Search

Recent literature has raised concerns about the sensitivity and accuracy of radiographs at diagnosing rib fractures. Studies have shown that chest computed tomography (CT) has far greater sensitivity at detecting rib fractures than radiographs. The purpose of this study was to evaluate the sensitivity of skeletal survey (SS) radiographs at diagnosis of rib fractures compared to CT in the pediatric population. This retrospective review included 57 patients who had undergone both a SS and a CT chest or CT chest/abdomen/pelvis within 30 days of each other for the indication of either non-accidental (NAT) or accidental trauma between 2009 and 2017. Images and reports were analyzed by a pediatric radiology fellow for presence/absence of fracture, evidence of healing and location of rib fracture, including rib level, location within the rib (anterior, lateral, and posterior), and side (right versus left). 225 rib fractures were identified in 25 patients on CT. 38 of those fractures were missed on the preceding SS, yielding a miss rate of 17%. Acute fractures were more likely to be missed than chronic or healing fractures (p ≤ 0.01). Location within the rib did not impact rib detection on radiographs. Left-sided rib fractures were not more common in NAT patients, compared to accidental trauma. SS miss approximately 17% of all rib fractures and CT is more sensitive modality in the detection of rib fractures, particularly acute rib fractures, regardless of location. Low-dose Chest CT could be a helpful modality in the work-up of NAT trauma.

Colonoscopy screening test plays a crucial role in identifying and classifying possible cancerous polyps. The automation of polyp classification is challenging due to the limitations associated with the traditional handcrafted image features. This paper proposes a deep learning model for colorectal polyp classification, referred to as DeepCPC. This model consists of three main stages: image pre-processing using patch extraction and data augmentation; model initialization using six pre-trained CNNs to generate the fine-tuned baseline model; and formulating and learning generic yet discriminating image descriptors extracted and fused from the convolutional layers of two efficient CNNs architectures to classify colorectal polyps. The DeepCPC architecture has been fine-tuned on the CVC-Clinic dataset through a complete end-to-end training in which the patch extraction and image augmentation have been applied to generate more colonoscopy images for the patients. The experimental results show that the baseline fine-tuned model achieves an accuracy of 97.6%, and the final DeepCPC model achieves an accuracy of 98.4%. The reported results also demonstrate the capability of the proposed approach in identifying polyps in terms of precision, recall, and f-score. The DeepCPC helps the endoscopic physicians in classifying polyps and decreasing the colorectal polyp miss rate.

Colorectal cancer (CRC) was the second-ranked worldwide type of cancer during 2020 due to the crude mortality rate of 12.0 per 100000 inhabitants. It can be prevented if glandular tissue (adenomatous polyps) is detected early. Colonoscopy has been strongly recommended as a screening test for both early cancer and adenomatous polyps. However, it has some limitations that include the high polyp miss rate for smaller (< 10 mm) or flat polyps, which are easily missed during visual inspection. Due to the rapid advancement of technology, artificial intelligence (AI) has been a thriving area in different fields, including medicine. Particularly, in gastroenterology AI software has been included in computer-aided systems for diagnosis and to improve the assertiveness of automatic polyp detection and its classification as a preventive method for CRC. This article provides an overview of recent research focusing on AI tools and their applications in the early detection of CRC and adenomatous polyps, as well as an insightful analysis of the main advantages and misconceptions in the field.

Impaired cognitive control has been associated with the occurrence of attentional errors in those with schizophrenia. However, the extent of altered proactive or reactive control underlying such errors is still unknown. Twenty-two patients with schizophrenia and 21 healthy matched controls performed a detection task (i.e., the continuous temporal expectancy task). Electrophysiological measures of proactive and reactive control were based on two periods of interest: during the target presentation (the critical window) and four trials before the critical window. Regarding the proactive mode, patients with schizophrenia exhibited a specific decrease in frontal midline theta power during the critical window before a miss compared to a correct detection. In contrast, the contingent negative variation amplitude was altered regardless of the response type, four trials before the critical window. Regarding the reactive mode, a reduced P3 amplitude was revealed later before a miss than a correct detection with differences apparent only two trials before the critical window in patients with schizophrenia, whereas it was observable up to four trials prior in healthy controls. Moreover, only the P3 amplitude reduction in patients with schizophrenia predicted the miss rate and was anti-correlated with the clinical symptoms. Thus, our results revealed a specific impairment of the proactive goal-updating process before an error and an altered implementation of the endogenous proactive mode engagement regardless of the response type. The results also highlighted the strong relationship between the disrupted reactive mode and the increased rate of attentional errors and severity of the clinical symptoms of schizophrenia.

Background Adenoma detection rate (ADR) varies significantly between endoscopists, with adenoma miss rates (AMRs) up to 26 %. Artificial intelligence (AI) systems may improve endoscopy quality and reduce the rate of interval cancer. We evaluated the efficacy of an AI system in real-time colonoscopy and its influence on AMR and ADR. Methods This prospective, nonrandomized, comparative study analyzed patients undergoing diagnostic colonoscopy at a single endoscopy center in Germany from June to October 2020. Every patient was examined concurrently by an endoscopist and AI using two opposing screens. The AI system, overseen by a second observer, was not visible to the endoscopist. AMR was the primary outcome. Both methods were compared using McNemar test. Results 150 patients were included (mean age 65 years [standard deviation 14]; 69 women). There was no significant or clinically relevant difference (P = 0.75) in AMR between the AI system (6/197, 3.0 %; 95 % confidence interval [CI] 1.1–6.5) and routine colonoscopy (4/197, 2.0 %; 95 %CI 0.6–5.1). The polyp miss rate of the AI system (14/311, 4.5 %; 95 %CI 2.5–7.4) was not significantly different (P = 0.72) from routine colonoscopy (17/311, 5.5 %; 95 %CI 3.2–8.6). There was no significant difference (P = 0.50) in ADR between routine colonoscopy (78/150, 52.0 %; 95 %CI 43.7–60.2) and the AI system (76/150, 50.7 %; 95 %CI 42.4–58.9). Routine colonoscopy detected adenomas in two patients that were missed by the AI system. Conclusion The AI system performance was comparable to that of experienced endoscopists during real-time colonoscopy with similar high ADR (> 50 %).

We have developed the computer-aided detection (CADe) system using an original deep learning algorithm based on a convolutional neural network for assisting endoscopists in detecting colorectal lesions during colonoscopy. The aim of this study was to clarify whether adenoma miss rate (AMR) could be reduced with CADe assistance during screening and surveillance colonoscopy. This study was a multicenter randomized controlled trial. Patients aged 40 to 80 years who were referred for colorectal screening or surveillance at four sites in Japan were randomly assigned at a 1:1 ratio to either the “standard colonoscopy (SC)-first group” or the “CADe-first group” to undergo a back-to-back tandem procedure. Tandem colonoscopies were performed on the same day for each participant by the same endoscopist in a preassigned order. All polyps detected in each pass were histopathologically diagnosed after biopsy or resection. A total of 358 patients were enrolled and 179 patients were assigned to the SC-first group or CADe-first group. The AMR of the CADe-first group was significantly lower than that of the SC-first group (13.8% vs. 36.7%, P

To systematically determine the miss rate and risk factors for polyps, adenomas and advanced adenomas in the same population, and to further analyze the impact of colonoscopy with anesthesia on miss rate.We retrospectively analyzed the information of the patients undergoing the second colonoscopy within 1 year after their first. The patient and lesion miss rate were calculated. The patient and lesion features of missed lesion were compared with non-missed lesion. Finally, the patients were divided into anesthesia group and without anesthesia group, and the impact of colonoscopy with anesthesia on missed lesions was further analyzed.The patient miss rate of polyps, adenomas and advanced adenomas was 32.8, 25.6 and 10.4%, and the lesions miss rate was 19.6, 15.8 and 7.2%. In multivariable logistic regression analysis, lesion-related factors (large number of lesions, small lesion size, flat shape and location at the right colon) and patient-related factors (male, elder, abdominal symptoms, surgical history, diverticulum, colonoscopy without anesthesia and suboptimal bowel preparation) were found to be independently associated with missed polyps and adenomas (Many factors of patients and lesions can affect the lesions miss rate. Colonoscopy with anesthesia can reduce the PMR and male and elderly patients are more likely to be missed during colonoscopy without anesthesia.

In this paper, an automated layer defect detection system for construction 3D printing is proposed. Initially, a step-by-step procedure is implemented to develop a deep convolutional neural network that receives images as input and is able to distinguish concrete layers from other surrounding objects through semantic pixel-wise segmentation. Using data augmentation techniques, 1M labeled images are generated and used to train and test the CNN model. Then, a defect detection module is developed which is able to detect deformations in the printed concrete layers extracted from the images using the CNN model. The evaluation results based on metrics such as accuracy, F1 score, and miss rate verify the acceptable performance of the developed system.

Several studies have shown a significant adenoma miss rate up to 35% during screening colonoscopy, especially in patients with diminutive adenomas. The use of artificial intelligence (AI) in colonoscopy has been gaining popularity by helping endoscopists in polyp detection, with the aim to increase their adenoma detection rate (ADR) and polyp detection rate (PDR) in order to reduce the incidence of interval cancers. The efficacy of deep convolutional neural network (DCNN)-based AI system for polyp detection has been trained and tested in ex vivo settings such as colonoscopy still images or videos. Recent trials have evaluated the real-time efficacy of DCNN-based systems showing promising results in term of improved ADR and PDR. In this review we reported data from the preliminary ex vivo experiences and summarized the results of the initial randomized controlled trials.

Lesions missed by colonoscopy are one of the main reasons for post-colonoscopy colorectal cancer, which is usually associated with a worse prognosis. Because the adenoma miss rate could be as high as 26%, it has been noted that endoscopists with higher adenoma detection rates are usually associated with lower adenoma miss rates. Artificial intelligence (AI), particularly the deep learning model, is a promising innovation in colonoscopy. Recent studies have shown that AI is not only accurate in colorectal polyp detection but can also reduce the miss rate. Nevertheless, the application of AI in real-time detection has been hindered by heterogeneity of the AI models and study design as well as a lack of long-term outcomes. Herein, we discussed the principle of various AI models and systematically reviewed the current data on the use of AI on colorectal polyp detection and miss rates. The limitations and future prospects of AI on colorectal polyp detection are also discussed.

Polyps are a group of cells growing on the inner surface of the colon. Over time, some polyps can lead to colon cancer, which is often fatal if found in its later stages. Colon cancer can be prevented if the polyps are identified and removed in their early stages. Colonoscopy is a very effective screening method to remove polyps and it largely prevents colon cancer. However, some polyps may not be detected during a colonoscopy due to human error. Over the past two decades, many studies have been conducted on computer-aided detection to reduce the miss rate of polyps. This study consists of two distinct parts, the detection of frames containing polyps and polyp segmentation. In the first section, a new convolutional neural network based on the VGG network is proposed. The proposed network has an accuracy of 86% on a newly collected dataset. In the polyp segmentation section, a fully convolutional network and an effective post-processing algorithm are presented. An evaluation of the proposed polyp segmentation system on the ETIS-LARIB database achieves an overall 82.00% F2 score, which outperforms the methods that participated in the sub-challenge of MICCAI.

Background Adenoma detection rate (ADR) is a quality indicator for colonoscopy. However, many missed adenomas have subsequently been identified after colonoscopies performed by endoscopists with ADR ≥ 25 %. Adenomas per positive participant (APP; mean number of adenomas detected by an endoscopist among screenees with positive findings) correlates well inversely with adenoma miss rate. This study aimed to evaluate whether APP added additional information on the detection rate for advanced adenomas (AADR) and proximal adenomas (pADR) and among endoscopists with acceptable ADRs (≥ 25 %). Methods A total of 47 endoscopists performed 7339 screening colonoscopies that were retrospectively reviewed. Using a cutoff APP value of 2.0, endoscopist performance was classified as high or low APP. Endoscopist ADRs were also classified as acceptable (25 % – 29 %), high standard (30 % – 39 %) and aspirational (≥ 40 %). Generalized linear models were used to assess the relationship between AADR or pADR, and ADR and APP, after adjusting for potential confounders. Results After adjusting for endoscopist performance and patient characteristics, endoscopists with high APP had a significant 2.1 percentage point increase in AADR (95 %CI 0.3 to 3.9; P = 0.02) and a 2.1 percentage point increase in pADR (95 %CI – 0.8 to 5.1; P = 0.15) compared to endoscopists with low APP. In total, 11 (24 %), 18 (38 %), and 18 (38 %) endoscopists were classified as having acceptable, high standard, and aspirational ADRs, respectively. APP values higher than the cutoff were found in 18 %, 44 %, and 72 % of endoscopists with acceptable, high standard, and aspirational ADRs, respectively (P = 0.02). Conclusion APP is helpful for identifying more meticulous endoscopists who can detect a greater number of advanced adenomas. Endoscopists who achieved an only acceptable ADR had the lowest APP.

BACKGROUND AND AIM We aimed to investigate whether linked color imaging (LCI) improves endoscopic visibility of early gastric cancers (EGC) after Helicobacter pylori eradication, which are often difficult to detect, and reduces the miss rate when compared with white-light imaging (WLI). METHODS The visibility study used two images, one each with WLI and LCI, from 84 consecutive EGC after H. pylori eradication. Endoscopic visibility was evaluated using a visibility score and color difference (CD) value. To analyze miss rates, we studied a library of recorded videos using both WLI and LCI for 70 other consecutive patients after H. pylori eradication, among whom 19 had EGC. Endoscopic screening was done using the same protocol to map the entire stomach. Six endoscopists reviewed the videos in a randomized order. Miss rates of EGC were compared among the modalities. RESULTS Mean [(±standard deviation) visibility scores with LCI were significantly higher than those with WLI (3.19 ± 0.84 vs 2.52 ± 0.98, P

Aims Despite the widespread use of endoscopy, upper gastrointestinal (UGI) cancers still present at an advanced stage. As survival is closely linked to stage and time of diagnosis, failure to detect subtle precancerous changes at endoscopy may compound poor prognoses. This study calculates the miss rate of UGI cancer over a 5-year period at a district general hospital. Methods All patients diagnosed with UGI cancer between January 2015 - December 2019 were identified from a prospectively collected cancer registry. Electronic health records and Unisoft GI reporting tool were used to identify patient demographics and previous UGI endoscopies. ‘Missed cancers’ were defined as patients who had a normal endoscopy within 3 years of their cancer diagnosis. Results The median age at diagnosis was 72.2 years (age range 24 - 98, n = 408) with a male predominance of 2:1 (66.6% male vs 33.4% female) in keeping with UK statistics. Within this 5 year study period, there were 22 missed cancers (5.4%, n = 408). A year by year break down shows miss rate in 2015 of 3% ( 3,n=100), 2016 of 4.2% (5,n=120), 2017 5.5% (5,n=91), 2018 6.4% (6,n=94) and most recently in 2019 3.2% (3,n=94). Conclusions In 2014, a meta-analysis by S.Menon et al recorded a miss rate of 11.3%. More recently published UK studies report miss rates between 6% - 7.3%, more in keeping with our local rate of 5.4%. Further assessment is required to assess whether the 2017 BSG and AUGIS UGI endoscopy quality standard statement will improve this rate.

Haustral folds are colon wall protrusions implicated for high polyp miss rate during optical colonoscopy procedures. If segmented accurately, haustral folds can allow for better estimation of missed surface and can also serve as valuable landmarks for registering pre-treatment virtual (CT) and optical colonoscopies, to guide navigation towards the anomalies found in pre-treatment scans. We present a novel generative adversarial network, FoldIt, for feature-consistent image translation of optical colonoscopy videos to virtual colonoscopy renderings with haustral fold overlays. A new transitive loss is introduced in order to leverage ground truth information between haustral fold annotations and virtual colonoscopy renderings. We demonstrate the effectiveness of our model on real challenging optical colonoscopy videos as well as on textured virtual colonoscopy videos with clinician-verified haustral fold annotations. All code and scripts to reproduce the experiments of this paper will be made available via our Computational Endoscopy Platform at https://github.com/nadeemlab/CEP., Comment: MICCAI 2021 (Early Accept) (Oral Presentation), *Saad Nadeem and Shawn Mathew contributed equally

Diagnosticians report that autism spectrum disorder (ASD) is immediately apparent in some, but not all, children ultimately diagnosed. Clinicians' initial diagnostic impressions have implications for ASD early detection, yet the literature raises questions about their accuracy. This study explores diagnostic impressions of ASD specialists made within the first 5 minutes of meeting a young child and investigates factors associated with the match between initial impressions and final diagnoses. Participants were children (n=294, aged 12-53 months) referred for an ASD evaluation as part of multi-site ASD screening studies. After 5 minutes observing each child, clinicians with expertise diagnosing ASD recorded if they thought the child would meet criteria for ASD following a complete evaluation, and recorded their confidence in this impression. Clinicians' initial impressions matched the final diagnosis in 81% of cases. Ninety-two percent of cases initially thought to have ASD met criteria following a full evaluation; however, 24% of cases initially thought not to have ASD also met criteria, suggesting a high miss rate. Clinicians were generally confident in their initial impressions, reporting highest confidence for children initially thought correctly not to have ASD. ASD behavioral presentation, but not demographic characteristics or developmental level, were associated with matching initial impression and final diagnosis, and confidence. Brief observations indicating ASD should trigger referral to intervention services, but are likely to under-detect positive cases and should not be used to rule out ASD, highlighting the need to incorporate information beyond initial clinical impression. LAY SUMMARY: When children come in for an autism evaluation, clinicians often form early impressions-before doing any formal testing-about whether the child has autism. We studied how often these early impressions match the final diagnosis, and found that clinicians could not easily rule out autism (many children who initially appeared not to have autism were ultimately diagnosed), but were generally accurate ruling in autism (when a child appeared to have autism within 5 minutes, they were almost always so diagnosed).

Non-Maximum Suppression (NMS) is essential for object detection and affects the evaluation results by incorporating False Positives (FP) and False Negatives (FN), especially in crowd occlusion scenes. In this paper, we raise the problem of weak connection between the training targets and the evaluation metrics caused by NMS and propose a novel NMS-Loss making the NMS procedure can be trained end-to-end without any additional network parameters. Our NMS-Loss punishes two cases when FP is not suppressed and FN is wrongly eliminated by NMS. Specifically, we propose a pull loss to pull predictions with the same target close to each other, and a push loss to push predictions with different targets away from each other. Experimental results show that with the help of NMS-Loss, our detector, namely NMS-Ped, achieves impressive results with Miss Rate of 5.92% on Caltech dataset and 10.08% on CityPersons dataset, which are both better than state-of-the-art competitors., Comment: ICMR2021 camera ready version

Iceberg distribution, dispersion and melting patterns are fundamental aspects in the balance of heat and freshwater in the Southern Ocean; yet these features are not fully understood. This lack of understanding is, in part, due to the difficulties in accurately identifying icebergs in different environmental conditions. To improve the understanding, reliable iceberg detection tools are necessary to achieve a detailed picture of iceberg drift and disintegration patterns, an thus to gain further information on the freshwater input into the Southern Ocean. Here, we present an accurate automatic large-scale iceberg detection method using an alternative machine learning architecture applied to high resolution Synthetic Aperture Radar (SAR) images. Our method is based on the concept of adaptability and focuses on improving the performance of identifying icebergs in ambiguous environmental contexts with wide radiometric, textural, size and shape variability. The fundamentals of the method are centred on superpixel segmentation, ensemble learning and incremental learning. The method is applied to a dataset containing 586 ENVISAT Advanced SAR images acquired during 2003–2005 (Weddell Sea region) and to the Radarsat-1 Antarctic Mapping Project (RAMP) mosaic, covering the Antarctic wide near-coastal zone. These images cover scenes under heterogenous backscattering signatures for all seasons with variable meteorological, oceanographic and acquisition parameters (e.g. band, polarization). Our method is highly adaptable to distinguish icebergs from ambiguous objects hidden in the images. The average false positive rate and miss rate are 2.3 ± 0.4% and 3.3 ± 0.4%, respectively. Overall, 9512 icebergs with sizes varying from 0.1 to 4567.82 km2 are detected with average classification accuracy of 97.5 ± 0.6%. The results confirm that the method presented here is robust for widespread iceberg detection in the Antarctic seas.

The impact of virtual chromoendoscopy such as narrow-band imaging, Fujinon intelligent chromoendoscopy, blue-light imaging, linked-color imaging, and i-SCAN on adenoma detection rate has been variable. However, adenoma miss rate (another measure suggested as a quality indicator) of electronic chromoendoscopy modalities has not been systematically evaluated.This study aimed to perform a systematic review and meta-analysis to examine the adenoma miss rate of white-light endoscopy compared with electronic chromoendoscopy.Medline, Embase, Scopus, Web of Knowledge, and Cochrane were the data sources for this study.The studies selected were tandem randomized controlled trials of electronic chromoendoscopy modalities compared with white-light endoscopy.Electronic chromoendoscopy compared with white light endoscopy was used to measure the adenoma miss rate.Primary outcome was the pooled adenoma miss rate of electronic chromoendoscopy techniques compared with white-light endoscopy. Secondary outcomes were subgroup analysis, adenoma size analysis, and adenoma detection rate.A total of 3507 patients were evaluated from 7 eligible tandem randomized control trials. A total of 1423 patients had white-light endoscopy as the first of the tandem examinations. The rest of the patients had electronic chromoendoscopy as the first of the tandem examination (narrow-band imaging, 988 patients; Fujinon intelligent chromoendoscopy, 728 patients; i-SCAN, 233 patients; blue-light imaging, 64 patients; and linked-color imaging, 71 patients). The pooled adenoma miss rate for electronic chromoendoscopy was not different than white-light endoscopy (17.9% vs 21%; OR, 0.72 (0.67-1.11); I 67%; p = 0.13). When only narrow-band imaging, blue-light imaging, and linked-color imaging were considered, the pooled rate was statistically significant (OR, 0.60 (0.37-0.98); p = 0.04). The pooled adenoma detection rate was not statistically different with electronic chromoendoscopy than white-light endoscopy (OR, 1.02 (0.88-1.19); p = 0.78).The small number of studies to assess the impact of each modality limited stratified conclusions.Electronic chromoendoscopy is not associated with a significant reduction in adenoma miss rate compared with white-light colonoscopy.

Background and aim: Right colon adenomas can easily missed with significant adenoma miss rate particularly when they are located behind the haustral folds and anatomical flexures. The aim of this study was to assess the significance of retroflexion technique in the right colon for improving adenoma detection rates compared with forward-view technique. Methods: A total of 200 patients scheduled for elective colonoscopy were recruited. During the standard forward-view examination, the colonoscope was withdrawn from the cecum to the hepatic flexure while under meticulous observation. Subsequently, the cecum was reintubated and the right side of the colon was assessed in the retroflexed-view to the hepatic flexure. The primary outcome of this study was comparison of the adenoma detection rates and adenoma miss rate of the retroflexed-view versus forward-view examination of the right colon. Results: Of the 200 enrolled patients, retroflexion was successful in 182 (91%). Forward-view examination of the right colon detected 88 polyps, of which 66 were adenomas yielding a polyp and adenoma detection rates of 25.5% and 19%, respectively. Additionally, retroflexion detected a further 30 polyps of which 23 were adenomas, accordingly improving the polyp and adenoma detection rates to 29.5% and 25%, respectively. Furthermore, the polyp and adenoma miss rate in the right colon was 24.1% and 25.8% respectively. Conclusion: Colonoscopy with right-sided retroflexion after forward-view examination significantly increases the detection of adenomas compared with standard forward-view examination with a high success rate and small risk of adverse events. Resumo: Antecedentes e objetivo: Os adenomas do cólon direito podem não ser diagnosticados, apresentando uma elevada taxa de não detecção de adenoma, especialmente quando estão localizados atrás das pregas haustrais e flexões anatômicas. O objetivo deste estudo foi comparar, no cólon direito, a técnica de visualização retroflexionada com a técnica de visualização frontal quanto à taxa de detecção de adenoma. Métodos: Um total de 200 pacientes agendados para colonoscopia eletiva foram recrutados para o presente estudo. Durante o exame de visualização frontal padrão, o colonoscópio foi retirado do ceco em direção à flexão hepática sob observação meticulosa. Posteriormente, o ceco foi reintubado e o lado direito do cólon foi avaliado na visualização retroflexionada em direção à flexão hepática. O objetivo primário deste estudo foi comparar a taxa de detecção de adenoma e a taxa de não detecção de adenoma do exame de visualização retroflexionada e visualização frontal do cólon direito. Resultados: Dos 200 pacientes inscritos, a retroflexão foi bem sucedida em 182 (91%). A visualização frontal do cólon direito detectou 88 pólipos, dos quais 66 eram adenomas, com uma taxa de detecção de pólipos e taxa de detecção de adenoma de 25,5% e 19%, respectivamente. Além disso, a visualização retroflexionada detectou mais 30 pólipos, dos quais 23 eram adenomas, aumentando a taxa de detecção de pólipos e adenoma para 29,5% e 25%, respectivamente. Além disso, a taxa não detecção de pólipos e a taxa de não detecção de adenoma no cólon direito foram de 24,1% e 25,8%, respectivamente. Conclusão: A colonoscopia com visualização retroflexionada direita após visualização frontal aumenta significativamente a detecção de adenomas em comparação com a visualização frontal padrão, com uma alta taxa de sucesso e baixo risco de eventos adversos. Keywords: Colonoscopy, Forward view, Retroflexed view, Adenoma, Palavras-chave: Colonoscopia, Visualização frontal, Visualização retroflexionada, Adenoma

Background: The benefits of narrow band imaging (NBI) for improving the detection rate of colorectal polyps remain unclear. New generation NBI using the 290 system (290-NBI) provides an at least tw...

Recent breakthroughs in artificial intelligence (AI), specifically via its emerging sub-field "deep learning," have direct implications for computer-aided detection and diagnosis (CADe and/or CADx) for colonoscopy. AI is expected to have at least 2 major roles in colonoscopy practice-polyp detection (CADe) and polyp characterization (CADx). CADe has the potential to decrease the polyp miss rate, contributing to improving adenoma detection, whereas CADx can improve the accuracy of colorectal polyp optical diagnosis, leading to reduction of unnecessary polypectomy of non-neoplastic lesions, potential implementation of a resect-and-discard paradigm, and proper application of advanced resection techniques. A growing number of medical-engineering researchers are developing both CADe and CADx systems, some of which allow real-time recognition of polyps or in vivo identification of adenomas, with over 90% accuracy. However, the quality of the developed AI systems as well as that of the study designs vary significantly, hence raising some concerns regarding the generalization of the proposed AI systems. Initial studies were conducted in an exploratory or retrospective fashion by using stored images and likely overestimating the results. These drawbacks potentially hinder smooth implementation of this novel technology into colonoscopy practice. The aim of this article is to review both contributions and limitations in recent machine-learning-based CADe and/or CADx colonoscopy studies and propose some principles that should underlie system development and clinical testing.

BackgroundOur objective was to perform a systematic review of studies reporting the accuracy of termination of resuscitation rules (TORRs) for out-of-hospital cardiac arrest (OHCA).MethodsWe performed a comprehensive search of the literature for studies evaluating the accuracy of TORRs, with two investigators abstracting relevant data from each study regarding study design, study quality and the accuracy of the TORRs. Bivariate meta-analysis was performed using the mada procedure in R.ResultsWe identified 14 studies reporting the performance of 9 separate TORRs. The sensitivity (proportion of eventual survivors for whom the TORR recommends resuscitation and transport) was generally high: 95% for the European Resuscitation Council (ERC) TORR, 97% for the basic life support (BLS) TORR and 99% for the advanced life support (ALS) TORR. The BLS and ERC TORR were more specific, which would lead to fewer futile transports, and all three of these TORRs had a miss rate of ≤0.13% (defined as a case where a patient is recommended for termination but survives). The pooled proportion of patients for whom each rule recommends TOR was much higher for the ERC and BLS TORRs (93.5% and 74.8%, respectively) than for the ALS TORR (29.0%).ConclusionsThe BLS and ERC TORRs identify a large proportion of patients who are candidates for termination of resuscitation following OHCA while having a very low rate of misclassifying eventual survivors (

ObjectiveFaecal immunochemical test (FIT) shows promise as a non-invasive triage test for colorectal cancer (CRC) in the symptomatic population. The aim of this study was to assess the use of FIT within the recent NG12 and DG30 National Institute for Health and Care Excellence (NICE) guidelines.DesignA single-centre prospective study of patients referred to University Hospitals Coventry and Warwickshire NHS Trust via the 2-week wait (TWW) pathway between January 2015 and March 2016 was conducted. 612 patients were reviewed, of which 519 were found to meet the NG12 criteria and 79 met the DG30 criteria. Data included age, sex, symptoms, colonoscopy or CT colonography, histology and FIT results.Main outcome measuresFIT was performed in all patients and sensitivity, specificity, positive predictive value and negative predictive value, with 95% CI, for cancers and adenomas within each pathway (TWW, NG12 and DG30) was calculated.ResultsCRC sensitivity in TWW pathway patients, NG12 and DG30 group was 86.84% (95% CI 71.91% to 95.59%), 84.85% (95% CI 68.1% to 94.89%) and 100% (95% CI 47.82% to 100%), respectively. Specificity was 82.23% (95% CI 78.85% to 85.27%), 81.28% (95% CI 77.52% to 84.65%) and 91.89% (95% CI 83.18% to 96.97%), respectively. Adenoma sensitivity in the groups was 30.69% (95% CI 29.9% to 40.66%), 30.77% (95% CI 21.51% to 41.32%) and 25% (95% CI 3.19% to 65.09%), respectively.ConclusionUse of FIT within the remit of the NG12 NICE guidelines shows a high sensitivity and specificity and may be an effective triage tool when considering whether to perform investigations. However, there is still a miss rate. FIT within DG30 has excellent sensitivity and improved specificity; however, DG30 targets lower risk groups and accounts for only 13% of the entire referrals for suspected cancer. Therefore, managing the larger, higher risk NG12 group may require the addition of another test or marker to ensure that CRC is not missed.

BACKGROUND AND AIM To assess the visibility of colorectal lesions using blue laser imaging (BLI)-bright and linked-color imaging (LCI) with an eye-tracking system. METHODS Eleven endoscopists evaluated 90 images of 30 colorectal lesions. The lesions were randomly selected. Three images of each lesion comprised white light imaging (WLI), BLI-bright, and LCI in the same position. Participants gazed at the images, and their eye movements were tracked by the eye tracker. We analyzed whether the participants could detect the lesion and how long they took to detect the lesion. We assessed the miss rate and detection time among the imaging modalities. RESULTS One endoscopist was excluded, and 10 endoscopists were assessed. Overall, 12.6% of lesions were missed with WLI, 6.0% with BLI-bright, and 4.3% with LCI; the miss rate of BLI-bright and LCI was significantly lower than that of WLI (P

Contains fulltext : 205162.pdf (Publisher’s version ) (Open Access) BACKGROUND AND AIMS: Colorectal cancer (CRC) is largely preventable with routine screening and surveillance colonoscopy; however, interval cancers arising from precancerous lesions missed by standard colonoscopy still occur. An increased adenoma detection rate (ADR) has been found to be inversely associated with interval cancers. The G-EYE device includes a reusable balloon integrated at the distal tip of a standard colonoscope, which flattens haustral folds, centralizes the colonoscope's optics, and reduces bowel slippage. The insufflated balloon also aims to enhance visualization of the colon during withdrawal, thereby increasing the ADR. METHODS: In this randomized, controlled, international, multicenter study (11 centers), patients (aged >/=50 years) referred to colonoscopy for screening, surveillance, or changes in bowel habits were randomized to undergo either balloon-assisted colonoscopy by using an insufflated balloon during withdrawal or standard high-definition colonoscopy. The primary endpoint was the ADR. RESULTS: One thousand patients were enrolled between May 2014 and September 2016 to undergo colonoscopy by experienced endoscopists; 803 were finally analyzed (standard colonoscopy n = 396; balloon-assisted colonoscopy n = 407). Baseline parameters were similar in both groups. Balloon-assisted colonoscopy provided a 48.0% ADR compared with 37.5% in the standard colonoscopy group (28% increase; P = .0027). Additionally, balloon-assisted colonoscopy provided for a significant increase in detection of advanced (P = .0033) flat adenomas (P < .0001) and sessile serrated adenomas/polyps (P = .0026). CONCLUSION: Balloon-assisted colonoscopy yielded a higher ADR and increased the detection of advanced, flat, and sessile serrated adenomas/polyps when compared with standard colonoscopy. Improved detection by the G-EYE device could impact the quality of CRC screening by reducing miss rates and consequently reducing interval cancer incidence. (Clinical trial registration number: NCT01917513.).

ObjectiveThe effect of colonoscopy on colorectal cancer mortality is limited by several factors, among them a certain miss rate, leading to limited adenoma detection rates (ADRs). We investigated the effect of an automatic polyp detection system based on deep learning on polyp detection rate and ADR.DesignIn an open, non-blinded trial, consecutive patients were prospectively randomised to undergo diagnostic colonoscopy with or without assistance of a real-time automatic polyp detection system providing a simultaneous visual notice and sound alarm on polyp detection. The primary outcome was ADR.ResultsOf 1058 patients included, 536 were randomised to standard colonoscopy, and 522 were randomised to colonoscopy with computer-aided diagnosis. The artificial intelligence (AI) system significantly increased ADR (29.1%vs20.3%, pConclusionsIn a low prevalent ADR population, an automatic polyp detection system during colonoscopy resulted in a significant increase in the number of diminutive adenomas detected, as well as an increase in the rate of hyperplastic polyps. The cost–benefit ratio of such effects has to be determined further.Trial registration numberChiCTR-DDD-17012221; Results.

Background and aims Colorectal cancer is a leading cause of death. Colonoscopy is the criterion standard for detection and removal of precancerous lesions and has been shown to reduce mortality. The polyp miss rate during colonoscopies is 22% to 28%. DEEP DEtection of Elusive Polyps (DEEP2) is a new polyp detection system based on deep learning, which alerts the operator in real-time to the presence and location of polyps. The primary outcome was the performance of DEEP2 on the detection of elusive polyps. Methods The DEEP2 system was trained on 3,611 hours of colonoscopy videos derived from 2 sources, and was validated on a set comprising 1,393 hours, from a third unrelated source. The ground truth labeling was provided by offline gastroenterologist annotators, who were able to watch the video in slow-motion and pause/rewind as required. To assess the applicability, stability, user experience and in order to obtain some preliminary data on performance in a real-life scenario, a preliminary prospective clinical validation study was performed, comprising 100 procedures (clinicaltrial.gov ID: NCT04693078). Results DEEP2 achieved a sensitivity of 97.1% at 4.6 false alarms per video for all polyps, 88.5% and 84.9% for polyps that are in the field of view for less than 5 and 2 seconds, respectively. DEEP2 was able to detect polyps, not seen by live real-time endoscopists or offline annotators in an average of 0.22 polyps per sequence. In the clinical validation study the system detected an average of 0.89 additional polyps per procedure. No adverse events occurred. Conclusions DEEP2 has a high sensitivity for polyp detection and was effective in increasing the detection of polyps both in colonoscopy videos and in real procedures with a low number of false alarms.

Autism spectrum disorder is a neurodevelopmental disorder that characterizes by reducing concentration on social activities and improving interest in non-social tasks. The aim of this work is to investigate eye gazing images and identify autism applying various machine learning techniques. Therefore, we collected eye-tracking data from the Figshare data repository. But, these scanpath images were almost similar for normal and autistic children. To obtain similar groups, k-means clustering method was used and generated four clusters. Further, several classifiers were applied into primary data and these clusters and evaluated the performance of them using various metrics. After the assessment of overall results, MLP shows the highest 87% accuracy in cluster 1. In addition, it shows the best area under curve, f-measure, g-mean, sensitivity, specificity, fall out and miss rate respectively. This predictive model could notably useful to forecast ASD status at early stages.

Background. We frequently encounter early gastric cancer (EGC) that could not be detected in the previous esophagogastroduodenoscopy even if the procedure was annually performed. However, little evidence exists regarding the characteristics of false-negative EGCs. Our aim was to reveal the clinical features of false-negative EGCs. Methods. We retrospectively reviewed cases of endoscopic submucosal dissection (ESD) for EGCs in Fukuchiyama City Hospital between January 2013 and May 2019. False-negative EGCs were defined as EGCs within 3 years of negative endoscopy. We evaluated the clinical characteristics of false-negative and initially detected EGCs and the difference in the detected and last missed endoscopy in false-negative EGCs. The miss rates of false-negative EGCs were compared between trainees (nonboard-certified endoscopists) and experienced endoscopists (board-certified endoscopists); thereafter, the characteristics of false-negative EGCs missed by trainees were investigated. Results. Of 219 cases, 119 were classified as false-negative EGCs. False-negative EGCs were characterized as smaller lesions, which presented with normal color or gastritis-like appearance, and were diagnosed after ESD and H. pylori eradication ( P < 0.01 ). The rate of trainees in the last missed endoscopy was significantly higher than that in the detected endoscopy. The miss rate of false-negative EGC by trainees was higher than that of experienced endoscopists but not significantly different (0.70% vs. 0.57%, P = 0.08 ). The false-negative EGCs missed by trainees were characterized as reddish or well-differentiated lesions, which were located in the lower or lesser curvature of the stomach ( P < 0.05 ). Conclusion. The characteristics of false-negative EGCs were similar to those of H. pylori-eradicated EGC. Procedures with shortened examination time and those performed by trainees were risk factors of missing false-negative EGCs. Trainees should pay attention to reddish or well-differentiated EGCs located in the lower or lesser curvature of the stomach.

Background Colorectal cancer is the 3rd most common cancer in the world, with about 1.2 million new cases reported annually. It is one of the three most common causes of cancer related mortality in Europe and North America. Thus, prevention and detection are critical aspects in managing colorectal cancer. Colonoscopy remains the gold standard for screening of colorectal cancer, as it is valuable not only for detection but also prevention with polyp identification. Adenoma detection rate remains a pivotal part of a good endoscopic exam. While various factors have been known to influence it, data regarding ideal screen distance for adenoma detection remains unclear. The aim of this study was to assess the rate of polyp detection and estimate the size of diminutive (

CMOS image sensor cameras are integral part of modern hand held devices. Traditionally, CMOS image sensors are affected by many types of noises which reduce the quality of image generated. These spatially and temporally varying noises alter the pixel intensities, leading to corrupted pixels, also known as “bad” pixels. The proposed method uses a simple neural network approach to detect such bad pixels on a Bayer sensor image so that it can be corrected and overall image quality can be improved. The results show that we are able to achieve a defect miss rate of less than 0.045% with the proposed method.

Background Use of artificial intelligence may increase detection of colorectal neoplasia at colonoscopy by improving lesion recognition (CADe) and reduce pathology costs by improving optical diagnosis (CADx). Methods A multicenter library of ≥ 200 000 images from 1572 polyps was used to train a combined CADe/CADx system. System testing was performed on two independent image sets (CADe: 446 with polyps, 234 without; CADx: 267) from 234 polyps, which were also evaluated by six endoscopists (three experts, three non-experts). Results CADe showed sensitivity, specificity, and accuracy of 92.9 %, 90.6 %, and 91.7 %, respectively. Experts showed significantly higher accuracy and specificity, and similar sensitivity, while non-experts + CADe showed comparable sensitivity but lower specificity and accuracy than CADe and experts. CADx showed sensitivity, specificity, and accuracy of 85.0 %, 79.4 %, and 83.6 %, respectively. Experts showed comparable performance, whereas non-experts + CADx showed comparable accuracy but lower specificity than CADx and experts. Conclusions The high accuracy shown by CADe and CADx was similar to that of experts, supporting further evaluation in a clinical setting. When using CAD, non-experts achieved a similar performance to experts, with suboptimal specificity.

Esophageal cancer is the eight most common cancer in the world and is associated with a poor prognosis. Significant efforts are necessary to improve the detection of early squamous cell cancer such that curative endoscopic therapy can be offered. Studies have shown an overall miss rate of esophageal cancer of up to 6.4%. Human factors including fatigue and lack of attention may be a contributory factor. Computer aided detection and characterisation of early squamous cell cancer can be a second reader which potentially offsets these factors. Recent studies developing artificial intelligence systems show real promise in the detection of early squamous cell cancer and predicting depth of invasion to aid in the management of patients in the same endoscopic session. This has the potential to revolutionise this area of endoscopy.

Human detection is needed for various applications such as advanced driver assistance systems and autonomous driving, security and surveillance etc. Thermal imaging is often adopted for night time because of its capability of capturing the energy emitted from human body where visible light camera fails. In this paper, we employ YOLOv3 for an accurate real-time human detection using thermal images. We modified the network parameters according to the characteristics of the human, making this method more suitable for detecting human. Subset of Korea Advanced Institute of Science and Technology (KAIST) multispectral dataset consisting of 47650 thermal images is used for training and testing of YOLOv3. During experimentation, it is observed that humans are detected at 17 millisecond which is much faster than a local machine detection. Our test result also shows improved performance of the detector with thermal image with average precision of 95.5% and miss rate of 4.7%.

In recent years, many new decomposition methods for identification of motor unit (MU) firings from high-density surface electromyograms (HDEMG) have been developed, with recent attempts focused on the use of different neural networks (NN). In this study, we evaluated the need for HDEMG signal whitening in NN-based MU identification. For this purpose, we analyzed the learning efficiency of two different types of NN, namely dense NN and long short-term memory (LSTM) NN, on the same HDEMG signals, with and without spatial whitening applied to them. All the HDEMG signals used were simulated with advanced HDEMG simulator, providing a full control of MU firing patterns and MU characteristics in our test environment. Spatial whitening of HDEMG signals significantly improved the precision of MU identification, regardless of the type of NN tested. For dense NN, precision of identified MU increased from 32.2 ± 20.2% to 93.1 ± 8.7%, whereas miss rate decreased from 48.4 ± 23.9% to 12.0 ± 13.3% when whitening of HDEMG signals was employed. For LSTM NN the precision of MU identification increased from 59.7 ± 19.7% to 99.4 ± 2.0% whereas miss rate decreased from 43.1 ± 22.3% to 12.7 ± 9.7% with whitening.

This paper proposes a two-phase solution for aerial inspecting. First phase focuses on removing images with no abnormal, modified YOLOv3 is used in this phase. Second phase focuses on target locating and identifying, modified SSD is applied in this phase. The experiment result shows the modified YOLOv3 get 2.2% higher accuracy than original design, and the miss rate of detecting images with no abnormal is only 2.6%.

Background & Goals: We observed that the number of colorectal polyps found intraoperatively was often higher than that encountered preoperatively during elective colonoscopic polypectomy. To evaluate whether more polyps can be detected when they are purposely sought than when they are routinely examined during colonoscopy. Materials and methods: Patients undergoing colonoscopy were randomized into groups A and B. Before colonoscopy was performed, endoscopists were instructed to seek polyps for group A purposely but not for group B. Polypectomy was electively completed. In groups A and B, the cases of elective polypectomy were named groups AR and BR, including groups AR-1 and BR-1, during the first colonoscopy and groups AR-2 and BR-2 during the second colonoscopy for polypectomy, respectively. The following data were calculated: the number of polyps detected (NPD) and the polyp detection rate (PDR) in all cases and the number of polyps missed (NPM) and partial polyp miss rate (PPMR) in the cases of colorectal polyps. Results: A total of 419 cases were included in group A, 421 in group B, 43 in group AR, and 35 in group BR. No significant differences in PDR were found between groups A and B and in PPMR between groups AR-1 and BR-1 (P > .05), although PPMR in group AR-1 was higher than in group AR-2 (P

Background Adenomas may be missed in up to 40% of screening colonoscopies. Although the water exchange (WE) method can improve ADR, as shown in several RCTs, it remains uncertain whether it can increase the detection of missing adenomas compared with standard air-insufflated (AI) colonoscopy. Methods Patients aged 18–80 years who underwent selective polypectomy were randomly allocated to the WE or AI group. The primary endpoint was the adenoma miss rate (AMR), defined as the number of patients with one or more additional adenomas during the polypectomy procedure divided by the total number of patients in each group. Results A total of 450 patients were enrolled, with 225 in each group. The overall AMRs were 45.8% (103/225) in the WE group and 35.6% (80/225) in the AI group (p = 0.035). More patients in the WE group had at least one missed adenoma in the proximal colon (38.2% vs 24.4%, p = 0.002). The adenoma-level miss rate was also higher in the WE group than in the AI group (35.1% vs 29.0%, p = 0.036). Subgroup analysis showed that patients in the WE group had more missed adenomas located in the proximal colon or with flat shapes. Conclusions This study confirmed that substantial adenomas were missed in patients undergoing selective polypectomy. The WE method significantly improved the detection of missed adenomas, especially those located in the proximal colon or with flat shapes. (ClnicalTrials.gov number: NCT02880748)

Background Current guidelines regarding surveillance after screening colonoscopy assume adequate bowel preparation. However, follow-up intervals after suboptimal cleansing are highly heterogeneous. We aimed to determine the diagnostic yield of early repeat colonoscopy in patients with suboptimal bowel preparation in fecal immunochemical test (FIT)-based screening colonoscopy. Methods An observational study including patients who underwent colonoscopy with suboptimal bowel preparation after positive FIT screening and then repeat colonoscopy within 1 year. Suboptimal preparation was defined as a Boston Bowel Preparation Scale (BBPS) score of 1 in any segment. Patients with a BBPS score of 0 in any segment or incomplete examination were excluded. The adenoma detection rate (ADR), advanced ADR (AADR), and colorectal cancer rate were calculated for the index and repeat colonoscopies. Results Of the 2474 patients with FIT-positive colonoscopy at our center during this period, 314 (12.7 %) had suboptimal preparation. Of the 259 (82.5 %) patients who underwent repeat colonoscopy, suboptimal cleansing persisted in 22 (9 %). On repeat colonoscopy, the ADR was 38.7 % (95 %CI 32.6 % to 44.8 %) and the AADR was 14.9 % (95 %CI 10.5 % to 19.4 %). The per-adenoma miss rate was 27.7 % (95 %CI 24.0 % to 31.6 %), and the per-advanced adenoma miss rate was 17.6 % (95 %CI 13.3 % to 22.7 %). After repeat colonoscopy, the post-polypectomy surveillance recommendation changed from 10 to 3 years in 14.7 % of the patients with previous 10-year surveillance recommendation. Conclusions Patients with suboptimal bowel preparation on FIT-positive colonoscopy present a high rate of advanced adenomas in repeat colonoscopy, with major changes in post-polypectomy surveillance recommendations.

ObjectivesOne in four cases of acute aortic syndrome are missed. This national survey examined Canadian Emergency physicians’ opinion on risk stratification, the need for a clinical decision aid to risk stratify patients, and the required sensitivity of such a tool.MethodsWe surveyed 1,556 members of the Canadian Association of Emergency Physicians. We used a modified Dillman technique with a prenotification email and up to three survey attempts using electronic mail. Physicians were asked 21 questions about demographics, importance of certain high-risk features, investigation options, threshold for investigation, and if a clinical decision tool is requiredResultsWe had a response rate of 32%. Respondents were 66% male, and 49% practicing >10 years, with 59% in an academic teaching hospital. A total of 93% reported a need for a clinical decision aid to risk stratify for acute aortic syndrome. A total of 99.6% of physicians were pragmatic accepting a non-zero miss-rate, two-thirds accepting ConclusionsOur national survey determined that emergency physicians would use a highly sensitive clinical decision aid to determine which patients are at low, medium, or high-risk for acute aortic syndrome. The majority of clinicians have a low threshold (

BackgroundEvidence has shown that deep learning computer aided detection (CADe) system achieved high overall detection accuracy for polyp detection during colonoscopy.AimThe detection performance of CADe system on non-polypoid laterally spreading tumors (LSTs) and sessile serrated adenomas/polyps (SSA/Ps), with higher risk for malignancy transformation and miss rate, has not been exclusively investigated.MethodsA previously validated deep learning CADe system for polyp detection was tested exclusively on LSTs and SSA/Ps. 1451 LST images from 184 patients were collected between July 2015 and January 2019, 82 SSA/Ps videos from 26 patients were collected between September 2018 and January 2019. The per-frame sensitivity and per-lesion sensitivity were calculated.Results(1) For LSTs image dataset, the system achieved an overall per-image sensitivity and per-lesion sensitivity of 94.07% (1365/1451) and 98.99% (197/199) respectively. The per-frame sensitivity for LST-G(H), LST-G(M), LST-NG(F), LST-NG(PD) was 93.97% (343/365), 98.72% (692/701), 85.71% (324/378) and 85.71% (6/7) respectively. The per-lesion sensitivity of each subgroup was 100.00% (71/71), 100.00% (64/64), 98.31% (58/59) and 80.00% (4/5). (2) For SSA/Ps video dataset, the system achieved an overall per-frame sensitivity and per-lesion sensitivity of 84.10% (15883/18885) and 100.00% (42/42), respectively.ConclusionsThis study demonstrated that a local-feature-prioritized automatic CADe system could detect LSTs and SSA/Ps with high sensitivity. The per-frame sensitivity for non-granular LSTs and small SSA/Ps should be further improved.

Pedestrian detection is one of the important fields of computer vision. Computer vision is getting more information from the digital images and videos and distinguishes between objects and classifies the object. Pedestrian detection is being applied in a wide range of applications such as video surveillance, automated driving, etc. This pedestrian detection has been a growing research area and then so many techniques have been used for detection. This paper focuses on various techniques applied to pedestrian detection and discusses the outcome of every technique and its accuracy and miss rate.

Due to the big variance between infrared images in day and night, the detection of infrared pedestrians in full-time has been a difficult point and has great significance both in theory and practice. Based on the deep learning method, we proposed three full-time infrared pedestrian detection models based on CSP (Center and Scale Prediction) network, namely, night model, daytime model, full-time model. Then, the performance of the three models are analyzed and compared quantitatively, and the results show that the performance of the full time model is the optimal one, the detection accuracy rate in day and night reach 83.86% and 80.22% respectively. Compared with the state-of-the-art method, the miss rate in day and night detection of the full time model in this paper is 12.23% and 4.56% lower respectively, thus the performance of the full-time model in our work is more effective.

The miss rate of flat advanced colorectal neoplasia is still unacceptably high, especially in the Western setting, notwithstanding the widespread implementation of quality improvement programs and training. It is well known that flat morphology is associated with miss rate of colorectal neoplasia, and that this subset of lesions often shows a more aggressive biological behaviour. Artificial intelligence (AI) applied to the detection of colorectal neoplasia has been shown to increase adenoma detection rate, consistently across all lesion sizes and locations in the colon. However, there is still uncertainty whether AI can reduce the miss rate of flat advanced neoplasia, mainly because all published trials report a low number of flat colorectal lesions in their training sets, and this could reduce AI accuracy for this subset of lesions. In addition, flat lesions have different morphologies with variable prevalence and potentially different accuracy in their detection. For example, the subtle appearance and rarer frequency of a non-granular laterally spreading tumor (LST) could be much harder to identify than a granular mixed LST. In this review, we present a summary of the evidence on the role of AI in the identification of colorectal flat neoplasia, with a focus on data regarding presence of LSTs in the training/validation sets of the AI systems currently available on the market.

Background and aim Right colon polyps can especially be overlooked when they are located on the backs of haustral folds. Previous studies have reported that repeated forward-view examinations in the right colon were effective in reducing adenoma miss rates. The aim of this study was to clarify the impact of retroflexion in the right colon after repeated forward-view examinations. Methods This multicenter, prospective, observational study was conducted at three institutions in Kumamoto, Japan, between February 2014 and December 2015. Subjects who were over 40 years old and scheduled for colonoscopy were recruited. For the forward view, after cecal intubation, the colonoscope was withdrawn to the hepatic flexure. The colonoscope was sequentially reinserted to the cecum and then withdrawn to the hepatic flexure. For the retroflexion view (RV), the colonoscope was reinserted to the cecum, retroflexed, and then withdrawn to the hepatic flexure. All polyps were resected at the time of detection. The primary outcome of this study was the adenoma miss rate for the repeated forward-view examinations. Results Of the 777 enrolled participants, retroflexion was successful in 730 (94.0%). The repeated forward-view withdrawal technique detected 291 adenomas, while the third withdrawal in the RV detected 53. The adenoma miss rate for the repeated forward-view withdrawal was 15.4%. No severe adverse events occurred during retroflexion. Conclusion Because adenomas located on potential blind spots can be missed when only using forward-view examinations, retroflexion in the right colon after repeated forward-view examinations might improve colonoscopy detection rates.

Current endoscopic surveillance protocols for Barrett's oesophagus have several limitations, mainly the poor cost-effectiveness and high miss rate. However, there is sufficient evidence that patients enrolled in a surveillance program have better survival chances of oesophageal cancer due to earlier tumor stages at diagnosis compared to patients with de novo diagnosed oesophagus cancer. Risk stratifications aim to identify patients at highest risk of developing adenocarcinoma of the oesophagus; most of them base on the length of the Barrett's segment and the presence of dysplasia. This review discusses prognostic factors and provides practical guidance on how to improve the efficacy and outcome in Barrett's surveillance programs.

The major bottleneck of pedestrian detection lies on the different scales of pedestrians and confusing overlapped pedestrians. To solve these problems, an improved detection method based on Faster R-CNN is proposed in this work. On the one hand, the region proposals are selected according to the usual size-distribution of pedestrians, which can moderately improve the detecting accuracy for different-scale pedestrians. On the other hand, the Soft-NMS algorithm is employed when pruning the redundant detection boxes. It can help to reduce the miss rate caused by overlapped dense pedestrians. With such strategies, the detection performance for multi-scale pedestrians and dense pedestrians can be improved. Finally, experiments to verify the effectiveness of the proposed method are conducted over the challenging Caltech Dataset. It shows that the proposed method outperforms 3.5% better MAP than original Faster R-CNN at the expense of a minor increase in detecting time.