Evaluation of narrow-band imaging in the diagnosis of colorectal lesions: is a learning curve involved?

BACKGROUND: The usefulness of endoscopy narrow-band imaging (NBI) in differentiating colorectal lesions has been demonstrated. However, the learning curve associated with this technique is a concern for endoscopists. METHODS: Prior to carrying out these colonoscopies, four endoscopists attended a training course designed to teach the principles of NBI and application of the Sano Capillary Pattern (CP) classification criteria. Following a pre-test, endoscopists used NBI with magnification and CP analysis for real-time colonoscopy exams to predict lesion histology. Three sets of 15 lesions were imaged. These three sets included both lesions requiring endoscopic treatment (e.g. target lesions) and lesions that were not, or could not be, treated by endoscopy (e.g. non-target lesions). The diagnostic accuracy of each endoscopist for each set of lesions was evaluated to assess the learning curve associated with the application of NBI. RESULTS: Overall accuracy, sensitivity, and specificity for differentiating neoplastic and non-neoplastic lesions were 95.4%, 98.0%, and 92.0%, respectively. For target lesions versus non-target lesions, the diagnostic accuracy associated with the second set of lesions was better than that achieved with the first set of lesions (78.3% vs 96.7% (P = 0.02) and 70.0% vs 96.7% ( P < 0.01), respectively in each case). In contrast, the difference in diagnostic accuracy between the second and third sets of lesions was not significant. CONCLUSION: NBI with magnification is a useful tool for the diagnosis of colorectal lesions. Moreover, following a short training program and with minimal clinic practice, less experienced endoscopists were able to become competent in the method.