Is additional acetic acid chromoendoscopy effective for prediction of histology of submucosal invasion depth in colonic neoplasms?

Aims Acetic acid visualizes the detailed surface of the colonic polyps when combined with image-enhanced endoscopy (IEE). We have reported its ability to enhance pit pattern as an acetoelectronic chromoendoscopy (AEC). Although AEC is expected to be helpful in the pretreatment inspection of colonic neoplasms, including the degree of submucosal invasion, the efficacy of AEC has not been thoroughly investigated. Our aim is to validate the diagnostic value of adding AEC to conventional observation (CO) by physicians with various background experiences in pretreatment assessment of colonic neoplasms, including submucosal invasive lesions. Methods Endoscopic still images of 30 colonic neoplastic lesions were collected between April 2020 and July 2023 and were assessed on pretreatment diagnosis by nine endoscopists: three Western expert endoscopists (WE) with a lot of experience in endoscopic treatment but with few experiences of magnifying pit pattern observation using Christal Violet (CV), three Japanese expert endoscopists (JE) with a lot of experience with both treatment and observation using CV, and three residents in training (RT) with few experience of both of them. Of 30 lesions, 63% (19/30) were lesions indicated for endoscopic resection (LER) (3 low-grade dysplasia (10%), 13 high-grade dysplasia (43%), 3 T1a (10%)), and 37% (11/30) were lesions indicated for surgical resection (LSR) (9 T1b (30%), and 2 T2 (6.7%)). Median size was 20mm (range 5-90). 86% were elevated or flat elevated, including lateral spreading tumours, and 14% were depressed lesions. For each lesion, two sets of endoscopic images were prepared with representative images; one was with white light imaging (WLI) and IEE (conventional observation, CO), and the other was with WLI, IEE, and AEC (CO+AEC). Participants were asked to classify each lesion into LER or LSR and specify high/low confidence of assessment. The primary outcome was an overall accuracy with or without AEC, and the kappa value was calculated for interobserver agreement. Results Overall accuracy of CO and CO+AEC was 75.5% (70-83.3) and 76.7% (73.3-80), 65.6% (56.7-76.7) and 72.2% (70-73.3), and 68.9% (63.3-73.3) and 66.7% (56.7-76.7) for JE, WE, and RT, respectively. Sensitivity and specificity for LSR (CO/CO+AEC) were 59.3%/61.8% and 77.3%/81.5%. Among evaluations with high confidence, JE showed the highest accuracy at 89.3% with CO and 86.7% with CO+AEC compared to WE (69.8% and 81.2%) and RT (77.5% and 74.9%). Interobserver agreement (CO/CO+AEC) was moderate ( k =0.58/0.57) in JE and poor in WE ( k =0.26/0.28) and in RT( k =0.17/0.3) [ 1 ] [ 2 ]. Conclusions Additional effect of acetic acid for predicting colonic neoplasms, including submucosal invasion depth, was limited in all three groups of physicians with different background experiences. Highly experienced with magnifying observation might have contributed to higher interobserver agreement in pretreatment diagnosis. Publication History Article published online: 15 April 2024 © 2024. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany