Probe-based confocal laser endomicroscopy aids the detection of residual colorectal neoplasia and small colorectal polyps

Two studies from Michael Wallace and colleagues indicate that probe-based confocal laser endomicroscopy (pCLE) combined with narrow band imaging (NBI) could aid the assessment of patients who have residual colorectal neoplasia after endoscopic mucosal resection (EMR) and those with small (<10 mm) colorectal polyps. If validated, histological examination of biopsy samples may not be needed and the ability to make real-time management decisions could become a reality.

In the first study, researchers from the USA, France and The Netherlands assessed whether pCLE could be used to make a specific diagnosis of residual colorectal neoplasia after EMR and hence guide therapy. “Approximately 10–20% of individuals have some amount of residual polyp after EMR and thus there is a need for close follow-up and re-resection or ablation in these individuals,” explains Wallace. At follow-up it can be difficult to assess if there is residual neoplasia that needs re-treating, owing to scar tissue and potential granulation or inflammatory tissue at the EMR site. Currently, histologic examination of biopsy samples is the only definitive way to detect residual neoplasia, but treatment decisions have to be made independently, because biopsy results are not available at that time.

 Adenomatous colorectal polyp (top) and non-neoplastic colon tissue (bottom) seen by pCLE. Courtesy of M. Wallace.

92 patients with 129 colorectal EMR scars took part (EMR was in the previous year). Histologic examination identified residual neoplasia at 29 sites (22%). Individually, pCLE and NBI were quite accurate, but their combination proved best—with 90% accuracy, 100% sensitivity, 87% specificity, a 67% positive predictive value and a 100% negative predictive value. “If NBI plus pCLE were negative the physician could feel very reassured when making a decision to perform no further treatment and also avoid unnecessary empiric treatment ... it would also allow a significant reduction in the number of repeat colonoscopies,” clarifies Wallace.

As three academic referral centers with considerable imaging expertise took part in this study, Wallace notes the need to validate the findings in other centers to determine the true accuracy in clinical practice. “We also need to further evaluate the implications of an in vivo microscopy approach given the fact that there are some important trade-offs involved between repeat colonoscopies versus the downside of a slight increase in over-treatment of non-neoplastic tissue.”

In the second study, a US-based team assessed whether pCLE could distinguish non-neoplastic from neoplastic small colorectal polyps, which can be visualized thanks to advances in endoscopic imaging. Histologic examination of small colorectal polyps is the standard; however, it is costly and its value is unclear as the risk of neoplasia is extremely low. “Several groups have proposed a strategy where small polyps are diagnosed based on endoscopic appearance and then removed and discarded without further histologic characterization,” explains Wallace, but no candidate imaging technology is currently accurate enough to replace histology.

65 patients with a total of 130 colorectal polyps were included in this study. Histological examination identified 58 polyps as neoplastic and 72 as non-neoplastic. When compared with the histopathology findings, pCLE had a higher sensitivity but a lower specificity than NBI, but both had similar overall accuracy. For 65 polyps, pCLE findings could be combined with those of NBI (high-quality pCLE videos were available and the same diagnosis was made with both techniques)—this combination conferred 95% accuracy, 94% sensitivity, 97% specificity, a 97% positive predictive value and a 94% negative predictive value compared with histopathology findings.

“When used in combination ... these techniques achieved the current threshold level of accuracy as recommended by the American Society for Gastrointestinal Endoscopy of a >90% negative predictive value for adenomas,” says Wallace.

Again, these results need to be reproduced at other centers and with larger sample sizes, and Wallace notes that improvements in probe stabilization and a reduction in the costs of the device are also needed. Nonetheless, the potential of combined pCLE and NBI in this setting is clear—a reduction in the risk of removing non-neoplastic polyps and in the costs associated with colonoscopy and polypectomy.