Development of multi-class computer-aided diagnostic systems using the NICE/JNET classifications for colorectal lesions.

BACKGROUND AND AIM: Diagnostic support using artificial intelligence may contribute to the equalization of endoscopic diagnosis of colorectal lesions. We developed computer-aided diagnosis (CADx) support system for diagnosing colorectal lesions using the NBI International Colorectal Endoscopic (NICE) classification and the Japan NBI Expert Team (JNET) classification. METHODS: Using Residual Network as the classifier and NBI images as training images, we developed a CADx based on the NICE classification (CADx-N) and a CADx based on the JNET classification (CADx-J). For validation, 480 non-magnifying and magnifying NBI images were used for the CADx-N and 320 magnifying NBI images were used for the CADx-J. The diagnostic performance of the CADx-N was evaluated using the magnification rate. RESULTS: The accuracy of the CADx-N for Types 1, 2, and 3 was 97.5%, 91.2%, and 93.8%, respectively. The diagnostic performance for each magnification level was good (no statistically significant difference). The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the CADx-J were 100%, 96.3%, 82.8%, 100%, and 96.9% for Type 1; 80.3%, 93.7%, 94.1%, 79.2%, and 86.3% for Type 2A; 80.4%, 84.7%, 46.8%, 96.3%, and 84.1% for Type 2B; and 62.5%, 99.6%, 96.8%, 93.8%, and 94.1% for Type 3, respectively. CONCLUSIONS: The multi-class CADx systems had good diagnostic performance with both the NICE and JNET classifications and may aid in educating non-expert endoscopists and assist in diagnosing colorectal lesions.

Functional and Structural Changes in the Corticospinal Tract of Streptozotocin-Induced Diabetic Rats.

This study aimed to reveal functional and morphological changes in the corticospinal tract, a pathway shown to be susceptible to diabetes. Type 1 diabetes was induced in 13-week-old male Wistar rats administered streptozotocin. Twenty-three weeks after streptozotocin injection, diabetic animals and age-matched control animals were used to demonstrate the conduction velocity of the corticospinal tract. Other animals were used for morphometric analyses of the base of the dorsal funiculus of the corticospinal tract in the spinal cord using both optical and electron microscopy. The conduction velocity of the corticospinal tract decreased in the lumbar spinal cord in the diabetic animal, although it did not decrease in the cervical spinal cord. Furthermore, atrophy of the fibers of the base of the dorsal funiculus was observed along their entire length, with an increase in the g-ratio in the lumbar spinal cord in the diabetic animal. This study indicates that the corticospinal tract fibers projecting to the lumbar spinal cord experience a decrease in conduction velocity at the lumbar spinal cord of these axons in diabetic animals, likely caused by a combination of axonal atrophy and an increased g-ratio due to thinning of the myelin sheath.

High-Resolution Representation for Mobile Mapping Data in Curved Regular Grid Model.

As mobile mapping systems become a mature technology, there are many applications for the process of the measured data. One interesting application is the use of driving simulators that can be used to analyze the data of tire vibration or vehicle simulations. In previous research, we presented our proposed method that can create a precise three-dimensional point cloud model of road surface regions and trajectory points. Our data sets were obtained by a vehicle-mounted mobile mapping system (MMS). The collected data were converted into point cloud data and color images. In this paper, we utilize the previous results as input data and present a solution that can generate an elevation grid for building an OpenCRG model. The OpenCRG project was originally developed to describe road surface elevation data, and also defined an open file format. As it can be difficult to generate a regular grid from point cloud directly, the road surface is first divided into straight lines, circular arcs, and and clothoids. Secondly, a non-regular grid which contains the elevation of road surface points is created for each road surface segment. Then, a regular grid is generated by accurately interpolating the elevation values from the non-regular grid. Finally, the curved regular grid (CRG) model files are created based on the above procedures, and can be visualized by OpenCRG tools. The experimental results on real-world data show that the proposed approach provided a very-high-resolution road surface elevation model.

Neck and trunk representations in the primary motor cortex in rats.

[Purpose] The neck and trunk play crucial roles in body movement and are extremely important areas of treatment for physical therapists. However, many aspects of the neural basis of this motor control remain unknown. Therefore, we investigated the distribution and electrophysiological properties of the neck and trunk in the primary motor cortex in rats. [Subjects and Methods] Using intracortical microstimulation, we investigated the somatotopic representation and movements induced by electrical stimulation of the neck and truck areas of the motor cortex in 8 Wistar rats. [Results] We determined that the neck and trunk areas are located separately on the rostral and caudal sides of the motor cortex, respectively. The neck area was significantly larger in size, while the threshold was significantly larger for the trunk area. Stimulation of the neck area with a current higher than the threshold induced movement of the forelimbs, jaw, trunk, and whiskers. However, stimulation of the trunk area did not result in movement in sites other than the trunk. [Conclusion] During movement, the respective activities of the neck and trunk are interdependent. However, due to the separate locations of these areas in the motor cortex, their properties differ greatly.

Motor cortex somatotopic presentation after restriction of neck movement in rats.

[Purpose] In this study, we aimed to investigate the effects of neck movement restriction on somatotopic mapping of the motor cortex. We restricted cervical extension for two weeks and investigated the effects on motor cortex somatic representation in rats. [Subjects and Methods] We placed six Wistar rats into each of three groups: (i) the experimental group, in which cervical extension was restricted; (ii) the sham group, in which cervical movement was not restricted, but a splint was placed in the shoulder girdle; and (iii) the control group. After cervical immobilization for two weeks, we evaluated the motor cortex somatic representation using intra-cortical micro-stimulation. [Results] In the experimental group, the areas of the cervical and vibrissal domains of the motor cortex decreased by approximately 50%, and the forelimb domain showed slight reduction. In addition, a trunk domain formed at the locus of the vibrissal area. There were no differences between the sham and control groups. [Conclusion] Restriction of cervical extension for two weeks resulted in changes in motor cortex somatic representation. Reversible changes occurred in cortical areas that controlled the neck and parts of the body involved in cervical movement.

Computer-aided diagnosis of colorectal polyp histology by using a real-time image recognition system and narrow-band imaging magnifying colonoscopy.

BACKGROUND AND AIMS: It is necessary to establish cost-effective examinations and treatments for diminutive colorectal tumors that consider the treatment risk and surveillance interval after treatment. The Preservation and Incorporation of Valuable Endoscopic Innovations (PIVI) committee of the American Society for Gastrointestinal Endoscopy published a statement recommending the establishment of endoscopic techniques that practice the resect and discard strategy. The aims of this study were to evaluate whether our newly developed real-time image recognition system can predict histologic diagnoses of colorectal lesions depicted on narrow-band imaging and to satisfy some problems with the PIVI recommendations. METHODS: We enrolled 41 patients who had undergone endoscopic resection of 118 colorectal lesions (45 nonneoplastic lesions and 73 neoplastic lesions). We compared the results of real-time image recognition system analysis with that of narrow-band imaging diagnosis and evaluated the correlation between image analysis and the pathological results. RESULTS: Concordance between the endoscopic diagnosis and diagnosis by a real-time image recognition system with a support vector machine output value was 97.5% (115/118). Accuracy between the histologic findings of diminutive colorectal lesions (polyps) and diagnosis by a real-time image recognition system with a support vector machine output value was 93.2% (sensitivity, 93.0%; specificity, 93.3%; positive predictive value (PPV), 93.0%; and negative predictive value, 93.3%). CONCLUSIONS: Although further investigation is necessary to establish our computer-aided diagnosis system, this real-time image recognition system may satisfy the PIVI recommendations and be useful for predicting the histology of colorectal tumors.

A computer system to be used with laser-based endoscopy for quantitative diagnosis of early gastric cancer.

GOALS: To evaluate the usefulness of a newly devised computer system for use with laser-based endoscopy in differentiating between early gastric cancer, reddened lesions, and surrounding tissue. BACKGROUND: Narrow-band imaging based on laser light illumination has come into recent use. We devised a support vector machine (SVM)-based analysis system to be used with the newly devised endoscopy system to quantitatively identify gastric cancer on images obtained by magnifying endoscopy with blue-laser imaging (BLI). We evaluated the usefulness of the computer system in combination with the new endoscopy system. STUDY: We evaluated the system as applied to 100 consecutive early gastric cancers in 95 patients examined by BLI magnification at Hiroshima University Hospital. We produced a set of images from the 100 early gastric cancers; 40 flat or slightly depressed, small, reddened lesions; and surrounding tissues, and we attempted to identify gastric cancer, reddened lesions, and surrounding tissue quantitatively. RESULTS: The average SVM output value was 0.846 ± 0.220 for cancerous lesions, 0.381 ± 0.349 for reddened lesions, and 0.219 ± 0.277 for surrounding tissue, with the SVM output value for cancerous lesions being significantly greater than that for reddened lesions or surrounding tissue. The average SVM output value for differentiated-type cancer was 0.840 ± 0.207 and for undifferentiated-type cancer was 0.865 ± 0.259. CONCLUSIONS: Although further development is needed, we conclude that our computer-based analysis system used with BLI will identify gastric cancers quantitatively.

Effects of low-intensity exercise on contractile property of skeletal muscle and the number of motor neurons in diabetic rats.

The mode of diabetes-induced muscle and motor neuron damage depends on the type of muscle and motor neuron. One of the purposes of exercise therapy for diabetes is to improve blood glucose levels; however, information on the effects of low-intensity exercise on muscle and motor neuron disorders remain unknown. Therefore, this study aimed to examine the effects of low-intensity exercise on diabetes-induced muscle and motor neuron damage in a rat model of type 1 diabetes mellitus. We subjected adult male Wistar rats treated with streptozotocin to develop type 1 diabetes and age-matched rats to low-intensity treadmill exercise for 12 weeks. We recorded electrically evoked maximum twitch tension in leg muscles, and examined the number of motor neurons and cell body sizes. Low-intensity exercise ameliorated the prolonged half-relaxation time and the decreased numbers of the retrograde-labeled motor neurons observed in the soleus muscle of type 1 diabetic rats. However, no effect was observed in the diabetic group, as atrophy was not improved and the twitch force in the medial gastrocnemius muscle was decreased in the diabetic group. In addition, there was no improvement in the blood glucose levels after exercise. These data indicate that low-intensity exercise may relieve the onset of muscle and motor neuron damage in the soleus muscle of type 1 diabetic rats.

Quantitative identification of mucosal gastric cancer under magnifying endoscopy with flexible spectral imaging color enhancement.

BACKGROUND AND AIM: Magnifying endoscopy with flexible spectral imaging color enhancement (FICE) is clinically useful in diagnosing gastric cancer and determining treatment options; however, there is a learning curve. Accurate FICE-based diagnosis requires training and experience. In addition, objectivity is necessary. Thus, a software program that can identify gastric cancer quantitatively was developed. METHODS: A bag-of-features framework with densely sampled scale-invariant feature transform descriptors to magnifying endoscopy images of 46 mucosal gastric cancers was applied. Computer-based findings were compared with histologic findings. The probability of gastric cancer was calculated by means of logistic regression, and sensitivity and specificity of the system were determined. RESULTS: The average probability was 0.78 ± 0.25 for the images of cancer and 0.31 ± 0.25 for the images of noncancer tissue, with a significant difference between the two groups. An optimal cut-off point of 0.59 was determined on the basis of the receiver operating characteristic curves. The computer-aided diagnosis system yielded a detection accuracy of 85.9% (79/92), sensitivity for a diagnosis of cancer of 84.8% (39/46), and specificity of 87.0% (40/46). CONCLUSION: Further development of this system will allow for quantitative evaluation of mucosal gastric cancers on magnifying gastrointestinal endoscopy images obtained with FICE.

Motor skills training-induced activation of descending pathways mediating cortical command to hindlimb motoneurons in experimental diabetic rats.

Diabetes disrupts the corticospinal tract (CST) system components that control hindlimb and trunk movement, resulting in weakness of the lower extremities. However, there is no information about a method to improve these disorders. This study aimed to investigate the rehabilitative effects of 2 weeks of aerobic training (AT) and complex motor skills training (ST) on motor disorders in streptozotocin-induced type 1 diabetic rats. In this study, electrophysiological mapping of the motor cortex showed that the diabetes mellitus (DM)-ST group had a larger motor cortical area compared to the DM-AT group and sedentary diabetic animals. Moreover, hand grip strength and rotarod latency increased in the DM-ST group; however, these two parameters did not change in the DM-AT group, as well as in control and sedentary diabetic rats. Furthermore, in the DM-ST group, cortical stimulation-induced and motor-evoked potentials were preserved after the interception of the CST; however, this potential disappeared after additional lesions were made on lateral funiculus, suggesting that their function extends to activating motor descending pathways other than the CST locating lateral funiculus. According to immunohistochemical analysis, the larger fibers present on the dorsal part of the lateral funiculus, which corresponds to the rubrospinal tract of the DM-ST group, expressed the phosphorylated growth-associated protein, 43 kD, which is a specific marker of axons with plastic changes. Additionally, electrical stimulation of the red nucleus revealed expansion of the hindlimb-responsible area and increased motor-evoked potentials of the hindlimb in the DM-ST group, suggesting a strengthening of synaptic connections between the red nucleus and spinal interneurons driving motoneurons. These results reveal that ST induces plastic changes in the rubrospinal tract in a diabetic model, which can compensate for diabetes by disrupting the CST system components that control the hindlimb. This finding suggests that ST can be a novel rehabilitation strategy to improve motor dysfunctions in diabetic patients.

Computer-aided system for predicting the histology of colorectal tumors by using narrow-band imaging magnifying colonoscopy (with video).

BACKGROUND: Narrow-band imaging (NBI) classification of colorectal lesions is clinically useful in determining treatment options for colorectal tumors. There is a learning curve, however. Accurate NBI-based diagnosis requires training and experience. In addition, objective diagnosis is necessary. Thus, we developed a computerized system to automatically classify NBI magnifying colonoscopic images. OBJECTIVE: To evaluate the utility and limitations of our automated NBI classification system. DESIGN: Retrospective study. SETTING: Department of endoscopy, university hospital. MAIN OUTCOME MEASUREMENTS: Performance of our computer-based system for classification of NBI magnifying colonoscopy images in comparison to classification by two experienced endoscopists and to histologic findings. RESULTS: For the 371 colorectal lesions depicted on validation images, the computer-aided classification system yielded a detection accuracy of 97.8% (363/371); sensitivity and specificity of types B-C3 lesions for a diagnosis of neoplastic lesion were 97.8% (317/324) and 97.9% (46/47), respectively. Diagnostic concordance between the computer-aided classification system and the two experienced endoscopists was 98.7% (366/371), with no significant difference between methods. LIMITATIONS: Retrospective, single-center in this initial report. CONCLUSION: Our new computer-aided system is reliable for predicting the histology of colorectal tumors by using NBI magnifying colonoscopy.

Quantitative analysis and development of a computer-aided system for identification of regular pit patterns of colorectal lesions.

BACKGROUND: Because pit pattern classification of colorectal lesions is clinically useful in determining treatment options for colorectal tumors but requires extensive training, we developed a computerized system to automatically quantify and thus classify pit patterns depicted on magnifying endoscopy images. OBJECTIVE: To evaluate the utility and limitations of our automated pit pattern classification system. DESIGN: Retrospective study. SETTING: Department of endoscopy at a university hospital. MAIN OUTCOME MEASUREMENTS: Performance of our automated computer-based system for classification of pit patterns on magnifying endoscopic images in comparison to classification by diagnosis of the 134 regular pit pattern images by an endoscopist. RESULTS: For type I and II pit patterns, the results of discriminant analysis were in complete agreement with the endoscopic diagnoses. Type IIIl was diagnosed in 29 of 30 cases (96.7%) and type IV was diagnosed in 1 case. Twenty-nine of 30 cases (96.7%) were diagnosed as type IV pit pattern. The overall accuracy of our computerized recognition system was 132 of 134 (98.5%). CONCLUSIONS: Our system is best characterized as semiautomated but is a step toward the development of a fully automated system to assist in the diagnosis of colorectal lesions based on classification of pit patterns.

Quest for the best endoscopic imaging modality for computer-assisted colonic polyp staging.

BACKGROUND: It was shown in previous studies that high definition endoscopy, high magnification endoscopy and image enhancement technologies, such as chromoendoscopy and digital chromoendoscopy [narrow-band imaging (NBI), i-Scan] facilitate the detection and classification of colonic polyps during endoscopic sessions. However, there are no comprehensive studies so far that analyze which endoscopic imaging modalities facilitate the automated classification of colonic polyps. In this work, we investigate the impact of endoscopic imaging modalities on the results of computer-assisted diagnosis systems for colonic polyp staging. AIM: To assess which endoscopic imaging modalities are best suited for the computer-assisted staging of colonic polyps. METHODS: In our experiments, we apply twelve state-of-the-art feature extraction methods for the classification of colonic polyps to five endoscopic image databases of colonic lesions. For this purpose, we employ a specifically designed experimental setup to avoid biases in the outcomes caused by differing numbers of images per image database. The image databases were obtained using different imaging modalities. Two databases were obtained by high-definition endoscopy in combination with i-Scan technology (one with chromoendoscopy and one without chromoendoscopy). Three databases were obtained by high-magnification endoscopy (two databases using narrow band imaging and one using chromoendoscopy). The lesions are categorized into non-neoplastic and neoplastic according to the histological diagnosis. RESULTS: Generally, it is feature-dependent which imaging modalities achieve high results and which do not. For the high-definition image databases, we achieved overall classification rates of up to 79.2% with chromoendoscopy and 88.9% without chromoendoscopy. In the case of the database obtained by high-magnification chromoendoscopy, the classification rates were up to 81.4%. For the combination of high-magnification endoscopy with NBI, results of up to 97.4% for one database and up to 84% for the other were achieved. Non-neoplastic lesions were classified more accurately in general than non-neoplastic lesions. It was shown that the image recording conditions highly affect the performance of automated diagnosis systems and partly contribute to a stronger effect on the staging results than the used imaging modality. CONCLUSION: Chromoendoscopy has a negative impact on the results of the methods. NBI is better suited than chromoendoscopy. High-definition and high-magnification endoscopy are equally suited.

Effects of streptozotocin-induced diabetes on leg muscle contractile properties and motor neuron morphology in rats.

Skeletal muscle fiber subtypes are differentially sensitive to diabetes-related pathology; For example, fast-twitch muscles exhibit severe decreases in contraction force while slow-twitch muscles demonstrate prolonged half-relaxation time. However, such alterations have only been examined after a relatively short period following diabetes onset, with no information available regarding muscle damage caused by longer disease periods (>20 weeks). This study examined alterations in the contractile properties of the medial gastrocnemius (fast-twitch) and soleus (slow-twitch) muscles, as well as morphological changes in their motor neurons 12 and 22 weeks after diabetes onset. Adult male Wistar rats were divided into diabetic (12- or 22-week post-streptozotocin injection) and age-matched control groups. Electrically evoked maximum twitch and tetanic tension were recorded from leg muscles. Additionally, motor neuron number and cell body size were examined. At 12 weeks after diabetes onset, decreases in twitch force were observed predominantly in medial gastrocnemius muscles, while soleus muscles exhibited prolonged half-relaxation time. However, these differences became ambiguous at 22 weeks, with decreased twitch force and prolonged half-relaxation time observed in both muscles. On the other hand, reduction in soleus motor neurons was observed 12 weeks after diabetes onset, while medial gastrocnemius motor neurons were diminished at 22 weeks. These data indicate that experimental diabetes induces differential damage to medial gastrocnemius and soleus muscles as well as motor neurons. These diabetes-induced differences may partly underlie the differential deficits observed in gastrocnemius and soleus.

Effect of streptozotocin-induced diabetes on motor representations in the motor cortex and corticospinal tract in rats.

Motor disorders in patients with diabetes are associated with diabetic peripheral neuropathy, which can lead to symptoms such as lower extremity weakness. However, it is unclear whether central motor system disorders can disrupt motor function in patients with diabetes. In a streptozotocin-induced rat model of type 1 diabetes, we used intracortical microstimulation to evaluate motor representations in the motor cortex, recorded antidromic motor cortex responses to spinal cord stimulation to evaluate the function of corticospinal tract (CST) axons, and used retrograde labeling to evaluate morphological alterations of CST neurons. The diabetic rats exhibited size reductions in the hindlimb area at 4 weeks and in trunk and forelimb areas after 13 weeks, with the hindlimb and trunk area reductions being the most severe. Other areas were unaffected. Additionally, we observed reduced antidromic responses in CST neurons with axons projecting to lumbar spinal segments (CST-L) but not in those with axons projecting to cervical segments (CST-C). This was consistent with the observation that retrograde-labeled CST-L neurons were decreased in number following tracer injection into the spinal cord in diabetic animals but that CST-C neurons were preserved. These results show that diabetes disrupts the CST system components controlling hindlimb and trunk movement. This disruption may contribute to lower extremity weakness in patients.

Training of polyp staging systems using mixed imaging modalities.

BACKGROUND: In medical image data sets, the number of images is usually quite small. The small number of training samples does not allow to properly train classifiers which leads to massive overfitting to the training data. In this work, we investigate whether increasing the number of training samples by merging datasets from different imaging modalities can be effectively applied to improve predictive performance. Further, we investigate if the extracted features from the employed image representations differ between different imaging modalities and if domain adaption helps to overcome these differences. METHOD: We employ twelve feature extraction methods to differentiate between non-neoplastic and neoplastic lesions. Experiments are performed using four different classifier training strategies, each with a different combination of training data. The specifically designed setup for these experiments enables a fair comparison between the four training strategies. RESULTS: Combining high definition with high magnification training data and chromoscopic with non-chromoscopic training data partly improved the results. The usage of domain adaptation has only a small effect on the results compared to just using non-adapted training data. CONCLUSION: Merging datasets from different imaging modalities turned out to be partially beneficial for the case of combining high definition endoscopic data with high magnification endoscopic data and for combining chromoscopic with non-chromoscopic data. NBI and chromoendoscopy on the other hand are mostly too different with respect to the extracted features to combine images of these two modalities for classifier training.

A case report of the therapeutic effect of cryofiltration in a patient with glucocorticoid-resistant ulcerative colitis.

Ulcerative colitis (UC) is characterized by chronic inflammation of the colon and its cause and pathogenesis have not been fully clarified. Although UC is treated with various drugs, including 5-amino-salycilate and glucocorticoids, some patients are resistant to them. It was recently reported that apheresis, such as leukocytapheresis and granulocytapheresis, improves intestinal inflammation in refractory cases of UC. On the other hand, cryofiltration, in which plasma apheresis is used to remove immunoglobulin and immune complexes, has been used for the treatment of autoimmune diseases. We herein report a case of glucocorticoid-resistant UC successfully treated with cryofiltration. Interestingly, the level of interleukin-10 (IL-10) in the patient's serum was markedly increased after eight sessions of cryofiltration. This suggests that cryofiltration suppresses intestinal inflammation, in part via up-regulation of IL-10.

Primal-dual approach to optical tomography with discretized path integral with efficient formulations.

We propose an efficient optical tomography with discretized path integral. We first introduce the primal-dual approach to solve the inverse problem formulated as a constraint optimization problem. Next, we develop efficient formulations for computing Jacobian and Hessian of the cost function of the constraint nonlinear optimization problem. Numerical experiments show that the proposed formulation is faster ([Formula: see text]) than the previous work with the log-barrier interior point method ([Formula: see text]) for the Shepp-Logan phantom with a grid size of [Formula: see text], while keeping the quality of the estimation results (root-mean-square error increasing by up to 12%).

Effect of streptozotocin-induced diabetes on motoneurons and muscle spindles in rats.

This study examined the alterations in the number and size of motoneurons innervating the medial gastrocnemius (MG) and biceps femoris (BF) motor nuclei in diabetic rats (12 or 22 weeks after injection of streptozotocin) and age-matched controls using retrograde labeling technique. Additionally, morphological alterations of muscle spindles in BF and MG muscles were tested. Significantly fewer labeled MG motoneurons were found in 12- and 22-week diabetic rats as compared with age-matched control animals. In contrast, the number of BF motoneurons was preserved in each group. Compared to control animals, the ratio of larger motoneurons of MG and BF muscle were decreased at 12 weeks, and smaller MG motoneurons were drastically decreased at 22 weeks. Moreover, MG muscle spindle showed reduction of its number and increase of intrafusal muscle fibers; however, BF muscle spindles showed little or no difference from control animals. We conclude that there is an early loss of alpha motoneurons for both MG and BF muscles followed by a later loss of gamma motoneurons in MG muscle in diabetic animals. Moreover, loss of gamma motoneuron might induce atrophy of MG muscle spindles.

Defocus-aware Dirichlet particle filter for stable endoscopic video frame recognition.

BACKGROUND AND OBJECTIVE: A computer-aided system for colorectal endoscopy could provide endoscopists with important helpful diagnostic support during examinations. A straightforward means of providing an objective diagnosis in real time might be for using classifiers to identify individual parts of every endoscopic video frame, but the results could be highly unstable due to out-of-focus frames. To address this problem, we propose a defocus-aware Dirichlet particle filter (D-DPF) that combines a particle filter with a Dirichlet distribution and defocus information. METHODS: We develop a particle filter with a Dirichlet distribution that represents the state transition and likelihood of each video frame. We also incorporate additional defocus information by using isolated pixel ratios to sample from a Rayleigh distribution. RESULTS: We tested the performance of the proposed method using synthetic and real endoscopic videos with a frame-wise classifier trained on 1671 images of colorectal endoscopy. Two synthetic videos comprising 600 frames were used for comparisons with a Kalman filter and D-DPF without defocus information, and D-DPF was shown to be more robust against the instability of frame-wise classification results. Computation time was approximately 88ms/frame, which is sufficient for real-time applications. We applied our method to 33 endoscopic videos and showed that the proposed method can effectively smoothen highly unstable probability curves under actual defocus of the endoscopic videos. CONCLUSION: The proposed D-DPF is a useful tool for smoothing unstable results of frame-wise classification of endoscopic videos to support real-time diagnosis during endoscopic examinations.