Adding Narrow-Band Imaging to Chromoendoscopy for the Evaluation of Tumor Response to Neoadjuvant Therapy in Rectal Cancer

BACKGROUND: Endoscopic assessment is crucial in diagnosing clinical complete response after neoadjuvant therapy in rectal cancer. OBJECTIVE: The purpose of this research was to evaluate the benefits of adding narrow-band imaging endoscopy to conventional chromoendoscopy in predicting pathologic complete response in the surgical specimen. DESIGN: This was a prospective nonrandomized study. SETTINGS: This was an ad hoc study of a prospective phase II trial at a single comprehensive cancer center that evaluated oncologic outcomes of a neoadjuvant therapy for rectal cancer. PATIENTS: Patients with high-risk stage II to III low rectal cancer who received neoadjuvant modified folinic acid, fluorouracil, and oxaliplatin plus bevacizumab followed by chemoradiotherapy and surgery were included. INTERVENTION: Tumor response after neoadjuvant therapy was evaluated using conventional white light endoscopy plus chromoendoscopy then followed by using narrow-band imaging based on a predefined diagnostic protocol. MAIN OUTCOME MEASURES: Diagnostic accuracy for predicting pathologic complete response and inter-rater agreement between an expert and trainee endoscopists were compared between the assessments using conventional white light endoscopy plus chromoendoscopy and the assessment adding narrow-band imaging. RESULTS: In total, 61 patients were eligible for the study, and 19 had pathologic complete response (31.1%). Although the addition of narrow-band imaging correctly converted the diagnosis in 3 patients, overall diagnostic improvement in predicting pathologic complete response was limited (conventional chromoendoscopy vs adding narrow-band imaging: accuracy, 70.5% vs 75.4%; sensitivity, 63.2% vs 73.7%; specificity, 73.8% vs 76.2%; positive predictive value, 52.2% vs 58.3%; and negative predictive value, 81.6% vs 86.5%). A κ value for the inter-rater agreement improved from 0.599 to 0.756 by adding narrow-band imaging. LIMITATIONS: This was a single-center study with a relatively small sample size. CONCLUSIONS: Despite the limited improvement in diagnostic accuracy, adding narrow-band imaging to chromoendoscopy improved inter-rater agreement between the expert and nonexpert endoscopists. Narrow-band imaging is a reliable and promising modality for universal standardization of the diagnosis of clinical complete response.