Polyp cancers: size matters!

The risk of cancer in a polyp increases with increasing polyp size. The prevalence of cancer in polyps < 5 mm is negligible [1]. This has led to the development of “diagnose-and-leave” and “resect-and-discard” strategies, underpinned by optical biopsy [2]. Attempts have been made to extend these strategies beyond diminutive polyps. The DISCOUNT study group explored optical diagnosis in real time and the implications of extending the practice from diminutive to small polyps. The findings suggested that the Preservation and Incorporation of Valuable Endoscopic Innovations criteria could be met for diminutive as well as small (6 – 9 mm) polyps, resulting in big cost savings [3]. However, the group found that 7/2115 (0.33 %) small polyps harbored cancer and, under the resect-and-discard strategy, these polyps would be discarded leading to inappropriate management of these cancers. Is this risk low enough to be acceptable? The interesting fact is that optical diagnosis could only identify 1 /7 cancers (14 %) before resection, indicating that optical diagnosis works well for differentiating hyperplastic polyps from adenomas but not from cancers. This could be due to the very low prevalence of cancer, as highlighted in the DISCOUNT study, or a lack of training or understanding of cancers. The incidence of polyp cancers stratified by size is useful information for both clinicians and policy makers.

In this issue of Endoscopy, Parsa et al. report on their review of 5093 lesions ≥ 10 mm from 4112 endoscopies performed by 48 endoscopists between 2002 and 2016 [4]. In total, 3068/4904 polyps (62.6 %) were 10 – 19 mm and 1836/4904 (37.4 %) were ≥ 20 mm. Of the polyps sized 10 – 19 mm, 1997/3068 (65.1 %) were adenomas, 1071/3068 (34.9 %) were serrated lesions, and 28/3068 (0.9 %) were cancers, 25 of which arose from adenomatous polyps and only 3/28 from serrated polyps. Among 1836 polyps sized ≥ 20 mm, 1487 (81.0 %) were adenomas, 349 (19.0 %) were serrated polyps, and 110 (6.0 %) were cancers, 103 of which were from adenomatous polyps and 7 from serrated polyps. This demonstrates a more than sixfold increase in the risk of cancer in polyps ≥ 20 mm in size compared with those of 10 – 19 mm. These data further add to our knowledge of the link between polyp size and cancer. It is also clear from these data that the majority of polyp cancers come from adenomas and not from serrated lesions. It would have been really helpful if the authors had stratified the cancer risk by morphology of polyps, as the risk and clinical approach differs between pedunculated and nonpedunculated polyps.

“The relatively low incidence of polyp cancers and an over-reliance on histology and imaging has hampered the development of optical diagnosis in Western endoscopists.”

A recent study by Turner et al. [5] reviewed 550 811 polyps from a pathology database of private practices across the USA. A total of 447 343 (81 %) were 1 – 9 mm in size, and 103 517 (19 %) were ≥ 10 mm. The incidence of cancer in polyps 1 – 5 mm was 0 %. This was followed by 6 – 9 mm, with an incidence of 0.1 %, increasing to 0.5 % for polyps 10 – 19 mm and 2.2 % for polyps ≥ 20 mm. The incidence of cancer in this study was even lower than that reported by Parsa et al. and this could be due to the differences between the endoscopy-based database in the Parsa et al. study and the histology-based database used in the Turner et al. study. It should also be noted that Parsa et al. excluded inflammatory bowel disease and polyposis, and that their center is a tertiary referral university center compared with the private practices from across the USA that participated in the Turner et al. study. The differences in the population base between the two studies could explain the difference in the incidence of polyp cancers. It is also possible that the study size could influence the reported risk of polyp cancers; the study from Turner et al. was very large, thus increasing the denominator, leading to a drop in prevalence of polyp cancers. It appears that the true incidence of polyp cancers lies between 0.5 % and 0.9 % for polyps 10 – 19 mm in size, and between 2.2 % and 5.9 % for those ≥ 20 mm.

The other interesting aspect of the Parsa et al. study is the optical diagnosis of cancers prior to resection. The authors found that 13 (52.0 %) of the 25 polyps sized 10 – 19 mm for which endoscopic imaging was available had overt endoscopic features of cancer, but this rose to 79.2 % (84/106) for cancer polyps ≥ 20 mm. These figures were based on a post hoc image review by a single senior endoscopist who had an extensive track record and expertise in optical diagnosis. Although these figures cannot be generalized, it does suggest that it is a lot easier to recognize cancer features in the larger polyps.

The focus of optical diagnosis in Western countries has been on distinguishing hyperplastic from adenomatous polyps, but identification of cancer and distinguishing superficial from deep cancer is possible and is well reported by Japanese colleagues [6]. The relatively low incidence of polyp cancers and an over-reliance on histology and imaging has hampered the development of these skills in Western endoscopists. This can lead to mismanagement of polyp cancers, and this issue needs to be addressed. However, the data from the Parsa et al. study suggest that only 4/25 cancers of 10 – 19 mm in size were resected in a piecemeal fashion, and it is reassuring to note that this amounts to only 0.13 % of the total polyps treated in this size category.

Parsa et al. have made an important contribution to the literature and have highlighted the importance of size and optical diagnosis for polyps ≥ 10 mm in size. The challenge is that polyp size is visually estimated by endoscopists and is therefore prone to subjective errors. Polyp size has been shown to be both overestimated and underestimated [7] [8]. Despite not having an accurate and standardized method of measuring polyp size, the importance of size is recognized, especially for determination of surveillance interval and risk of cancer, and the need for en bloc resection. Research is urgently required to develop a robust way of measuring lesion size. The issue of optical diagnosis remains a challenge but computer-aided diagnosis is looking very promising [9]. The current focus of optical diagnosis centers on diminutive polyps but I believe that this will soon shift toward the identification of polyp cancers. Once this has been achieved, then the issue of size may no longer be as relevant, but until then, size really does matter!

• References

• 1 Kamiński MF, Hassan C, Bisschops R. et al. Advanced imaging for detection and differentiation of colorectal neoplasia: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2014; 46: 435-449
• 2 Rex DK, Kahi C, O’Brien M. et al. The American Society for Gastrointestinal Endoscopy PIVI (Preservation and Incorporation of Valuable Endoscopic Innovations) on real-time endoscopic assessment of the histology of diminutive colorectal polyps. Gastrointest Endosc 2011; 73: 419-422
• 3 Vleugels JLA, Hazewinkel Y, Dijkgraa MGW. Optical diagnosis expanded to small polyps: post-hoc analysis of diagnostic performance in a prospective multicenter study. Endoscopy 2019; 51: 244-252
• 4 Parsa N, Ponugoti P, Broadley H. et al. Risk of cancer in 10–19 mm endoscopically detected colorectal lesions. Endoscopy 2019; 51: 452-457
• 5 Turner KO, Genta RM, Sonnenberg A. Lesions of all types exist in colon polyps of all sizes. Am J Gastroenterol 2018; 113: 303-306
• 6 Matsuda T, Fujii T, Saito Y. et al. Efficacy of the invasive/non-invasive pattern by magnifying chromoendoscopy to estimate the depth of invasion of early colorectal neoplasms. Am J Gastroenterol 2008; 103: 2700-2706
• 7 Eichenseer PJ, Dhanekula R, Jakate S. et al. Endoscopic mis-sizing of polyps changes colorectal cancer surveillance recommendations. Dis Colon Rectum 2013; 56: 315-321
• 8 Anderson BW, Smyrk TC, Anderson KS. et al. Endoscopic overestimation of colorectal polyp size. Gastrointest Endosc 2016; 83: 201-208
• 9 Byrne MF, Chapados N, Soudan F. et al. Real-time differentiation of adenomatous and hyperplastic diminutive colorectal polyps during analysis of unaltered videos of standard colonoscopy using a deep learning model. Gut 2019; 68: 94-100