Sa1834 Identification and Characterization of Subsquamous Intestinal Metaplasia Pre- and Post-Radiofrequency Ablation Using Three-Dimensional Optical Coherence Tomography

Chao Zhou,Tsung-Han Tsai,Hsiang-Chieh Lee,Tejas Kirtane,Marisa Figueiredo,Yuankai K. Tao,Osman O. Ahsen,Desmond C. Adler,Joseph M. Schmitt,Qin Huang,James G. Fujimoto,Hiroshi Mashimo
DOI: https://doi.org/10.1016/s0016-5085(12)61270-0
IF: 29.4
2012-01-01
Gastroenterology
Abstract:criteria to differentiate squamous mucosa from gastric cardia and BE, a training set composed of 19 images was compiled with features distinguishing these three types of mucosa.The 10 readers reviewed the training set and subsequently interpreted images in a test set of 45 different images as squamous mucosa (n=15), gastric cardia (n=15), or BE (n=15).All readers had no formal training prior to viewing the training set.Four readers had experience with OFDI images prior to the exam.Interobserver agreement for differentiating the three tissue types and BE vs. non-BE mucosa was determined using multi-rater Fleiss's κ value.The images were later randomized and four readers repeated the test three weeks later to assess intraobserver reliability.Results: For all 10 readers, the mean time to completion of the test was 25.1±18.4minutes or 33.5±11.2seconds/image.There was excellent agreement between all 10 readers for OFDI diagnosis of BE vs. non-BE (κ= 0.811, p<0.0001) and for OFDI diagnosis of BE vs. gastric cardia vs. squamous mucosa (κ= 0.866, p<0.0001).Interobserver agreement for those with prior OFDI experience was excellent (κ =0.872, p<0.001).For those with limited OFDI exposure, interobserver variability between the six readers was also good (κ= 0.765, p<0.001).The 4 readers who repeated the test showed excellent intraobserver agreement of diagnosing BE vs. non-BE (median κ= 0.975, IQR: 0.94, 1.0).Heterogeneous backscatter and irregular surface morphology were the two features most commonly associated with the diagnosis of BE.Conclusions: In this study we have demonstrated that trained readers have a high interobserver agreement for using OFDI to discriminate BE, squamous mucosa, and gastric cardia and BE from non-BE tissues.Future studies will assess real time accuracy of OFDI during endoscopy to detect BE and grade dysplasia.
What problem does this paper attempt to address?