We read with keen interest the study by Ortiz et al. evaluating the diagnostic utility of linked color imaging (LCI) in detecting recurrence following piecemeal endoscopic mucosal resection (p-EMR) of large nonpedunculated lesions [11]. This study reinforces the diagnostic performance of optical endoscopic techniques, which we find commendable. However, we have some concerns worth addressing.

First, the histologic recurrence rate of 32% is higher compared with previous studies. While the authors attribute this to indistinct lesion margins, the p-EMR technique itself for larger lesions may increase local recurrence risk, necessitating early endoscopic surveillance [22] [33]. Providing the R0 resection rate after p-EMR at baseline could better explain these findings. Moreover, the lack of clarity regarding postoperative endoscopic surveillance timing, coupled with the higher pathologic grades in the included population, could potentially inflate the positive histologic rate. Statistically, with increased disease prevalence, the positive predictive value and negative predictive value may be impacted, affecting the overall diagnostic performance assessment.

Furthermore, optical diagnosis techniques demand specialized training to achieve optimal diagnostic accuracy, which may be challenging in regular endoscopy centers relying on pathologic diagnosis. Artificial intelligence (AI) has demonstrated promising diagnostic performance for gastrointestinal adenomas, polyps, and laterally spreading tumors, and the combination of LCI and AI (LCA) has proven to improve adenoma detection rates for all endoscopists in colon examinations [44]. Therefore, further research evaluating the diagnostic performance of LCA in detecting recurrence after polypectomy may be more clinically applicable and may potentially maximize the value of optical diagnosis for a broader range of endoscopists.

Article published online:
30 October 2024

© 2024. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany