Summary of gastroscopy quality indicators applicable in clinical practice

Fasting time

For UGI endoscopy, patients should fast for at least 6 h for solids and at least 2 h for clear liquids before the procedure. This approach has been widely studied and is a standard of care supported by scientific societies [6, 7].

This fasting period may need to be extended in certain cases, such as for patients with gastroparesis or treated with glucagon like peptide-2 antagonists [8]. The American Society for Gastrointestinal Endoscopy (ASGE) emphasizes the need for longer fasting periods for patients with gastroparesis and achalasia to minimize the risk of aspiration [9]. Besides anatomical changes, drugs such as the recently very popular semaglutide also influence delayed gastric emptying [10]. It has been observed that patients treated with this drug exhibit increased residual gastric content, especially those with prior symptoms of nausea or vomiting. Additionally, the quantity of food consumed should be taken into consideration when determining the fasting period [8].

In contrast to the commonly used 6-hour fasting period, Cai et al. proposed a 4-hour semifluid and 2-hour water (4 + 2) fasting protocol [11]. Their study showed that the reduced 4 + 2 protocol did not diminish mucosal visibility and, moreover, increased EGD patient satisfaction (86.8% vs. 63.9%, p = 0.002) [11]. In 2021, Marsman et al. examined the volume of gastric fluid during EGD [12]. They found that patients who drank a small amount of clear fluid (< 150 ml of water) up to an average of 39 min before the EGD did not have more than the physiological average amount of fluid in their stomachs. Besides different fasting regimens, other pre-procedural modalities have also been assessed. In 2022, van Noort et al. reported that specific fasting instructions given to patients by nurses increased gastric visibility and reduced discomfort related to the procedure compared to a group of patients who did not receive such strict instructions [13]. Further research is required to reach a clear consensus on the fasting period before EGD and necessary actions in pre-procedural periods.

Premedication

Preparation before endoscopy, also termed premedication, plays an important role in ensuring good mucosal visibility, which is essential for detecting lesions in the digestive tract. It is widely recognized that bowel preparation before colonoscopy has a significant impact on the adenoma detection rate [14–16]. Similarly, recent studies have been focusing on the impact of mucosal visibility and lesion detection rate [17, 18].

Premedication is one of the most popular issues, with an increasing amount of new investigations in recent years. Numerous studies assessing premedication agents for enhancing UGI mucosal visibility during EGD have been published [19–23]. The use of premedication has been recommended in Japanese guidelines [24]. Premedication strategies aimed at enhancing mucosal visualization typically involve the administration of mucolytics, defoaming agents, and antispasmodic medications. While premedication with N-acetylcysteine (NAC) and simethicone (SIM) has shown improvements in mucosal visualization (excellent visualization in stomach SIM 76.2% and NAC + SIM 74.5% vs. water 38.8%; p < 0.001), its impact on the detection of gastric neoplasia remains inconclusive across various studies [25].

A meta-analysis conducted by Burke et al. in 2021 concluded that SIM (133 mg being the most effective dose) was more effective than a single preparation agent, and SIM + NAC was not superior to SIM alone [19]. Another meta-analysis, by Li et al. in 2019, revealed that preparation with SIM + NAC was superior to SIM alone (MD = –0.14 (–0.25, –0.03), p = 0.01, I² = 0%) in terms of mucosal visibility. Additionally, SIM + NAC vs. water improved the overall pathology detection (RR = 1.31, 95% CI: 1.12–1.53, p = 0.0006), while SIM alone led to shorter procedure times [22]. A meta-analysis of randomized controlled trials evaluating the impact of preparation with NAC + SIM for EGD published in 2018 showed improved mucosal visibility compared to EGDs without premedication [23]. These discrepancies underscore the need for further investigation in this area.

Krishnamurthy et al. observed in 2023 that preparation with SIM + NAC resulted in better mucosal visibility during EGD compared to SIM and NAC alone, as well as placebo. This combination also led to shorter procedure time (mean time: 5.27 ± 1.28 min) [20]. Similarly, Stepan et al. demonstrated in 2023 that preparation with NAC + SIM doses of 600 + 400 mg showed better mucosal visualization compared to no preparation or water. However, the improvement in visualization with NAC + SIM doses of 400 mg + 20 mg was not statistically significant [26]. In a study conducted in 2023 by Cao et al. preparation with SIM (S), pronase (P), or both improved mucosal visibility compared to saline (NS) (total visibility score – higher is better – 11.86 ±3.36 in group P vs. 14.52 ±2.57 in group NS, p < 0.001; 12.36 ±2.93 in group S vs. 14.52 ±2.57 in group NS, p = 0.006). However, this preparation did not significantly affect the detection rate of diminutive lesions [27]. A study published in 2021 by Liu et al. found that premedication with pronase and simethicone did not increase lesion detection during EGD [28]. In conclusion, all studies except one showed improved mucosal visibility when premedication was used, even when drug doses were low.

The importance of cleanliness during EGD has been increasingly recognized. Recently, three research groups aimed to evaluate upper gastrointestinal cleanliness scales, namely the PEACE scale, the Toronto Upper Gastrointestinal Cleaning Score (TUGCS), and the Barcelona scale [19, 29, 30].

The PEACE scale, assessing preparation in the esophagus, stomach, and duodenum, uses a four-point scale from 0 to 3, with a minimum acceptable value of 2 for each segment [17, 19]. The scale has shown good inter- and intra-observer agreement [17]. Proper cleanliness has also been associated with increased detection of clinically significant lesions [19]. The TUGCS contains the evaluation of four sections of the UGI (fundus, body, antrum, and duodenum), each from 0 to 3. The total number of points is therefore from 0 to 12 [29]. The more elaborate Barcelona scale assesses 5 segments (esophagus, gastric fundus, body, antrum, and duodenum), each scored from 0 to 2 [30, 31].

In conclusion, the growing awareness of the importance of cleanliness during EGD has led to the development of scales that assess the degree of cleanliness. The effectiveness of different premedication regimens for improving mucosal visibility during EGD has been proven. However, the correlation between cleanliness and neoplastic lesions detection requires further exploration.

Sedation

Sedation plays a crucial role in enhancing the patient’s comfort during endoscopy [32]. Propofol is commonly used for sedation, with continuous monitoring of blood pressure, heart rate, and oxygenation to ensure patient safety [33].

Sedation aids patients in overcoming their anxiety about the procedure, thus improving overall satisfaction. Comparatively, propofol-based sedation tends to result in higher satisfaction rates among both endoscopists and patients when compared to midazolam-based sedation [34]. Additionally, sedation with propofol leads to a shorter recovery time compared to midazolam, which can impact the overall duration of the procedure [35, 36].

The data regarding the impact of sedation on the detection of pathologies during EGD are conflicting. Lee et al. found that sedation does not increase the detection rate of reflux esophagitis, Barrett’s esophagus, early or advanced gastric cancer, or gastric ulcer. Moreover, in the sedation group, hernias and minimal esophagitis were diagnosed less frequently [37]. In contrast, Wu et al. proved that EDG performed with sedation (propofol) gives a better chance of detecting small neoplasms (2.80% vs. 2.02%; p < 0.001) of the upper gastrointestinal tract, particularly in the antrum. Sedation was also associated with a higher biopsy rate, which positively correlated with lesion detection (41.4% vs. 36.4%, p < 0.001) [38]. In 2021, Zhou et al. in a study which included 432,202 patients, found that sedation improved the detection of early cancer and high-grade intraepithelial neoplasia in the esophagus and stomach [39]. However, Liang et al. found no significant difference in the detection of early esophageal squamous cell cancer between EGD procedures performed with or without anesthesia [40].

The ESGE recommends sedation in the ultra-long Barrett’s esophagus surveillance program [41]. Nonetheless, anaesthesia is associated with sedation-related adverse events [42, 43]. This highlights the importance of carefully weighing the benefits and risks of sedation options when performing EGD.

Examination time

Adequate timing of the test appears to be an important determinant of EGD quality. A study by Teh et al. demonstrated that endoscopists who worked at a slower pace (> 7 min) had a significantly higher ability to detect high-risk gastric lesions compared to their faster counterparts, with an odds ratio of 2.50 (with a 95% confidence interval of 1.52–4.12). This was observed regardless of whether they were part of the endoscopy staff or trainees [44]. Moreover, these slower endoscopists were able to identify three times as many gastric neoplastic lesions (cancer or dysplasia) as their faster counterparts, with an odds ratio of 3.42 (and a 95% CI of 1.25–10.38).

It was also reported that endoscopists who performed EGD in under 5 min (without biopsy) had a lower neoplastic lesions detection rate than slower endoscopists (under 5 min 0.57% (13/2288) vs. 5–7 min 0.97% (99/10 180), and > 7 min 0.94% (31/3295)) [45].

Park came to similar conclusions in a study published in 2017, which analyzed more than 111,000 cases. A longer EGD (>3 min) testing time was associated with a higher rate of detection of neoplastic lesions (0.20% in the faster group, vs. 0.28% in the slower group, p = 0.0054) [46]. After introducing a protocol for extended testing time (> 3 min during instrument withdrawal after reaching the duodenum) and analyzing more than 30,000 EGD cases, this research group aimed to compare these EGDs and the group from the previous study. The average examination time was significantly longer (3:35 ±0:50 vs. 2:38 ±0:21), which translated into a 33% increase in the detection of precancerous lesions [47]. In a prospective observational study, Romańczyk et al. found that an EGD time exceeding 4.2 min translated into a higher composite detection rate (26.3% vs. 11.8%; p < 0.0001) – a proposed new quality indicator. Furthermore, all neoplastic lesions were found within longer procedures [48].

In another prospective study published in 2023, Gao et al. proposed a minimum EGD time of 6 min, as this was associated with increased detection of focal lesions (average time less than 6 min 33.6% vs. more than 6 min 39.3%, p = 0.011) [49].

In a Korean retrospective study from 2023, it was found that independent risk factors for missed gastric adenomas during EGD included shorter observation time and the presence of gastric intestinal metaplasia [50].

The 7-min threshold seems to be adequate, as Yoshimizu et al. found no statistically significant differences in this examination outcomes between 10-minute and 7-minute procedures [51]. Regardless of different EGD time thresholds, it seems that “slower” operators tended to detect more neoplastic lesions.

Photodocumentation

Various scientific societies propose different approaches for photodocumentation of the upper gastrointestinal (UGI) tract. The broadest is the one proposed by World Endoscopy Organization (WEO) [52]. According to their protocol, 28 images should be taken, starting with the larynx, followed by four in the esophagus, then the pylorus, four images each (large curvature, small curvature, anterior and posterior wall) in the antrum, distal and middle part of the body, 4 images of the cardia and proximal body in inversion, and finally the bulb and descending part of the duodenum [52].

According to the ESGE guidelines published in 2016, the UGI tract photodocumentation should consist of at least 10 photos of the main anatomical points: distal and proximal esophagus, squamocolumnar junction, antrum, angulus, corpus, retroflex of the fundus, diaphragmatic indentation, upper end of the gastric folds, duodenum, major papilla, and any observed abnormalities [7].

To avoid blind spots, precise protocols for photodocumentation have been developed in Japan. In 2013, a schema for EGD photodocumentation was published, the Systematic Stomach Screening (SSS) protocol, with over 22 images, to detect early gastric cancer. It includes four quadrant photos of the body in an antegrade view (upper body, lower body, and antrum) and ten photos in a retrograde view (three from the incisura and upper body and four from the fundus [53].

To increase the detection of neoplastic lesions during EGD, Quea et al. developed the alphanumeric SACE protocol, providing accurate and reproducible photographic documentation, covering 8 regions and 28 areas [54]. It should be mentioned that no data support the use of any of the protocols, and the rationale to use them is based primarily on good clinical practice.

Virtual chromoendoscopy

Virtual chromoendoscopy (VC) is modern alternative for dye-based chromoendoscopy. Nowadays, VC is an integral part of the endoscopic examination, allowing better assessment of the mucosa and visible lesions. One of the most popular options is narrow-band imaging (NBI – Olympus), the use of which is an integral part of the scales applied during colonoscopy to evaluate polypoid lesions, such as NICE and JNET [55, 56]. Other popular types of VC are i-SCAN (PENTAX) and flexible spectral imaging color enhancement (FICE-Fujinon) [57–59]. They are easily accessible and do not require demanding equipment; therefore, VC has been extensively assessed for enhancing the detection of various gastrointestinal lesions [60–65].

In 2005, it was reported that EGD with NBI had an 86% sensitivity rate in detecting high-grade dysplasia (HGD) or early cancer in Barrett’s esophagus [62].

A prospective study from 2008 found that among patients with Barrett’s esophagus, HGD was detected in 18% of patients when using NBI, while the detection rate was 0% in the group using only white light endoscopy [63]. Furthermore, in the NBI group, any grade of dysplasia was detected in 57% of biopsies compared to 43% without VC [63].

Jayasekera et al. demonstrated that NBI was more sensitive than high-definition white light endoscopy (HD WLE) in detecting HGD or early cancer (89% vs. 79.1% respectively), though with lower specificity (80.1% vs. 83.1%, respectively) [64]. A systematic review conducted by ASGE in 2016 showed that NBI had a sensitivity of 94.2% and specificity of 94.4% for Barrett’s esophagus-related dysplasia, leading to the recommendation of this method in practice [64, 65].

In 2009, Sharma et al. reported that the use of NBI reduces the number of biopsies required to detect dysplasia [66]. The appearance of dysplasia within Barrett’s esophagus has been summarized in the classification called BING, which uses dysplasia features and allows VC to be easily used in daily practice [67].

VC has also been shown to be effective in the detection of esophageal squamous cell carcinomas (SCC). In 2010, Muto et al., in a multicenter, prospective, randomized controlled trial, found a statistically significant increase in the detection of SCC using NBI compared to white light, with a sensitivity of 97.2% (100% vs. 8%, p < 0.001; 97% vs. 55%, p < 0.001 depending on region) [68].

A meta-analysis published in 2017 by Morita et al. showed that NBI was more specific than Lugol-based chromoendoscopy in distinguishing SCC from other esophageal lesions, with comparable sensitivity [69]. A prospective, randomized multicenter trial conducted in France comparing dye-based chromoendoscopy (Lugol) vs. NBI showed statistically significantly higher specificity (66% vs. 79.9% respectively) in the detection of SCC [70].

Another study published in 2023 by Chaber-Ciopinska et al. showed that NBI imaging, compared to Lugol’s liquid staining, resulted in better endoscopy tolerance, shorter examination time and fewer necessary biopsies (12.75% vs. 41.11%; p = 0.003) [71].

Gastric intestinal metaplasia and dysplasia are pre-cancerous conditions leading to the development of gastric cancer. However, the endoscopic detection of these conditions remains suboptimal [72]. A multicenter prospective study showed that NBI imaging had higher sensitivity than WLE in detecting gastric intestinal metaplasia (87% vs. 53%; p < 0.001) and dysplasia (92% vs. 74%) [73]. The diagnostic accuracy of NBI was also higher than in WLE (NBI 94% vs. WLE 83%; p < 0.001).

Two metanalyses from 2021 and 2023 found that the use of NBI increases the detection of intestinal metaplasia in the stomach compared to EGD with WLE (70% vs. 38%, RR = 1.79; 95% CI: 1.34-2.37; p < 0.01) [74]. The endoscopic grading of gastric intestinal metaplasia (EGGIM) is a useful tool for identifying and assessing the severity of gastric intestinal metaplasia, making it applicable in daily practice [75].

Quality metrics

There are several research-based indicators that reflect the quality of the EGD. In 2014, Park et al. sought an indicator associated with increased detection of gastric cancer lesions. They demonstrated that a scoring system based on the detection of gastric diverticula and gastric subepithelial lesions was linked to higher detection of early gastric neoplasm [76].

In a study published in 2018, it was found that endoscopists who frequently performed photodocumentation of the Vater papilla were significantly more likely to detect upper GI neoplasms [77]. In 2019 Januszewicz et al. reported that the endoscopist biopsy rate was strongly associated with detection of gastric precancerous conditions (p = 0.83; p < 0.001). They also found that patients examined by endoscopists with higher EBR had a 56% lower risk of missing a gastric cancer while undergoing EGD [78].

In a multicenter, prospective observational study, involving nearly 3000 patients, Romańczyk et al. validated the composite detection rate (CDR) [79]. The CDR, which includes the detection of at least one of three types of lesions – post-ulcer duodenal bulb deformation, cervical inlet patch, or gastric polyp – was correlated with the detection of upper GI neoplasms. Interestingly, the endoscopist biopsy rate (EBR) did not show a similar correlation [79].

Biopsy sampling

Biopsy sampling is a critical factor influencing the detection rate of benign, premalignant or malignant changes in the upper gastrointestinal tract (UGT) [80].

To address this issue, the ESGE has proposed guidelines [81].

Strict biopsy sampling protocols are well established for monitoring, detecting and suspecting diseases, depending on the type. In the case of Barrett esophagus (BE), it is recommended to take biopsies from every visible lesion and four-quadrant biopsy for every 2 cm of BE length, dedicating at least 1 minute per centimeter for proper inspection [41, 82–84].

The Sydney protocol has been implemented to improve the detection of atrophic gastritis, intestinal metaplasia, and dysplasia [85]. The procedure and biopsy scheme are outlined in the MAPS III guidelines, which recommend taking biopsies from the antrum, angle, and the corpus from the lesser and greater curvature (each) into separate containers (minimum of five biopsies) [86]. A specimen should also be taken from every focal lesion into a separate container. Samples should be evaluated using the OLGA and OLGIM histological classifications [3]. Performing EGD according to the Sydney protocol significantly increases the detection of gastric intestinal metaplasia (GIM) and gastric atrophy (GA) [87]. The latest ESGE guidelines recommend taking at least two samples from the antrum and two from the corpus, while assessing GA and GIM [81].

Referring to the updated consensus from 2011, suspicion or evaluation of treatment results of eosinophilic esophagitis requires taking at least 6 biopsies from the proximal and distal esophagus into separate vials [88, 89]. When advanced gastric or esophageal cancer is suspected, at least 6 biopsies should be taken in cases of advanced disease. Only 1–2 samples should be obtained if an endoscopic resection seems to be possible [90–93]. It is recommended to obtain at least 6 biopsies from the duodenum (including 2 from the bulb) to diagnose celiac disease, due to the patchy distribution of atrophy [81, 94–96]. Overall, adequate biopsy sampling is essential to accurately establish the diagnosis.

Coordinated care and training

Several reports have indicated that coordinated care can enhance the detection of early gastric cancer. In 2017, a paper published by Lianjun Di et al. based on 60,800 EGDs demonstrated that a multi-disciplinary team (MDT; consisting of physicians from the Departments of Digestive Endoscopy, Gastroenterology, Gastrointestinal Surgery, Anesthesiology, and Pathology) approach improves the detection rate of early gastric cancer. The MDT was based on medical consilia, training and high-quality endoscopy using NBI, magnifying endoscopy, antifoaming, mucus decomposing, and spasmolytic agents in a high-risk population. The endoscopists performed intensive EGD for high-risk patients, were educated through lectures, review of images and videos, and on-site teaching, and also took part in discussions within and between MDTs [97].

A study conducted by Zhang showed that training, NBI, and magnifying endoscopy could improve detection of early gastric cancer (EGC). The training consisted of four components. The first was to raise awareness of gastric cancer, taking part in symposiums, participating in treatment committees with case discussions including evaluation of video recordings, and on-site instruction with endoscopy experts. The second was to perform EGD according to strict Japanese endoscopic standards described in the book Standard Gastroscopy, edited by Hosoi Tozo. The third was learning and application of magnifying endoscopy. The last one was training in endoscopic submucosal dissection [98].

A multicenter study from China reported that following the developed protocol resulted in increased detection of gastric lesions, which was associated with longer examination times and more extensive photodocumentation [99].

In a 2023 retrospective study of 3485 cases, Manfredi demonstrated that following a specific quality protocol during EGD increased the likelihood of detecting intestinal metaplasia (IM) [100].

There are also reports suggesting that dedicated care for patients with Barrett esophagus has positive outcomes.

During a 5-year British study, 921 patients with BE were surveyed by two groups of doctors. Statistically, more dysplasia was detected in the group of patients undergoing EGD by a group of doctors trained in BE assessment (consisting of two trainees and a one gastroenterology specialist trained in detecting BE-related lesions, who performs more than 100 surveillance BE examinations annually), compared to patients not undergoing EGD as part of the dedicated services provided by doctors without training (6.3% (36/568) vs. 2.7% (9/337) (p = 0.014) [101]. A higher dysplasia detection rate (DDR) was corelated with the number of biopsies, target biopsy rate and photodocumentation of the lesions [101].

In a 2017 prospective comparative cohort study, patients diagnosed with BE were divided into two groups: one group of 217 patients received dedicated care, while the second group of 78 patients was assessed during routine EGD. The first group, aimed at detecting BE related lesions, investigated by an endoscopist interest in BE, was significantly more likely to follow the Seattle protocol with the proper number of biopsies and locations, using the Prague classification, and following the appropriate surveillance interval [102].

A similar study conducted in the Netherlands on a group of 1244 patients with BE found that during EGDs performed as part of dedicated services, the examination was significantly more likely to follow correct time intervals (60% vs. 47%, p < 0.01) and take the correct number of biopsies (85% vs. 66%, p < 0.001). It is noteworthy that the endoscopists in the BE-dedicated group were not experts in the treatment of BE lesions, and had not received specific training prior to the study. There was no significant difference in the detection of focal lesions and dysplasia [103].