PROSpective Assessment of Post EmR Recurrence

Colonoscopy and polypectomy reduces the anticipated incidence of colorectal malignancy in patients with significant adenomatous polyps by approximately 80% in long term follow up. Most endoscopists routinely perform removal of small polyps. However, removal of flat colonic neoplasia 20mm in size or larger is more complex and requires specific endoscopic techniques, one such technique being termed wide field endoscopic mucosal resection (WF-EMR). Traditionally these lesions were treated surgically at significant expense to the healthcare system. Endoscopic treatment of large colonic polyps reduces health care costs by approximately $11,000 per patient treated, saves bed days and avoids surgery in more than 90% of patients.

EMR describes the endoscopic technique of treating colorectal adenomatous polyps with submucosal lifting and careful piecemeal snare resection. This procedure has been shown to be safe and effective at resecting lesions limited to the mucosa. Clearly the importance of predicting lesions that are unlikely to have invaded the deeper layers is of the utmost importance here, and significant improvements in our ability to assess this have been made in large volume centres such as Westmead

An important longer-term complication of EMR of large flat colonic neoplasia is the phenomenon of residual polyp tissue or polyp recurrence, which will be the focus of the proposed research study. Recurrence is detected by surveillance colonoscopies (SC), which are performed at defined intervals after the index procedure. Our current standard for safe surveillance interval at Westmead after >= 20mm EMR is 5 months (SC1). In the largest study to date, the Australian Colonic EMR (ACE) study, recurrence at SC1 stands at 16.5% for all patients.

We have recently come to the end of a randomised trial (SCAR) of thermal treatment (snare tip soft coagulation, STSC) of the EMR defect margin (SCAR technique), with the findings to be presented at the Digestive Diseases Week in San Diego. The results of this trial (n=353) are promising with a risk of adenoma recurrence in the treated group of 6.4% vs 20.7% in the non-treated group (relative risk 0.3, p < .001). There have been no adverse events related to this treatment. The investigators plan to perform this intervention on all patients in the proposed study.

In addition, there is increasing evidence, both in the scientific literature and at our institution, that there are identifiable factors at the initial EMR that predict recurrence at SC1. Current data from the ACE cohort suggests that increasing size of lesion (>=40mm), presence of high grade dysplasia (HGD) in the resection specimen and intra-procedural bleeding requiring endoscopic control predict greater likelihood of recurrence. The investigators have created, and have submitted for publication, a risk score for recurrence after EMR known as the Sydney EMR Recurrence Tool (SERT), (figure 1). SERT was created by binomial logistic regression analysis (figure 2) on 692 patients (model cohort) from the ACE study and validated on the remaining 691 patients (validation cohort) who had undergone SC1. Kaplan Meier curves (figure 3) were used to determine the predicted incidence of recurrence at various points in time after EMR on the validation cohort (figure 4). The strength of SERT lies in predicting the absence of recurrence, with negative predictive value of 92.6% at first surveillance colonoscopy. While this score was not derived from patients treated with SCAR technique, it is expected that the stratification of lesions by SERT will continue to be valid with lower overall rates.

Crucial in ensuring the absence of polyp recurrence in the long term is surveillance colonoscopy. In the endoscopy department at Westmead the current standard of follow-up is SC1 at 5 months. This short interval is costly and inconvenient for patients, but has been thought necessary due to previous anxiety over recurrence rates.

The investigators therefore propose a study to prospectively evaluate recurrence after WF-EMR by triaging follow up based on SERT, with all lesions treated by SCAR technique. Patients with SERT 0 lesions would be triaged to first follow up at 18 months (the usual time for SC2) since their rate of recurrence until this point is predicted to be very low (8.5%) and such recurrence is likely to be easily treated endoscopically after this time (in the validation cohort recurrence was commonly diminutive (<5mm, 67%), uni-focal (75%), with no high grade dysplasia (HGD). All recurrences were treatable endoscopically. This compares favourably to the 90.3% successful endoscopic treatment of recurrence rate in a recent meta-analysis.

Higher risk lesions (SERT 1-4) would be triaged to earlier follow up at 6 then 18 months, since their predicted risk of recurrence at 6 months is 14.8% and at 18 months is 31.8%. All lesions would require a follow up procedure at approximately 36 months after the original EMR since both groups rate of recurrence increases by this time (figure 4) and this will be suggested to the referring specialist.

Risk factor Score Laterally Spreading lesion Size >= 40, score 2 Intraprocedural bleeding requiring endoscopic control, score 1 High grade dysplasia, score 1