Fig 2 - uploaded by Anderson Maciel
Content may be subject to copyright.
![Model of the liver functional segments distribution according to Couinaud [3]. Segment I is the caudate lobe, segments II, III, IVa and IVb compose the left lobe, and segments V, VI, VII and VIII compose the right lobe.](profile/Anderson-Maciel-2/publication/49628487/figure/fig2/AS:305710207389705@1449898401827/Model-of-the-liver-functional-segments-distribution-according-to-Couinaud-3-Segment-I.png)
Model of the liver functional segments distribution according to Couinaud [3]. Segment I is the caudate lobe, segments II, III, IVa and IVb compose the left lobe, and segments V, VI, VII and VIII compose the right lobe.
Source publication
Anatomic hepatectomies are resections in which compromised segments or sectors of the liver are extracted according to the topological structure of its vascular elements. Such structure varies considerably among patients, which makes the current anatomy-based planning methods often inaccurate. In this work we propose a strategy to efficiently and s...
Context in source publication
Context 1
... After this contour is entered, it is sampled for control points and a Bézier spline is created which is a smooth vector information that eliminates the dent effects caused by image resolution (1b). Such curve is copied to the next slice and the live-wires algorithm is applied to its control points to fit the new contour. It is usually necessary to manually adjust a few points between slices, but rarely more than 2 or 3 points (figure 1c and 1d). In a short time, every slice will contain a contour which is used to separate pixels inside the liver from pixels outside. These pixels are then recolored to white and black respec- tively. This new black-and-white dataset (1e) can be used as a mask to define a segmented liver in 3D (figure 1f) and to measure the total volume of the organ. chances III. II. V that ESSEL L IVER they C S LASSIFICATION are EGMENTATION used directly WITH WITH by the L SmartContour IVER surgeons. S EGMENTS We In We the developed developed remaining the the of SmartContour LiverSegments this paper, section program as a II tool describes as to a tool classify our for semi-automatic semi-automatic regions of a volumetric segmentation approach model for of image of a the CT segmentation, liver dataset. according The and program to sec- the tion has anatomic been III introduces regions designed defined to our require liver by Couinaud segments the minimum [3]. classification Such possible classifica- user tool based intervention. tion is on made vessel It interactively combines branching. by automatic In selecting section edge IV vessels we detection explain directly with how in the parametric proposed curves tools to are quickly applied create on actual organic planning contours of of liver the resections, segmented which areas. is followed by results in section V. Finally, a discussion The input and is a our stack conclusions of ordinary are CT presented images in in section grey scale. VI. The program first processes the images to increase contrast. An implementation of the live-wires algorithm [1] is then used in every slice of the dataset. It is based on the algorithm of Dijkstra [4] which solves the problem of the shortest path in a graph. When applied to an image, the algorithm finds a path in which every pixel represents a graph node with weighted edges connecting the neighborhood according to the relative neighbors tonality. The lowest cost node represents a suggestion of optimal step between the two nodes. The combination of a sequence of steps is a path representing the contour of a segmented region. In practice, a user is required to click somewhere on the liver border to define a start for the contour. After, by moving the cursor, the program automatically proposes a contour line from the starting point to the current cursor position. New clicks have to be applied when the proposed contour does not follow the organ edge. This occurs whenever a neighboring organ presents a very similar density with the liver. With standard CT images, about 10 clicks are required to define a fit contour for one slice (figure 1a). After this contour is entered, it is sampled for control points and a Bézier spline is created which is a smooth vector information that eliminates the dent effects caused by image resolution (1b). Such curve is copied to the next slice and the live-wires algorithm is applied to its control points to fit the new contour. It is usually necessary to manually adjust a few points between slices, but rarely more than 2 or 3 points (figure 1c and 1d). In a short time, every slice will contain a contour which is used to separate pixels inside the liver from pixels outside. These pixels are then recolored to white and black respec- tively. This new black-and-white dataset (1e) can be used as a mask to define a segmented liver in 3D (figure 1f) and to measure the total volume of the organ. III. V ESSEL C LASSIFICATION WITH L IVER S EGMENTS We developed the LiverSegments as a tool to classify regions of a volumetric model of the liver according to the anatomic regions defined by Couinaud [3]. Such classification is made interactively by selecting vessels directly in three-dimensions on the segmented volume data described in section II. The segment classification employed here divides the liver in 8 functional segments. They are defined according to the distribution of the hepatic vessels. Such anatomy-based classification schema has been first described by Claude Couinaud [3] and is also known as Couinaud segmentation. In general, two branches of arterial and portal blood enter the liver. They are referred as right and left as they supply the right and left sides of the organ. A simplified classification divide the liver in left and right lobe, which are supplied by each of the two main vessel branches, plus the caudate lobe, which receives blood from the two main branches and is considered separately. Such simplified view aids in understanding the Couinaud segmentation depicted in figure 2. The caudate lobe, which is in the central-posterior part of the liver, represents the segment I. The left lobe contains the segments II, III, IVa and IVb. The right lobe contains the segments V, VI, VII and VIII. The input to the program is the segmented data from SmartContour (sectio II) and the original CT. To increase contrast between the parenchyma and the blood vessels, a pipe of image filters is used: gaussian blur; noise reduction, sharpen, brightness decrease. This helps to apply trans- parency between the parenchyma and the vessels, making them visually more solid and uniform as in figure 3c. Having a 3D view of the vessels, the user is required to insert points on the vessel tree to label the veins as belonging to one of the 8 segments of Couinaud. Then the system labels all voxels of the liver according to their distances to those points. We used the same strategy of a Voronoi diagram which describes a spatial decomposition by means of the proximity of the regions with a given set of points. The closest regions to given points are associated to them. In our implementation, multiple points may define a unique segment. IV. S URGERY P LANNING After patient specific liver segmentation and segment classification processes are done with the programs described in sections II and III, damaged areas can be identified in 3D and the doctors can decide which segments will be removed surgically. They can also plan how to make the incisions in such a way that they avoid unnecessarily sectioning of important vessels. Another important information that can be obtained is related to total and functional volumes. Total liver volume and each functional segment volumes are automatically calculated by the LiverSegments tool. These volumes, together with the identification of potential areas of ischemia and venous stasis in consequence of surgery, are taken into account to estimate the total volume of the remaining functional tissue. Diagnostic and planning start with the CT acquisition. Conventional contrast injections are used, and the three phases of the liver circulation (portal, venous and arterial) can be acquired separately along time, as usual. They can also be acquired all together after a sufficient amount of contrast perfused through the 3 systems. In either case, datasets are then exported in DICOM format from the CT scanner. They can be imported into the SmarContour for segmentation. After liver segmentation, SmartContour exports both the original images and a 3D segmentation mask. This information is the input to LiverSegments. In LiverSegments the user (radiologist or surgeon) inspects the 3D volume of the liver, showing or not the surrounding organs (Figure 3a). At this point, the total volume of the liver is also calculated and displayed. Then, interactively, the user sets the center and threshold of the density window for visualization. Selected density ranges are set to transparent allowing the vessels to be highlighted (Figure 3b). At this moment the user is able to classify the liver segments (for example, Couinaud’s) by clicking on vessel branches directly on the 3D view. Such branch selection eventually produces a color distribution not only on the vessels but also in the neighboring regions of the parenchyma (Figures 3c and Figure 3d). Each colored region corresponds to one functional segment. Borders between these regions are suggested as incision lines for surgery. When segments are selected for removal, the functional volume of the remaining liver is also calculated and displayed for consideration. V. R ESULTS We performed comparative experiments to evaluate the SmartContour and the LiverSegments in the context of volume estimation for hepatectomy planning. The tests are based on 4 CT datasets and have been performed on a Core 2 desktop PC. Our hypothesis is that liver segmentation and volume estimation from CT using our methods are at least as accurate as the ones obtained with the workstation attached to the CT scanner. Figure 4 compares a tipical CT workstation segmentation with our methods segmentation for the same dataset. Notice that the very organic shape of our result is closer to the actual anatomy. Table I, in turn, presents the volume data obtained during the tests for a set of 4 liver datasets. For decision making, the CT workstation calculated volume is trusted to have an error margin of 10% . Notice that the volume calculated with our methods is close to the CT workstation by a monotonous margin inferior to 5% . The interactive data visualization for preoperative planning provided by our tools allows for thoroughly analysis of the liver directly in 3D, which is far more comprehensive than usual methods. In addition, the 3D viewer is available in the operation room for verification during surgery. VI. D ISCUSSION AND C ONCLUSION In this paper we present a new strategy for liver surgery planning which takes into account the blood vessel branching within the organ. ...
Similar publications
Automatic segmentation techniques, despite demonstrating excellent overall accuracy, can often produce inaccuracies in local regions. As a result, correcting segmentations remains an important task that is often laborious, especially when done manually for 3D datasets. This work presents a powerful tool called Intelligent Learning-Based Editor of S...
Citations
... Clinically, one of the procedures that considerably benefit from visual aid is liver resection, because portal blood vessels may be difficult to visualize under the liver, residual liver volume is hard to estimate, and there may be hidden tumours that are not detected initially. 25,26,29,30 Patient-specific virtual reality anatomical models may be built through preoperative or intraoperative images depending on data availability. 26,30,31 During planning, surgeons may freely rotate the model to detect unexpected tumours and adjust the resection plane accordingly; based on the resection plane, the software returns the residual liver volume as well as security margins around the liver for surgeons to analyze the validity of the plan. ...
... 25,26,29,30 Patient-specific virtual reality anatomical models may be built through preoperative or intraoperative images depending on data availability. 26,30,31 During planning, surgeons may freely rotate the model to detect unexpected tumours and adjust the resection plane accordingly; based on the resection plane, the software returns the residual liver volume as well as security margins around the liver for surgeons to analyze the validity of the plan. 26,30 Patient-specific digital models can be generated to support valve replacement in heart surgeries. ...
... 26,30,31 During planning, surgeons may freely rotate the model to detect unexpected tumours and adjust the resection plane accordingly; based on the resection plane, the software returns the residual liver volume as well as security margins around the liver for surgeons to analyze the validity of the plan. 26,30 Patient-specific digital models can be generated to support valve replacement in heart surgeries. 32,33 Trajectory optimization done by machine learning can reduce the risks in cannulation and catheterization procedures. ...
Background
In recent years, numerous innovative yet challenging surgeries, such as minimally invasive procedures, have introduced an overwhelming amount of new technologies, increasing the cognitive load for surgeons and potentially diluting their attention. Cognitive support technologies (CSTs) have been in development to reduce surgeons’ cognitive load and minimize errors. Despite its huge demands, it still lacks a systematic review.
Methods
Literature was searched up until May 21st, 2021. Pubmed, Web of Science, and IEEExplore. Studies that aimed at reducing the cognitive load of surgeons were included. Additionally, studies that contained an experimental trial with real patients and real surgeons were prioritized, although phantom and animal studies were also included. Major outcomes that were assessed included surgical error, anatomical localization accuracy, total procedural time, and patient outcome.
Results
A total of 37 studies were included. Overall, the implementation of CSTs had better surgical performance than the traditional methods. Most studies reported decreased error rate and increased efficiency. In terms of accuracy, most CSTs had over 90% accuracy in identifying anatomical markers with an error margin below 5 mm. Most studies reported a decrease in surgical time, although some were statistically insignificant.
Discussion
CSTs have been shown to reduce the mental workload of surgeons. However, the limited ergonomic design of current CSTs has hindered their widespread use in the clinical setting. Overall, more clinical data on actual patients is needed to provide concrete evidence before the ubiquitous implementation of CSTs.
... Provided that portal and hepatic veins are extracted, liver segments are defined with respect to voxel distances to specific branches [104,105], voxel projections onto vascular intersections [106], or categorical search by Voronoi diagram [107][108][109]. However, these methods suffer from computationally intensive voxel sorting. ...
The earliest and most accurate detection of the pathological manifestations of hepatic diseases ensures effective treatments and thus positive prognostic outcomes. In clinical settings, screening and determining the extent of a pathology are prominent factors in preparing remedial agents and administering appropriate therapeutic procedures. Moreover, in a patient undergoing liver resection, a realistic preoperative simulation of the subject-specific anatomy and physiology also plays a vital part in conducting initial assessments, making surgical decisions during the procedure, and anticipating postoperative results. Conventionally, various medical imaging modalities, e.g. , computed tomography, magnetic resonance imaging, and positron emission tomography, have been employed to assist in these tasks. In fact, several standardized procedures, such as lesion detection and liver segmentation, are also incorporated into prominent commercial software packages. Thus far, most integrated software as a medical device typically involves tedious interactions from the physician, such as manual delineation and empirical adjustments, as per a given patient. With the rapid progress in digital health approaches, especially medical image analysis, a wide range of computer algorithms have been proposed to facilitate those procedures. They include pattern recognition of a liver, its periphery, and lesion, as well as pre- and postoperative simulations. Prior to clinical adoption, however, software must conform to regulatory requirements set by the governing agency, for instance, valid clinical association and analytical and clinical validation. Therefore, this paper provides a detailed account and discussion of the state-of-the-art methods for liver image analyses, visualization, and simulation in the literature. Emphasis is placed upon their concepts, algorithmic classifications, merits, limitations, clinical considerations, and future research trends.
... In the past two decades, a lot of research work has been done in computer-assisted liver segment segmentation [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21], most of them using traditional machine learning techniques [4,[7][8][9][10][11][12][13][14][15][16][17][18][19][20][21], which cannot meet the needs of clinical applications in terms of segmentation performance and efficiency. The method proposed by Lebre et al. [8] requires first segmentation of the hepatic vessels using the skeletonization process, and then the main direction of the largest vessels was extracted to achieve separation of different liver segments. ...
... In the past two decades, a lot of research work has been done in computer-assisted liver segment segmentation [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21], most of them using traditional machine learning techniques [4,[7][8][9][10][11][12][13][14][15][16][17][18][19][20][21], which cannot meet the needs of clinical applications in terms of segmentation performance and efficiency. The method proposed by Lebre et al. [8] requires first segmentation of the hepatic vessels using the skeletonization process, and then the main direction of the largest vessels was extracted to achieve separation of different liver segments. ...
Objective
We aim to develop and validate a three-dimensional convolutional neural network (3D-CNN) model for automatic liver segment segmentation on MRI images.
Methods
This retrospective study evaluated an automated method using a deep neural network that was trained, validated, and tested with 367, 157, and 158 portal venous phase MR images, respectively. The Dice similarity coefficient (DSC), mean surface distance (MSD), Hausdorff distance (HD), and volume ratio (RV) were used to quantitatively measure the accuracy of segmentation. The time consumed for model and manual segmentation was also compared. In addition, the model was applied to 100 consecutive cases from real clinical scenario for a qualitative evaluation and indirect evaluation.
Results
In quantitative evaluation, the model achieved high accuracy for DSC, MSD, HD and RV (0.920, 3.34, 3.61 and 1.01, respectively). Compared to manual segmentation, the automated method reduced the segmentation time from 26 min to 8 s. In qualitative evaluation, the segmentation quality was rated as good in 79% of the cases, moderate in 15% and poor in 6%. In indirect evaluation, 93.4% (99/106) of lesions could be assigned to the correct segment by only referring to the results from automated segmentation.
Conclusion
The proposed model may serve as an effective tool for automated anatomical region annotation of the liver on MRI images.
... Couinaud segments of the liver were finally labelled following [16]. In addition to categorical search, Voronoi algorithm was adopted by Debarba [19] and Chen [20], et al., to classify voxels with their segments. Their difference was in calculating a Voronoi diagram. ...
... Some remarks may be drawn from the above surveys. The methods in the former group [16][17][18][19][20][21] partitioned a liver based on membership values associated with interior voxels. Thus, their major hinderance was evaluation of membership function for each voxel with respect to relevant features, such as vascular structures, was computationally intensive. ...
Successful liver resection relies on accurate estimation of future liver remnant volume (FLRV). According to Couinaud’s scheme, a liver is composed of eight functionally independent segments, each of which has its own vascular in- and out-flow tracks. Segmenting a liver by this scheme is vital to postoperative regeneration and hence prognosis outcome. Conventionally, estimation of liver segments was often done by hand on 3D computed tomography. The process is generally tedious, time consuming, and prone to observer variability. Alternatively, computerized methods had been proposed but impeded by anatomically irrelevant approximation and manually specified markers. To resolve the issues, this paper presents a novel method for functional liver segmentation. Its main contribution was performing analyses of differential geometry directly on a liver surface and interior venous system. Except for a few points being placed on major vessels, anatom-ical references required for defining all separating surfaces were automatically identified. To demonstrate its merits, virtual liver resection was implemented on the standard MICCAI SLIVER07 dataset, and the resultant segments were benchmarked against four most related works. Visual and numerical assessments reported herein indicated that our method could faithfully label all Couinaud’s segments, especially the caudate, with lesser degree of user interaction. The preliminary findings suggested that it can be integrated into augmented surgical planning and intervention.
... In Computer-Aided Design, segmentation were reported based on the features of surface mesh (Xu, Zhou, Wu, Shui, & Ali, 2015;Park, Lee, Chae, & Kwon, 2019). One noteworthy work was to segment the CT image of LV using that of CA: In this work, a user manually picks some voxel points belonging to CA so that each voxel of LV can be assigned to its closest picked point (Debarba, Zanchet, Fracaro, Maciel, & Kalil, 2010;Kurata et al., 2015). Image-based approach was also used for analyzing the morphometry of CA, e.g. the diameter, branching pattern, etc. (Wischgoll, Choy, Ritman, & Kassab, 2008). ...
The role of coronary arteries is to supply sufficient blood to myocardium. Obstruction of coronary arteries limits blood supply and causes myocardial ischemia or acute myocardial infarction, a major cause of human death. Hence, the quantification of the regional amount of heart muscle subtended by obstructed coronary arteries is of critical value in clinical medicine. However, conventional methods are inaccurate and frequently disagree with clinical practice. This study proposes a novel medial-axis-based correspondence (Medial-ABC) algorithm to find the correspondence between myocardium and coronary artery in order to segment regional myocardium at risk subtended by any potentially obstructed coronary artery. Given the triangular mesh models of coronary artery and myocardium, the proposed algorithm (i) computes the medial axis of coronary artery, (ii) finds the correspondence using the medial axis of coronary artery, and (iii) segments the coronary artery and myocardium. The proposed algorithm provides a robust mathematical linkage between myocardium at risk and supplying coronary arteries so that ischemic myocardial regions can be accurately identified. Hence, both the extent and severity of myocardial ischemia can be quantified effectively, efficiently, and accurately. Furthermore, the constructed mesh model of segmented coronary artery and myocardium can be post-processed for applications such as building optimization models of cardiac systems. The CardiacVis program, which implements the Medial-ABC algorithm, is freely available at Voronoi Diagram Research Center (http://voronoi.hanyang.ac.kr/software/cardiacvis) and will be an invaluable tool for quantitative patient-specific risk stratification in clinical practice.
... Reitinger interactively. These selected points are used for liver segmental anatomy similar to the approach proposed by Selle et al. [16]. ...
Purpose
For the liver to remain viable, the resection during hepatectomy procedure should proceed along the major vessels; hence, the resection planes of the anatomic segments are defined, which mark the peripheries of the self-contained segments inside the liver. Liver anatomic segments identification represents an essential step in the preoperative planning for liver surgical resection treatment.
Method
The method based on constructing atlases for the portal and the hepatic veins bifurcations, the atlas is used to localize the corresponding vein in each segmented vasculature using atlas matching. Point-based registration is used to deform the mesh of atlas to the vein branch. Three-dimensional distance map of the hepatic veins is constructed; the fast marching scheme is applied to extract the centerlines. The centerlines of the labeled major veins are extracted by defining the starting and the ending points of each labeled vein. Centerline is extracted by finding the shortest path between the two points. The extracted centerline is used to define the trajectories to plot the required planes between the anatomical segments.
Results
The proposed approach is validated on the IRCAD database. Using visual inspection, the method succeeded to extract the major veins centerlines. Based on that, the anatomic segments are defined according to Couinaud segmental anatomy.
Conclusion
Automatic liver segmental anatomy identification assists the surgeons for liver analysis in a robust and reproducible way. The anatomic segments with other liver structures construct a 3D visualization tool that is used by the surgeons to study clearly the liver anatomy and the extension of the cancer inside the liver.
... The authors used the work of Selle et al. for liver segmental anatomy. Debarba et al. used the liver segmentation tool for vessel classification by annotating the vessels by placing points interactively [110]. Pamulapati et al. used a graph representation of the vessel tree of the main hepatic and portal veins. ...
Preoperative planning for liver surgical treatments is an essential planning tool that aids in reducing the risks of surgical resection. Based on the computed tomography (CT) images, the resection can be planned before the actual tumour resection surgery. The computer-aided system provides an overview of the spatial relationships of the liver organ and its internal structures, tumours, and vasculature. It also allows for an accurate calculation of the remaining liver volume after resection. The aim of this paper was to review the main stages of the computer-aided system that helps to evaluate the risk of resection during liver cancer surgical treatments. The computer-aided system assists with surgical planning by enabling physicians to get volumetric measurements and visualise the liver, tumours, and surrounding vasculature. In this paper, it is concluded that for accurate planning of tumour resections, the liver organ and its internal structures should be segmented to understand the clear spatial relationship between them, thus allowing for a safer resection. This paper presents the main proposed segmentation techniques for each stage in the computer-aided system, namely the liver organ, tumours, and vessels. From the reviewed methods, it has been found that instead of relying on a single specific technique, a combination of a group of techniques would give more accurate segmentation results. The extracted masks from the segmentation algorithms are fused together to give the surgeons the 3D visualisation tool to study the spatial relationships of the liver and to calculate the required resection planning parameters.
... Additionally, the presumed clinical applicability and anatomical integrity of a 3D model may improve our diagnostic skills and interventional planning. Three-dimensional models play an important role in neurosurgery, vascular, hepatic and orthopaedic surgery [10][11][12][13][14][15][16]. Pelvic surgeons, however, remain primarily reliant on the interpretation of 2D crosssectional imaging [11]. ...
Aims:
Our aim was to assess the feasibility of rendering 3D pelvic models using magnetic resonance imaging (MRI) scans of patients with vaginal, urethral and paraurethral lesions and obtain additional information previously unavailable through 2D imaging modalities.
Methods:
A purposive sample of five female patients 26-40 years old undergoing investigations for vaginal or paraurethral mass was obtained in a tertiary teaching hospital. 3D volume renderings of the bladder, urethra and paraurethral masses were constructed using 3D-Slicer v.3.4.0. Spatial dimensions were determined and compared with findings from clinical, MRI, surgical and histopathological reports. The quality of information regarding size and location of paraurethral masses obtained from 3D models was compared with information from cross-sectional MRI and review of clinical, surgical and histopathological findings.
Results:
The analysis of rendered 3D models yielded detailed anatomical dimensions and provided information that was in agreement and in higher detail than information based on clinical examination, cross-sectional 2D MRI analysis and histopathological reports. High-quality pelvic 3D models were rendered with the characteristics and resolution to allow identification and detailed viewing of the spatial relationship between anatomical structures.
Conclusions:
To our knowledge, this is the first preliminary study to evaluate the role of MRI-based 3D pelvic models for investigating paraurethral masses. This is a feasible technique and may prove a useful addition to conventional 2D MRI. Further prospective studies are required to evaluate this modality for investigating such lesions and planning appropriate management.
... A rational liver surgery can be planned by analyzing the structures of the three liver vascular trees (portal vein, hepatic vein, and hepatic artery) in LDLT and the spatial relationship between a tumor(s) and the vascular trees in tumor resection. As shown in Figure 1, a 3D VLSP system needs to provide not only visual information for the location and size of a tumor(s), the liver vascular trees, the anatomical segments of the liver, and the location, orientation, and shape of a resection plane, but also quantitative information of the volumes of the liver, vessels, tumor(s), liver segments, remnant, and/or graft (Debarba et al., 2010;Reitinger et al., 2006;Sorantin et al., 2008). ...
Objective: The present study developed a user-centered 3D virtual liver surgery planning (VLSP) system called Dr. Liver to provide preoperative information for safe and rational surgery. Background: Preoperative 3D VLSP is needed for patients` safety in liver surgery. Existing systems either do not provide functions specialized for liver surgery planning or do not provide functions for cross-check of the accuracy of analysis results. Method: Use scenarios of Dr. Liver were developed through literature review, benchmarking, and interviews with surgeons. User interfaces of Dr. Liver with various user-friendly features (e.g., context-sensitive hotkey menu and 3D view navigation box) was designed. Novel image processing algorithms (e.g., hybrid semi-automatic algorithm for liver extraction and customized region growing algorithm for vessel extraction) were developed for accurate and efficient liver surgery planning. Usability problems of a preliminary version of Dr. Liver were identified by surgeons and system developers and then design changes were made to resolve the identified usability problems. Results: A usability testing showed that the revised version of Dr. Liver achieved a high level of satisfaction (6.1±0.8 out of 7) and an acceptable time efficiency (26.7±0.9 min) in liver surgery planning. Conclusion: Involvement of usability testing in system development process from the beginning is useful to identify potential usability problems to improve for shortening system development period and cost. Application: The development and evaluation process of Dr. Liver in this study can be referred in designing a user-centered system.
... Furthermore, the Voronoi diagram is a type of centerline method that divides space (i.e., volume) by seeded points or lines [8], and is utilized in industry and geography. In diagnostic imaging, a previous study has shown the usefulness of this algorithm for liver segmentation, based on CT portal venography [9], and a recent study has reported that ICA-based stenosis-related CT myocardial territory correlates with the SPECT-based myocardial area at risk (MAAR) [10]. Coronary CTA stenosis-related CT myocardial territory is an assumption of the maximum MAAR that is obtained from a resting coronary CTA dataset. ...
Purpose: To assess the clinical feasibility of automated segmentation of the myocardial area at risk (MAAR) using coronary computed tomography angiography (CT-MAAR), as compared to stress magnetic resonance myocardial perfusion imaging (MR-MPI). Materials and Methods: Thirty patients who underwent coronary computed tomography angiography (CTA) and stress MR-MPI were retrospectively evaluated. The myocardial territory of the left ventricle (LV) distal to coronary artery stenosis (≥50% or ≥70% stenosis on coronary CTA) was three-dimensionally quantified using a Voronoi diagram. The ratio of all stenosis-related territories to the LV volume was defined as CT-MAAR (%-LV volume). The proportion of segments with perfusion defects in stress MR-MPI to the total of 16 segments (range: 0% - 100%; with a 6.3%-interval scale) was defined as the reference. Correlation was assessed using Spearman’s test. The capability of CT-MAAR to predict the ischemic burden was assessed. Results: Stress MR-MPI depicted a median ischemic burden of 25.2% (range: 18.9% - 44.1%) in 30 patients without myocardial infarction. When CTA stenosis criteria of ≥50% (n = 30) and ≥70% (n = 27) were applied to estimate CT-MAAR, the median CT-MAAR values were 48.2% (31.6% - 64.3%) and 32.5% (23.7% - 51.9%), respectively. The correlations between the CT-MAAR values and the MR-based ischemic burden were significant (0.73 and 0.97 for ≥50% and ≥70% stenosis, respectively). CT-MAAR predicted the MR-based ischemic burden within ±1 segment of %-LV (6.3%) in 40% (12/30) of patients with ≥50% stenosis, and in 81.5% (22/27) of patients with ≥70% stenosis. Conclusions: Comprehensive assessment of resting coronary CTA combined with Voronoi diagram-based myocardial segmentation may help predict the myocardial ischemic burden in patients with severe coronary CTA stenosis.
