Optimizing referral practices for resection of large colorectal polyps

There has been an increasing focus on educating and encouraging referral of large complex colorectal polyps to an advanced endoscopist for endoscopic mucosal resection (EMR) or submucosal dissection (ESD) and on discouraging physicians from sending these patients for unnecessary surgery. However, practice patterns vary widely depending on the available resources, training, and preferences of the referring and accepting physician. Endoscopists who may not be comfortable removing large polyps or lack the equipment for these procedures will often take a biopsy on the initial examination to 1) rule out cancer, 2) provide results and reassurance to the patient, and 3) include pathology in the referral. We frequently hear frustration from endoscopists on the receiving end that these biopsies cause submucosal fibrosis and make the resection more challenging. However much of this is anecdotal, with few data to support these complaints. Finally, we have more insight into the value and detriment of biopsy prior to EMR.

In this issue of Endoscopy, Ma et al. report their experience from the Australian colonic EMR trial (2013–2016), which included 586 nonpedunculated colorectal polyps ≥ 2 cm referred for resection. [1] They evaluated the impact of pre-resection biopsy (PRB) during the index examination on EMR outcomes and final histology. About half the polyps had granular morphology, a third were nongranular, and the rest were serrated/other. Polyps that had PRB (58.6 %) were larger than those not biopsied (median 35 mm vs. 30 mm; P = 0.007) but otherwise had similar characteristics. PRB was associated with greater submucosal fibrosis (43.1 % vs. 30.0 %; P = 0.001) and intraprocedural bleeding during EMR (31.5 % vs. 23.5 %; P = 0.03) but did not influence EMR success, delayed bleeding or perforation, or recurrence rates.

The final histology after EMR was unchanged in two-thirds of polyps, upstaged in 26 %, and downstaged in 14 %. PRB was 77 % sensitive for low grade dysplasia but only 21 % for high grade dysplasia. There were 19 cancers found on EMR that were not initially identified (15 LGD and 4 HGD on PRB). The authors conclude that PRB has minimal value and should be used with caution given the low yield of detecting cancer, potential false reassurance against cancer, and detriment to the resection.

The main utility of a PRB is to rule out cancer as that would impact the clinical pathway. Confirming a cancer diagnosis up front prompts an appropriate, on-time referral to surgery and prevents a delay in care while the patient waits for an unnecessary second colonoscopy and possibly an inappropriate EMR. Although not quantified in the present study, we must acknowledge all the polyps that never made it into this cohort because the initial biopsy confirmed cancer. Nonetheless, the authors express concern that PRB was ineffective at predicting these 19 cancers. It is hard to put this finding into context, but a 5.5 % rate of unrecognized cancers on initial biopsy may be acceptable. It is possible that misclassification bias or sampling error impacted these initial pathology results. Additionally, EMR can sometimes serve as a diagnostic staging tool for borderline lesions. Finally, these results also indicate that experienced endoscopists still need more training in lesion recognition and image-enhanced endoscopy to identify submucosal invasion that precludes endoscopic resection.

“We should be cautious to not universally recommend against pre-resection biopsy because we want endoscopists to feel comfortable taking a biopsy of a large or complex lesion if this could potentially avoid an operation. There is significant variation in clinical experience, learning curves, and practice patterns, and the reality is that not all endoscopists across communities can appreciate the subtle nuances of an advanced polyp versus an early cancer.”

Submucosal fibrosis is the nemesis of endoscopic resection. It makes the procedure more difficult, more frustrating, and longer in duration. We typically think of it occurring after prior attempted resection, but theoretically any previous manipulation, including biopsy, could create scarring. In a recent Japanese study of 360 colorectal ESD cases, polyps that had PRB (15.7 %) had an 8 × greater likelihood of having severe fibrosis [2]. Ma et al. similarly showed this association and concluded that PRB should be avoided; however, associated data complicates this takeaway message. First, the typical consequences of submucosal fibrosis (failure to complete procedure, bleeding, perforation) were not higher in the PRB group. Additionally, a third of polyps without PRB had submucosal fibrosis, a fairly high rate for the “control” group, leaving an absolute difference of only 13 % between groups. These results likely reflect unmeasured or unaccounted for factors that contribute to submucosal fibrosis such as tumor infiltration/depth of invasion, inflammation, prolapse, and others that warrant further investigation.

This study also identified higher rates of intraprocedural bleeding in the PRB group, though most immediate bleeding was treated with snare tip soft coagulation without consequence and rates of post-procedure bleeding were similar among the two groups. There is no obvious mechanism where prior biopsy causes bleeding during EMR, whether related to or irrespective of fibrosis, and this is likely confounded by other noncontrolled factors such as polyp size, location, morphology, electrosurgical generator settings, and patient use of antiplatelet or anticoagulant drugs.

It is still useful to consider specific scenarios where PRB may help and how to maximize yield. Too many nonmalignant polyps continue to be sent for surgery. We should be cautious to not universally recommend against PRB because we want endoscopists to feel comfortable taking a biopsy of a large or complex lesion if this could potentially avoid an operation. There is significant variation in clinical experience, learning curves, and practice patterns, and the reality is that not all endoscopists across communities can appreciate the subtle nuances of an advanced polyp versus an early cancer. Other fundamental guiding principles for the index examination include: 1) never start a job that you do not feel confident you can complete; 2) use saline alone to test for adequate submucosal lift, given reports of fibrosis from lifting agents that may complicate subsequent resection [3]; and 3) only tattoo when necessary, as this can also cause fibrosis and impact EMR. There are few data to support taking a biopsy of the edge versus the middle of the polyp, or the bulky portion versus the flat portion to avoid fibrosis, but it stands to reason that if you are going to take a biopsy then make it worthwhile: target the most advanced portion (depressed or erythematous area) to maximize diagnostic yield. Finally, if you need to take a biopsy, results from this study are reassuring that this will not affect key outcomes such as success of EMR and recurrence rates.

Publication History

Article published online:
09 November 2022

• References

• 1 Ma MX, Tate DJ, Sidhu M. et al. Effect of pre-resection biopsy on detection of advanced dysplasia in large nonpedunculated colorectal polyps undergoing endoscopic mucosal resection. Endoscopy 2023; 55: 267-273
• 2 Kuroha M, Shiga H, Kanazawa Y. et al. Factors associated with fibrosis during colorectal endoscopic submucosal dissection: does pretreatment biopsy potentially elicit submucosal fibrosis and affect endoscopic submucosal dissection outcomes?. Digestion 2021; 102: 590-598
• 3 Kolb JM, Aihara H, Wani S. et al. Severe submucosal fibrosis and granuloma complicating endoscopic submucosal dissection: unintended consequences of a lifting agent (with video). Gastrointest Endosc 2021; 93: 524-526