Roel Bogie

Chapter 12

Quality and effectiveness of colonoscopy is more important than ever in the era of population based screening programs for colorectal cancer (CRC). Asymptomatic people are now receiving colonoscopy as a primary or a confirmatory test to diagnose CRC precursors and CRC at an early stage. Safety and effectiveness of colonoscopy are prerequisites to justify the use of this diagnostic and therapeutic tool in colon cancer screening programs. Studies around the world showed that post-colonoscopy complications, such as colonic perforation or bleeding, have a low occurence. 1 In the Dutch national CRC screening program a fatal complication rate of about two to nine per 100,000 colonoscopies was reported, depending on the estimation method used. 2, 3 However, missed lesions, incomplete procedures, recurrences and preventable additional colonoscopies or surgery should also be considered as an additional risk for patients, as they lead to additional burden and health risks for patients. A large part of this thesis was focused on CRCs that occur after colonoscopies that are negative for CRC, the so-called post-colonoscopy colorectal cancers (PCCRCs). 4 From the perspective of the endoscopist, the risk of PCCRC occurrence may be modifiable. Better detection of high-risk lesions, better therapeutic strategies and more strict surveillance, may reduce the occurrence of PCCRCs and may improve efficacy of nationwide colon cancer screening programs. Laterally spreading tumors (LSTs) are large flat neoplasms which are potential precursor lesions of PCCRCs. 5 LSTs are harder to detect and resect, 6 which could eventually lead to PCCRCs when managed incorrectly. 7 Furthermore, their treatment can be more challenging, posing a burden on patients and on colonoscopy capacity. 8, 9 In the end, better management strategies for LSTs may result in reduced risk of PCCRC and in increased treatment efficacy with reduced burden for patients. In this thesis, ways to optimize quality, safety and efficacy of colonoscopy (and eventually also of the CRC screening program) have been studied from the perspective of LSTs and PCCRCs. We searched for evidence to support our hypothesis that LSTs may be precursor lesions of PCCRCs. By studying LSTs we aimed to achieve better diagnosis, better endoscopic treatment, reduced need for surgery and reduced prevalence of (untreated) recurrence. Understanding the biology of PCCRCs could provide clues on how to reduce the PCCRC incidence. Laterally spreading tumors LSTs are defined as all non-polypoid colorectal neoplasms of minimal 10mm in size. LSTs can be subtyped into four morphological types; two main types based on surface, LST granular and non granular (LST-G and LST-NG resp.), with each two subtypes. The granular LSTs can consist of evenly sized granules, called homogeneous granular, or can have a dominant nodule (sessile component), called nodular-mixed granular (LST-G-H and LST-G-NM resp.). The non-granular LSTs can have a complete flat surface, called flat-elevated, or can have a central depression in the elevated tissue, called pseudo-depressed (LST-NG-FE and LST-NG-PD resp.). This classification is called the Kudo LST classification (see Figure 2.2 ). 5 Subtyping of LSTs has therapeutic consequences, as the risk of containing submucosal invasion differs among subtypes. Therefore, not only experts but all colonoscopists should be able to apply the LST classification in daily practice. In Chapter 2 the Kudo LST classification was tested on interobserver agreement and practical usefulness. Among experts, moderate to substantial interobserver agreement was found (Gwet’s AC1 coefficient: 0.62, 95%CI: 0.55 – 0.69), with overall mean pairwise agreement of 71.0%. Classifying LSTs only into granular versus non-granular resulted in a higher interobserver agreement (Gwet’s AC1 coefficient: 0.75, 95% CI: 0.66 – 0.83) and was considered to be more easy to distinguish. To investigate difficulties in the Kudo LST classification, mean pairwise agreement was calculated per

222