Kazunari Misawa

Automated Abdominal Multi-Organ Segmentation With Subject-Specific Atlas Generation

A robust automated segmentation of abdominal organs can be crucial for computer aided diagnosis and laparoscopic surgery assistance. Many existing methods are specialized to the segmentation of individual organs and struggle to deal with the variability of the shape and position of abdominal organs. We present a general, fully-automated method for multi-organ segmentation of abdominal computed tomography (CT) scans. The method is based on a hierarchical atlas registration and weighting scheme that generates target specific priors from an atlas database by combining aspects from multi-atlas registration and patch-based segmentation, two widely used methods in brain segmentation. The final segmentation is obtained by applying an automatically learned intensity model in a graph-cuts optimization step, incorporating high-level spatial knowledge. The proposed approach allows to deal with high inter-subject variation while being flexible enough to be applied to different organs. We have evaluated the segmentation on a database of 150 manually segmented CT images. The achieved results compare well to state-of-the-art methods, that are usually tailored to more specific questions, with Dice overlap values of 94%, 93%, 70%, and 92% for liver, kidneys, pancreas, and spleen, respectively.

Multi-organ Segmentation Based on Spatially-Divided Probabilistic Atlas from 3D Abdominal CT Images

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.

Association of prostate stem cell antigen gene polymorphisms with the risk of stomach cancer in Japanese

A recent whole‐genome association study identified a strong association between polymorphisms in the prostate stem cell antigen (PSCA) gene and stomach cancer risk. In this case‐control study, we aimed to validate this association, and further to explore environmental factors possibly interacting with PSCA polymorphisms in 708 incident stomach cancer cases and 708 age–sex matched controls. The association between PSCA polymorphisms and Helicobacter pylori infection was also examined. We found that rs2294008 and rs2976392, which were strongly linked to each other (D′ = 1.00), were significantly associated with stomach cancer risk. Per allele odds ratio for rs2994008 was 1.40 (95% confidence interval: 1.19–1.65; p = 3.7 × 10−5). We found significant interaction with a family history of stomach cancer in first‐degree relatives (p‐heterogeneity = 0.009). Similar to originally reported association, we found significant heterogeneity between diffuse and intestinal type (p‐heterogeneity = 0.007). No association was seen between PSCA polymorphisms and H. pylori infection. In conclusion, PSCA polymorphisms are associated with stomach cancer risk in Japanese. A possible interaction with family history warrants further evaluation.

Discriminative dictionary learning for abdominal multi-organ segmentation.

An automated segmentation method is presented for multi-organ segmentation in abdominal CT images. Dictionary learning and sparse coding techniques are used in the proposed method to generate target specific priors for segmentation. The method simultaneously learns dictionaries which have reconstructive power and classifiers which have discriminative ability from a set of selected atlases. Based on the learnt dictionaries and classifiers, probabilistic atlases are then generated to provide priors for the segmentation of unseen target images. The final segmentation is obtained by applying a post-processing step based on a graph-cuts method. In addition, this paper proposes a voxel-wise local atlas selection strategy to deal with high inter-subject variation in abdominal CT images. The segmentation performance of the proposed method with different atlas selection strategies are also compared. Our proposed method has been evaluated on a database of 150 abdominal CT images and achieves a promising segmentation performance with Dice overlap values of 94.9%, 93.6%, 71.1%, and 92.5% for liver, kidneys, pancreas, and spleen, respectively.

Multi-organ abdominal CT segmentation using hierarchically weighted subject-specific atlases

A robust automated segmentation of abdominal organs can be crucial for computer aided diagnosis and laparoscopic surgery assistance. Many existing methods are specialised to the segmentation of individual organs or struggle to deal with the variability of the shape and position of abdominal organs. We present a general, fully-automated method for multi-organ segmentation of abdominal CT scans. The method is based on a hierarchical atlas registration and weighting scheme that generates target specific priors from an atlas database by combining aspects from multi-atlas registration and patch-based segmentation, two widely used methods in brain segmentation. This approach allows to deal with high inter-subject variation while being flexible enough to be applied to different organs. Our results on a dataset of 100 CT scans compare favourable to the state-of-the-art with Dice overlap values of 94%, 91%, 66% and 94% for liver, spleen, pancreas and kidney respectively.

Geodesic Patch-Based Segmentation

Label propagation has been shown to be effective in many automatic segmentation applications. However, its reliance on accurate image alignment means that segmentation results can be affected by any registration errors which occur. Patch-based methods relax this dependence by avoiding explicit one-to-one correspondence assumptions between images but are still limited by the search window size. Too small, and it does not account for enough registration error; too big, and it becomes more likely to select incorrect patches of similar appearance for label fusion. This paper presents a novel patch-based label propagation approach which uses relative geodesic distances to define patient-specific coordinate systems as spatial context to overcome this problem. The approach is evaluated on multi-organ segmentation of 20 cardiac MR images and 100 abdominal CT images, demonstrating competitive results.

Organ segmentation from 3d abdominal CT images based on atlas selection and graph cut

This paper presents a method for segmenting abdominal organs from 3D abdominal CT images based on atlas selection and graph cut. The training samples are divided into multiple clusters based on the image similarity. The average image and atlas for each cluster are created. For an input image, we select the most similar atlas to the input image by measuring the image similarity between the input and average images. Segmentation of organs based on the MAP estimation using the selected atlas is then performed, followed by the precise segmentation by the graph cut algorithm. We applied the proposed method to a hundred cases of CT images. The experimental results showed that the extraction accuracy could be improved using multiple atlases, achieving more than 90% of the precision rate except for the pancreas.

Automated nomenclature of upper abdominal arteries for displaying anatomical names on virtual laparoscopic images

This paper presents a method for automated nomenclature of abdominal arteries that are extracted from 3D CT images based on the combination optimization approach for the displaying anatomical names on virtual laparoscopic images. It is important to understand the blood vessel network of a patient. Our proposed method recognizes the anatomical names of each arterial branch extracted from contrasted 3D images based on geometric features. We employ a combination optimization approach for treating the variations of branching patterns and overlay recognized anatomical names on virtual laparoscopic views for assisting the recognition of patient anatomy for surgeons. Experimental results using 89 cases of 3D CT images showed that the nomenclature accuracy for uncorrected blood vessel tree and corrected blood vessel tree were about 84.2% and 88.8% in average respectively and demonstrated anatomical name overlay on virtual laparoscopic images.

Clinical course of gastrointestinal stromal tumor diagnosed by endoscopic ultrasound-guided fine-needle aspiration.

Gastrointestinal stromal tumors (GIST) are the most common mesenchymal tumor of the gastrointestinal tract. However, little is known about the clinical presentation of GIST, especially small lesions. The purpose of the present study was to clarify the efficacy of endoscopic ultrasound‐guided fine‐needle aspiration (EUS‐FNA) for the diagnosis of GIST and to determine its clinical course.

Adequate length of the surgical distal resection margin in rectal cancer: from the viewpoint of pathological findings.

BACKGROUND Previous studies have not identified how to determine the optimal distal margin in rectal cancer based on histopathological diagnosis. We examined the surgical distal resection margin from a histopathological viewpoint. METHODS We enrolled 629 patients. The type of distal spread was evaluated, and the maximum length of distal spread was measured using a micrometer. RESULTS The frequencies of discontinuous spread type were 1.0%, 8.4%, 52.9%, and 81.5%, and the average lengths of distal spread were .5 ± 1.3 mm, 7 ± 1.8 mm, 2.7 ± 2.4 mm, and 10.0 ± 9.5 mm for well-differentiated adenocarcinomas, moderately differentiated adenocarcinomas, solid (por1)-type poorly differentiated adenocarcinomas, and nonsolid (por2)-type poorly differentiated adenocarcinomas, (moderately vs solid [por1] type: P = .004), respectively. CONCLUSIONS The surgical distal resection margin based on pathological diagnosis is longer somewhat than that based on macroscopic findings. Therefore, it is important to select surgical procedures with great care to ensure an adequate surgical distal resection margin.