Endoscopic Evaluation of Duodenal Polyposis in Patients With Familial Adenomatous Polyposis (FAP)

Overview

Familial adenomatous polyposis (FAP) is an autosomal dominant genetic disorder that predisposes to a number or malignant disorders \[1,2\]. Clinically, FAP presents with an abnormal number of colorectal polyps (100-5000), while it genetically is defined by mutations in the APC-gene \[1\]. Historically, colorectal cancer has been the major cause of deaths for FAP patient. However, as the incidence of colorectal cancer has decreased with the use of prophylactic colectomy, the incidence of duodenal cancer has increased \[3,4\]. It is estimated that the cumulative lifetime risk of duodenal polyposis exceeds 95% \[1,5\]. The predictor of duodenal cancer is duodenal polyposis, which is almost inevitable in patients with FAP. In 1989 the Spigelman score was introduced in order to assess the severity of duodenal polyposis and stratify patients according to risk of duodenal cancer (Table 1) \[6\]. It is a composite score that includes two endoscopic parameters (number and maximum size of polyps, respectively) and two histopathological parameters (histological subtype and grade of dysplasia). The score ranges from 0-12 and it has been classified in four stages. The 10-year risk of developing duodenal cancer corresponds with the Spigelman stage ranging from ≈0 for stage 0-1 to 36% for stage 4 \[7\]. Besides duodenal cancer, the indications of cancer prophylactic surgical resection are debatable, but generally recommended in the case of Spigelman stage 4 or high-grade dysplasia. Table 1 Spigelman Classification for duodenal polyposis Criterion 1 point 2 points 3 points Polyp number 1-4 5-20 \>20 Polyp size (mm) 1-4 5-10 \>20 Histology Tubular Tubulovillous Villous Dysplasia Low grade\* High grade\* Stage 0: 0 points; stage I: 1-4 points; stage II: 5-6 points; stage III: 7-8 points; stage IV: 9-12 points. \*Originally, 3 grades of dysplasia were incorporated. While the correlation to cancer has been explored in several studies, the validation and the reproducibility of the Spigelman score remains somewhat unclear. The primary aim of this study is to assess the inter- and intra-observer agreement of the Spigelman score for experienced endoscopists using state-of-the-art high-definition (HD) endoscopes. Hypothesis: The Spigelman score has perfect reproducibility for endoscopic experts (κ\>0.80 with 95% CI.).

The Danish Polyposis Register is sited at Copenhagen University Hospital Hvidovre. From this registry consecutive FAP patients referred for esophagogastroduodenoscopy (EGD) will prospectively be enrolled in this single-center study. Both patients referred for endoscopic surveillance and interventional endoscopy are eligible. All patients will be evaluated with both standard HD gastroscope and side-viewing duodenoscope; the latter to ensure proper visualization of the major papilla. Movie sequences will be saved for post-procedural evaluation. Any sedation or buscopan administered will be registered. A biopsy protocol according to the attached CRF will be followed carefully and furthermore, if any interventional procedures are carried out, they will be registered as well as the presence of any gastric adenomas. In general, the approach is considered standard care and the study does not differ from the treatment algorithm and the national guidelines. Post-procedural, the movies will be evaluated by three expert endoscopists all having long-term experience with FAP patients. In order to assess the inter-observer agreement of the Spigelman score, the endoscopic sub-scores by the three experts will be combined with the histopathological data. Furthermore, the inter-observer agreement of the endoscopic sub-scores will be analyzed separately. To assess the intra-observer agreement, after a minimum of one week one of the expert endoscopists will undertake a second evaluation of the movies after they have been randomized. The precise same methodology will be carried out for three novices. All clinical details will be erased from the movies before evaluation. Design: A blinded single-center prospective observation and methodology study with subsequent calculation of intra-and inter-observer variability. Statistical method: Intra-and inter-observer agreements between 3 operators at for the Spigelman score are calculated by weighted kappa statistics (inter class correlation). Patients: A complete sample-size calculation has been made for the inter-observer study of the 3 observers for the Spigelman Score. To estimate an expected ICC of 0.9 being significantly higher than a threshold at 0.8 with power of 0.8 and a one-sided significance level of 0.05, 33 patients need to be enrolled in the study per protocol.