Impact of Full Spectrum Endoscopy® (Fuse®, EndoChoice®) on adenoma detection: a prospective French pilot study

Introduction

Colorectal cancer (CCR) is a major public health problem. Risk factors such as personal and family history are well known. Colonoscopy improves the detection of lesions at a curable stage. Endoscopic resection of colonic adenomas reduces cancer-related mortality rate by 31 or 53% according to two main studies [1,2]. However, this protection is imperfect, especially for carcinomas of the right colon, and interval cancer represents 3-5% of CCRs [3,4]. To reduce the interval cancer rate, several recommendations have been proposed to improve the quality of colonoscopy. One of the most relevant colonoscopy quality indicators is the adenoma detection rate (ADR). The ADR is inversely proportional to the missed cancer rate [5]. The American Society for Gastrointestinal Endoscopy (ASGE) recommends achieving an ADR of higher than 20% and 30%, respectively, for female and male patients [6].

Several technologies have been evaluated to improve the ADR. The use of high-definition endoscopy had a low impact on the ADR [7]. Virtual chromoendoscopy shows no or only a small difference. Various techniques have been used with some interesting results [8-15]: cap-assisted colonoscopy [10], Endocuff®-assisted colonoscopy [10,14] and Third-Eye® Retroscope® colonoscopy [13]. Their main drawback is the change in the colonoscopy technique; passing through the working channel is a major limitation [10-15]. Moreover, proximal colon retroflexion was not recommended [16,17].

More recently, two new colonoscopes with wide angles of view were developed with promising results: “Extra Wide Angle View” (Olympus® medical systems, Tokyo, Japan) and Full Spectrum Endoscopy® (Fuse®, EndoChoice®, GA, USA) [18-21]. The “Extra Wide Angle View” is allowed on the lateral backward view of 16/28 adenomas [18]. Fuse® colonoscopy is safe [19,20] and has been shown in a back-to-back study to significantly decrease the miss rate of adenomas [21]. Both techniques have the advantage of avoiding any modification of the colonoscopy procedure and maintaining a free working channel.

Fuse® is equipped with three cameras that allow up to a 330-degree angle of view; each camera provides a view of 170° (Fig. 1). A tandem, randomized, multicenter study showed promising results: an increase in the adenoma rate of 41% during a second colonoscopy, including 15% of advanced adenomas [21]. The aim of this first French prospective observational study was to evaluate the efficacy and safety of the Fuse® system.

Figure 1 The Fuse® scope

Patients and methods

Study design and patients

This single-center prospective study enrolled patients over 18 years of age referred for a colonoscopy between January 2015 and March 2016. Exclusion criteria were a history of colectomy, genetic polyposis, a planned therapeutic colonoscopy (endoscopic mucosal resection or endoscopic submucosal resection), and bad bowel preparation (Boston bowel preparation scale [BBPS]<7). The BBPS is valid and reliable scale for bowel preparation: three regions of the colon receive a segment score from 0 to 3 for a total BBPS score ranging from 0 to 9 [22]. Patients with incomplete colonoscopies were excluded from the final analysis.

The primary endpoint was potentially omitted adenomas (POA), i.e., adenomas seen only on the side cameras during the withdrawal of the colonoscope and without angulation of the tip. Polyps potentially omitted (PPO) were defined as observable polyps on lateral screens during colonoscope withdrawal that required maneuvers to be recovered on the central screen. An example of a PPO is presented in Fig. 2. The POA was therefore defined as a PPO that was considered to be an adenoma by the pathologist. Secondary endpoints included ADR, Fuse® impact on ADR, time to cecal intubation and withdrawal time.

Figure 2 Example of potentially omitted adenoma on left screen

Results This study enrolled 141 patients between January 2015 and March 2016. Three patients were excluded because their colonoscopies were incomplete. The final analysis included 78 men and 60 women (sex ratio 1.3). The mean age was 60.4 years. The colonoscopy was performed in subjects with a familial history of CCR, a personal history of adenoma, or for digestive symptoms in 22, 46, and 70 subjects, respectively. Patient characteristics are presented in Table 1.

Table 1 Patient characteristics

Complete colonoscopy was achieved in 97.9% (138/141) patients. In two cases, the colonoscopy was incomplete because of a “fixed” sigmoid without visible stricture: the colonoscopy could be completed with a pediatric scope (PENTAX medical®, Japan). In one case, the colonoscopy was incomplete because of the formation of a non-reducible sigmoid loop despite several attempts. There was no colonoscopy-related perforation. The mean time to reach the cecum was 10 min, and the mean withdrawal time was 12 min. A total of 91 colonoscopies (66%) were performed by fellows under senior supervision, and 34% were performed by seniors. The main results are shown in Table 2.

Table 2 Main results

Three CCRs were found. One CCR (in the left colon) was seen mainly on a lateral screen in a diverticulum and specific maneuvers were required to visualize it in a standard forward view, but we did not consider it an omitted lesion. These three CCRs were not counted as adenomas. All patients underwent surgery because they had no synchronous metastasis.

A total of 130 polyps were resected without complications. One 10-mm adenoma could not be resected given the poor visibility on the central screen, despite multiple endoscopic movements. A second colonoscopy was proposed to the patient, but he refused. Among the resected polyps, 88/130 were adenomas (68%). The pathology of all adenomas was low grade dysplasia. In all, 52/130 polyps were classified as PPO, including 34 adenomas. Therefore, 34/88 adenomas (39%) were POA. The POA location included the left colon in 18/34 cases (53%), the transverse colon in 10 cases (29%) and the right colon in 6 cases (18%). The study design is shown in Fig. 3. In all, 8/34 (24%) POAs were between 10 and 15 mm in size, while the others were less than 10 mm in size (Table 3).

Figure 3 Study design

* A POA (potentially omitted adenoma) was an adenoma seen only on a side screen during withdrawal of the endoscope, requiring at least an ascent of the endoscope and lateral bumpers to be seen on the central screen. The presence of a POA was established by a senior endoscopist

Table 3 Size of potentially omitted adenoma

The global ADR was 34% for men (35% if one likely adenoma that could not be resected was included) and 31% for women. The estimated ADR without POA was 29% for men and 19% for women; in this situation, POA were not counted. These ADRs and estimated ADRs are presented in Table 4.

Table 4 Adenoma detection rate (ADR), estimated adenoma detection rate without lateral screens and ASGE guidelines

Discussion

The results of our study support the use of the Fuse® system for several reasons. First, POA represented 39% of all adenomas, which confirms our primary endpoint. The results are in agreement with a pervious study that described the safety of this new scope [19-21].

The time to achievement of the Fuse® colonoscopy was acceptable; many endoscopies (66%) were performed by fellows under senior supervision. Furthermore, the resection of a PPO could be time-consuming, even if polypectomy was performed in all cases except one. Regarding the detection of adenomas, the results of seniors and fellows were not analyzed separately. Indeed, the seniors assisted the fellows during all colonoscopies and it appears unlikely that the ADR was underestimated, given that the screens were analyzed by both operators. The only potential influence on colonoscopy was probably the duration of the examination, although these fellows had performed colonoscopies for at least 3 years.

The ADR for men was congruent with ASGE guidelines [6] since it was higher than 30%. The ADR for women was also in agreement with ASGE guidelines. The impact of the panoramic view on ADR was considered significant, given the 10% increase in both sexes (Table 2).

Our POA rate is compatible with back-to-back studies. Indeed, many studies have shown that up to 20% of adenomas are missed during colonoscopy because of poor visualization behind the folds and the inner curves of flexures [23]. Despite the most objective evaluation possible (the main limitation of this study), it is likely that some POAs would have been seen with a standard colonoscopy, but not all. Therefore, a rate of 20% missed adenomas is probably very close to reality. Before a polyp was described as a PPO, insufflation was of course used, so that a polyp located behind a badly insufflated fold would not be classed as a PPO. In any case, the feasibility of Fuse® colonoscopy has been demonstrated given the good ADR.

During the evolution of the technique, quality indicators of colonoscopy have been clearly defined by professional societies [6]. Some criteria are obvious and are recommended, including good bowel preparation, informed consent, withdrawal time longer than 6 min, appropriate indication, cecal intubation, perforation lower than 1/1000, etc. [6,22,24,25]. Although the results are not spectacular in the literature, high definition and virtual chromoendoscopy ameliorate the work quality of endoscopists [7,26]. Most of these quality indicators are intended to increase the ADR, because it is a major indicator of quality and it is measurable [5,6]. It has been estimated that 10% of the colonic surface remains to be observed under the standard forward-viewing colonoscope, even with good bowel preparation [27]. Many endoscopic techniques have shown excellent results in preliminary studies [10-15,28], but generalization is limited by changes in the technique of colonoscopy, occupation of the working channel, or increased time.

Fuse® colonoscopy has a low learning curve. The tandem colonoscopy trial of Gralnek et al demonstrated this, as did a second recently published study [19,21]. The cost and practice were the same, and it is easy to expand the use of colonoscopy to 330°, contrary to many other technologies. For example, the Third Eye® Retroscope® and Third Eye® Panoramic™ devices (Avantis Medical Systems, Sunnyvale, CA, USA) are very interesting and efficient (they exhibited an increased adenoma rate of 23.2%), but usual practice appears illusory on account of the working channel occupation and the learning curve [13].

However, our study has some limitations. First, is the possible subjectivity of the primary endpoint. Indeed, though operators described what they considered to be PPOs during scope withdrawal, a certain diagnosis was difficult. Nevertheless, a non-PPO polyp is also first seen on the side screen and then on the central screen. Of course, we did not consider this situation to resemble a PPO, because PPOs required oriented movements to find them on a central screen, but subjectivity remains. Other associated limitations are the lack of randomization or a control group.

Contrary to Gralnek et al, we do not have “advance” POA. One explanation could be the exclusion of patients referred for endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD). The resection of a polyp behind the folds could be time-consuming but resection is helped by a panoramic view. Among the POA, 8/39 (24%) had a size between 10 and 15 mm. This category of patients has thus benefited from the Fuse® technology, which can be questionable for patients with an infra-centimeter POA. The study by Brand et al describes this limit for new technologies designed to increase adenoma detection [29].

In the future, the development of lateral working channels could completely revolutionize colonoscopy, especially for EMR or ESD. The industry will probably develop other similar scopes in the coming years, and patients and doctors may agree to have a narrow view while a panoramic vision is possible.

Recently, a high-quality randomized study questioned the utility of Fuse® in the detection of adenomas in a population with a positive fecal immunochemical test: no difference was demonstrated in ADR [30]. This new result calls into question the superiority of the Fuse®, suggested by the feasibility studies and demonstrated by the tandem study of Gralnek et al. This very interesting study encourages further clinical research into this scope to assess whether it represents a minor or major improvement in CCR screening. Several types of endoscopy center (expert or not) should be involved in future randomized trials for “real-life” studies.