Search

Purposes: We performed a conversation analysis of the speech conducted among the surgical team during three-dimensional (3D)-printed liver model navigation for thrice or more repeated hepatectomy (TMRH)., Methods: Seventeen patients underwent 3D-printed liver navigation surgery for TMRH. After transcription of the utterances recorded during surgery, the transcribed utterances were coded by the utterer, utterance object, utterance content, sensor, and surgical process during conversation. We then analyzed the utterances and clarified the association between the surgical process and conversation through the intraoperative reference of the 3D-printed liver., Results: In total, 130 conversations including 1648 segments were recorded. Utterance coding showed that the operator/assistant, 3D-printed liver/real liver, fact check (F)/plan check (Pc), visual check/tactile check, and confirmation of planned resection or preservation target (T)/confirmation of planned or ongoing resection line (L) accounted for 791/857, 885/763, 1148/500, 1208/440, and 1304/344 segments, respectively. The utterance's proportions of assistants, F, F of T on 3D-printed liver, F of T on real liver, and Pc of L on 3D-printed liver were significantly higher during non-expert surgeries than during expert surgeries. Confirming the surgical process with both 3D-printed liver and real liver and performing planning using a 3D-printed liver facilitates the safe implementation of TMRH, regardless of the surgeon's experience., Conclusions: The present study, using a unique conversation analysis, provided the first evidence for the clinical value of 3D-printed liver for TMRH for anatomical guidance of non-expert surgeons., (© 2024. The Author(s).)

Aim: To develop an artificial intelligence (AI) algorithm that diagnoses cataracts/corneal diseases from multiple conditions using smartphone images., Methods: This study included 6442 images that were captured using a slit-lamp microscope (6106 images) and smartphone (336 images). An AI algorithm was developed based on slit-lamp images to differentiate 36 major diseases (cataracts and corneal diseases) into 9 categories. To validate the AI model, smartphone images were used for the testing dataset. We evaluated AI performance that included sensitivity, specificity and receiver operating characteristic (ROC) curve for the diagnosis and triage of the diseases., Results: The AI algorithm achieved an area under the ROC curve of 0.998 (95% CI, 0.992 to 0.999) for normal eyes, 0.986 (95% CI, 0.978 to 0.997) for infectious keratitis, 0.960 (95% CI, 0.925 to 0.994) for immunological keratitis, 0.987 (95% CI, 0.978 to 0.996) for cornea scars, 0.997 (95% CI, 0.992 to 1.000) for ocular surface tumours, 0.993 (95% CI, 0.984 to 1.000) for corneal deposits, 1.000 (95% CI, 1.000 to 1.000) for acute angle-closure glaucoma, 0.992 (95% CI, 0.985 to 0.999) for cataracts and 0.993 (95% CI, 0.985 to 1.000) for bullous keratopathy. The triage of referral suggestion using the smartphone images exhibited high performance, in which the sensitivity and specificity were 1.00 (95% CI, 0.478 to 1.00) and 1.00 (95% CI, 0.976 to 1.000) for 'urgent', 0.867 (95% CI, 0.683 to 0.962) and 1.00 (95% CI, 0.971 to 1.000) for 'semi-urgent', 0.853 (95% CI, 0.689 to 0.950) and 0.983 (95% CI, 0.942 to 0.998) for 'routine' and 1.00 (95% CI, 0.958 to 1.00) and 0.896 (95% CI, 0.797 to 0.957) for 'observation', respectively., Conclusions: The AI system achieved promising performance in the diagnosis of cataracts and corneal diseases., Competing Interests: Competing interests: TY: Grants (Novartis Pharma); honoraria for lectures (Alcon Japan, HOYA, Novartis Pharma, AMO Japan, Santen Pharmaceuticals, Senju Pharmaceutical, Johnson & Johnson), MM: Grants (Novartis Pharma); honoraria for lectures (Bayer Yakuhin, Kowa Pharmaceutical, Alcon Japan, HOYA, Novartis Pharma, AMO Japan, Santen Pharmaceutical, Senju Pharmaceutical, Johnson & Johnson K.K., Japan Ophthalmic Instruments Association). MA: Grants (Novartis, Santen), honoraria for lectures (Novartis, Takeda, Senju, Chugai, Kowa), Support for attending meetings and travel (Wakamoto, Novartis), Endowed (NIDEK)., (© Author(s) (or their employer(s)) 2024. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.)

Purpose: Accurate segmentation of tubular structures is crucial for clinical diagnosis and treatment but is challenging due to their complex branching structures and volume imbalance. The purpose of this study is to propose a 3D deep learning network that incorporates skeleton information to enhance segmentation accuracy in these tubular structures., Methods: Our approach employs a 3D convolutional network to extract 3D tubular structures from medical images such as CT volumetric images. We introduce a skeleton-guided module that operates on extracted features to capture and preserve the skeleton information in the segmentation results. Additionally, to effectively train our deep model in leveraging skeleton information, we propose a sigmoid-adaptive Tversky loss function which is specifically designed for skeleton segmentation., Results: We conducted experiments on two distinct 3D medical image datasets. The first dataset consisted of 90 cases of chest CT volumetric images, while the second dataset comprised 35 cases of abdominal CT volumetric images. Comparative analysis with previous segmentation approaches demonstrated the superior performance of our method. For the airway segmentation task, our method achieved an average tree length rate of 93.0%, a branch detection rate of 91.5%, and a precision rate of 90.0%. In the case of abdominal artery segmentation, our method attained an average precision rate of 97.7%, a recall rate of 91.7%, and an F-measure of 94.6%., Conclusion: We present a skeleton-guided 3D convolutional network to segment tubular structures from 3D medical images. Our skeleton-guided 3D convolutional network could effectively segment small tubular structures, outperforming previous methods., (© 2024. The Author(s).)

Voluntary behaviors such as sniffing, moving, and eating require decision-making accompanied by intentional respiration. Based on the study of respiration-coherent activity of rodent olfactory networks, we infer that during the inhalation phase of respiration, olfactory cortical areas process environmental odor information and transmit it to the higher multisensory cognitive areas via feedforward pathways to comprehensively evaluate the surrounding situation. We also infer that during the exhalation phase, the higher multisensory areas generate cognitive-signals and transmit them not only to the behavioral output system but also back to the olfactory cortical areas. We presume that the cortical mechanism couples the intentional respiration with the voluntary behaviors. Thus, in one respiratory cycle, the mammalian brain may transmit and process sensory information to cognize and evaluate the multisensory image of the external world, leading to one behavioral decision and one emotional expression. In this perspective article, we propose that one respiratory cycle provides a minimum time unit for decision making during wakefulness., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Mori and Sakano.)

Purpose: To clarify the differences between struma ovarii (SO) and mucinous carcinomas (MC) on CT and MRI, including T2*-based images, diffusion-weighted images (DWI), and time-intensity curve (TIC) patterns, which have not been previously reported., Methods: We retrospectively compared the presence of low intensity on T2-weighted and T2*-based images, high intensity on T1-weighted images, hyperattenuation on non-contrast CT, TIC pattern, T2 ratio, T1 ratio, CT value, and apparent diffusion coefficient (ADC) value in 15 patients with SO and 27 patients with MC., Results: SO exhibited a significantly higher frequency of low intensity on T2-weighted and T2*-based images, and hyperattenuation on non-contrast CT than MC (P < 0.001, <0.001, and 0.006, respectively). The T2 ratios and CT attenuation of the locules were also significantly different (P < 0.001, and 0.006, respectively). In SO, sites of low intensity on T2-weighted and T2*-based images and sites of hyperattenuation on CT images always coincided. Regarding the TIC pattern, most SO showed a high-risk pattern, with a significant difference (P = 0.003). The ADC values of SO were significantly lower, and only one case of SO showed high signal intensity on DWI., Conclusion: SO were more frequently with low intensity on T2-weighted and T2*-based images, and hyperattenuation on non-contrast CT, and showed high-risk TIC patterns without diffusion restriction., Advances in Knowledge: SO shows a high-risk TIC pattern, but can be specifically diagnosed in combination with the lack of diffusion restriction and loculi with marked hypointensity on T2-weighted and T2*-based images consistent with hyperattenuation on non-contrast CT., (© The Author(s) 2024. Published by Oxford University Press on behalf of the British Institute of Radiology. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Purpose: To compare the diagnostic performance of virtual monoenergetic imaging (VMI), computed tomography (CT), and magnetic resonance imaging (MRI) in patients with endometrial cancer (EC)., Material and Methods: This retrospective study analysed 45 EC patients (mean age: 62 years, range: 44-84 years) undergoing contrast-enhanced CT with dual-energy CT (DECT) and MRI between September 2021 and October 2022. Dual-energy CT generated conventional CT (C-CT) and 40 keV VMI. Quantitative analysis compared contrast-to-noise ratio (CNR) of tumour to myometrium between C-CT and VMI. Qualitative assessment by 5 radiologists compared C-CT, VMI, and MRI for myometrial invasion (MI), cervical invasion, and lymph node metastasis. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) were calculated and compared for each diagnostic parameter., Results: Virtual monoenergetic imaging showed significantly higher CNR than C-CT ( p < 0.001) and a higher sensitivity for MI than C-CT ( p = 0.027) and MRI ( p = 0.011) but lower specificity than MRI ( p = 0.018). C-CT had a higher sensitivity and AUC for cervical invasion than MRI ( p = 0.018 and 0.004, respectively)., Conclusions: The study found no significant superiority of MRI over CT across all diagnostic parameters. VMI demonstrated heightened sensitivity for MI, and C-CT showed greater sensitivity and AUC for cervical invasion than MRI. This suggests that combining VMI with C-CT holds promise as a comprehensive preoperative staging tool for EC when MRI cannot be performed., (© Pol J Radiol 2024.)

Objective  This article evaluates the ability of low-energy (40 keV) virtual monoenergetic images (VMIs) in the local diagnosis of cervical cancer compared with that of conventional computed tomography (C-CT) and magnetic resonance imaging (MRI), using clinicopathologic staging as a reference. Methods  This prospective study included 33 patients with pathologically confirmed cervical cancer who underwent dual-energy CT and MRI between 2021 and 2022. The contrast-to-noise ratio (CNR) of the tumor-to-myometrium was compared between C-CT and VMI. Additionally, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) for each local diagnostic parameter were compared between C-CT, VMI, and MRI. Interradiologist agreement was also assessed. Results  The mean CNR was significantly higher on VMI ( p  = 0.002). No significant difference in AUC was found between C-CT and VMI for all local diagnostic parameters, and the specificity of VMI was often significantly less than that of MRI. For parametrial invasion, mean sensitivity, specificity, and AUC for C-CT, VMI, and MRI were 0.81, 0.99, 0.93; 0.64, 0.35, 0.79; and 0.73, 0.67, 0.86, respectively, and MRI had significantly higher specificity and AUC than that of VMI ( p  = 0.013 and 0.008, respectively). Interradiologist agreement was higher for VMI than C-CT and for MRI than VMI. Conclusion  The CNR of VMI was significantly higher than C-CT and interradiologist agreement was better than with C-CT; however, the overall diagnostic performance of VMI did not significantly differ from C-CT and was inferior to MRI. VMI was characterized by low specificity, which should be understood and used for reading., Competing Interests: Conflict of Interest None declared., (Indian Radiological Association. This is an open access article published by Thieme under the terms of the Creative Commons Attribution-NonDerivative-NonCommercial License, permitting copying and reproduction so long as the original work is given appropriate credit. Contents may not be used for commercial purposes, or adapted, remixed, transformed or built upon. ( https://creativecommons.org/licenses/by-nc-nd/4.0/ ).)

Background: Pediatric minimally invasive surgery requires advanced technical skills. Off-the-job training (OJT), especially when using disease-specific models, is an effective method of acquiring surgical skills. To achieve effective OJT, it is necessary to provide objective and appropriate skill assessment feedback to trainees. We aimed to construct a system that automatically evaluates surgical skills based on forceps movement using deep learning (DL)., Methods: Using our original esophageal atresia OJT model, participants were tasked with performing esophageal anastomosis. All tasks were recorded for image analysis. Based on manual objective skill assessments, each participant's surgical skills were categorized into two groups: good and poor. The motion of the forceps in both groups was used as training data. Employing this training data, we constructed an automated system that recognized the movement of forceps and determined the quality of the surgical technique., Results: Thirteen participants were assigned to the good skill group and 32 to the poor skill group. These cases were validated using an automated skill assessment system. This system showed a precision of 75%, a specificity of 94%, and an area under the receiver operating characteristic curve of 0.81., Conclusions: We constructed a system that automatically evaluated the quality of surgical techniques based on the movement of forceps using DL. Artificial intelligence diagnostics further revealed the procedures important for suture manipulation., Levels of Evidence: Level IV., Competing Interests: Conflicts of interest None., (Copyright © 2024 Elsevier Inc. All rights reserved.)

Purpose: Parkinson disease (PD) is a common progressive neurodegenerative disorder in our ageing society. Early-stage PD biomarkers are desired for timely clinical intervention and understanding of pathophysiology. Since one of the characteristics of PD is the progressive loss of dopaminergic neurons in the substantia nigra pars compacta, we propose a feature extraction method for analysing the differences in the substantia nigra between PD and non-PD patients., Method: We propose a feature-extraction method for volumetric images based on a rank-1 tensor decomposition. Furthermore, we apply a feature selection method that excludes common features between PD and non-PD. We collect neuromelanin images of 263 patients: 124 PD and 139 non-PD patients and divide them into training and testing datasets for experiments. We then experimentally evaluate the classification accuracy of the substantia nigra between PD and non-PD patients using the proposed feature extraction method and linear discriminant analysis., Results: The proposed method achieves a sensitivity of 0.72 and a specificity of 0.64 for our testing dataset of 66 non-PD and 42 PD patients. Furthermore, we visualise the important patterns in the substantia nigra by a linear combination of rank-1 tensors with selected features. The visualised patterns include the ventrolateral tier, where the severe loss of neurons can be observed in PD., Conclusions: We develop a new feature-extraction method for the analysis of the substantia nigra towards PD diagnosis. In the experiments, even though the classification accuracy with the proposed feature extraction method and linear discriminant analysis is lower than that of expert physicians, the results suggest the potential of tensorial feature extraction., (© 2024. CARS.)

Background: We report a new real-time navigation system for laparoscopic hepatectomy (LH), which resembles a car navigation system., Material and Methods: Virtual three-dimensional liver and body images were reconstructed using the "New-VES" system, which worked as roadmap during surgery. Several points of the patient's body were registered in virtual images using a magnetic position sensor (MPS). A magnetic transmitter, corresponding to an artificial satellite, was placed about 40 cm above the patient's body. Another MPS, corresponding to a GPS antenna, was fixed on the handling part of the laparoscope. Fiducial registration error (FRE, an error between real and virtual lengths) was utilized to evaluate the accuracy of this system., Results: Twenty-one patients underwent LH with this system. Mean FRE of the initial five patients was 17.7 mm. Mean FRE of eight patients in whom MDCT was taken using radiological markers for registration of body parts as first improvement, was reduced to 10.2 mm ( p  = .014). As second improvement, a new MPS as an intraoperative body position sensor was fixed on the right-sided chest wall for automatic correction of postural gap. The preoperative and postoperative mean FREs of 8 patients with both improvements were 11.1 mm and 10.1 mm ( p  = .250)., Conclusions: Our system may provide a promising option that virtually guides LH.

Purpose: This paper considers a new problem setting for multi-organ segmentation based on the following observations. In reality, (1) collecting a large-scale dataset from various institutes is usually impeded due to privacy issues; (2) many images are not labeled since the slice-by-slice annotation is costly; and (3) datasets may exhibit inconsistent, partial annotations across different institutes. Learning a federated model from these distributed, partially labeled, and unlabeled samples is an unexplored problem., Methods: To simulate this multi-organ segmentation problem, several distributed clients and a central server are maintained. The central server coordinates with clients to learn a global model using distributed private datasets, which comprise a small part of partially labeled images and a large part of unlabeled images. To address this problem, a practical framework that unifies partially supervised learning (PSL), semi-supervised learning (SSL), and federated learning (FL) paradigms with PSL, SSL, and FL modules is proposed. The PSL module manages to learn from partially labeled samples. The SSL module extracts valuable information from unlabeled data. Besides, the FL module aggregates local information from distributed clients to generate a global statistical model. With the collaboration of three modules, the presented scheme could take advantage of these distributed imperfect datasets to train a generalizable model., Results: The proposed method was extensively evaluated with multiple abdominal CT datasets, achieving an average result of 84.83% in Dice and 41.62 mm in 95HD for multi-organ (liver, spleen, and stomach) segmentation. Moreover, its efficacy in transfer learning further demonstrated its good generalization ability for downstream segmentation tasks., Conclusion: This study considers a novel problem of multi-organ segmentation, which aims to develop a generalizable model using distributed, partially labeled, and unlabeled CT images. A practical framework is presented, which, through extensive validation, has proved to be an effective solution, demonstrating strong potential in addressing this challenging problem., (© 2024. The Author(s).)

This review aims to provide an overview of neoplastic lesions associated with genetic diseases affecting the female reproductive organs. It seeks to enhance our understanding of the radiological aspects in diagnosing genetic diseases including hereditary breast and ovarian cancer syndromes, Lynch syndrome, Peutz-Jeghers syndrome, nevoid basal cell carcinoma syndrome, and Swyer syndrome, and explores the patterns and mechanisms of inheritance that require elucidation. Additionally, we discuss the imaging characteristics of lesions occurring in other regions due to the same genetic diseases., (© 2024. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.)

A 38-year-old woman was admitted to our university hospital with loss of muscle strength. She was diagnosed with dermatomyositis and underwent contrast-enhanced computed tomography of the entire body to check for malignant tumors. Computed tomography revealed multiple enhanced hepatic nodules and an extrahepatic portosystemic shunt. Although a needle biopsy of the nodule could not diagnose definitive hepatocellular carcinoma, some nodules increased in size after three months. Because of the inconclusive results of the second biopsy, we performed shunt embolization using a vascular plug. After another three months, the hepatic nodules shrank markedly, as expected.

Case 1 : A 75-year-old man was emergently admitted to our hospital with a complaint of continuous bleeding from the ileal conduit. The conduit was constructed by a total pelvic resection for sigmoid colon cancer that invaded the urinary bladder 24 years ago. Swollen cutaneous mucosa was seen around the ileal conduit, but no obvious bleeding spot was observed. The contrast-enhanced computed tomographic (CT) scan and 3D visualization revealed varices extending to the abdominal wall. Percutaneous transhepatic embolization successfully stopped the bleeding, but it was needed again after two years. Case 2 : A 72-yearold man with a history of open cystectomy and ileal conduit for bladder cancer came to our hospital two years after the surgery, complaining of continuous bleeding from the conduit. The skin around the stoma site was discolored purple, but no obvious bleeding site or bloody urine was observed. The CT scan similar to Case 1 revealed varices in the ileal conduit, and percutaneous transhepatic embolization successfully stopped the bleeding, but it was needed again after five months. After that, three months passed without recurrence.

Purpose: Pancreatic duct dilation is associated with an increased risk of pancreatic cancer, the most lethal malignancy with the lowest 5-year relative survival rate. Automatic segmentation of the dilated pancreatic duct from contrast-enhanced CT scans would facilitate early diagnosis. However, pancreatic duct segmentation poses challenges due to its small anatomical structure and poor contrast in abdominal CT. In this work, we investigate an anatomical attention strategy to address this issue., Methods: Our proposed anatomical attention strategy consists of two steps: pancreas localization and pancreatic duct segmentation. The coarse pancreatic mask segmentation is used to guide the fully convolutional networks (FCNs) to concentrate on the pancreas' anatomy and disregard unnecessary features. We further apply a multi-scale aggregation scheme to leverage the information from different scales. Moreover, we integrate the tubular structure enhancement as an additional input channel of FCN., Results: We performed extensive experiments on 30 cases of contrast-enhanced abdominal CT volumes. To evaluate the pancreatic duct segmentation performance, we employed four measurements, including the Dice similarity coefficient (DSC), sensitivity, normalized surface distance, and 95 percentile Hausdorff distance. The average DSC achieves 55.7%, surpassing other pancreatic duct segmentation methods on single-phase CT scans only., Conclusions: We proposed an anatomical attention-based strategy for the dilated pancreatic duct segmentation. Our proposed strategy significantly outperforms earlier approaches. The attention mechanism helps to focus on the pancreas region, while the enhancement of the tubular structure enables FCNs to capture the vessel-like structure. The proposed technique might be applied to other tube-like structure segmentation tasks within targeted anatomies., (© 2024. The Author(s).)

Objectives: In this study we aimed to develop an artificial intelligence-based model for predicting postendoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP)., Methods: We retrospectively reviewed ERCP patients at Nagoya University Hospital (NUH) and Toyota Memorial Hospital (TMH). We constructed two prediction models, a random forest (RF), one of the machine-learning algorithms, and a logistic regression (LR) model. First, we selected features of each model from 40 possible features. Then the models were trained and validated using three fold cross-validation in the NUH cohort and tested in the TMH cohort. The area under the receiver operating characteristic curve (AUROC) was used to assess model performance. Finally, using the output parameters of the RF model, we classified the patients into low-, medium-, and high-risk groups., Results: A total of 615 patients at NUH and 544 patients at TMH were enrolled. Ten features were selected for the RF model, including albumin, creatinine, biliary tract cancer, pancreatic cancer, bile duct stone, total procedure time, pancreatic duct injection, pancreatic guidewire-assisted technique without a pancreatic stent, intraductal ultrasonography, and bile duct biopsy. In the three fold cross-validation, the RF model showed better predictive ability than the LR model (AUROC 0.821 vs. 0.660). In the test, the RF model also showed better performance (AUROC 0.770 vs. 0.663, P = 0.002). Based on the RF model, we classified the patients according to the incidence of PEP (2.9%, 10.0%, and 23.9%)., Conclusion: We developed an RF model. Machine-learning algorithms could be powerful tools to develop accurate prediction models., (© 2023 Japan Gastroenterological Endoscopy Society.)

This review focuses on inflammatory diseases of female and male genital organs and discusses their epidemiology, pathogenesis, clinical presentation, and imaging findings. The female section covers pelvic inflammatory disease (PID) primarily caused by sexually transmitted infections (STIs) that affect the uterus, fallopian tubes, and ovaries. Unusual causes such as actinomycosis and tuberculosis have also been explored. The male section delves into infections affecting the vas deferens, epididymis, testes, prostate, and seminal vesicles. Uncommon causes such as tuberculosis, and Zinner syndrome have also been discussed. In addition, this review highlights other conditions that mimic male genital tract infections such as vasculitis, IgG4-related diseases, and sarcoidosis. Accurate diagnosis and appropriate management of these inflammatory diseases are essential for preventing serious complications and infertility. Imaging modalities such as ultrasound, magnetic resonance imaging, and computed tomography play a crucial role in diagnosis. Understanding the diverse etiologies and imaging findings is vital for the effective management of inflammatory diseases of the genital organs., (© 2023. The Author(s).)

Since even subtle mucosal changes may be depicted using virtual endoscopy created by the three-dimensional reconstruction of MDCT images, we developed a novel diagnostic imaging system that integrates and displays virtual enteroscopy, curved planar reconstruction, and a virtual unfolded view, the width of which changes with increases/decreases in the inner luminal diameter. The system is also equipped with artificial intelligence that superimposes and displays depressed areas, generates an automatic small bowel centerline that connects fragmented small bowel regions, and performs electronic cleansing. We retrospectively evaluated the diagnostic performance of this system for small bowel lesions in Crohn's disease, which were divided into two groups: endoscopically-observable and endoscopically-unobservable. Lesion detection rates for stenoses, longitudinal ulcers with a cobblestone appearance, and scars were excellent in both groups. This system, when used in combination with endoscopy, shows slight mucosal changes in areas in which an endoscope cannot reach due to strictures, thereby extending the range of observation of the small bowel. This system is a useful diagnostic modality that has the capacity to assess mucosal healing and provide extraluminal information.

Background: Fundus fluorescein angiography (FFA) is an imaging method used to assess retinal vascular structures by injecting exogenous dye. FFA images provide complementary information to that provided by the widely used color fundus (CF) images. However, the injected dye can cause some adverse side effects, and the method is not suitable for all patients., Methods: To meet the demand for high-quality FFA images in the diagnosis of retinopathy without side effects to patients, this study proposed an unsupervised image synthesis framework based on dual contrastive learning that can synthesize FFA images from unpaired CF images by inferring the effective mappings and avoid the shortcoming of generating blurred pathological features caused by cycle-consistency in conventional approaches. By adding class activation mapping (CAM) to the adaptive layer-instance normalization (AdaLIN) function, the generated images are made more realistic. Additionally, the use of CAM improves the discriminative ability of the model. Further, the Coordinate Attention Block was used for better feature extraction, and it was compared with other attention mechanisms to demonstrate its effectiveness. The synthesized images were quantified by the Fréchet inception distance (FID), kernel inception distance (KID), and learned perceptual image patch similarity (LPIPS)., Results: The extensive experimental results showed the proposed approach achieved the best results with the lowest overall average FID of 50.490, the lowest overall average KID of 0.01529, and the lowest overall average LPIPS of 0.245 among all the approaches., Conclusions: When compared with several popular image synthesis approaches, our approach not only produced higher-quality FFA images with clearer vascular structures and pathological features, but also achieved the best FID, KID, and LPIPS scores in the quantitative evaluation., Competing Interests: Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-23-1304/coif). The authors have no conflicts of interest to declare., (2024 Quantitative Imaging in Medicine and Surgery. All rights reserved.)

Purpose: We propose a large-factor super-resolution (SR) method for performing SR on registered medical image datasets. Conventional SR approaches use low-resolution (LR) and high-resolution (HR) image pairs to train a deep convolutional neural network (DCN). However, LR-HR images in medical imaging are commonly acquired from different imaging devices, and acquiring LR-HR image pairs needs registration. Registered LR-HR images have registration errors inevitably. Using LR-HR images with registration error for training an SR DCN causes collapsed SR results. To address these challenges, we introduce a novel SR approach designed specifically for registered LR-HR medical images., Methods: We propose style-subnets-assisted generative latent bank for large-factor super-resolution (SGSR) trained with registered medical image datasets. Pre-trained generative models named generative latent bank (GLB), which stores rich image priors, can be applied in SR to generate realistic and faithful images. We improve GLB by newly introducing style-subnets-assisted GLB (S-GLB). We also propose a novel inter-uncertainty loss to boost our method's performance. Introducing more spatial information by inputting adjacent slices further improved the results., Results: SGSR outperforms state-of-the-art (SOTA) supervised SR methods qualitatively and quantitatively on multiple datasets. SGSR achieved higher reconstruction accuracy than recently supervised baselines by increasing peak signal-to-noise ratio from 32.628 to 34.206 dB., Conclusion: SGSR performs large-factor SR while given a registered LR-HR medical image dataset with registration error for training. SGSR's results have both realistic textures and accurate anatomical structures due to favorable quantitative and qualitative results. Experiments on multiple datasets demonstrated SGSR's superiority over other SOTA methods. SR medical images generated by SGSR are expected to improve the accuracy of pre-surgery diagnosis and reduce patient burden., (© 2023. CARS.)

In the mouse olfactory system, odor information is converted to a topographic map of activated glomeruli in the olfactory bulb (OB). Although the arrangement of glomeruli is genetically determined, the glomerular structure is plastic and can be modified by environmental stimuli. If the pups are exposed to a particular odorant, responding glomeruli become larger recruiting the dendrites of connecting projection neurons and interneurons. This imprinting not only increases the sensitivity to the exposed odor, but also imposes the positive quality on imprinted memory. External odor information represented as an odor map in the OB is transmitted to the olfactory cortex (OC) and amygdala for decision making to elicit emotional and behavioral outputs using two distinct neural pathways, innate and learned. Innate olfactory circuits start to work right after birth, whereas learned circuits become functional later on. In this paper, the recent progress will be summarized in the study of olfactory circuit formation and odor perception in mice. We will also propose new hypotheses on the timing and gating of olfactory circuit activity in relation to the respiration cycle., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Mori and Sakano.)

This study focuses on enhancing the inference speed of laparoscopic tool detection on embedded devices. Laparoscopy, a minimally invasive surgery technique, markedly reduces patient recovery times and postoperative complications. Real-time laparoscopic tool detection helps assisting laparoscopy by providing information for surgical navigation, and its implementation on embedded devices is gaining interest due to the portability, network independence and scalability of the devices. However, embedded devices often face computation resource limitations, potentially hindering inference speed. To mitigate this concern, the work introduces a two-fold modification to the YOLOv7 model: the feature channels and integrate RepBlock is halved, yielding the YOLOv7-RepFPN model. This configuration leads to a significant reduction in computational complexity. Additionally, the focal EIoU (efficient intersection of union) loss function is employed for bounding box regression. Experimental results on an embedded device demonstrate that for frame-by-frame laparoscopic tool detection, the proposed YOLOv7-RepFPN achieved an mAP of 88.2% (with IoU set to 0.5) on a custom dataset based on EndoVis17, and an inference speed of 62.9 FPS. Contrasting with the original YOLOv7, which garnered an 89.3% mAP and 41.8 FPS under identical conditions, the methodology enhances the speed by 21.1 FPS while maintaining detection accuracy. This emphasizes the effectiveness of the work., Competing Interests: The authors declare no conflicts of interest., (© 2024 The Authors. Healthcare Technology Letters published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.)

The task of segmentation is integral to computer-aided surgery systems. Given the privacy concerns associated with medical data, collecting a large amount of annotated data for training is challenging. Unsupervised learning techniques, such as contrastive learning, have shown powerful capabilities in learning image-level representations from unlabelled data. This study leverages classification labels to enhance the accuracy of the segmentation model trained on limited annotated data. The method uses a multi-scale projection head to extract image features at various scales. The partitioning method for positive sample pairs is then improved to perform contrastive learning on the extracted features at each scale to effectively represent the differences between positive and negative samples in contrastive learning. Furthermore, the model is trained simultaneously with both segmentation labels and classification labels. This enables the model to extract features more effectively from each segmentation target class and further accelerates the convergence speed. The method was validated using the publicly available CholecSeg8k dataset for comprehensive abdominal cavity surgical segmentation. Compared to select existing methods, the proposed approach significantly enhances segmentation performance, even with a small labelled subset (1-10%) of the dataset, showcasing a superior intersection over union (IoU) score., Competing Interests: The authors declare no conflicts of interest., (© 2024 The Authors. Healthcare Technology Letters published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.)

This paper focuses on a new and challenging problem related to instrument segmentation. This paper aims to learn a generalizable model from distributed datasets with various imperfect annotations. Collecting a large-scale dataset for centralized learning is usually impeded due to data silos and privacy issues. Besides, local clients, such as hospitals or medical institutes, may hold datasets with diverse and imperfect annotations. These datasets can include scarce annotations (many samples are unlabelled), noisy labels prone to errors, and scribble annotations with less precision. Federated learning (FL) has emerged as an attractive paradigm for developing global models with these locally distributed datasets. However, its potential in instrument segmentation has yet to be fully investigated. Moreover, the problem of learning from various imperfect annotations in an FL setup is rarely studied, even though it presents a more practical and beneficial scenario. This work rethinks instrument segmentation in such a setting and propose a practical FL framework for this issue. Notably, this approach surpassed centralized learning under various imperfect annotation settings. This method established a foundational benchmark, and future work can build upon it by considering each client owning various annotations and aligning closer with real-world complexities., Competing Interests: The authors declare no conflicts of interest., (© 2024 The Authors. Healthcare Technology Letters published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.)

Purpose: To compare MRI findings of high-grade serous carcinoma (HGSC) with and without breast cancer (BRCA) gene variants to explore the feasibility of MRI as a genetic predictor., Methods: We retrospectively reviewed MRI data from 16 patients with BRCA variant-positive (11 patients of BRCA1 and 5 patients of BRCA2 variant-positive) and 32 patients with BRCA variant-negative HGSCs and evaluated tumor size, appearance, nature of solid components, apparent diffusion coefficient (ADC) value, time-intensity curve, several dynamic contrast-enhanced curve descriptors, and nature of peritoneal metastasis. Age, primary site, tumor stage, bilaterality, presence of lymph node metastasis, presence of peritoneal metastasis, and tumor markers were also compared between the groups with the Mann-Whitney U and chi-square tests., Results: The mean tumor size of BRCA variant-positive HGSCs was 9.6 cm, and that of variant-negative HGSCs was 6.8 cm, with no significant difference (P = 0.241). No significant difference was found between BRCA variant-positive and negative HGSCs in other evaluated factors, except for age (mean age, 53 years old; range, 32-78 years old for BRCA variant-positive and mean age, 61 years old; range, 44-80 years old for BRCA variant-negative, P = 0.033). Comparing BRCA1 variant-positive and BRCA2 variant-positive HGSCs, BRCA1 variant-positive HGSCs were larger (P = 0.040), had greater Max enhancement (P = 0.013), Area under the curve (P = 0.013), and CA125 (P = 0.038), and had a higher frequency of lymph node metastasis (P = 0.049), with significance., Conclusion: There was no significant difference in the MRI findings between patients with HGSCs with and without BRCA variants. Although studied in small numbers, BRCA1 variant-positive HGSCs were larger and more enhanced than BRCA2 variant-positive HGSCs with higher CA125 and more frequent lymph node metastases, and may represent more aggressive features.

In the context of Minimally Invasive Surgery, surgeons mainly rely on visual feedback during medical operations. In common procedures such as tissue resection, the automation of endoscopic control is crucial yet challenging, particularly due to the interactive dynamics of multi-agent operations and the necessity for real-time adaptation. This paper introduces a novel framework that unites a Hierarchical Quadratic Programming controller with an advanced interactive perception module. This integration addresses the need for adaptive visual field control and robust tool tracking in the operating scene, ensuring that surgeons and assistants have optimal viewpoint throughout the surgical task. The proposed framework handles multiple objectives within predefined thresholds, ensuring efficient tracking even amidst changes in operating backgrounds, varying lighting conditions, and partial occlusions. Empirical validations in scenarios involving single, double, and quadruple tool tracking during tissue resection tasks have underscored the system's robustness and adaptability. The positive feedback from user studies, coupled with the low cognitive and physical strain reported by surgeons and assistants, highlight the system's potential for real-world application.

Background and Aims: Gastric cancer (GC) is associated with chronic gastritis. To evaluate the risk, the Operative Link on Gastric Intestinal Metaplasia Assessment (OLGIM) system was constructed and showed a higher GC risk in stage III or IV patients, determined by the degree of intestinal metaplasia (IM). Although the OLGIM system is useful, evaluating the degree of IM requires substantial experience to produce precise scoring. Whole-slide imaging is becoming routine, but most artificial intelligence (AI) systems in pathology are focused on neoplastic lesions., Methods: Hematoxylin and eosin-stained slides were scanned. Images were divided into each gastric biopsy tissue sample and labeled with an IM score. IM was scored as follows: 0 (no IM), 1 (mild IM), 2 (moderate IM), and 3 (severe IM). Overall, 5753 images were prepared. A deep convolutional neural network (DCNN) model, ResNet50, was used for classification., Results: ResNet50 classified images with and without IM with a sensitivity of 97.7% and specificity of 94.6%. IM scores 2 and 3, involved as criteria of stage III or IV in the OLGIM system, were classified by ResNet50 in 18%. The respective sensitivity and specificity values of classifying IM between scores 0 and 1 and 2 and 3 were 98.5% and 94.9%, respectively. The IM scores classified by pathologists and the AI system were different in only 438 images (7.6%), and we found that ResNet50 tended to miss small foci of IM but successfully identified minimal IM areas that pathologists missed during the review., Conclusions: Our findings suggested that this AI system would contribute to evaluating the risk of GC accuracy, reliability, and repeatability with worldwide standardization., Competing Interests: Disclosure All authors disclosed no financial relationships., (Copyright © 2023 American Society for Gastrointestinal Endoscopy. Published by Elsevier Inc. All rights reserved.)

Ossification of the ligaments progresses slowly in the initial stages, and most patients are unaware of the disease until obvious myelopathy symptoms appear. Consequently, treatment and clinical outcomes are not satisfactory. This study is aimed at developing an automated system for the detection of the thoracic ossification of the posterior longitudinal ligament (OPLL) using deep learning and plain radiography. We retrospectively reviewed the data of 146 patients with thoracic OPLL and 150 control cases without thoracic OPLL. Plain lateral thoracic radiographs were used for object detection, training, and validation. Thereafter, an object detection system was developed, and its accuracy was calculated. The performance of the proposed system was compared with that of two spine surgeons. The accuracy of the proposed object detection model based on plain lateral thoracic radiographs was 83.4%, whereas the accuracies of spine surgeons 1 and 2 were 80.4% and 77.4%, respectively. Our findings indicate that our automated system, which uses a deep learning-based method based on plain radiographs, can accurately detect thoracic OPLL. This system has the potential to improve the diagnostic accuracy of thoracic OPLL., Competing Interests: The authors declare that there is no conflict of interest regarding the publication of this paper., (Copyright © 2023 Sadayuki Ito et al.)

Purpose: Postoperative complication prediction helps surgeons to inform and manage patient expectations. Deep learning, a model that finds patterns in large samples of data, outperform traditional statistical methods in making predictions. This study aimed to create a deep learning-based model (DLM) to predict postoperative complications in patients with cervical ossification of the posterior longitudinal ligament (OPLL)., Methods: This prospective multicenter study was conducted by the 28 institutions, and 478 patients were included in the analysis. Deep learning was used to create two predictive models of the overall postoperative complications and neurological complications, one of the major complications. These models were constructed by learning the patient's preoperative background, clinical symptoms, surgical procedures, and imaging findings. These logistic regression models were also created, and these accuracies were compared with those of the DLM., Results: Overall complications were observed in 127 cases (26.6%). The accuracy of the DLM was 74.6 ± 3.7% for predicting the overall occurrence of complications, which was comparable to that of the logistic regression (74.1%). Neurological complications were observed in 48 cases (10.0%), and the accuracy of the DLM was 91.7 ± 3.5%, which was higher than that of the logistic regression (90.1%)., Conclusion: A new algorithm using deep learning was able to predict complications after cervical OPLL surgery. This model was well calibrated, with prediction accuracy comparable to that of regression models. The accuracy remained high even for predicting only neurological complications, for which the case number is limited compared to conventional statistical methods., (© 2023. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.)

COVID-19 is linked to endotheliopathy and coagulopathy, which can result in multi-organ failure. The mechanisms causing endothelial damage due to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) remain elusive. Here, we developed an infection-competent human vascular organoid from pluripotent stem cells for modeling endotheliopathy. Longitudinal serum proteome analysis identified aberrant complement signature in critically ill patients driven by the amplification cycle regulated by complement factor B and D (CFD). This deviant complement pattern initiates endothelial damage, neutrophil activation, and thrombosis specific to organoid-derived human blood vessels, as verified through intravital imaging. We examined a new long-acting, pH-sensitive (acid-switched) antibody targeting CFD. In both human and macaque COVID-19 models, this long-acting anti-CFD monoclonal antibody mitigated abnormal complement activation, protected endothelial cells, and curtailed the innate immune response post-viral exposure. Collectively, our findings suggest that the complement alternative pathway exacerbates endothelial injury and inflammation. This underscores the potential of CFD-targeted therapeutics against severe viral-induced inflammathrombotic outcomes., Competing Interests: Declaration of interests N.S. and T.T. are patent holders associated with the technology described in this project., (Copyright © 2023 Elsevier Inc. All rights reserved.)

Purpose: Silent corticotroph pituitary adenomas (SCAs)/pituitary neuroendocrine tumors (PitNETs) are common non-functioning pituitary adenomas (NFAs)/PitNETs with a clinically aggressive course. This study aimed to investigate the ability of time-intensity analysis of dynamic magnetic resonance imaging (MRI) for distinguishing adrenocorticotropic hormone (ACTH)-positive SCAs and ACTH-negative SCAs from other NFAs., Materials and Methods: We retrospectively evaluated the dynamic MRI findings of patients with NFAs. The initial slope of the kinetic curve (slope ini ) obtained by dynamic MRI for each tumor was analyzed using a modified empirical mathematical model. The maximum slope of the kinetic curve (slope max ) was obtained by geometric calculation., Results: A total of 106 patients with NFAs (11 ACTH-positive SCAs, 5 ACTH-negative SCAs, and 90 other NFAs) were evaluated. The kinetic curves of ACTH-positive SCAs had significantly lesser slope ini and slope max compared with ACTH-negative SCAs (P = 0.040 and P = 0.001, respectively) and other NFAs (P = 0.018 and P = 0.035, respectively). Conversely, the slope ini and slope max were significantly greater in ACTH-negative SCAs than in NFAs other than ACTH-negative SCAs (P = 0.033 and P = 0.044, respectively). In receiver operating characteristic analysis of ACTH-positive SCAs and other NFAs, the area under the curve (AUC) values for slope ini and slope max were 0.762 and 0748, respectively. In predicting ACTH-negative SCAs, the AUC values for slope ini and slope max were 0.784 and 0.846, respectively., Conclusions: Dynamic MRI can distinguish ACTH-positive SCAs and ACTH-negative SCAs from other NFAs., (© 2023. The Author(s).)

Background: Preoperative assessment of the histological type of ovarian cancer is essential to determine the appropriate treatment strategy. Tumor location may be helpful in this regard. The purpose of this study was to compare the position of endometriosis-associated (EAOCs) and non-associated (non-EAOCs) ovarian cancer relative to the uterus using MRI., Methods: This retrospective study included patients with pathologically confirmed malignant epithelial ovarian tumors who underwent MRI at our hospital between January 2015 and January 2023. T2-weighted images of the sagittal and axial sections of the long axis of the uterine body were used for the analysis. Three blinded experienced radiologists independently interpreted the images and assessed whether the ovarian tumor was attached to the uterus, and the angle between the uterus and the tumor was measured. The presence of attachment and the measured angles were compared for each histology. In addition, the angles between EAOCs, including endometrioid carcinomas (ECs) and clear cell carcinomas (CCCs), were compared with non-EAOCs., Results: In total, 184 women (mean age, 56 years; age range, 20-91 years) were evaluated. High-grade serous carcinomas (HGSCs) were significantly smaller than the others and had significantly less uterine attachment than CCCs (p < 0.01 for all readers). According to the mean of the measured angles, CCCs were positioned significantly more posteriorly than HGSCs and mucinous carcinomas (p < 0.02), and EAOCs were positioned significantly more posteriorly to the uterus than non-EAOCs (p < 0.01)., Conclusion: HGSCs are often not attached to the uterus, and EAOCs are positioned more posteriorly to the uterus than non-EAOCs., Critical Relevance Statement: High-grade serous carcinomas were often not attached to the uterus, and endometriosis-associated ovarian cancers were positioned more posteriorly to the uterus than non-endometriosis-associated ovarian cancers., Key Points: • The position of the ovarian tumor can be determined using MRI. • High-grade serous carcinomas had less attachment to the uterus. • Endometriosis-associated cancers were positioned more posteriorly to the uterus. • The location of ovarian tumors is helpful in estimating histology., (© 2023. European Society of Radiology (ESR).)

Spinal cord tumors are infrequently identified spinal diseases that are often difficult to diagnose even with magnetic resonance imaging (MRI) findings. To minimize the probability of overlooking these tumors and improve diagnostic accuracy, an automatic diagnostic system is needed. We aimed to develop an automated system for detecting and diagnosing spinal schwannomas and meningiomas based on deep learning using You Only Look Once (YOLO) version 4 and MRI. In this retrospective diagnostic accuracy study, the data of 50 patients with spinal schwannomas, 45 patients with meningiomas, and 100 control cases were reviewed, respectively. Sagittal T1-weighted (T1W) and T2-weighted (T2W) images were used for object detection, classification, training, and validation. The object detection and diagnosis system was developed using YOLO version 4. The accuracies of the proposed object detections based on T1W, T2W, and T1W + T2W images were 84.8%, 90.3%, and 93.8%, respectively. The accuracies of the object detection for two spine surgeons were 88.9% and 90.1%, respectively. The accuracies of the proposed diagnoses based on T1W, T2W, and T1W + T2W images were 76.4%, 83.3%, and 84.1%, respectively. The accuracies of the diagnosis for two spine surgeons were 77.4% and 76.1%, respectively. We demonstrated an accurate, automated detection and diagnosis of spinal schwannomas and meningiomas using the developed deep learning-based method based on MRI. This system could be valuable in supporting radiological diagnosis of spinal schwannomas and meningioma, with a potential of reducing the radiologist's overall workload.

Background: Locally advanced pancreatic ductal adenocarcinoma (PDAC), accounting for about 30% of PDAC patients, is difficult to cure by radical resection or systemic chemotherapy alone. A multidisciplinary strategy is required and our TT-LAP trial aims to evaluate whether triple-modal treatment with proton beam therapy (PBT), hyperthermia, and gemcitabine plus nab-paclitaxel is a safe and synergistically effective treatment for patients with locally advanced PDAC., Methods: This trial is an interventional, open-label, non-randomized, single-center, single-arm phase I/II clinical trial organized and sponsored by the University of Tsukuba. Eligible patients who are diagnosed with locally advanced pancreatic cancer, including both borderline resectable (BR) and unresectable locally advanced (UR-LA) patients, and selected according to the inclusion and exclusion criteria will receive triple-modal treatment consisting of chemotherapy, hyperthermia, and proton beam radiation. Treatment induction will include 2 cycles of chemotherapy (gemcitabine plus nab-paclitaxel), proton beam therapy, and 6 total sessions of hyperthermia therapy. The initial 5 patients will move to phase II after adverse events are verified by a monitoring committee and safety is ensured. The primary endpoint is 2-year survival rate while secondary endpoints include adverse event rate, treatment completion rate, response rate, progression-free survival, overall survival, resection rate, pathologic response rate, and R0 (no pathologic cancer remnants) rate. The target sample size is set at 30 cases., Discussion: The TT-LAP trial is the first to evaluate the safety and effectiveness (phases1/2) of triple-modal treatment comprised of proton beam therapy, hyperthermia, and gemcitabine/nab-paclitaxel for locally advanced pancreatic cancer., Ethics and Dissemination: This protocol was approved by the Tsukuba University Clinical Research Review Board (reference number TCRB22-007). Results will be analyzed after study recruitment and follow-up are completed. Results will be presented at international meetings of interest in pancreatic cancer plus gastrointestinal, hepatobiliary, and pancreatic surgeries and published in peer-reviewed journals., Trial Registration: Japan Registry of Clinical Trials, jRCTs031220160. Registered 24 th June 2022, https://jrct.niph.go.jp/en-latest-detail/jRCTs031220160 ., (© 2023. The Author(s).)

Objectives: Meticulous inspection of the mucosa during colonoscopy, represents a lengthier withdrawal time, but has been shown to increase adenoma detection rate (ADR). We investigated if artificial intelligence-aided speed monitoring can improve suboptimal withdrawal time., Methods: We evaluated the implementation of a computer-aided speed monitoring device during colonoscopy at a large academic endoscopy center. After informed consent, patients ≥18 years undergoing colonoscopy between 5 March and 29 April 2021 were examined without the use of the speedometer, and with the speedometer between 29 April and 30 June 2021. All colonoscopies were recorded, and withdrawal time was assessed based on the recordings in a blinded fashion. We compared mean withdrawal time, percentage of withdrawal time ≥6 min, and ADR with and without the speedometer., Results: One hundred sixty-six patients in each group were eligible for analyses. Mean withdrawal time was 9 min and 6.6 s (95% CI: 8 min and 34.8 s to 9 min and 39 s) without the use of the speedometer, and 9 min and 9 s (95% CI: 8 min and 45 s to 9 min and 33.6 s) with the speedometer; difference 2.3 s (95% CI: -42.3-37.7, p  = 0.91). The ADRs were 45.2% (95% CI: 37.6-52.8) without the speedometer as compared to 45.8% (95% CI: 38.2-53.4) with the speedometer ( p  = 0.91). The proportion of colonoscopies with withdrawal time ≥6 min without the speedometer was 85.5% (95% CI: 80.2-90.9) versus 86.7% (95% CI: 81.6-91.9) with the speedometer ( p  = 0.75)., Conclusions: Use of speed monitoring during withdrawal did not increase withdrawal time or ADR in colonoscopy., Clinicaltrials.gov Identifier: NCT04710251.

Purpose: Endoscopic submucosal dissection (ESD) is a minimally invasive treatment for early gastric cancer. However, perforations may happen and cause peritonitis during ESD. Thus, there is a potential demand for a computer-aided diagnosis system to support physicians in ESD. This paper presents a method to detect and localize perforations from colonoscopy videos to avoid perforation ignoring or enlarging by ESD physicians., Method: We proposed a training method for YOLOv3 by using GIoU and Gaussian affinity losses for perforation detection and localization in colonoscopic images. In this method, the object functional contains the generalized intersection over Union loss and Gaussian affinity loss. We propose a training method for the architecture of YOLOv3 with the presented loss functional to detect and localize perforations precisely., Results: To qualitatively and quantitatively evaluate the presented method, we created a dataset from 49 ESD videos. The results of the presented method on our dataset revealed a state-of-the-art performance of perforation detection and localization, which achieved 0.881 accuracy, 0.869 AUC, and 0.879 mean average precision. Furthermore, the presented method is able to detect a newly appeared perforation in 0.1 s., Conclusions: The experimental results demonstrated that YOLOv3 trained by the presented loss functional were very effective in perforation detection and localization. The presented method can quickly and precisely remind physicians of perforation happening in ESD. We believe a future CAD system can be constructed for clinical applications with the proposed method., (© 2023. CARS.)

Purpose: Minimally invasive surgery (MIS) using a thoraco- or laparoscope is becoming a more common surgical technique. In MIS, a magnified view from a thoracoscope helps surgeons conduct precise operations. However, there is a risk of the visible area becoming narrow. To confirm that the operation field is safe, the surgeon will draw the thoracoscope back to check the marginal area of the target and insert it again many times during MIS. To reduce the surgeon's load, we aim to visualize the entire thoracic cavity using a newly developed device called "panorama vision ring" (PVR)., Method: The PVR is used instead of a wound retractor or a trocar. It is a ring-type socket with one big hole for the thoracoscope and four small holes for tiny cameras placed around the big hole. The views from the tiny cameras are fused into one wider view that visualizes the entire thoracic cavity. A surgeon can proceed with an operation by checking what exists outside of the thoracoscopic view. Also, she/he can check whether or not bleeding has occurred from the image of the entire cavity., Results: We evaluated the view-expansion ability of the PVR by using a three-dimensional full-scale thoracic model. The experimental results showed that the entire thoracic cavity could be visible in a panoramic view generated by the PVR. We also demonstrated pulmonary lobectomy in virtual MIS using the PVR. Surgeons could perform a pulmonary lobectomy while checking the entire cavity., Conclusion: We developed the PVR, which uses tiny auxiliary cameras to create a panoramic view of the entire thoracic cavity during MIS. We aim to make MIS safer for patients and more comfortable for surgeons through the development of the PVR., (© 2023. CARS.)

Purpose: Segmentation tasks are important for computer-assisted surgery systems as they provide the shapes of organs and the locations of instruments. What prevents the most powerful segmentation approaches from becoming practical applications is the requirement for annotated data. Active learning provides strategies to dynamically select the most informative samples to reduce the annotation workload. However, most previous active learning literature has failed to select the frames that containing low-appearing frequency classes, even though the existence of these classes is common in laparoscopic videos, resulting in poor performance in segmentation tasks. Furthermore, few previous works have explored the unselected data to improve active learning. Therefore, in this work, we focus on these classes to improve the segmentation performance., Methods: We propose a class-wise confidence bank that stores and updates the confidence scores for each class and a new acquisition function based on a confidence bank. We apply confidence scores to explore an unlabeled dataset by combining it with a class-wise data mixture method to exploit unlabeled datasets without any annotation., Results: We validated our proposal on two open-source datasets, CholecSeg8k and RobSeg2017, and observed that its performance surpassed previous active learning studies with about [Formula: see text] improvement on CholecSeg8k, especially for classes with a low-appearing frequency. For robSeg2017, we conducted experiments with a small and large annotation budgets to validate situation that shows the effectiveness of our proposal., Conclusions: We presented a class-wise confidence score to improve the acquisition function for active learning and explored unlabeled data with our proposed class-wise confidence score, which results in a large improvement over the compared methods. The experiments also showed that our proposal improved the segmentation performance for classes with a low-appearing frequency., (© 2022. CARS.)

Objective: To develop a convolutional neural network to recognize the seminal vesicle and vas deferens (SV-VD) in the posterior approach of robot-assisted radical prostatectomy (RARP) and assess the performance of the convolutional neural network model under clinically relevant conditions., Methods: Intraoperative videos of robot-assisted radical prostatectomy performed by the posterior approach from 3 institutions were obtained between 2019 and 2020. Using SV-VD dissection videos, semantic segmentation of the seminal vesicle-vas deferens area was performed using a convolutional neural network-based approach. The dataset was split into training and test data in a 10:3 ratio. The average time required by 6 novice urologists to correctly recognize the SV-VD was compared using intraoperative videos with and without segmentation masks generated by the convolutional neural network model, which was evaluated with the test data using the Dice similarity coefficient. Training and test datasets were compared using the Mann-Whitney U-test and chi-square test. Time required to recognize the SV-VD was evaluated using the Mann-Whitney U-test., Results: From 26 patient videos, 1 040 images were created (520 SV-VD annotated images and 520 SV-VD non-displayed images). The convolutional neural network model had a Dice similarity coefficient value of 0.73 in the test data. Compared with original videos, videos with the generated segmentation mask promoted significantly faster seminal vesicle and vas deferens recognition (P < .001)., Conclusion: The convolutional neural network model provides accurate recognition of the SV-VD in the posterior approach RARP, which may be helpful, especially for novice urologists., (Copyright © 2022 Elsevier Inc. All rights reserved.)

Purpose: This paper aims to propose a deep learning-based method for abdominal artery segmentation. Blood vessel structure information is essential to diagnosis and treatment. Accurate blood vessel segmentation is critical to preoperative planning. Although deep learning-based methods perform well on large organs, segmenting small organs such as blood vessels is challenging due to complicated branching structures and positions. We propose a 3D deep learning network from a skeleton context-aware perspective to improve segmentation accuracy. In addition, we propose a novel 3D patch generation method which could strengthen the structural diversity of a training data set., Method: The proposed method segments abdominal arteries from an abdominal computed tomography (CT) volume using a 3D fully convolutional network (FCN). We add two auxiliary tasks to the network to extract the skeleton context of abdominal arteries. In addition, our skeleton-based patch generation (SBPG) method further enables the FCN to segment small arteries. SBPG generates a 3D patch from a CT volume by leveraging artery skeleton information. These methods improve the segmentation accuracies of small arteries., Results: We used 20 cases of abdominal CT volumes to evaluate the proposed method. The experimental results showed that our method outperformed previous segmentation accuracies. The averaged precision rate, recall rate, and F-measure were 95.5%, 91.0%, and 93.2%, respectively. Compared to a baseline method, our method improved 1.5% the averaged recall rate and 0.7% the averaged F-measure., Conclusions: We present a skeleton context-aware 3D FCN to segment abdominal arteries from an abdominal CT volume. In addition, we propose a 3D patch generation method. Our fully automated method segmented most of the abdominal artery regions. The method produced competitive segmentation performance compared to previous methods., (© 2022. CARS.)

Purpose: A surgical navigation system helps surgeons understand anatomical structures in the operative field during surgery. Patient-to-image registration, which aligns coordinate systems between the CT volume and a positional tracker, is vital for accurate surgical navigation. Although a point-based rigid registration method using fiducials on the body surface is often utilized for laparoscopic surgery navigation, precise registration is difficult due to such factors as soft tissue deformation. We propose a method that compensates a transformation matrix computed using fiducials on the body surface based on the analysis of positional information in the database., Methods: We built our database by measuring the positional information of the fiducials and the guidance targets in both the CT volume and positional tracker coordinate systems through previous surgeries. We computed two transformation matrices: using only the fiducials and using only the guidance targets in all the data in the database. We calculated the differences between the two transformation matrices in each piece of data. The compensation transformation matrix was computed by averaging these difference matrices. In this step, we selected the data from the database based on the similarity of the fiducials and the configuration of the guidance targets., Results: We evaluated our proposed method using 20 pieces of data acquired during laparoscopic gastrectomy for gastric cancer. The locations of blood vessels were used as guidance targets for computing target registration error. The mean target registration errors significantly decreased from 33.0 to 17.1 mm before and after the compensation., Conclusion: This paper described a registration error compensation method using a database for image-guided laparoscopic surgery. Since our proposed method reduced registration error without additional intraoperative measurements during surgery, it increases the accuracy of surgical navigation for laparoscopic surgery., (© 2022. CARS.)

Objective: The objective of this study was to investigate the mid-term outcomes of embolization procedures for type II endoleak after endovascular abdominal aortic repair, and clarify the risk factors for aneurysm enlargement after embolization procedures., Methods: This was a retrospective multicenter registry study enrolling patients who underwent embolization procedures for type II endoleaks after EVAR from January 2012 to December 2018 at 19 Japanese centers. The primary end point was the rate of freedom from aneurysm enlargement, more than 5 mm in the aortic maximum diameter, after an embolization procedure. Demographic, procedural, follow-up, and laboratory data were collected. Continuous variables were summarized descriptively, and Kaplan-Meier analyses and a Cox regression model were used for statistical analyses., Results: A total of 315 patients (248 men and 67 women) were enrolled. The average duration from the initial embolization procedure to the last follow-up was 31.6 ± 24.6 months. The rates of freedom from aneurysm enlargement at 3 and 5 years were 55.4 ± 3.8% and 37.0 ± 5.2%, respectively. A multivariate analysis revealed that a larger aortic diameter at the initial embolization procedure and the presence of a Moyamoya endoleak, defined as heterogeneous contrast opacity with an indistinct faint border, were associated with aneurysm enlargement after embolization management., Conclusions: The embolization procedures were generally ineffective in preventing further expansion of abdominal aortic aneurysms in patients with type II endoleaks after EVAR, especially in patients with a large abdominal aortic aneurysm and/or a presence of a Moyamoya endoleak., (Copyright © 2022 Society for Vascular Surgery. All rights reserved.)

This study presents a novel framework for classifying and visualizing pneumonia induced by COVID-19 from CT images. Although many image classification methods using deep learning have been proposed, in the case of medical image fields, standard classification methods are unable to be used in some cases because the medical images that belong to the same category vary depending on the progression of the symptoms and the size of the inflamed area. In addition, it is essential that the models used be transparent and explainable, allowing health care providers to trust the models and avoid mistakes. In this study, we propose a classification method using contrastive learning and an attention mechanism. Contrastive learning is able to close the distance for images of the same category and generate a better feature space for classification. An attention mechanism is able to emphasize an important area in the image and visualize the location related to classification. Through experiments conducted on two-types of classification using a three-fold cross validation, we confirmed that the classification accuracy was significantly improved; in addition, a detailed visual explanation was achieved comparison with conventional methods., (© 2022. The Author(s).)

Purpose: Precise polyp detection and localisation are essential for colonoscopy diagnosis. Statistical machine learning with a large-scale data set can contribute to the construction of a computer-aided diagnosis system for the prevention of overlooking and miss-localisation of a polyp in colonoscopy. We propose new visual explaining methods for a well-trained object detector, which achieves fast and accurate polyp detection with a bounding box towards a precise automated polyp localisation., Method: We refine gradient-weighted class activation mapping for more accurate highlighting of important patterns in processing a convolutional neural network. Extending the refined mapping into multiscaled processing, we define object activation mapping that highlights important object patterns in an image for a detection task. Finally, we define polyp activation mapping to achieve precise polyp localisation by integrating adaptive local thresholding into object activation mapping. We experimentally evaluate the proposed visual explaining methods with four publicly available databases., Results: The refined mapping visualises important patterns in each convolutional layer more accurately than the original gradient-weighted class activation mapping. The object activation mapping clearly visualises important patterns in colonoscopic images for polyp detection. The polyp activation mapping localises the detected polyps in ETIS-Larib, CVC-Clinic and Kvasir-SEG database with mean Dice scores of 0.76, 0.72 and 0.72, respectively., Conclusions: We developed new visual explaining methods for a convolutional neural network by refining and extending gradient-weighted class activation mapping. Experimental results demonstrated the validity of the proposed methods by showing that accurate visualisation of important patterns and localisation of polyps in a colonoscopic image. The proposed visual explaining methods are useful for the interpreting and applying a trained polyp detector., (© 2022. CARS.)

Background and Aims: Because of a lack of reliable preoperative prediction of lymph node involvement in early-stage T2 colorectal cancer (CRC), surgical resection is the current standard treatment. This leads to overtreatment because only 25% of T2 CRC patients turn out to have lymph node metastasis (LNM). We assessed a novel artificial intelligence (AI) system to predict LNM in T2 CRC to ascertain patients who can be safely treated with less-invasive endoscopic resection such as endoscopic full-thickness resection and do not need surgery., Methods: We included 511 consecutive patients who had surgical resection with T2 CRC from 2001 to 2016; 411 patients (2001-2014) were used as a training set for the random forest-based AI prediction tool, and 100 patients (2014-2016) were used to validate the AI tool performance. The AI algorithm included 8 clinicopathologic variables (patient age and sex, tumor size and location, lymphatic invasion, vascular invasion, histologic differentiation, and serum carcinoembryonic antigen level) and predicted the likelihood of LNM by receiver-operating characteristics using area under the curve (AUC) estimates., Results: Rates of LNM in the training and validation datasets were 26% (106/411) and 28% (28/100), respectively. The AUC of the AI algorithm for the validation cohort was .93. With 96% sensitivity (95% confidence interval, 90%-99%), specificity was 88% (95% confidence interval, 80%-94%). In this case, 64% of patients could avoid surgery, whereas 1.6% of patients with LNM would lose a chance to receive surgery., Conclusions: Our proposed AI prediction model has a potential to reduce unnecessary surgery for patients with T2 CRC with very little risk. (Clinical trial registration number: UMIN 000038257.)., (Copyright © 2022 American Society for Gastrointestinal Endoscopy. Published by Elsevier Inc. All rights reserved.)

Competing Interests: Shin-ei Kudo, Masashi Misawa, and Yuichi Mori have received speaking honoraria from Olympus Corporation (Tokyo, Japan) and have ownership interest in the products of Cybernet Systems (Tokyo, Japan). Masashi Misawa, Shin-ei Kudo, and Yuichi Mori have patents (Japan Patent JP 6059271 and JP 6580446) licensed to Cybernet Systems and Showa University.

Background: Artificial intelligence (AI) for polyp detection is being introduced to colonoscopy, but there is uncertainty how this affects endoscopists' ability to detect polyps and neoplasms. We performed a video-based study to address whether AI improved the endoscopists' performance to detect polyps., Methods: We established a dataset of 200 colonoscopy videos (length 5 s; 100 without polyps and 100 with one polyp). About 33 early-career endoscopists (50-400 colonoscopies performed) from 10 European countries classified each video as either 'polyp present' or 'polyp not present'. The video assessment was performed twice with a four-week interval. The first assessment was performed without any AI tool, whereas the second was performed with an AI tool for polyp detection. The primary endpoint was early-career endoscopists' sensitivity to detect polyps. Gold standard for presence and histology of polyps were confirmed by two expert endoscopists and pathologists, respectively. McNemar's test was used for statistical significance., Results: There were 86 neoplastic and 14 non-neoplastic polyps (mean size 5.6 mm) in the 100 videos with polyps. Early-career endoscopists' sensitivity to detect polyps increased from 86.3% (95% confidence interval [CI]: 85.1-87.5%) to 91.7% (95%CI: 90.7-92.6%) with the AI aid ( p  < .0001). Their sensitivity to detect neoplastic polyps increased from 85.4% (95% CI: 84.0-86.7%) to 92.1% (95%CI: 91.1-93.1%) with the AI aid ( p  < .0001)., Conclusion: The polyp detection AI tool helped early-career endoscopists to increase their sensitivity to identify all polyps and neoplastic polyps during colonoscopy.