We thank Saito et al for their interest in our paper that prospectively examined the safety and efficacy of high magnification chromoscopic colonoscopy for the diagnosis of neoplasia in flat and depressed lesions of the colorectum (Gut 2004;53:284–90).

Their abstract data,¹ using the modified Paris classification of superficial neoplastic lesions,² as determined by the newly designated 1000m vertical mucosal extension limit, offers impressive sensitivity and overall accuracy rates when distinguishing sm1 from sm2 invasion. We look forward to publication of the full manuscript soon.

However, fundamental differences still exist between Japanese and Western histopathologists that make comparative studies between East and West difficult.³ Our histopathology was interpreted using the Vienna classification of gastrointestinal epithelial neoplasia, as proposed by Schlemper et al following the consensus classification workshop.⁴ The consensus terminology differentiates between non-invasive low grade neoplasia, non-invasive high grade neoplasia (which includes suspicion of invasive neoplasia), and invasive neoplasia (intramucosal carcinoma, submucosal carcinoma, or beyond) into Vienna categories 3, 4, and 5, respectively.⁴ Hence in table 4 (Gut 2004;53:284–90), our neoplastic invasive group includes both Vienna 4 and 5 lesions together. This “collection” of data reflects current histopathological practice in the UK where sm invasion rates (1–3) are not routinely specified. While we accept that the modified Paris guidelines² are extremely helpful, our current practice is aimed at overall patient “risk stratification” where the natural history of low grade dysplasia (LGD) has yet to be determined but that of high grade dysplasia is more defined; assuming a high risk of progression to invasive neoplasia.⁵ Differentiation of hyperplastic (non-neoplastic/non-invasive) from LGD adenoma (neoplastic/non-invasive) therefore still remains an important asset to the endoscopists (sensitivity 98%/specificity 92% in our series), as the number of inappropriate biopsies and resections can be safely limited while allowing for appropriate risk stratification permitting suitable colonoscopic follow up intervals.⁶

We agree with the recent data of Hurlstone et al who have demonstrated that sm superficial lesions without lymphovascular invasion or poorly differentiated histological features have a low incidence of local nodal metastasis.⁷

While we acknowledge the concerns raised by Saito et al, we feel that despite adopting the Paris guidelines, our data still provide a comprehensive guide to high magnification chromoscopic colonoscopy, as is practiced within the UK.