Colorectal cancer (CRC) is the third most commonly diagnosed malignancy and the fourth leading cause of cancer related mortality worldwide. Most colorectal cancers arise form premalignant colorectal polyps. Colorectal polyps can be roughly subdivided into hyperplastic, non-malignant polyps and neoplastic, pre-malignant polyps, called adenomas. The development of CRC from adenomas is thought to occur over 10 years on average, creating a window of opportunity to detect and treat colorectal adenomas. Adenomas can be detected and removed during colonoscopy, which has been shown to reduce the incidence and mortality of CRC. This reduction depends on the adenoma detection rate, but also on the completeness of endoscopic resection. In this thesis we investigated the detection, differentiation and endoscopic treatment of colorectal adenomas and early carcinomas.
In the first part of the thesis, we investigated routine colonoscopies and detected significant differences in quality of colonoscopy between hospitals. Adequate bowel preparation was the most important predictor of adenoma detection and improving bowel preparation should therefore be a main target when aiming to improve quality of colonoscopy. As it is not yet possible to reliably classify detected polyps during colonoscopy, all resected colorectal polyps have to be retrieved for histologic evaluation. International guidelines recommend retrieval rates of at least 90%. We successfully improved the polyp retrieval rate in one endoscopy center from 88% to 93% by using competitive feedback to endoscopists. We also investigated whether it would be possible to differentiate between hyperplastic polyps and adenomas during colonoscopy, using probe-based confocal laser endomicroscopy (pCLE). pCLE uses laser light to create enlarged images of polyps in vivo. The accuracy for two endoscopists starting to use pCLE was however not high enough to allow reliable differentiation during colonoscopy.
In the second part of the thesis, we performed a meta-analysis to assess outcomes after endoscopic mucosal resection (EMR) of non-pedunculated adenomas and early carcinomas. We found that the risk of local recurrence was higher after EMR in multiple fragments (piecemeal, 20%) than after EMR in one piece (3%). Follow-up after piecemeal EMR should therefore be performed after 6 and 12 months. In addition, we showed that in a nationwide cohort study, the risk of CRC arising from incompletely resected adenomas was significant, occurring after 1 of 400 adenomas and 1 of 170 advanced adenomas. Furthermore, we investigated long-term outcomes after resection of submucosal invasive (T1) CRC. We found that histologic risk factors predicted the presence of lymph node metastasis and that a positive margin of the resected specimen was the most important risk factor for long-term recurrence. Therefore, additional surgery after initial endoscopic resection of T1 CRC should be performed in case of histologic risk factors and/or positive resection margins.
In the third part of this thesis, we conducted a placebo-controlled trial to investigate the preventive effects of mesalazine in sporadic colorectal adenoma patients. Compared to placebo, mesalazine had no significant effect on proliferation or apoptosis in normal colorectal mucosa, but the increased apoptosis lower crypt regions in mesalazine-treated patients might still suggest chemopreventive properties of mesalazine.