This article is more than 5 years old. Information may no longer be current.
NBI meets ASGE PIVI thresholds for real-time endoscopic assessment of colorectal polyps
Click Here to Manage Email Alerts
The American Society for Gastrointestinal Endoscopy has determined that its Preservation and Incorporation of Valuable endoscopic Innovations, or PIVI, performance thresholds for real-time endoscopic assessment of diminutive colorectal polyp histology, have been met by endoscopists using narrow-band imaging.
Aiming to determine whether the PIVI performance thresholds — which the ASGE has determined advanced imaging technologies should meet before they are implemented in clinical practice — have been met, researchers on the ASGE Technology Committee performed a systematic review and meta-analysis of the existing literature. The meta-analysis included 20 studies evaluating narrow-band imaging (NBI; Olympus), eight evaluating i-SCAN (Pentax) and eight evaluating Fujinon Intelligent Color Enhancement (FICE; Fujinon Inc.), all published between 2008 and 2014.
The researchers calculated the pooled negative predictive value (NPV) for these technologies when used for optical biopsy to predict adenomatous polyp histology of small or diminutive colorectal polyps, and also calculated the pooled percentage agreement with histopathology for postpolypectomy surveillance interval assignment based on real-time optical biopsy of colorectal polyps ≤ 5 mm combined with histopathologic assessment of polyps > 5 mm.
Meta-analyses of NBI studies
The studies included in the meta-analysis evaluating NPV for adenomatous polyp histology with NBI collectively evaluated 4,013 in situ small or diminutive colorectal polyps. The pooled NPV of optical biopsy with NBI was 91% (95% CI, 88-94). Subgroup analysis revealed pooled NPV was:
- 91.8% (95% CI, 89-94) at academic centers vs. 88.3% (95% CI, 82-94) at community practices;
- 93% (95% CI, 91-96) for expert operators vs. 87% (95% CI, 83-91) for novice operators;
- 93% (95% CI, 90-96) with high-confidence assessment vs. 88% (95% CI, 84-92) with no information on confidence level; and
- 90% (95% CI, 86-94) for novice operators with high-confidence assessment vs. 95% (95% CI, 92-98) for experienced operators with high-confidence assessment.
The studies included in the meta-analysis evaluating agreement in assignment of postpolypectomy surveillance intervals based on optical biopsy with NBI vs. histopathology including 3,082 patients. The pooled percentage of agreement was 89% (95% CI, 85-93). Subgroup analysis revealed agreement was:
- 91% (95% CI, 86-95) in academic centers vs. 82% (95% CI, 74-90) in community practices;
- 92% (95% CI, 88-96) for expert operators vs. 82% (95% CI, 75-88) for novice operators;
- 91% (95% CI, 88-95) with high-confidence assessment vs. 79% (95% CI, 71-86) with no information on confidence level; and
- 87% (95% CI, 82-93) for novice operators with high-confidence assessment vs. 93% (95% CI, 90-96) for expert operators with high-confidence assessment.
Meta-analyses of i-SCAN and FICE studies
The studies included in the meta-analysis evaluating NPV for adenomatous polyp histology with i-SCAN collectively evaluated 979 small or diminutive colorectal polyps. The pooled NPV of optical biopsy with i-SCAN was 84% (95% CI, 76-91). Subgroup analysis revealed pooled NPV was 96% (95% CI, 94-98) for expert operators vs. 72% (95% CI, 69-76) for novice operators. Only one study evaluated agreement in assignment of postpolypectomy surveillance intervals based on optical biopsy with i-SCAN vs. histopathology, and the percentage of agreement was 69.5% (95% CI, 63-75).
The studies included in the meta-analysis evaluating NPV for adenomatous polyp histology with FICE collectively evaluated 1,234 small or diminutive colorectal polyps. The pooled NPV of optical biopsy with FICE was 80% (95% CI, 76-85). Subgroup analysis revealed pooled NPV did not improve with operator experience, but use of magnification did (85% CI; 95% CI, 79-91). Only two studies evaluated agreement in assignment of postpolypectomy surveillance intervals based on optical biopsy with FICE vs. histopathology, and the percentages of agreement were 100% (95% CI, 91-100) and 97% (95% CI, 89-100).
Data on confocal laser endomicroscopy (CLE) was limited, but of the studies that assessed adenomatous polyp histology in real time with CLE, NPVs were 79%, 88%, 92% and 100%.
This study “confirms that the thresholds established by the ASGE PIVI for real-time endoscopic assessment of the histology of diminutive polyps have been met, at least for NBI, with endoscopists who are experts in using these advanced imaging technologies and when assessments are made with high confidence,” the committee researchers wrote. “The ASGE Technology Committee therefore endorses the use of NBI for both the ‘diagnose-and-leave’ strategy for diminutive rectosigmoid hyperplastic polyps and the ‘resect-and-discard’ strategy for diminutive adenomatous polyps by endoscopists trained in using this technology for polyp characterization, making assessments with high confidence.
“Our analyses indicate that the PIVI thresholds can potentially also be met by other advanced imaging technologies including i-SCAN, FICE and CLE,” they added. “However, given the limited data on the use of these technologies in the characterization of diminutive colorectal polyps, further studies need to be performed.” – by Adam Leitenberger
Disclosure: Some of the committee researchers report financial relationships with Mauna Kea Technologies, Olympus and Pentax. Please see the study for a full list of all authors’ relevant financial disclosures.
Perspective
Back to Top
Douglas K. Rex, MD
The ASGE PIVI process is designed to focus research in endoscopic technologies toward clinically meaningful purposes. The meta-analysis by the ASGE Technology Assessment Committee begins to bring the colorectal polyp PIVI full circle by evaluating whether available evidence indicates that specific technologies have reached the performance thresholds originally proposed in the PIVI. The committee found that the performance thresholds have been met, particularly for narrow band imaging (NBI). Part of the explanation for why the meta-analysis could reach the firmest conclusions about NBI is that the volume of data for NBI far exceeds the other technologies. In the case of resect-and-discard, the document found that the PIVI thresholds have been met in NBI studies from academic but not community centers.
As real-time prediction of polyp histology stands on the threshold of entering clinical practice, community colonoscopists can be reassured that paradigms such as resect-and-discard will not be forced on them. Rather, the new management paradigms for diminutive polyp management that rely on real-time pathology interpretation will be an accepted alternative to current management. Clinicians may choose to participate based on their inclination, training and financial incentives. Further, because of the variation with which the PIVI thresholds have been met, we can anticipate that the thresholds may change from technology based to individual performance based thresholds. Thus, if colonoscopists choose to practice resect-and-discard, let them prove they are capable, regardless of what technology they use.
Click Here to Manage Email Alerts