How is narrow-band imaging (NBI) used to evaluate and manage Barrett's esophagus, including detection of dysplasia and biopsy strategy?

Narrow-Band Imaging in Barrett's Esophagus

NBI should be used during Barrett's esophagus surveillance to enhance detection of dysplasia and guide targeted biopsies, but cannot replace Seattle protocol random biopsies. 1, 2

Role of NBI in Barrett's Surveillance

NBI is an image-enhanced endoscopy technology that improves visualization of mucosal glandular structures and vascular patterns in Barrett's esophagus. The evidence demonstrates:

Detection Performance

NBI detects 10-20% more dysplastic lesions compared to high-definition white light endoscopy alone 1. In the most rigorous tandem endoscopy study, NBI detected dysplasia in 57% of patients versus 43% with standard endoscopy, and identified higher grades of dysplasia in 18% of patients that were completely missed by white light examination 3. The meta-analysis confirms excellent per-patient sensitivity (91%) and specificity (95%) for high-grade dysplasia detection 4.

Biopsy Strategy

NBI enables targeted biopsies with fewer samples while maintaining or improving diagnostic yield 3. The tandem study showed NBI required an average of 4.7 biopsies versus 8.5 with random protocol, yet detected more dysplasia 3. However, NBI-guided biopsies cannot safely replace Seattle protocol four-quadrant random biopsies every 2 cm 2, 5. The systematic review conclusively demonstrates that neither NBI nor other advanced imaging can substitute for random biopsies 2.

Practical Implementation Algorithm

During Surveillance Endoscopy:

1. Initial inspection with high-definition white light endoscopy - ensure adequate mucosal visualization by aspirating all fluid/debris 1

2. Switch to NBI for systematic inspection - use as "red flag" technique to identify suspicious areas 6

3. Apply validated classification systems when abnormalities detected:

   • BING criteria (Barrett's International NBI Group) - identifies dysplasia with 85% overall accuracy, improving to 92% when assessed with high confidence 7
   • Simplified NBI classification - provides 95% accuracy with substantial inter-observer agreement (κ = 0.80) 8

4. Biopsy protocol:

   • Take targeted biopsies of all NBI-suspicious areas first 6, 5
   • Then perform mandatory Seattle protocol random four-quadrant biopsies every 2 cm regardless of NBI findings 2, 5

NBI Pattern Recognition

The validated BING criteria distinguish:

• Regular mucosal and vascular patterns = non-dysplastic Barrett's
• Irregular mucosal patterns with irregular microvasculature = high-grade dysplasia/early cancer 7, 6

Critical Caveats

Operator expertise matters significantly - the 2008 AGA guideline notes methylene blue staining was operator-dependent, and similar considerations apply to NBI interpretation 9. The simplified classification systems were developed specifically to improve reproducibility 8.

NBI works best for visible lesions - preliminary studies from 2008 showed NBI enhances visualization of surface glandular structures, but this was early evidence 9. The more recent 2016 BING validation and 2024 AGA update provide stronger support for clinical use 7, 1.

Cannot eliminate random biopsies - despite NBI's superior detection, the 2025 systematic review definitively shows it cannot replace Seattle protocol 2. The 2023 ESGE guideline maintains the requirement for random four-quadrant biopsies during surveillance 5.

Post-Eradication Surveillance

After successful endoscopic eradication therapy, NBI should be used to carefully inspect the neo-squamocolumnar junction and neo-squamous epithelium 5. Take targeted biopsies of any visible abnormalities, plus four-quadrant random biopsies just distal to the neo-squamocolumnar junction to detect occult dysplasia 5.

Minimum Standards

The 2023 ESGE guideline mandates minimum 1-minute inspection time per cm of Barrett's length during surveillance, with photodocumentation of the segment including one picture per cm 5. Use NBI or other image-enhancement technologies at minimum to characterize abnormalities seen on white light examination 1.