Keywords

Introduction

The accurate identification and consistent endoscopic removal of intraepithelial neoplasia within the luminal gastrointestinal (GI) tract are a crucial responsibility of the modern endoscopist. When considered alone, colorectal cancer (CRC) represents the fourth most common neoplasm and is the third leading-cause of cancer-related mortality in the United States (USA), according to the National Cancer Institute Surveillance, Epidemiology, and End Results Program (SEER) database [1]. The predictable development of most CRCs from adenomatous or serrated precursor lesions provides the conceptual basis for modern endoscopy-based screening guidelines [2] and likely explains the consistent evidence that CRC incidence and mortality decline following the adoption of such screening [3,4,5]. Screening on its own does not prevent or cure cancer. Rather, the linchpin of colorectal cancer prevention through colonoscopy is the effective and complete removal of precursor lesions via endoscopic resection [6, 7]. Despite lesser prevalence among Western populations, successful endoscopic management of superficial neoplastic lesions within the esophagus [8], stomach [9], and small intestine [10, 11] is of equivalent value when it comes to providing high-quality care for individual patients.

Broadly, techniques employed in endoscopic resection can be subdivided into traditional polypectomy , endoscopic mucosal resection (EMR) , and endoscopic submucosal dissection (ESD) . Regardless of the technique chosen by an endoscopist, the complete and durable resection of neoplastic or precursor tissue is critical to a desirable outcome. Failure to achieve complete polypectomy or curative resection can have dire consequences, such as the development of cancer between surveillance intervals [12,13,14].

Despite these stakes, and the near-ubiquitous practice of colonoscopy among gastroenterologists [15], there is often wide quality variation in the effectiveness of polypectomy among practitioners. In a landmark study done at Dartmouth University and the nearby Veterans’ Administration Hospital, Pohl et al. [12] demonstrated that marginal residual neoplastic tissue was routinely left behind (10.1% of the time) during polypectomy, despite the appearance of a macroscopically “complete” excision. More significant, however, was the incredible observed heterogeneity among individual proceduralists for this outcome. In this unblinded study, rates of incomplete resection ranged from 6.5% to nearly 23%, dependent solely upon the endoscopist performing the polypectomy.

This staggering level of variability in colonoscopy quality is not confined to polypectomy alone, but rather includes other metrics of endoscopist performance as well [7]. In an era where the successful removal of superficial mucosal neoplasia increasingly falls to the endoscopist rather than the surgeon, assuring the delivery of consistent, high-quality mucosal resection is of paramount importance to society. These complex procedures require significant training , valuable experience, and meticulous attention to detail in clinical practice. In this chapter, we will review paradigms for institutional training in mucosal resection, the challenges associated with measuring competency, and the importance of feedback and discuss some considerations inherent to training “nontraditional” endoscopic students.

Experiential Learning in Endoscopic Resection

The goal of training in endoscopic resection, be it polypectomy, EMR, or ESD, is the acquisition of a new skill by the learner, along with subsequent refinement, to the level of consistently demonstrated competence. The process of skill acquisition through experiential learning has been extensively described, modeled, and observed to occur in definable phases [16,17,18,19]. The stages of competence model [18] , introduced by Noel Burch in the 1970s, subdivides learners into four sequential categories: unconscious incompetence , conscious incompetence , conscious competence , and unconscious competence . In its purest form, this model implies that all learners arrive as novices within the first category and proceed sequentially to the fourth category in linear fashion. At the unconscious incompetence stage, a learner does not understand how to perform a given task but also does not recognize that he or she has a skill deficit. With experience and instruction, the learner first recognizes the deficit in skill (conscious incompetence) and then becomes able to perform the task with effort and concentration (conscious competence). Finally, with practice dependent upon task difficulty and aptitude, the learner refines and automates the skill so that it no longer requires conscious cognitive involvement to perform (unconscious competence).

This progressive model of learning can be applied to a myriad of activities, including the practice of endoscopy. This model does have several limitations: learners may not universally arrive at phase one, not all subjects progress through each of the four phases, and skill regression is entirely possible without maintenance [20]. For example, it is possible for technical skills to be acquired by observation and persistent repetition alone, without conscious understanding by the subject of the precise movements needed to achieve success.

In addition, the achievement of competence in endoscopic resection typically involves ensuring gains in a number of parallel domains, which are supportive of the psychomotor skills required to actually perform the task. Complementary nontechnical domains include both cognitive and less tangible integrative skills [21]. In the context of mucosal resection, cognitive skills include a thorough understanding of the steps involved in EMR or ESD, indications for each procedure, contraindications, selection of equipment, as well as the ability to readily recognize both success and adverse events . Integrative skills are less discrete and prevail at the intersection of cognitive skills, physical skills, and communication. Such ability often requires leadership, team interaction, judgment, adaptability, and situational awareness. It can be successfully argued that the importance of these nontechnical skills is paramount, as a failure to develop them alongside psychomotor prowess can result in significant adverse events regardless of technical expertise [22].

The Apprenticeship Model

Historically, training in EMR in Western countries has followed a traditional mentor-apprentice model, commonplace in both surgical and endoscopic pedagogy. The stage at which trainees are typically first exposed to mucosal resection is variable and may occur either during the course of a structured 3-year gastroenterology fellowship program or as a component of additional, dedicated advanced endoscopy training (AET) [23]. The decision to include learners of different educational levels is center specific and is typically determined by local faculty with expertise in mucosal resection. Frequently, this decision may be the result of long-standing, top-down curriculum design rather than the needs or interests of the individual learner. In addition, exposure of trainees to appropriate cases and overall volume can be unpredictable, often reflective of clinical assignments and the particulars of cases that are scheduled on a given day [24].

Traditional training under the apprenticeship model generally incorporates the learner in a progression of graduated responsibility, beginning with observation of endoscopic resection performed by an expert mentor. Over the course of training, often defined by the term of enrollment, learners are expected to progress to the level of independent practice under direct supervision by their proctor. Put simply, the paradigm of most EMR training programs is “see one, do one, teach one” where one is replaced by an integer determined mostly by the instructor. Importantly, the acquisition of competence under this model is frequently defined in nebulous fashion: gestalt on the part of the instructor or the completion of a time-based curriculum. Rarely, if ever, is competence in EMR currently measured among trainees with the use of a validated objectives, observational tools, or other reproducible outcomes [21].

The most frequent objective measure used in order to obtain professional certification in most endoscopic procedures is often caseload. Under the traditional model, this can be defined as a “critical mass ” or minimal threshold of procedural volume above which competence can be reliably expected. Such a threshold is frequently variable, and therefore potentially dubious, even for basic procedures such as routine colonoscopy. As an example, the American Board of Surgery recommends that trainees perform a total of 50 colonoscopies during a surgical residency program to obtain basic competence [25]. The UK Joint Advisory Group recommends that learners perform of at least 200 cases [26]. These numbers are widely discordant, and, if both standards are acceptable for certification, such inconsistency raises questions about the concept of using an arbitrary minimum threshold as the sole measure of success [26]. Even so, there exists no well-established algorithm or even an identifiable minimal case volume necessary to confirm competence in EMR under a traditional instructional approach [27].

Western training in ESD is even more problematic than EMR when considered under a traditional model. In Japan, where the practice of ESD is well-established, most training programs fall into the category of a traditional mentor-apprenticeship . These training opportunities are physically located at institutions which possess a discernible history of performing submucosal dissection [28]. In Europe and the United States, no such mature infrastructure of programmatic experience exists, and there are comparatively few endoscopists with sufficient ESD experience to proctor any form of widespread training [23]. Those experts who do possess experience have frequently received their instruction via markedly nontraditional means, usually after extensive experience as a therapeutic endoscopist. This model is disparate from other forms of Western endoscopy training, which are typically incorporated as a part of graduate medical education or structured AET [29]. Further compounding the problem of dissemination, there is a significant paucity of gastric dysplasia and superficial carcinoma in Western populations, which has long been considered by Japanese experts to be the preferred target upon which to begin one’s training in ESD [30]. This combined scarcity, of both mentors and disease-state, serves to dramatically inhibit any training model, thereby making experiential learning extremely difficult.

Competency-Based Education

Over the past two decades, there has been a paradigm shift throughout graduate medical education away from traditional structured or time-driven curricula toward competency-based education (CBE). In contrast to traditional approaches, CBE may be distinguished by several features : (1) a focus on outcomes, (2) emphasis on ability, (3) reduced emphasis on time-based learning, and (4) the promotion of learner-centeredness [31]. This model promotes the incorporation of predetermined competencies and the measurement of definable outcomes, rather than assuming that learning has occurred due to time spent in training. Defined competencies should be targeted to reflect the needs of stakeholders (in the case of endoscopic resection, patient outcomes such as residual neoplasia would be logical), and feedback during learning is provided through formal means. This feedback takes the form of assessments and per-competency measures to highlight deficiencies and improve learner performance [31]. When well designed, CBE curricula support the development of learning, target the strengths and weaknesses of individual trainees, and provide both attainable and applicable goals. Such a design could offer significant advantages over traditional training methods for many procedural skills, including mucosal resection.

Even though a CBE model is based upon sound educational principles, there are numerous barriers to the adoption of this format with respect to training in EMR and ESD. First and foremost, there must be clearly defined and validated criteria for identifying competency among learners (“outcomes”). At the present time, no established metrics for measuring competence in mucosal resection training exist. Despite the conspicuous lack of concrete procedure-based learning objectives, professional societies such as the American Society for Gastrointestinal Endoscopy (ASGE) do provide suggestions for the development of a competency-based mucosal resection training program [27].

The “Core curriculum in EMR and ablative techniques ,” published in 2012 by the ASGE Training Committee [27], outlines potential goals of training in EMR and sets basic facilities and faculty requirements for an aspiring program. The ASGE also defines expected prerequisites for an incoming trainee (i.e., completion of a 2–3-year GI fellowship program with basic competency in upper and lower diagnostic endoscopy, submucosal injection, and management of complications) and describes the training process. They also present an overall strategy for assessment—as the ASGE adopts the core competency model employed in general by the ACGME for graduate medical education. The ASGE suggests that programs evaluate learners within the established GME competencies of patient care, medical knowledge, interpersonal and communication skills, professionalism, practice-based learning, and systems-based practice [32]. For each competency, the ASGE presents expectations and goals for trainees which, essentially, focus on the important cognitive and integrative funds necessary to successfully perform EMR.

Despite the usefulness of the proposed core curriculum, what is conspicuously lacking are directly observable benchmarks for the technical portion of learning and performing EMR. No formal criteria are presented or discussed in order to delineate a trainee’s success in terms of the critical steps of mucosal resection. Rather, instructors are advised to “determine… (the number of procedures necessary)… based upon the trainee’s individual performance,” to expect competence, and that “objective criteria for competence should be developed and met [27].” In this way, the ASGE Core Curriculum falls short of transitioning entirely away from the traditional apprenticeship model, but does offer an inroad toward the subsequent goal of a CBE curriculum.

Directly Observable Skills as Potential Metrics of Competence in Mucosal Resection

The ASGE has defined competence in endoscopy as “… the minimum level of skill, knowledge, and/or expertise, derived through training and experience, required to safely and proficiently perform a task or procedure.” [32] Mucosal resection is considered an advanced technique [23], and, as previously discussed, the metrics by which this specialized procedure should be measured are at best unclear. Although raw procedure numbers have been used in the past, reliance on this metric may be subject to inconsistency. Other metrics of outcome have been proposed including the rate of en bloc resection, the rate of residual or recurrent neoplasia, total procedure time, and adverse events. Unfortunately, none of these metrics have been validated independently as measures of competency in EMR [33]. Also, there are nuances to each of these potential competency measures . For example, the rate of en bloc resection for EMR generally applies only to smaller lesions which do not require piecemeal resection. Residual neoplasia, an important EMR outcome , may be predicted by lesional characteristics that are separate from an endoscopist’s capability at performing the procedure [34]. In the following section, we will discuss a few of the various procedural steps involved in mucosal resection and review where and how the inclusion of direct observational assessments could be employed in progressing toward competency-based learning for the technical component of mucosal resection training.

Lesion Identification and Characterization

Regardless of the technique employed, performing high-quality resection of intraepithelial neoplasia begins with a comprehensive and accurate assessment of the lesion to be removed. A variety of inspection and classification schema have been extensively described for lesion morphology, topography, and pit or vascular pattern(s). Perhaps the most straightforward and widely adopted means of describing lesion morphology is the “Paris classification of superficial neoplastic lesions [35].” In principle, first-stage accurate morphological diagnosis can assist a learner in determining which lesions are most suitable for traditional polypectomy vs. those that require EMR or ESD vs. those that should be referred for laparoscopic surgery. For example, Paris Classification 0-Ip lesions are pedunculated and may be removed without a lift or advanced intervention. Paris 0-IIa + IIc lesions are both superficially elevated and depressed (Fig. 11.1), and such lesions might require more complex intervention and harbor an increased risk of neoplasia associated with submucosal invasion (SMI). Despite its widespread use, there are no reports available that succinctly define the learning curve, inter-rater reliability, or intra-rater reliability of the Paris classification for lesion morphology, particularly among trainees. If such data were available, and there were consensus regarding acceptable proficiency levels, morphologic characterization of lesions would represent a useful and directly observable skill that could be included in CBE for EMR.

Fig. 11.1
figure 1

Two large Paris 0-IIa + IIc colonic lesions were found in the same person. High-definition white-light (a) and NBI (b) images are shown of the lesion in the cecum. White-light images of the second large polyp (c) with a mucous cap that remained despite irrigation were found in the ascending colon. Both polyps were removed by underwater EMR and found to be serrated sessile polyps without dysplasia

Full size image

The acquisition of other skills in lesion assessment has been evaluated, though these are frequently more complicated. Togashi et al. [36] have described the learning curve and accuracy of optical diagnosis of neoplastic and nonneoplastic polyps using the Kudo pit pattern [37] . In this study, sequential observation of lesions under chromo- and magnification endoscopy demonstrated an improvement to over 90% sensitivity for correctly identifying neoplastic lesions after at least 200 sequential assessments. While this does suggest a threshold for competence, assuming the benchmark of 90% is appropriate, implementation outside of Japan may be challenging. Many endoscopic platforms which are routinely available in the West do not offer a true optical magnification capability . Use of chromoendoscopy for optical diagnosis of the mucosal pit pattern using a high-definition endoscope that lacks optical magnification may introduce issues with fine-detail image resolution and therefore might not be sufficient to perform this task consistently [38].

More applicable to Western training programs, the learning curve for assessment of lesions using narrow-band imaging (NBI) and the meshed capillary pattern [38, 39] has been described [40] (Fig. 11.2). In this study [40], four experienced endoscopists with no NBI experience underwent a 4-hour training course in both NBI principles and the capillary pattern classification system . After as few as 30 cases, the subjects were able to distinguish between lesions appropriate for endoscopic resection (adenomas and superficial neoplasms) and those which required only biopsy (hyperplastic lesions and those with overt deep cancer). The subjects were highly discriminatory, with over 95% diagnostic accuracy . When combining the advantages of a short learning curve, excellent threshold for accurate diagnosis (>95%), and the widespread availability of NBI-capable equipment, assessment of vascular capillary patterns represents an ideal target as a directly observable skill and potential metric that can be included in a mucosal resection training program .

Fig. 11.2
figure 2

A 1.5-cm sessile polyp (Paris 0-Is) was found in the splenic flexure and high-definition white-light images (a) do not show the surface pattern well. NBI (b) better demonstrated the meshed capillary pattern , which in this case showed disordered vessels with some areas of avascularity. While there was concern for early invasive cancer (Sano type IIIB), given the relatively small size and lack of other surface features that might suggest deeper invasion (tenting, bridging folds, etc.), en bloc underwater EMR was performed (c) with complete resection and no bleeding or perforation (d). Pathology showed a well-differentiated adenocarcinoma with invasion into the superficial submucosa but negative margins (>1-mm negative deep margin). In this case, the risk of lymph node metastasis is likely around 5%, and the patient was directed to meet a colorectal surgeon to discuss the risks and benefits of further laparoscopic surgical resection, with the alternative being close endoscopic surveillance and yearly CT scans

Full size image

Mucosal Resection Technique

Following lesion identification and characterization, the endoscopist then prepares to tackle the fundamental act of removing the superficial neoplastic lesion. As the practice of mucosal resection has matured, understanding of the technical aspects of both EMR and ESD has become increasingly developed [41, 42]. For each technique that could be employed, there are fundamental steps associated with performance of an optimal resection. These may vary based upon lesion location, characteristics, technique, or equipment utilized. An extensive discussion of how to perform these procedures is beyond the scope of this chapter and is discussed elsewhere in the textbook.

In the case of traditional mucosal resection, simply the process of creating a submucosal lift has numerous technical considerations: choosing the proper submucosal injectate solution , injecting the lesion appropriately to facilitate a lift without obscuring visualization, assessing for an adequate or inadequate lift, and interpreting if the presence of poor lifting represents tumor invasion or benign submucosal fibrosis [43]. Snare resection is no less complex with regard to optimal technique and has several facets: lesion and scope orientation to facilitate technical success, incorporating a normal mucosal border within the snare (i.e., a 2–3-mm rim of normal tissue), maintaining a submucosal plane in piecemeal resection using the snare edge , and the application of cautery by using a modern electrosurgical generator. Each and any of these technical steps could be observed and evaluated as a measure of competence, were there a consistent methodology available to do so. Additionally, there are a number of other higher-level decisions made by endoscopists during EMR: attempting en bloc resection whenever possible, removing large or dysplastic appearing nodules in a single piece (given an increased risk of focal malignancy), and minimizing the number of overall fragments when en bloc resection is not possible [43, 44].

Learning Curves and Direct Observation of Polypectomy Skills (DOPyS)

Although no societal recommendation for a threshold number of EMR cases in training has been established, there does exist limited data regarding skill acquisition and the learning curves for EMR among practicing academic interventional endoscopists. In a retrospective series by Bhurwal et al. [34], a total of 578 consecutive colonic EMR procedures, performed by 3 endoscopists over a 9-year period, were tabulated and analyzed. Three relatively narrow outcomes were included: residual neoplasia upon interval surveillance, immediate assessment that an EMR was incomplete, and the occurrence of immediate bleeding as an adverse event. For each of three physicians, the occurrence of residual neoplasia (grossly and by surveillance biopsies) fell below 20% and plateaued by procedure number 100. Immediate bleeding was generally rare throughout the study and was acceptably below 5% by case 100 for all of the endoscopists. Although there were several limitations inherent in this series that may have prolonged the learning curve (referral bias to a tertiary care center, self-teaching environment, retrospective series, etc.), the observed number of cases required to establish a plateau of residual neoplasia was higher than expected. To date, this study represents the only published learning curve data with respect to EMR of large laterally spreading lesions in the colon.

Published data evaluating the rate of acquiring skills necessary to perform ESD in Japan exist and are more robust. In 2005, Gotoda et al. [45] reported that early proficiency in gastric ESD could be seen after 30 cases during intensive training. In 2012, Yamamoto et al. [28] provided observational evidence that by 40 gastric ESDs, trainees may have sufficient skill to reliably remove superficial mucosal lesions, without ulceration, that were less than 2 cm in size. With continued instruction and experience through 80 cases, trainees routinely demonstrated outcomes that approximated their expert instructors. In 2010, Hotta et al. [46] described the first learning curve for colonic ESD in a single endoscopist from Saku Central Hospital in Japan. They demonstrated that 40 ESD procedures were required to avoid an unacceptable rate of perforation (rate dropped from 12.5% to 5%) and that a total of 80 procedures were required to establish an acceptable rate of en bloc, R0 resection (rate increased from 85% to 92.5%). Translation of gastric ESD experience to colonic lesions appears more straightforward. In 2011, Sakamoto et al. [47] reported that trainees with experience in gastric ESD could successfully perform supervised colorectal ESD safely after approximately 30 procedures.

The Western experience with ESD skill acquisition has been more challenging, as the methods used thus far to disseminate ESD skill in Europe and the United States have been fractured and inconsistent. While reports of learning curves for Western endoscopists exist [29, 48, 49], these experiences draw upon training that has been universally disparate. Each published experience is similar only in that the individual pathways to competence have been distinctly different than others. As such, these data likely have only limited translational value to typical Western graduate (fellowship level) or postgraduate endoscopy training concerning ESD.

While learning curves may provide insight into the degree of experience necessary to attain competency in mucosal resection, they may not be reliable measures at the level of an individual trainee. This problem, and how it pertains to routine colonoscopy with polypectomy, has been widely recognized [21]. Out of concerns for inadequate training in routine polypectomy, an expert working group within the United Kingdom deconstructed the process of polypectomy into a 33-item checklist entitled the Direct Observation of Polypectomy Skills (DOPyS) tool [50, 51]. Skills were divided into several sections including (1) optimizing the view of/access to the polyp, (2) stalked polyps, (3) small sessile lesions, and (4) post-polypectomy. Endoscopic nontechnical skills were also included at the end of the assessment form. For each item on the checklist, a 4-point score is given which was intended to evaluate the subject by maneuver: 1 (standards not met), 2 (some uncorrected errors), 3 (competent and safe), and 4 (highly skilled). This methodology was developed, validated, and shown to be reliable in two sequential publications by Gupta et al. [50, 51] These studies demonstrated that assessors who had been formally trained in DOPyS could successfully and consistently delineate between procedures performed by expert endoscopists and those performed by trainees, so long as the assessor was able to observe at least five separate polypectomies performed by each individual.

Despite its promising value, the DOPyS has not been validated for use in more advanced mucosal resection. Polypectomies performed during the DOPyS validation studies were uniformly less than 18 mm in size, which is smaller than lesions typically considered for EMR or ESD (usually >20 mm in size). Further, the tool as written is not applicable to techniques apart from conventional injection-lift EMR . Nonetheless, a similar instrument could be of great value in establishing uniform competency-based mucosal resection training. If validated appropriately, such a method could be applied throughout a graduate or postgraduate training program, and used at intervals guided by data extracted from published learning curves, in order to establish an individual trainee’s progression to competence.

The Importance of Feedback

Well-developed CBE is based upon the principle of self-regulated learning. Perhaps the most critical aspect of this process is the capacity of an individual to assess the results of learning (e.g., trainee’s performance) and make adjustments, prospectively [52]. Feedback is central to this process of assessment and adaptability, as it allows the trainee to identify areas of weakness that are in need of additional attention. Feedback may be internal (self-assessment) or external, with the latter frequently delivered by one’s instructor, mentor, or program-director. Internal feedback is important, but sometimes can be unreliable, and is, in any case, outside the control of an institutional training program. External feedback is more likely to be accurate [53] and, when delivered by a subject matter expert, can include guidance for incremental improvement.

Despite its extremely high value, external feedback typically suffers from poor delivery and inadequate, poorly-timed, or generalized content. Fundamentally, external feedback is difficult because it takes place as an exchange within interpersonal relationships of varying resiliency. A strong relationship, with trust and credibility by both partners, allows for valuable insight and communication. A weak relationship may be harmed by the delivery of criticism or corrective feedback, making this either less effective or less likely to be delivered in the first place. While not explicitly stated, one of the major conceptual advantages of a traditional apprenticeship is this relationship. In the classical sense, a mentor and apprentice are expected to have both a close and invested working relationship built on trust and common interest. Such a relationship should allow for the provision of regular and honest formative feedback, which may even be bilateral (How can you/I learn better? How can you/I teach better?).

One potential drawback to an educational model with explicitly defined competencies and learner-directed focus is this need to provide frequent and accurate feedback. Evaluating procedural skills seems particularly susceptible to failures, especially at the outset of mucosal resection training, when trainees are expected to be at the first stage of learning (unconsciously incompetent). During this time they may possess only limited insight in order to make self-assessments, and the relationship with a mentor may be nascent or underdeveloped. Great care should be taken at all times to ensure that feedback is given effectively, frequently, and within a nonjudgmental and constructive environment.

Training Considerations for “Nontraditional Students ” and Continuing Medical Education (CME)

Much of what has been discussed previously is based upon the assumption that training in mucosal resection occurs in the milieu of a traditional graduate GI fellowship or postgraduate educational program in advanced endoscopy (formal advanced endoscopy fellowship training). While this may be the standard setting in which such skills are developed for new learners, most practicing endoscopists who perform EMR did not develop this technique under such formal circumstances. Regardless of how experience in mucosal resection is acquired, societal and professional expectations would suggest that the practicing endoscopist should demonstrate competency that is on par with graduates of established training programs (and vice versa). This remains inherently problematic, as there are no societally endorsed criteria for judging competency , no validated observational assessment tools, and limited data on learning curves even to suggest a minimum threshold volume of procedures necessary to use as a surrogate for competence.

The previous example of Western experts who perform ESD is perhaps the most striking conglomeration of skill acquisition by nontraditional means [29, 49, 54] and may be beneficial in devising a methodology for practicing endoscopists to acquire EMR skills at the CME level (during active clinical practice and after completion of all endoscopic training). Draganov et al.29 specifically highlight the impact that observation of live cases can have on acquiring new skills for an already experienced endoscopist. In his published experience, Dr. Draganov’s ESD learning curve was divided into three phases: (1) pre-observation , during which ESD was performed on animal models prior to observation; (2) observation , during which Dr. Draganov visited Japan and observed live ESD cases at experienced centers over a 5-week period; and (3) post-observation , where additional ex vivo animal models were again used to garner additional experience. Following the observation period, resections were completed on models with significant gains in efficiency (shorter time) and observable trends toward reduced incomplete resection rates and adverse events. This was notable, especially considering that ESD requires significant technical expertise distinctly disparate from traditional polypectomy, while traditional EMR is more akin to an extension or refinement of traditional polypectomy skills [23, 27].

Observation can be accomplished by practicing endoscopists under a number of different circumstances. There are various national and international endoscopy conferences at which practicing endoscopists may attend didactic lectures, breakout sessions, and live demonstrations which showcase EMR or ESD techniques [55, 56]. In keeping with the prior example, these courses may provide a valuable adjunctive experience to solidify skills among practicing physicians who already have extensive polypectomy experience and wish to gain skill in EMR or become exposed to ESD. These courses have several limitations , as many provide short-duration hands-on training limited to a few hours and utilize ex vivo animal models that cannot simulate intraprocedural hemorrhage or clinical instability in the setting of perforation. While these courses offer an important and convenient method for endoscopists to be taught by experts, they may be best digested by the endoscopic journeyman rather than the apprentice.

Professional societies, including the ASGE, American College of Gastroenterology (ACG) , and Society for Surgery of the Alimentary Tract (SSAT), offer several opportunities for hands-on training in mucosal resection which vary in scope, duration, and intensity. These range from introductory 3-hour workshops in EMR and ESD at annual meetings [57] to the formalized ASGE Skills, Training, Assessment, and Reinforcement (STAR) Certificate Programs in Lower GI and Upper GI EMR [58] and the joint ASGE-Japan Gastrointestinal Endoscopy Society Masters Course in ESD. The STAR EMR courses are multifaceted and are intended to be completed over the course of 3–6 months by enrollees. Enrollment is usually limited to practicing gastroenterologists with at least 2 years of experience, greater than 500 independent colonoscopies, and “proficiency in basic polypectomy, hemostasis, and injection techniques.” Each course initially includes a self-directed online curriculum which includes a baseline knowledge assessment (pretest), reading materials, online videos, and a summative assessment upon completion (posttest). The live portion of the course includes 10 hours of EMR-specific didactic and hands-on training, proctored by expert instructors using ex vivo animal models. On the subsequent day, a 4-hour hands-on summative assessment is performed and candidates who pass successfully are awarded a certificate of completion by the ASGE. While the STAR Certification is not a guarantee competency in EMR, it is constructed based upon competency-based educational principles, and the program offers a valuable opportunity for established endoscopists to learn this endoscopic technique (Fig. 11.3).

Fig. 11.3
figure 3figure 3

A third-year fellow was being taught to perform conventional EMR . A 15-mm cecal polyp was found and its morphology described using white-light endoscopy as Paris 0-IIa + Is (a). The lesion borders were hard to define, and NBI was used to delineate the borders of the polyp, which were then marked by using APC (b). No features of invasive cancer were found on high-definition white-light or NBI endoscopy. The lesion was submucosally lifted using a commercially available lifting solution tinted with methylene blue (c). En bloc resection using a 15-mm stiff, braided, oval-shaped snare was performed. The underside of the resected specimen showed no target sign (d) nor did the remaining portion of the resected colonic wall. However, a significant amount of mucosa intended for resection remained, as delineated by the marking dots (e). This mucosa was removed for completeness sake (f). After ablation of the edges using APC, the lesion was closed using two endoclips (g). Pathology showed a serrated sessile polyp. This example highlights many of the key steps necessary to teach, learn, and perform EMR, and this is also the standardized approach that is taught in the ASGE STAR EMR course

Full size image

Competency Thresholds by Volume and Outcome

As mentioned earlier, there is no established clear-cut methodology to teach trainees EMR, let alone ESD—particularly in countries where the prevalence of gastric cancer is low and where prevalence-based models are adopted where ESD experience (and also often EMR) typically begins in the colorectum [59]. It has been suggested that at least 250–300 procedures are required before competency can be assessed for conventional polypectomy [60]. Similarly, 100 procedures may be required for EMR, 200 procedures for chromoendoscopy , and 30 procedures for NBI prior to assessing or achieving competency [60]. After rigorous theoretical and experimental preparation, a skilled interventional endoscopist can achieve competence after 20–30 untutored ESD procedures, but outcomes data from twice that caseload are often required to support ongoing competence [59]. Importantly, volume thresholds do not necessarily equate to competency [27]. Furthermore, it is understood that to maintain competency a skill should be performed regularly, potentially weekly, or at least 1–2 times each month.

Irrespective of the training methodology, it is critical to remember that high-quality clinical outcomes are paramount. Those seeking to acquire skills in endoscopic resection must remember that these superficial neoplastic lesions are found in patients who present, not to provide a model for gaining experience or for research, but rather to attain a result equivalent to and less risky than surgical resection.

We suggest certain achievable competency thresholds, which can be applied to newly graduated trainees or to experienced endoscopists who have acquired new skills in endoscopic resection. For EMR, the rate of residual neoplasia should be no greater than 20%–25% [61] when only lesions ≥15–20 mm are considered (i.e., for lesions necessitating piecemeal resection). For endoscopically curable lesions (generally those without invasive cancer or risk of lymph node metastasis that do not require surgery, which can vary by organ type), curative resection should be achieved in 90%–95% of cases by the third follow-up procedure [62,63,64]. For example, if residual neoplasia is found on biopsies during the first follow-up endoscopy, a second endoscopic procedure for eradication of residual neoplasia would be indicated (alternatively, if treatment of suspected macroscopically visible residual neoplasia was carried out at the first follow-up procedure, then a second surveillance procedure would be necessary 3–6 months later). It should be kept in mind that variations in recurrence rates may exist, as the odds of residual or recurrent neoplasia increase with larger lesions that are removed in a greater number of pieces [65]. Thus, a 20% rate of residual/recurrent neoplasia for an endoscopist who only removes lesions up to 2 cm in size might be considered high, whereas the same 20% recurrence rate for an endoscopists who mainly removes very large lesions (>4 cm in size) necessitating piecemeal EMR might be very reasonable.

For ESD, particularly in the West, the focus should probably be on rates of en bloc resection and adverse events such as perforation. Less focus should be applied to other outcomes (e.g., rates of R0 or curative resection), as Western endoscopists who perform ESD are often put in a difficult predicament by referring surgeons and oncologists and asked to resect lesions that fall outside of traditional guidelines. Competency in ESD should be demonstrated by en bloc resection rates of ≥80% with perforation rates of ≤10%; expertise in ESD is associated with en bloc resection rates of ≥90% with perforation rates of ≤5% [59, 66].

Incumbent with use of outcome metrics , patients should be followed closely, and it is recommended not to rely solely on endoscopic visual assessment to rule out recurrent/residual lesions but to also take biopsies from the center and periphery of the resection scar and of any nodularity that might be found.

Conclusion

EMR and by extension ESD have revolutionized the approach treating non-pedunculated, superficial intraepithelial neoplasia at various sites along the GI tract over the past two to three decades. Together, these mucosal resection techniques represent a collection of advanced endoscopic procedures that require significant technical skill, experience, and cognitive training to master. Although there is a long history of producing capable and even expert endoscopists through traditional training pathways, significant variation among practicing physicians suggests the need to establish a mature, competency-based educational process to ensure that both technical and nontechnical skill development has been successful prior to independent practice.

There are several barriers to implementation of such standards and curricula at the present time, which include paucity in information regarding the learning curve to perform EMR, the lack of accepted measures of intraprocedural competency, and the absence of an applicable direct observational tool such as the DOPyS for advanced mucosal resection. Despite these challenges, training in mucosal resection appears to be set on an inexorable tack toward competency-based education, in lockstep with the remainder of undergraduate and graduate medical education. Given the adoption of ACGME principles by professional GI societies, it is likely that a formalized CBE curriculum for mucosal resection could be developed and adopted.