Narrow band imaging avoids potential chromoendoscopy risks

We read with interest the letter by Davies et al (Gut 2007;56:155–6) extending their observations regarding the potential carcinogenicity of methylene blue from the oesophagus to the colon. The use of methylene blue for pan‐colonic chromoendoscopy has been proposed to increase dysplasia detection in ulcerative colitis, and it is also used for polyp characterisation in vivo, producing clearer pit patterns than indigo carmine, potentially providing “endohistology”.¹ A safe, rapid, simple method for in vivo endoscopic polyp characterisation could allow substantial reductions in the histopathological workload if formal histopathology were abandoned for small polyps (<10 mm), as 90% of polyps resected are <10 mm and rarely contain cancer. Although no cases of carcinoma were reported in diminutive polyps (<5 mm) in a large series, adenomas 5–10 mm contained carcinoma in approximately 1% of cases.² Carcinoma can be differentiated by the presence of Kudo type V pit pattern.²

The apparently ideal properties of methylene blue for polyp characterisation are offset by the risk of possible DNA damage. In practice, the requirement for mucolytics and dye‐spray catheters for optimal results, and the permanent staining of all surfaces contacted, make the routine use of methylene blue impractical. Narrow band imaging (NBI) “electronic dye‐spray”, a new optical technology with rapid push‐button activation, can provide similar images to methylene blue dye‐spray. NBI may help with dysplasia detection at colonoscopy, even in colitis, and Chiu et al have recently reported in Gut that NBI can characterise polyps with a high degree of accuracy, without the disadvantages of dye‐spray.³,⁴ This corresponds to our own early experience with NBI for polyp characterisation, where false positives for larger adenomas were very rare and sensitivity was maintained even for diminutive lesions (range 77–91%, polyps ⩽6 mm).⁵ However, Chiu et al included only one carcinoma of unknown size, and it therefore remains unclear whether NBI can accurately characterise small polyp carcinomas where endoscopic management is different and formal histopathology is mandatory.³ We report a case where NBI was able to characterise an 8 mm carcinoma in vivo with appropriate modification of the endoscopic approach.

A 65‐year‐old man who had had a low anterior resection for a Duke's B rectal carcinoma 4 years previously presented for his third surveillance colonoscopy. A high definition (HDTV) colonoscope was used which had the capacity to magnify up to 100 times with a prototype narrow band imaging system (XCF‐H240FZL/I video colonoscope, XCLV‐260HP xenon light source and XCV‐260HP video system centre; Olympus, Japan). During withdrawal an 8 mm polyp was seen in the sigmoid colon on white light examination. The polyp was sub‐pedunculated on a broad‐based stalk with a slightly depressed area on one side of the polyp head (fig 1A). NBI with magnification of the depressed area showed a Kudo type V non‐structural pit pattern with type III_L pit pattern adjacent (fig 1B). In view of the likelihood of invasive carcinoma, the polyp was carefully snare resected lower on the stalk than usual to try to ensure a clear excision margin and two tattoos were placed immediately distal to the polypectomy site. Histological examination confirmed an adenocarcinoma with a clear resection margin G2, pT1, pV1, Nx, Mx, R0, Haggitt level 2 (invasion of the polyp neck), with high risk features of vascular invasion and tumour budding (fig 1C). A CT scan of the thorax and abdomen revealed no local lymphadenopathy or distant metastasis. In view of multiple co‐morbidities, it was decided to view the endoscopic resection as definitive treatment.

Figure 1 (A) An 8 mm sigmoid polyp with depressed area on the polyp head, high definition (HDTV) white light image, snare catheter adjacent to polyp. (B) Narrow band imaging with magnification showing loss of pit pattern in depressed area (Kudo type V, non‐structural, arrow) adjacent to type III_L pattern (arrowhead). (C) Histology of the polyp showing that adenocarcinoma had developed within the polyp head but the resection margin was clear (H&E, magnification ×25). The inset shows tumour budding, indicating that the polyp was high risk (H&E, magnification ×400).

This case suggests, in combination with the report of Chiu et al, that NBI is able to fulfil the characterisation qualities of methylene blue for small polyps in a more time‐efficient way and without the need for potentially carcinogenic dyes.³ Colonoscopists should prepare themselves for the era of routine “endohistology” as NBI becomes widespread.

Acknowledgements

The authors thank Olympus Keymed, UK and Olympus Medical Systems Corporation, Japan for the loan of the prototype narrow band imaging system.