Screening Colonoscopy to Reduce the Incidence and Mortality of Colorectal Cancer

Colorectal cancer (CRC) is a leading cause of cancer-related death worldwide. Recently, in Japan, the incidence and mortality rates of CRC have been increasing, and age-adjusted mortality rates have remained unchanged. Thus, further measures are necessary to reduce CRC mortality rates. Total colonoscopy is an effective screening tool for CRC that allows detailed evaluations of the entire colon and rectum, as well as simultaneous treatment. Improving the adenoma detection rate (ADR) and subsequent endoscopic resection of colorectal neoplasias can effectively reduce mortality associated with CRC [1]. However, the ADR varies depending on the endoscopist and examination environment, and the rate of missed polyps during colonoscopy is as high as 25% [2]. Therefore, it is necessary to improve the sensitivity of lesion detection and enhance the quality of colonoscopy. This review provides an overview of screening colonoscopy and colonoscopy quality indicators (QIs). Additionally, it describes factors that can influence the ADR associated with colonoscopy.

Randomized controlled trials (RCTs) and meta-analyses have reported CRC mortality reductions with flexible sigmoidoscopy [3]. A prospective observational study of the efficacy of colonoscopy to reduce CRC mortality that included 22 years of follow-up found that patients who had undergone screening colonoscopy at least once had lower CRC-specific mortality rates than those who had never undergone screening colonoscopy (multivariate hazard ratio, 0.32; 95% confidence interval [CI], 0.24–0.45), and that screening colonoscopy reduced mortality associated with CRC in both the distal colon and proximal colon, whereas sigmoidoscopy reduced mortality associated with CRC only in the distal colon [4]. Although several ongoing RCTs are evaluating whether colonoscopy can reduce CRC mortality (Table 1) [5‒9], the evidence regarding the efficacy of colonoscopy in reducing mortality is not yet clear. However, the interim results (observed during the 10-year follow-up evaluation) of the NordICC trial, which was designed to evaluate the effects of screening colonoscopy on the incidence and mortality of CRC during a 15-year follow-up period, have been reported [9]. The NordICC trial targeted healthy men and women between 55 and 64 years of age and was conducted in four countries in Europe between 2009 and 2014. Participants were randomly assigned to either the colonoscopy group (invited group) or the usual care group. Although the invited group who underwent colonoscopy screenings exhibited statistically significant reductions of 18% (95% CI, 0.70–0.90) in the CRC incidence and 10% (95% CI, 0.64–1.16) in CRC mortality during the 10-year follow-up evaluation, the differences between these reductions and those of the usual care group who did not undergo colonoscopy screenings were not statistically significant. Furthermore, the intention-to-screen analyses of the NordICC trial indicated that factors that contributed to the efficacy of colonoscopy, which was lower than expected, included low screening colonoscopy participation rates (only 42% of the invited group in this study underwent colonoscopy), a relatively low ADR of 31%, and the possibility that the results of the 10-year follow-up evaluation were premature because of the natural history of adenomas. However, an adjusted per-protocol analysis that assumed that all patients who were registered in the invited group underwent colonoscopy indicated a 31% (95% CI, 0.55–0.83) reduction in the CRC incidence and a 50% (95% CI, 0.63–1.12) reduction in CRC mortality. These results demonstrated the importance of increasing the participation rate in screening colonoscopies, as well as increasing the quality of colonoscopies. In Germany, where screening colonoscopy has been introduced into the national cancer screening program, the cumulative participation rates for the 6-year period of 2003 to 2008 were as low as 15.5% for men and 17.2% for women (age range: 55–74 years) [10]. To increase the screening participation rate, it is important to inform the target population by repeatedly sending personal invitations or launching public campaigns.

Several studies have reported the efficacy of using CRC risk scoring systems for advanced colorectal neoplasia (ACN). Chiu et al. reported the usefulness of the algorithm of the Asia-Pacific Colorectal Screening scoring system, which is based on age, sex, family history, and smoking, as well as the effectiveness of the fecal immunochemical test (FIT) for detecting ACN during CRC screening [11]. This algorithm is a type of triage method that is performed before the FIT and colonoscopy to divide the CRC risk into low, medium, and high. The risks of ACN of the medium-risk group and high-risk group were 3.4-fold higher and 7.8-fold higher than that of the low-risk group. Additionally, early colonoscopy was appropriately indicated for 70.6% of ACNs and 95.1% of invasive cancers. Similarly, Sekiguchi et al. [12] reported that an ACN risk scoring system that includes age, sex, family history, body mass index, and smoking habits can increase the sensitivity of ACN screening compared to that of the FIT alone (46.4%–56.3% vs. 17.9%–33.9%). Thus, the use of algorithms that indicate the need for the FIT according to risk scores may effectively reduce the rate of unnecessary colonoscopies and inform patients of their own CRC risk, thus leading to increased rates of necessary colonoscopies.

An analysis performed in 2013 in the USA compared the FIT, sigmoidoscopy, and colonoscopy and found that the FIT was the most cost-effective screening strategy because it outperformed sigmoidoscopy and colonoscopy; however, the results were dependent on higher performance and adherence rates than those observed in the USA [13]. If the FIT adherence rate is less than 50%, then sigmoidoscopy and colonoscopy are more cost-effective because each costs less than USD 50,000 per quality-adjusted life-year (QALY). Colonoscopy costs USD 56,800 per QALY; however, it costs less than USD 100,000 per QALY if it reduces the proximal CRC risk by 50%. The cost-effectiveness of colonoscopy compared to that of sigmoidoscopy is dependent on its ability to provide significant protection against proximal CRC. The cost-effectiveness of CRC screening in Japan was assessed by evaluating the following three strategies: (1) FIT; (2) colonoscopy; and (3) FIT with colonoscopy at age 50 years [14]. All three strategies were more effective than no screening. The colonoscopy-based strategy was the most cost-effective in Japan; however, it required more than double the number of colonoscopy procedures compared to that of the other strategies. The strategy comprising FIT with colonoscopy at age 50 years offered a balance between cost-effectiveness and a manageable colonoscopy workload, thus making it a potentially optimal solution. The cost-effectiveness of screening colonoscopy is dependent on its effectiveness compared to that of FIT and sigmoidoscopy and the quality of the examination. Therefore, further studies are necessary.

The adverse event rate associated with screening colonoscopy ranges between 0.28% and 0.50% [10, 15]. The major adverse events are bleeding and perforation, with rates of 0.14%–0.25% and 0.01%–0.05%, respectively [10, 15, 16]. During polypectomy, 0.47% of patients experienced bleeding and 0.05% experienced perforation [10]. Other adverse events include cardiopulmonary events such as vasovagal reactions during examinations and postpolypectomy electrical coagulation syndrome, which rarely require additional interventions. The risk of complications increased with age. Compared to that of patients 55–59 years of age, the odds ratio (OR) of patients 79 years or older was 3.4 (range, 2.8–4.1) [10]. Colonoscopy-related deaths are rare. However, one study performed in Germany reported a total of seven deaths (0.245/100,000 colonoscopies); of these seven deaths, three were caused by perforation and one was caused by hemorrhage.

QIs that have been proposed for colonoscopy [3, 17, 18] are crucial because they can be used to assess procedural effectiveness and safety. QIs are mainly categorized as technical quality, patient safety, and patient acceptability. These indicators serve as benchmarks to ensure the clinical effectiveness of colonoscopies. The American Society for Gastrointestinal Endoscopy and the European Society for Gastrointestinal Endoscopy have indicated the target performance rate of each QI (Table 2) [17, 18]. Among these, the ADR, which refers to the percentage of patients with at least one histologically proven adenoma or carcinoma, is well-known as an established QI for the early detection and prevention of CRC. Additionally, the cecal intubation rate, withdrawal time, and bowel preparation quality have been proposed as the main QIs. Surveillance intervals determined using evidence-based guidelines have been recommended, and adherence to those intervals, which are crucial to reducing postcolonoscopy CRC (PCCRC), has been measured [3]. The surveillance period after colonoscopy is based on the performance of a high-quality colonoscopy. Colonoscopy screening and surveillance guidelines in Japan suggest a shorter surveillance period than that suggested by the American Society for Gastrointestinal Endoscopy guidelines to not only reduce CRC mortality but also avoid colorectal surgery to maintain the quality of life of individuals [3]. In particular, the ADR, cecal intubation with photodocumentation, and postneoplastic lesion treatment surveillance are considered important indicators by the American Society for Gastrointestinal Endoscopy guidelines. Patients who underwent colonoscopy performed by endoscopists with low cecal intubation rates had a higher risk of PCCRC in both the proximal and distal colon [17]; therefore, photographic documentation is recommended to confirm successful cecal intubation. Appropriate surveillance based on colonoscopy performed in accordance with QIs helps avoid unnecessary reexaminations and excessive follow-up evaluations.

A higher ADR can contribute to the early detection and removal of potentially cancerous polyps, thereby reducing the risk of CRC [1]. The ADR was proposed in 2002 and established as a QI for colonoscopy based on the results of a 2010 study that found that a lower ADR resulted in a higher number of PCCRC cases [19]. For every 5% increase in the ADR, the incidence and mortality rates of CRC decreased by 11.4% and 12.8%, respectively [1]. Therefore, although the ADR is an effective and objective measurement, several concerns have been highlighted. Initially, the ADR applied to patients older than 50 years of age who had not yet undergone their first colonoscopy. However, it changed over time; by 2015, the ADR applied to asymptomatic individuals who underwent their first screening colonoscopy and patients 50 years of age and older at average risk for CRC, resulting in target ADR values of 30% or more for men and 20% or more for women [18]. However, these ADR targets are not sufficient for polyp detection and provide minimal quality assurance. Additionally, it is difficult to determine an individual’s ADR because its calculation can be influenced by the target population being studied in clinical practice. Patient populations vary among institutions, and the definition of the ADR does not apply to all patients. This report serves as a reference for such issues. One study that compared screening ADR, surveillance ADR, and diagnostic ADR reported that the surveillance ADR was the highest and 10% higher than the screening ADR [20]. Based on these results, it may be possible to determine the screening ADR of each endoscopist.

Withdrawal time is a key indicator used to ensure the examination quality. A multicenter RCT of 1,027 individuals 40–85 years of age was performed [21]. Participants were assigned to the 9-min or the 6-min group, and the ADR was evaluated as the primary endpoint. The colon was divided into segments and time measurements were performed. During that trial, detailed time control was performed. The area from the cecum to the transverse colon was observed for 4 min for patients in the 6-min observation group; however, the same area was observed for 6 min for patients in the 9-min observation group. As a result, it was found that extending the withdrawal time from 6 to 9 min significantly increased the ADR (36.6% vs. 27.1%; p = 0.001). In particular, the ADR of the proximal colon (21.4% vs. 11.9%; p < 0.001) and that of endoscopists with less colonoscopy experience (<5,000 cases) (36.8% vs. 23.5%; p = 0.001) significantly increased. A multicenter, randomized, tandem trial demonstrated that a 9-min observation period significantly reduced both the adenoma miss rate (10.9% vs. 25.9%; p < 0.001) and the advanced adenoma miss rate (5.3% vs. 46.9%; p = 0.002) [22]. These observation times could also reduce the incidence of PCCRC, which primarily occurs because of missed lesions and is more common in the right colon [23] because polyps that occur in the right colon are more likely to be flat and serrated and have the potential to be precancerous lesions [24], thus reducing their detection rate. Therefore, the right colon, which has a higher risk of missed lesions because of its deeper folds and high prevalence of flat and serrated lesions, requires a longer observation time than that of the left colon. To reduce the incidence of PCCRC, it is important to not only increase the ADR achieved by detecting a single adenoma but also reduce the miss rate of neoplasias as much as possible, especially in the right colon. Therefore, although a withdrawal time of at least 6 min was previously recommended, recent studies have demonstrated that 9 min is optimal.

Insufficient bowel preparation leads to an increased rate of missed lesions and a reduced rate of cecal intubation, resulting in an increased risk of PCCRC [3]. According to one study that found an association between bowel preparation and ADR, patients with Boston Bowel Preparation Scale scores of 6 and 9 points had ADRs of 21.8% and 32.6%, respectively, and advanced ADRs of 8.0% and 17.1%, respectively [25]. In other words, there was a difference of more than 10% in the ADR of patients with Boston Bowel Preparation Scale scores of 2 points and 3 points for each section of the colon.

Laxatives are important in improving the quality of bowel preparation. Although polyethylene glycol (PEG) has been used as the standard laxative for bowel preparation, low-volume PEG with ascorbic acid is as safe and effective as PEG and is well-tolerated. A recent review article summarized several studies that demonstrated no significant difference between high-volume PEG and low-volume PEG in terms of bowel preparation efficacy. However, low-volume PEG was associated with better patient tolerability compared to that of high-volume PEG. Moreover, split-dose regimens improved the bowel cleansing quality, and patients were better able to tolerate them than single-dose regimens [26]. A recent meta-analysis of eight RCTs that included 2,059 patients compared the effectiveness of oral sulfate solution with that of PEG with ascorbic acid [27] and found that oral sulfate solution significantly increased both the polyp detection rate (47.34% vs. 40.14%; p = 0.01) and ADR (44.60% vs. 38.14%; p = 0.01), and that this level of efficacy was consistently observed among outpatients with a mean age younger than 55 years and body mass index <25 kg/m² who underwent morning colonoscopy after the use of a 2 L bowel preparation protocol. Although oral sulfate solution is not recommended for patients with several specific conditions, such as severe renal failure or congestive heart failure, it is an effective option because of its clinical efficacy.

The effects of image-enhanced endoscopy, including endoscopy with narrow-band imaging (NBI), linked color imaging, blue laser imaging, and texture and color enhancement imaging, on the ADR have been assessed by several studies. Of these techniques, NBI is the most widely used. The second-generation NBI was equipped with a brighter light source than that of first-generation NBI, resulting in improved image quality. A meta-analysis of 11 RCTs, including three RCTs comprising second-generation NBI, found a significant increase in the ADR with NBI compared to that with white light imaging (OR, 1.14; 95% CI, 1.01–1.29; p = 0.04). In particular, NBI resulted in a significant increase in the ADR of patients with optimal bowel preparation (OR, 1.30; 95% CI, 1.04–1.62; p = 0.02) and in that of patients who underwent colonoscopy with the use of second-generation NBI (OR, 1.28; 95% CI, 1.05–1.56; p = 0.02) [28]. The detection of more flat polyps with NBI than with white light imaging (OR, 1.24; 95% CI, 1.02–1.51; p = 0.03) may have contributed to the decreased incidence of PCCRC, which may have been caused by nonpolypoid lesions [23]. Because the ADR is influenced by several factors, the efficacy of NBI to improve the ADR may vary depending on the clinical setting.

Computer-aided endoscopy with artificial intelligence (AI) has been developed using deep learning systems and could reduce human error and enhance the efficiency and effectiveness of endoscopic examinations. There are two types of colonoscopy with AI: that with computer-aided detection (CADe) and that with computer-aided diagnosis. The meta-analysis of randomized trials have reported significant increases in the ADR and polyp detection rate with the use of colonoscopy with CADe compared to those of colonoscopy without CADe [29]. However, it is important to note that an increase of approximately 10% in the ADR was observed with colonoscopy with CADe, but the detection rate of advanced adenomas did not increase. An RCT of tandem colonoscopy revealed that the adenoma miss rate was significantly lower with colonoscopy with CADe than with colonoscopy without CADe [30]. These results suggest that CADe improves the detection of small polyps; however, its validity with regard to clinically significant lesions has not yet been confirmed. Because the use of CADe is expected to increase the detection of small polyps and increase medical costs, its cost-effectiveness and efficacy in decreasing the incidence and mortality of CRC require further investigation.

Based on previous prospective data regarding the efficacy of sigmoidoscopy and data from cohort studies of colonoscopy, screening colonoscopy could contribute to reduced incidence and mortality rates associated with CRC. However, no RCTs have adequately demonstrated the safety of colonoscopy and its efficacy as a screening tool. Further studies of large populations with specific background factors and long-term observation periods are necessary to assess the efficacy of colonoscopies.

Additionally, to establish evidence of the usefulness of screening colonoscopy, endoscopists must focus attention on the quality of total colonoscopy to improve the ADR and avoid missing lesions. Furthermore, the use of tools such as image-enhanced endoscopy and AI as well as QIs such as withdrawal time and bowel preparation quality should be applied. The continuous efforts of endoscopists to improve the quality of colonoscopy are essential to enhancing its efficacy.

We would like to thank Editage (www.editage.jp) for English language editing.

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

Conceptualization and methodology: Naoya T. and Naoto T.; data curation and writing – original draft preparation: Naoya T.; administration and supervision: K.S.; writing – review and editing: Naoto T.