Learning models for endoscopic ultrasonography in gastrointestinal endoscopy

Application value of an artificial intelligence-based diagnosis and recognition system in gastroscopy training for graduate students in gastroenterology: a preliminary study

OBJECTIVE

This study aimed to discuss the application value of an artificial intelligence-based diagnosis and recognition system (AIDRS) in the teaching activities for Bachelor of Medicine and Bachelor of Surgery (MBBS) in China. The learning performance of graduate students in gastroenterology during gastroscopy training with and without AIDRS was assessed.

METHODS

The study recruited 32 graduate students of the gastroenterology program at Jiangsu province hospital of Chinese medicine and Xiangyang No. 1 People's Hospital from March 2018 to March 2022 and randomly divided them into AIDRS (n = 16) and non-AIDRS (n = 16) groups. The AIDRS software was used for real-time monitoring of blind spots of gastroscopy to aid in lesion diagnosis and recognition in the AIDRS group. Only a conventional gastroscopic procedure was implemented in the non-AIDRS group. The final performance score, success rate of gastroscopy, lesion detection rate, and pain score of patients were compared between the two groups during gastroscopy. A self-prepared teaching and learning satisfaction questionnaire was administered to the two groups of students.

RESULTS

The AIDRS group had a higher final performance score (92.60 ± 2.83 vs. 89.21 ± 3.57, t = 2.98, P < 0.05), a higher success rate of gastroscopy (448/480 vs. 417/480, χ2 = 11.23, P < 0.05), and a higher detection rate of lesions (51/52 vs. 41/53, χ2 = 8.56, P < 0.05) compared with the non-AIDRS group. The pain scores of patients were lower in the AIDRS group than in the non-AIDRS group (3.40 [2.23, 3.98] vs. 4.45 [3.72, 4.75], Z = 3.04, P < 0.05). Besides, the average time for gastroscopy was lower in the AIDRS group than in the non-AIDRS group (7.15 ± 1.24 vs. 8.21 ± 1.26, t = 2.38, P = 0.02). The overall satisfaction level with the teaching program was higher in the AIDRS group (43.51 ± 2.29 vs. 40.93 ± 2.07, t = 3.33, P < 0.05).

CONCLUSION

In the context of medicine-education cooperation, AIDRS offered valuable assistance in gastroscopy training and increased the success rate of gastroscopy and teaching and learning satisfaction. AIDRS is worthy of wider-scale promotion.

Second-look upper endoscopy as the initial approach to subepithelial lesions: a reassuring and reliable strategy

BACKGROUND

Although endoscopic ultrasound (EUS) plays a critical role in the management of subepithelial lesions (SEL) of upper gastrointestinal tract many can be classified solely by a thorough upper gastrointestinal endoscopy (UGE) which can reduce the burden of additional studies.

AIMS

Analyze the impact of a stepwise approach starting with a second-look UGE before the decision of EUS in patients referred to our center with suspected SEL.

METHODS

Retrospective cohort study which included all adult patients referred to our center between 2015 and 2020 with suspected SEL.Second-look UGE evaluated the location, size, color, surface characteristics, movability and consistency of the SEL and bite-on-bite biopsies were performed. Decisions on SEL management and follow-up were collected.

RESULTS

A total of 193 SEL (190 patients) were included. At the index-UGE, stomach was the most frequent location (n = 115;59.6%). Most patients performed a second-look UGE (n = 180; 94.7%). A minority was oriented directly to EUS (n = 8;4.2%) or endoscopic resection (n = 2; 1.1%). In patients who underwent a second-look UGE, SEL were excluded in 25 (13.9%) and 21 (11.7%) did not need further work-up. The remaining patients were submitted to EUS (n = 88;48.9%), surveillance by UGE (n = 44; 24.4%) or endoscopic resection (n = 2; 1.1%).

CONCLUSION

Systematically performing a second-look UGE, in patients referred with suspected SEL, safely preclude the need for subsequent investigation in approximately one-fourth of the patients. As UGE is less invasive and more readily available, we suggest that a second-look UGE should be the initial approach in SEL management.

Curriculum for diagnostic endoscopic ultrasound training in Europe: European Society of Gastrointestinal Endoscopy (ESGE) Position Statement

The European Society of Gastrointestinal Endoscopy (ESGE) has recognized the need to formalize and enhance training in diagnostic endoscopic ultrasound (EUS). This manuscript represents the outcome of a formal Delphi process resulting in an official Position Statement of the ESGE and provides a framework to develop and maintain skills in diagnostic EUS. This curriculum is set out in terms of the prerequisites prior to training; the recommended steps of training to a defined syllabus; the quality of training; and how competence should be defined and evidenced before independent practice. 1:  Trainees should have achieved competence in upper gastrointestinal endoscopy before training in diagnostic EUS. 2:  The development of diagnostic EUS skills by methods that do not involve patients is advisable, but not mandatory, prior to commencing formal training in diagnostic EUS. 3:  A trainee's principal trainer should be performing adequate volumes of diagnostic EUSs to demonstrate maintenance of their own competence. 4:  Training centers for diagnostic EUS should offer expertise, as well as a high volume of procedures per year, to ensure an optimal level of quality for training. Under these conditions, training centers should be able to provide trainees with a sufficient wealth of experience in diagnostic EUS for at least 12 months. 5:  Trainees should engage in formal training and supplement this with a range of learning resources for diagnostic EUS, including EUS-guided fine-needle aspiration and biopsy (FNA/FNB). 6:  EUS training should follow a structured syllabus to guide the learning program. 7:  A minimum procedure volume should be offered to trainees during diagnostic EUS training to ensure that they have the opportunity to achieve competence in the technique. To evaluate competence in diagnostic EUS, trainees should have completed a minimum of 250 supervised EUS procedures: 80 for luminal tumors, 20 for subepithelial lesions, and 150 for pancreaticobiliary lesions. At least 75 EUS-FNA/FNBs should be performed, including mostly pancreaticobiliary lesions. 8:  Competence assessment in diagnostic EUS should take into consideration not only technical skills, but also cognitive and integrative skills. A reliable valid assessment tool should be used regularly during diagnostic EUS training to track the acquisition of competence and to support trainee feedback. 9:  A period of supervised practice should follow the start of independent activity. Supervision can be delivered either on site if other colleagues are already practicing EUS or by maintaining contacts with the training center and/or other EUS experts. 10:  Key performance measures including the annual number of procedures, frequency of obtaining a diagnostic sample during EUS-FNA/FNB, and adverse events should be recorded within an electronic documentation system and evaluated.

Impact and assessment of training models in interventional endoscopic ultrasound

Interventional endoscopic ultrasound (IEUS) has gained significant popularity in recent years because of its diagnostic and therapeutic capabilities. The proper training of endoscopists is critical to ensure safe and effective procedures. This review study aims to assess the impact of different training models on the competence of trainees performing IEUS. Eight studies that evaluated simulators for IEUS were identified in the medical literature. Various training models have been used, including the EASIE-R, Mumbai EUS, EUS Magic Box, EndoSim, Thai Association for Gastrointestinal Endoscopy model, and an ex vivo porcine model (HiFi SAM). The trainees underwent traditional didactic lectures, hands-on training using simulators, and direct supervision by experienced endoscopists. The effectiveness of these models has been evaluated based on objective and subjective parameters such as technical proficiency, operative time, diagnostic success, and participant feedback. As expected, the majority of skills were improved after the training sessions concluded, although the risk of bias is high in the absence of external validation. It is difficult to determine the ideal simulator among the existing ones because of the wide variation between them in terms of costs, reusability, design, fidelity of anatomical structures and feedback, and types of procedures performed. There is a need for a standardized approach for the evaluation of IEUS simulators and the ways skills are acquired by trainees, as well as a clearer definition of the key personal attributes necessary for developing a physician into a skilled endoscopist capable of performing basic and advanced therapeutic EUS interventions.

A deep learning-based system for mediastinum station localization in linear EUS (with video)

Background and Objectives

EUS is a crucial diagnostic and therapeutic method for many anatomical regions, especially in the evaluation of mediastinal diseases and related pathologies. Rapidly finding the standard stations is the key to achieving efficient and complete mediastinal EUS imaging. However, it requires substantial technical skills and extensive knowledge of mediastinal anatomy. We constructed a system, named EUS-MPS (EUS-mediastinal position system), for real-time mediastinal EUS station recognition.

Methods

The standard scanning of mediastinum EUS was divided into 7 stations. There were 33 010 images in mediastinum EUS examination collected to construct a station classification model. Then, we used 151 videos clips for video validation and used 1212 EUS images from 2 other hospitals for external validation. An independent data set containing 230 EUS images was applied for the man-machine contest. We conducted a crossover study to evaluate the effectiveness of this system in reducing the difficulty of mediastinal ultrasound image interpretation.

Results

For station classification, the model achieved an accuracy of 90.49% in image validation and 83.80% in video validation. At external validation, the models achieved 89.85% accuracy. In the man-machine contest, the model achieved an accuracy of 84.78%, which was comparable to that of expert (83.91%). The accuracy of the trainees' station recognition was significantly improved in the crossover study, with an increase of 13.26% (95% confidence interval, 11.04%-15.48%; P < 0.05).

Conclusions

This deep learning-based system shows great performance in mediastinum station localization, having the potential to play an important role in shortening the learning curve and establishing standard mediastinal scanning in the future.

The use of simulation in medical ultrasound: Current perspectives on applications and practical implementation (WFUMB state-of-the-art paper)

Simulation has been shown to improve clinical learning outcomes, speed up the learning process, and improve trainee confidence, while taking the pressure off initial face-to-face patient clinical areas. The second part of The World Federation for Ultrasound in Medicine and Biology state-of-the-art paper on the use of simulators provides a general approach on the practical implementation. The importance of needs assessment before developing a simulation-based training program is outlined. We describe the current practical implementation and critically analyze how simulators can be integrated into complex task scenarios to train small or large groups. A wide range of simulation equipment is available especially for those seeking interventional ultrasound training, ranging from animal tissue models, simple synthetic phantoms, to sophisticated high-fidelity simulation platforms using virtual reality. Virtual reality simulators provide feedback and thereby allow trainees to not only to practice their motor skills and hand eye coordination but also to interact with the simulator. Future developments will integrate more elements of automated assessment and artificial intelligence, thereby enabling enhanced realistic training experience and improving skill transfer into clinical practice.

Training in advanced bilio-pancreatic endoscopy

As the clinical applications of endoscopic retrograde cholangiopancreatography (ERCP) and endoscopic ultrasound (EUS) take more and more root in everyday practice and these endoscopic procedures become increasingly more complex and interventional with a higher chance for complications, the demand for expert bilio-pancreatic endoscopists continues to increase. In response to this growing need, specific postgraduate fellowships have been developed, however the standardization of training programs and the assessment of the achieved competence still remains an open debate. ERCP and EUS competency often requires training beyond the scope of a standard GI fellowship program, which lasts at least 1-2 years, and there are some differences in the way to face the issues of advanced bilio-pancreatic endoscopy training between Europe, America and the Asian regions. Today there is no role for the self-teaching of ERCP and EUS through trial and error without supervision and, in the near future, it is necessary to critically revise current training guidelines, to establish a standardized curriculum for advanced bilio-pancreatic endoscopists and to implement universally validated skill assessment tools, able to ensure constant and targeted feedback to trainees.

A deep learning-based system for bile duct annotation and station recognition in linear endoscopic ultrasound

BACKGROUND

Detailed evaluation of bile duct (BD) is main focus during endoscopic ultrasound (EUS). The aim of this study was to develop a system for EUS BD scanning augmentation.

METHODS

The scanning was divided into 4 stations. We developed a station classification model and a BD segmentation model with 10681 images and 2529 images, respectively. 1704 images and 667 images were applied to classification and segmentation internal validation. For classification and segmentation video validation, 264 and 517 videos clips were used. For man-machine contest, an independent data set contained 120 images was applied. 799 images from other two hospitals were used for external validation. A crossover study was conducted to evaluate the system effect on reducing difficulty in ultrasound images interpretation.

FINDINGS

For classification, the model achieved an accuracy of 93.3% in image set and 90.1% in video set. For segmentation, the model had a dice of 0.77 in image set, sensitivity of 89.48% and specificity of 82.3% in video set. For external validation, the model achieved 82.6% accuracy in classification. In man-machine contest, the models achieved 88.3% accuracy in classification and 0.72 dice in BD segmentation, which is comparable to that of expert. In the crossover study, trainees' accuracy improved from 60.8% to 76.3% (P < 0.01, 95% C.I. 20.9-27.2).

INTERPRETATION

We developed a deep learning-based augmentation system for EUS BD scanning augmentation.

FUNDING

Hubei Provincial Clinical Research Center for Digestive Disease Minimally Invasive Incision, Hubei Province Major Science and Technology Innovation Project, National Natural Science Foundation of China.

EUS-guided Gastroenterostomy: Assessing the Competence of the Endoscopist

Assessing competence is of utmost importance for complex EUS procedures like EUS-guided gastroenterostomy (EUS-GE) that are associated with higher risk of procedural complications. It is a matter of intense debate on what is the training needed to carry out these interventions, how to judge for competence, and who can perform these newer interventions. The procedure has several variations including the methods to identify the jejunal loops on EUS like direct puncture, balloon-assisted, and hybrid rendezvous with ultrathin endoscope. Therefore, there is lack of clarity on the role and ways of assessing the learning curve for EUS-GE. In this news and views, we are discussing two studies that assessed the learning curve for EUS-GE.

Validation of a novel swine model for training in EUS-FNA (with videos)

Background and Objectives:

EUS-FNA is applied widely in clinical practice, but there remains a lack of authentic training models. The present study aimed to develop a novel swine training model and to perform a preliminary assessment of its feasibility and efficacy.

Materials and Methods:

To create an internal lesion-like target, empty shells of iodine-125 seeds were implanted into the caudate lobe of the liver in Bama minipigs. A training program involving 10 trainees was subsequently carried out, in which a total of 60 needlings were performed, composed of 6 for each trainee obtained during two training steps. Comparisons of procedure-related variables were conducted between the two. Trainees completed a questionnaire to assess their basic endoscopic experiences and reasonability of the model.

Results:

A target region of 2.0 cm × 2.0 cm in diameter was successfully established on the caudate lobe in all implanted pigs. In the training program, the average procedure time decreased from the first to the second step and the average time for the total 30 needlings' obtainment was significantly shorter for the second training step (23.8 ± 4.5 min vs. 40.9 ± 9.0 min, P < 0.001). For the second step, there was also a significant improvement in total success rate (86.7% vs. 56.7%, P = 0.020) and accuracy rate (76.7% vs. 43.3%, P = 0.017). All trainees scored the effectiveness of the model highly and all reported improved confidence after the training.

Conclusion:

This novel swine training model could authentically mimic clinical EUS-FNA, providing an effective in vivo practice tool for novices before clinical practice.

A new minimally invasive porcine model for the study of intrahepatic bile duct dilatation

BACKGROUND

Endoscopic ultrasound (EUS) procedures are becoming more frequent nowadays and novel techniques are on the rise. These procedures require high technical experience and complex endoscopic skills. The goal of this study was to develop a new minimally invasive animal model of bile duct dilatation in the pig, in order to offer a new tool for endoscopic and surgical therapy training and to test new therapeutic strategies.

METHODS

Twenty-five female pigs underwent laparoscopic surgery in order to perform a common hepatic duct ligation. A pre- and postoperative biochemical analyses were performed: glucose, albumin, total bilirubin (TBil), gamma glutamyl transferase (GGT), alkaline phosphatase, and alanine aminotransferase were measured. Surgical time and intra- and postoperative complications were registered. Five to six days after surgery, an EUS was performed to measure intrahepatic duct size (mm). Distance from the bile duct to the EUS transductor was also recorded (mm). T-student for quantitative variables was applied. Statistical significance was defined as p value ≤ 0.05.

RESULTS

The mean surgical time was 29.5 ± 14.9 min. In five pigs (20%), some mild intraoperative problems occurred. A severe postoperative complication occurred in one animal (4%). No postoperative mortality was registered. Postoperative serum analyses showed an increase in total bilirubin (p = 0.005) and gamma glutamyl transferase levels (p = 0.001). Postoperative EUS showed dilatation of the intrahepatic bile duct in 76% of pigs, with a mean diameter of 9.6 ± 3.6 mm (distance from the gastric wall of 17.0 ± 6.4 mm).

CONCLUSION

The surgical procedure described here is a safe technique to induce dilatation of the intrahepatic bile ducts in the pig, with a minimally invasive approach and a high efficacy rate. This animal model might be useful for EUS techniques training and for evaluating new therapeutic approaches.

Training in bariatric and metabolic endoscopy

The limited penetration of bariatric surgery and the scarce outcome of pharmacological therapies created a favorable space for primary bariatric endoscopic techniques. Furthermore, bariatric endoscopy is largely used to diagnose and treat surgical complications and weight regain after bariatric surgery. The increasingly essential role of endoscopy in the management of obese patients results in the need for trained professionals. Training methods are evolving, and the apprenticeship method is giving way to the simulation-based method. Existing simulation platforms include mechanical simulators, ex vivo and in vivo models, and virtual reality simulators. This review analyzes current training methods for bariatric endoscopy and available training programs with dedicated bariatric core curricula, giving a glimpse of future perspectives.

EUS anatomy of the pancreatobiliary system in a swine model: The WISE experience

Background and Objectives

EUS training is recognized to have a substantial learning curve. To date, few dedicated training programs for EUS have been described. The swine model has been highlighted as a realistic tool to enhance EUS training. Studies extensively describing EUS swine anatomy are lacking in the current literature. The article aims to describe both radial and linear EUS pancreatobiliary swine anatomy.

Materials and Methods

Four live pigs were endoscoped under general anesthesia using both radial and linear array echoendoscopes. Relevant images and videos were recorded.

Results

It was possible to effectively image aorta, crus of the diaphragm, celiac trunk, superior mesenteric artery, pancreas, common bile duct, gallbladder, portal vein, kidneys, spleen, and hepatic hilum. Images were comparable to human EUS findings, with some remarkable differences. The pancreas was relatively larger in swine and in contrast to humans has three segments (duodenal, splenic, and connecting lobe).

Conclusions

The swine model was a highly realistic teaching model for linear and radial pancreatobiliary EUS and a useful tool for training in the setting of in vivo hands-on sessions.

Use of simulator for EUS training in the diagnosis of pancreatobiliary diseases

Background and Objectives:

EUS has been widely used in the diagnosis of pancreatobiliary diseases. However, improvements in the conventional training pattern of EUS are needed urgently. In this study, we compared the results achieved after use of clinical practice training patterns combined with or without simulator training and evaluated the effectiveness of simulator use in EUS training.

Subjects and Methods:

The trainees were randomly divided into two groups: the experimental group was trained with both clinical practice and simulator training system and the control group was only trained through clinical practice. After 1 month of training, trainees of both groups were tested with an established technical evaluation procedure that aimed to assess trainees’ ability to examine the normal anatomical structure. Then, trainees in the experimental group completed a questionnaire.

Results:

The mean test score of the experimental group (64.53 ± 4.91) was significantly greater than that of the control group (60.09 ± 5.49; P = 0.028). Moreover, the individual test score of trainees in the experimental group was positively correlated with the frequency of simulator use (P = 0.242).

Conclusion:

Simulator training can promote trainees’ ability to evaluate the normal anatomical structure, and thus, can improve the efficiency of the EUS training program.

Progress in Endoscopic Ultrasonography: Training in Therapeutic or Interventional Endoscopic Ultrasonography

Therapeutic endoscopic ultrasound is a rapidly expanding field, requiring training beyond the 3-year gastroenterology fellowship and at least an additional year in a structured advanced endoscopy fellowship program with mentorship from an expert at a sufficiently high-volume center. Simulation models can provide initial instruction on technique and increase familiarity with the rapidly changing devices. Trainees must also be given a graduated level of independence to perform each step and, eventually, be able to practice on a variety of endoscopic targets. With structured competency markers, trainees can learn methods to maximize success and minimize the risk of complications.

Training in Endoscopy: Endoscopic Ultrasound

Endoscopic ultrasound (EUS) has been recently established as an indispensable modality for the diagnosis and management of pancreatobiliary and gastrointestinal (GI) disorders. EUS proficiency requires both cognitive and technical abilities, including an understanding of the appropriate indications, the performance of appropriate evaluations before and after the procedure, and the management of procedure-related complications. An increasing demand for skills to handle a growing range of interventional EUS procedures and a continual shortage of EUS training programs are two major obstacles for EUS training. Acquiring the skills necessary to comprehend and conduct EUS often requires training beyond the scope of a standard GI fellowship program. In addition to traditional formal EUS training and preceptorships, regular short-term intensive EUS training programs that provide training at various levels may help EUS practitioners improve and maintain EUS-related knowledges and skills. Theoretical knowledge can be acquired from lectures, textbooks, atlases, slides, videotapes, digital video discs, interactive compact discs, and websites. Informal EUS training is generally based on 1- or 2-day intensive seminars, including didactic lectures, skills demonstrated by expert practitioners through live video-streaming of procedures, and hands-on learning using animal or phantom models.

Training in endoscopic ultrasonography: An Asian perspective

Training of endoscopic ultrasonography (EUS) in Asia faces two major challenges: (i) the ever-increasing demand for skills to handle a growing range of interventional EUS procedures; and (ii) a continual shortage of EUS training programs. As the therapeutic application of EUS continues to expand, the need to train more new endosonographers and upgrade skills of existing ones has become more critical than ever before. A formal fellowship to acquire EUS knowledge and skills in an advanced endoscopy center has always been perceived as the best way of training novices, but such opportunities remain limited in most Asian countries. To keep up with the pace of development in EUS, more short-term EUS programs have been conducted across Asia in recent years. Such programs are generally intensive and may combine didactic lectures, live-case demonstrations, and hands-on training on phantoms, or live animal models for teaching. Although not as rigorous as conventional full-time EUS fellowships, such short-term programs are not necessarily inferior in quality. With courses offered from basic to advanced levels, and at regular intervals, busy practising endoscopists have the flexibility to attend the course that best matches their individual levels of experience, learn at their own pace and acquire EUS knowledge and skills over as many courses as desired. This open-ended progressive learning model is more agile than established fixed-term learning models and is expected to adapt better to future needs.

Learning curve for endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) of pancreatic lesions in a novel ex-vivo simulation model

Background: Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) is essential in the management of digestive cancers. However, teaching and learning this technique remain challenging due to the lack of cost-effective models. Material and methods: This was a prospective experimental study using a complete porcine upper gastrointestinal ex-vivo organ package, placed in an Erlangen Active Simulator for Interventional Endoscopy (EASIE-R), and prepared with one cyst and two solid masses (2 cm). Five fellows inexperienced in EUS-FNA were enrolled, performing 10 procedures on each lesion, alternatively. The total time, number of attempts for success, of needle view losses, and of scope handling were recorded, associated with an independent skills rating by procedure. We compared the first 15 procedures with the last 15 for each fellow. Results: The fellows successfully performed all procedures in 2 to 40 minutes, requiring 1 to 6 attempts. All (5/5) improved their total time taken (P < 0.001), number of times when the EUS view of the needle was lost (P < 0.05), scope handling (P < 0.005), and skills rating (P < 0.001), whereas 4/5 (80 %) improved their number of attempts. The overall evaluation showed a significant decrease (P < 0.001) in the total time taken (11.2 ± 7.8 vs 4.3 ± 2.2 minutes), number of attempts (2.6 ± 1.2 vs 1.2 ± 0.7), number of times when the EUS view of the needle was lost (2.3 ± 2 vs 0.5 ± 0.7), and need for scope handling (1.1 ± 1.7 vs 0.1 ± 0.2). We also observed an improvement in skills rating (5 ± 1.9 vs. 7.7 ± 1.1). Conclusion: This newly designed ex-vivo model seems to be an effective way to improve the initial learning of EUS-FNA, by performing 30 procedures.