**4.2 Strategy of 'DISCARD-ME' policy**

Where the 'DISCARD-ME' policy has been adopted, when colonoscopists have detected a colorectal polyp during a screening colonoscopy, they can predict the polyp type (nonneoplastic, low grade adenoma, suspicious of high grade adenoma or carcinoma) by careful observation using NBI-ME and the above-described criteria. In addition to predicting histopathology, colonoscopists can make the following decisions for polyp management on the basis of the optical diagnosis using NBI-ME (Fig. 11): (1) whether to 'resect and discard' polyps (for serrated lesions in the proximal colon or low-grade adenomas; no formal histopathology required); (2) whether to 'resect and send' them for histopathology (if they cannot decide on the type of polyp or are concerned about high-grade adenoma or carcinoma), or (3) whether to 'leave it *in situ*' (for diminutive recto-sigmoid non-neoplastic lesions).

Fig. 11. Strategy of the 'DISCARD-ME' policy

#### **4.3 A 'proof of principle' pilot study for the 'DISCARD-ME' policy**

A prospective 'proof-of-principle' pilot study was conducted by the present authors to investigate the feasibility of the 'DISCARD-ME' policy. Forty-one patients undergoing colonoscopy for investigation of a positive screening FOBT, or who had been referred for surveillance colonoscopy after endoscopic resection of colorectal neoplasms, were enrolled.

When the microvascular architecture cannot be assessed, the pit pattern classification of surface pattern is applied, because Hirata et al. have reported that determination of the pit patterns of colorectal neoplasia by NBI-ME is almost the same as that achieved by standard magnification with chromo-endoscopy (Hirata, et al., 2007, Fig. 10). According to the pit pattern classification, lesions with Type I and II pit patterns are categorized as nonneoplastic, and lesions with Type III, IV and V pit patterns as neoplastic (Kudo, et al., 1994). Neoplastic lesions with a Type III pit pattern are categorized as low-grade adenomas and lesions with Type IV and V pit pattern as high-grade adenomas, villous adenomas or carcinomas. In cases where different histologic categories have been assigned by the CP and

Where the 'DISCARD-ME' policy has been adopted, when colonoscopists have detected a colorectal polyp during a screening colonoscopy, they can predict the polyp type (nonneoplastic, low grade adenoma, suspicious of high grade adenoma or carcinoma) by careful observation using NBI-ME and the above-described criteria. In addition to predicting histopathology, colonoscopists can make the following decisions for polyp management on the basis of the optical diagnosis using NBI-ME (Fig. 11): (1) whether to 'resect and discard' polyps (for serrated lesions in the proximal colon or low-grade adenomas; no formal histopathology required); (2) whether to 'resect and send' them for histopathology (if they cannot decide on the type of polyp or are concerned about high-grade adenoma or carcinoma), or (3) whether to 'leave it *in situ*' (for diminutive recto-sigmoid non-neoplastic

pit pattern classifications, the more severe category is adopted.

**4.2 Strategy of 'DISCARD-ME' policy** 

Fig. 11. Strategy of the 'DISCARD-ME' policy

**4.3 A 'proof of principle' pilot study for the 'DISCARD-ME' policy** 

A prospective 'proof-of-principle' pilot study was conducted by the present authors to investigate the feasibility of the 'DISCARD-ME' policy. Forty-one patients undergoing colonoscopy for investigation of a positive screening FOBT, or who had been referred for surveillance colonoscopy after endoscopic resection of colorectal neoplasms, were enrolled.

lesions).

In this pilot study, 105 lesions were detected. The histopathological diagnoses of two lesions were not obtained, histological diagnosis being available for the other 103 lesions (24 nonneoplastic lesions, 77 low grade adenomas, 1 high-grade adenoma, and 1 non-invasive carcinoma).

In 13 lesions (13%) which were endoscopically diagnosed as suspicious for high-grade adenoma or carcinoma, a decision was made to 'resect and send'. Of these 13 lesions, one was histopathologically diagnosed as high-grade adenoma and one as intramucosal carcinoma. Among the lesions for which the endoscopically made decisions were to 'resect and discard' or 'leave *in situ*', there were no high-grade adenomas or carcinomas. Therefore, it was concluded that decisions for management without formal histopathology could safely be made in 88% of small polyps (Fig. 12). The sensitivity of 'resect and send' for high-grade adenoma and carcinoma was 100%, and its specificity was 90%.

Fig. 12. Flow diagram of the pilot study. In this pilot study, 'resect and send' could safely have been selected for the 13 lesions that included the 2 high-risk lesions, and histopathological examination was omitted for the remaining 88% of lesions

Minimally invasive submucosal cancer is morphologically similar to intramucosal carcinoma, from which it is sometimes difficult to distinguish. Submucosal cancer should be assessed by histopathological examination for lymphovascular involvement and the vertical margin of the resected specimen to determine the need for additional surgery to prevent lymph node metastasis. With the 'DISCARD' policy without NBI-ME, there is a risk of small submucosal carcinomas being discarded, whereas the 'DISCARD-ME' policy could prevent inappropriate discarding. Furthermore, the 'DISCARD-ME' policy could be adopted in countries supporting the US guidelines, because these countries do not discard high-risk lesions.
