the operation time, improve the procedural efficiency and therapeutic effect and significantly reduce patients’ clinical symptoms.

In this paper we propose a new classification of neurogenic peri-articular heterotopic ossification (HO) of the hip based on three-dimensional (3D) CT, with the aim of improving pre-operative planning for its excision. A total of 55 patients (73 hips) with clinically significant HO after either traumatic brain or spinal cord injury were assessed by 3D-CT scanning, and the results compared with the intra-operative findings. At operation, the gross pathological anatomy of the HO as identified by 3D-CT imaging was confirmed as affecting the peri-articular hip muscles to a greater or lesser extent. We identified seven patterns of involvement: four basic (anterior, medial, posterior and lateral) and three mixed (anteromedial, posterolateral and circumferential). Excellent intra- and inter-observer agreement, with kappa values > 0.8, confirmed the reproducibility of the classification system. We describe the different surgical approaches used to excise the HO which were guided by the 3D-CT findings. Resection was always successful. 3D-CT imaging, complemented in some cases by angiography, allows the surgeon to define the 3D anatomy of the HO accurately and to plan its surgical excision with precision.

16-11 Automatic metastatic liver tumor extraction processing by 3D abdominal CT Images

Objective. We report the technique of three-dimensional computed tomography angiography (3D CTA) + two-dimensional computed tomographic (2D CT) imaging as an adjunct in early surgery for a ruptured anterior communicating artery (ACoA) aneurysm by adopting an anterior interhemispheric approach. These combined imaging modalities provide accurate intraoperative anatomical information. Methods. To produce images for an anterior interhemispheric approach, 3D CTA + 2D coronal CT images, which are perpendicular to the direction of the surgical approach at three levels (brain surface, genu of the corpus callosum and aneurysm neck), were constructed. We also produced two 3D CTA + 2D CT images of the lamina terminalis, with a horizontal 10-degree difference, to clarify the vascular architecture around the aneurysm stereotactically, as well as the dissection point and direction to open the lamina terminalis. Furthermore, we produced a 3D CTA + 2D sagittal CT image at the midline, which allowed us to understand the anatomical architecture of the aneurysm, planum sphenoidale and tuberculum sellae. In addition, four different 3D CTA aneurysm images were produced for deciding the clip size preoperatively. Results. The imaging findings in 28 patients with 28 ACoA aneurysms facilitated early clipping. Based on these 3D CTA + 2D CT images, we conducted aneurysm surgery, and successfully performed neck clipping via an anterior interhemispheric approach. Conclusion. The combination of 3D CTA and 2D CT images is a feasible and useful method of image guidance for ACoA aneurysm microsurgery.

Automated extraction of abdominal organs from uncontrasted 3D abdominal X-ray CT images based on anatomical knowledge

A method for automated segmentation of the stomach and its application for navigated diagnosis

Objective To compare the accuracy of plain X-my fdms and reconstructed 3D CT images in assessing characteristics of intra-articular distal radius fractures.Methods X-ray films and reconstructed 3D CT images of 16 cases of intra-articular distal radius fractures were retrospectively analyzed using Paces System and MIMICS software.The parameters that were evaluated included the ability to detect intra-articular step and gap,fracture displacements,the number of articular fragments,stretch of the fracture line to the sigmoid notch and the extensor tendon tunnel.Resuts The average value for intm-articular step and gap were (0.92 ±0.75) mm and (0.71 ± 0.87) mm,respectively,on posteroanterior and lateral radiographys.The average value for intra-articular step and gap were (2.58 ± 1.23) mm and (2.89 ± 1.13) mm,respectively,on coronal and sagittal 3D CT images (P ＜ 0.05).The average number of artic~ar fragments was (3.25 ± 1.18) detected by X-ray and (4.68 ±0.49) detected by 3D CT images (P ＜0.05).The fracture line was found stretched to the sigrnoid notch and the extensor tendon tunnel in 2 cases and 4 cases,respectively,on radiographs.On 3D CT images it was found to be present in 14 cases and 15 cases,respectively(P＜0.05).Conclusion Standard radiographs (anteroposterior and lateral views) are not sugicient to measure anterior displacement nor accurate enough to measure medial displacement of distal radius fractures.Three-dimensional computed tomography improves both the mhability and the accuracy of radiographic characterization of the intra-articular distal radius fracture and influences treatment decisions. Key words: Radius fracture; Tomography; Three dimensional reconstruction image

Elastography is a medical imaging modality to visualize the elasticity of soft tissues. Ultrasound and MRI have been exclusively used for elastography of soft tissues since they can sensitize the tissues' minute displacements of an order of μm. It is known that ultrasound and MRI elastography show cancerous tissues with much higher contrast than conventional ultrasound and MRI. To evaluate possibility of combining elastography with x-ray imaging, we have calculated strain images of a breast-mimicking phantom from its 3D CT image data. We first simulated the x-ray elastography using a FEM model which incorporated both the elasticity and x-ray attenuation behaviors of breast tissues. After validating the x-ray elastography scheme by simulation, we made a breast-mimicking phantom that contained a hard inclusion against soft background. With a micro-CT, we took 3D images of the phantom twice, changing the compressing force to the phantom. From the two 3D phantom images taken with two different compression ratios, we calculated the displacement vector maps that represented the compression-induced pixel displacements. In calculating the displacement vectors, we tracked the movements of image feature patterns from the less-compressed-phantom images to the more-compressed-phantom images using the 3D image correlation technique. We obtained strain images of the phantom by differentiating the displacement vector maps. The FEM simulation has shown that x-ray strain imaging is possible by tracking image feature patterns in the 3D CT images of the breast-mimicking phantom. The experimental displacement and strain images of a breast-mimicking phantom, obtained from the 3D micro-CT images taken with 0%-3% compression ratios, show behaviors similar to the FEM simulation results. The contrast and noise performance of the strain images improves as the phantom compression ratio increases. We have experimentally shown that we can improve x-ray strain image quality by applying 3D image correlation to the two sets of 3D CT images taken with different compression ratios. But, we need further investigations to evaluate the strain imaging performance considering the noise and decorrelation effects as well as the extra dose caused by two scans.

Deep neural networks have been widely used in medical image segmentation and they can achieve good results in segmentation of some big organs. However, for some small organs, such as the pancreas in 3D CT images, the segmentation results are usually not satisfactory due to the low proportion. In this paper, we present a novel network --- Dense Attentional Network (DA-Net), to improve the pancreas segmentation in abdominal CT scans. In DA-Net, dense connection is used to combine low-level features of encoder with corresponding features of decoder, which can help to improve the utilization of feature maps (FMs). In addition, a new module for recombination and recalibration of feature maps (RRFM) and a new attentional mechanism --- deep attentional features (DAF), are used, which can excite the most discriminating features. Pancreas is segmented from 3D CT images by a coarse-to-fine mode, in which pancreas is firstly located from 3D CT images by a coarse segmentation network, and then pancreas is further finely segmented by the DA-Net. We evaluate the proposed method with 129 CT images from NIH pancreas dataset and BTCV segmentation challenge, and compare it with several mainstream segmentation networks. Comparing with these networks, our DA-Net has the higher mean DSC of 81.39%. This shows the effectiveness and advantage of the proposed method.

The purpose of this retrospective study was to first compare the clinical outcome of anatomical double-bundle (DB) anterior cruciate ligament reconstruction (ACLR) with hamstring tendon (HT) autografts and anatomical rectangular tunnel ACLR with bone-patellar tendon-bone (BPTB) autografts. Secondly, we aimed to demonstrate the quantitative locations of the femoral and tibial tunnel apertures using postoperative three-dimensional computed tomography (3D CT). Twenty-five patients underwent anatomical rectangular tunnel ACLR using BPTB grafts (Group B) and 23 patients underwent anatomical DB ACLR using HT grafts (Group H). All patients underwent subjective postoperative evaluations using the Lysholm score and Knee Injury and Osteoarthritis Outcome Score. Patients also underwent objective evaluations by the International Knee Documentation Committee score, the Lachman test, the pivot-shift test, and range of motion. In addition, we quantitatively assessed anterior knee stability using a KneeLax3 arthrometer and thigh strength. All evaluations except for thigh strength were assessed for a minimum 2 years of follow-up period. Femoral and tibial tunnel aperture locations were quantitatively evaluated postoperatively using 3D CT images in all patients. BPTB grafts showed significantly better anterior knee stability than HT grafts (0.1 mm versus 1.1 mm, p = 0.01), although there were no significant differences in other objective and all subjective evaluations between the two graft types. Morphometric analysis of femoral and tibial tunnel locations revealed that the two procedures were based on the same anatomical concept. In conclusion, BPTB grafts showed significantly better anterior knee stability than HT grafts, although no significant differences in other objective evaluations and all subjective evaluations were detected between the two graft types in anatomical ACLR. Additional 3D CT data validated the anatomical concepts of these two procedures.

ObjectiveWe sought to examine anatomic variations of the atlas and the clinical significance of these variations.MethodsWe retrospectively reviewed 1029 cervical 3-dimensional (3D) CT images. Cervical 3D CT was performed between November 2011 and August 2014. Arcuate foramina were classified as partial or complete and left and/or right. Occipitalization of the atlas was classified in accordance with criteria specified by Mudaliar et al. Posterior arch defects of the atlas were classified in accordance with criteria specified by Currarino et al.ResultsOne hundred and eight vertebrae (108/1029, 10.5%) showed an arcuate foramen. Bilateral arcuate foramina were present in 41 of these vertebrae and the remaining 67 arcuate foramina were unilateral (right 31, left 36). Right-side arcuate foramina were partial on 18 sides and complete on 54 sides. Left-side arcuate foramina were partial on 24 sides and complete on 53 sides. One case of atlas assimilation was found. Twelve patients (12/1029, 1.17%) had a defect of the atlantal posterior arch. Nine of these patients (9/1029, 0.87%) had a type A posterior arch defect. We also identified one type B, one type D, and one type E defect.ConclusionPreoperative diagnosis of occipitalization of the atlas and arcuate foramina using 3D CT is of paramount importance in avoiding neurovascular injury during surgery. It is important to be aware of posterior arch defects of the atlas because they may be misdiagnosed as a fracture.

DrasCLR: A self-supervised framework of learning disease-related and anatomy-specific representation for 3D lung CT images

Carbonate reservoirs have complex pore structures and serious heterogeneity, which increase the difficulty to achieve efficient oilfield development. Conventional core plugs cannot accurately depict the heterogeneity characteristics that may beyond the scale of plugs. Thus, whole core experiments are conducted to better study this issue. The methodologies involved in this study include experimental study, outcrop observation, statistical analysis, and numerical modeling. 18 whole core sections are collected from a giant cretaceous carbonate reservoir located in the Middle East, and the OOIP of this carbonate is 4000 MM bbls. We conducted 3D CT, whole core experiments, and plug experiments for the total of 18 sections of the core. Furthermore, numerical simulation is conducted to better evaluate the impact of heterogeneity. After integrating results from all these methodologies and technologies, 2D and 3D CT images, horizontal permeability, vertical permeability, and density of whole core, horizontal plug, and vertical plug are obtained. Then, pore structure visualization and experimental results are combined, and analysis was carefully conducted. Results show that whole core can better depict the reservoir heterogeneity, and some interesting observations were found, such as when the permeability of the core section is low, permeability of whole core is greater than that of plugs; whereas when the permeability of the core section is high, permeability of whole core is lower than that of plugs. Simulation results indicate that heterogeneity is one of the major factors determining the performance of carbonate reservoirs. This study offers detailed evidence to support engineers and geologists for better understanding the heterogeneity and optimizing production plan on the carbonate reservoir.