Endoscopy 2018; 50(01): 6-7
DOI: 10.1055/s-0043-122389
© Georg Thieme Verlag KG Stuttgart · New York

Non-lifting colorectal neoplasia – shall we CAST it away?

Referring to Tate DJ et al. p. 52–62
Michael J. Bartel
Section of Gastroenterology, Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA
Jeffrey L. Tokar
› Author Affiliations
Further Information

Publication History

Publication Date:
21 December 2017 (online)

Tate and colleagues report in this issue of Endoscopy their experience of endoscopic mucosal resection for non-lifting laterally spreading colorectal lesions [1]. These lesions are particularly challenging to manage, not only from the endoscopic technical perspective, but also from the oncologic perspective, as they have a higher likelihood of harboring invasive cancer [2].

The authors propose a method for resecting non-lifting portions of laterally spreading lesions (LSLs) that have previously failed conventional submucosal lift-and-snare excision, using cold forceps avulsion of all visible neoplastic tissue followed by the application of soft coagulation current to the avulsion bed using the snare tip (CAST). The study included two types of non-lifting lesions: first, lesions in which resection had been previously attempted (PANLs); second, treatment-naïve lesions (NNLs), which are lesions where endoscopists usually pause given the risk of submucosal cancer invasion.

The CAST resections were compared with 540 laterally spreading lesions that had adequately lifted, thereby permitting complete snare excision. It is paramount to note that most of the evaluated lesions were tubular and tubulovillous adenomas in both the PANL and NNL cohorts. This correlated with mostly Paris 0-IIa and 0-II + Is lesions, as well as Kudo III and Kudo IV pit patterns. None of the non-lifting lesions had a Paris 0-IIc component and no Kudo V pit patterns were reported.

In that setting 6.3 % of NNLs, 2.6 % of PANLs, and 5.9 % of lifting LSLs harbored submucosal invasive cancer, which highlights the importance of accurate neoplasia assessment prior to resection, as invasive cancer in LSLs has been reported to be as high as 55.5 % in prior studies [3]. All 101 non-lifting colorectal LSLs were completely resected with the proposed CAST technique, including 63 NNLs and 38 PANLs. PANLs were smaller than NNLs and lifting LSLs, which is to be expected because resection of these polyps was previously attempted. Furthermore, the PANL and NNL cohorts had a higher proportion of 0-IIa lesions than the lifting LSL cohort, which had more 0-IIa + Is lesions. The PANL group also had a higher proportion of non-granular lesions compared with the lifting LSL group.

“…CAST is a very promising technique that, for obvious reasons, should not be “CAST away,” but it needs to be applied selectively and only for appropriate colorectal lesions.”

Recurrence of neoplasia during surveillance colonoscopy and avoidance of surgery were the primary study outcomes. Data on the first surveillance colonoscopy were available for most patients. Endoscopically evident recurrence was identified in 15.3 % of the lifting LSLs, 15.2 % of the PANLs, and 27.5 % of the NNL cohort. The higher proportion of recurrence in the NNL cohort remained significantly different in patients with the result of their second surveillance colonoscopy available (20.8 % NNL, 6.3 % PANL, and 4.6 % lifting LSL). Of note, one patient in the NNL cohort developed invasive cancer that was eventually diagnosed at the second surveillance colonoscopy.

From a safety perspective, all deep mural injuries, which were more common in the PANL cohort (18.4 % vs. 7.9 % and 3.7 %), and all reported full-thickness perforations were treated successfully with endoscopic clip closure.

This study provides important information about a relatively easy technique (CAST) that can help endoscopists treat non-malignant, non-lifting LSLs that could not be resected by conventional submucosal lift-and-snare resection, thereby sparing patients unnecessary surgical resection.

The CAST technique has clear advantages over other reported endoscopic techniques. Argon plasma coagulation (APC) was initially considered to be a favorable technique for the ablation of remaining colorectal neoplasia following piecemeal EMR [4], but was shown to be an independent predictor for residual neoplasia [3] and thus has become a less favored treatment option for visible residual neoplasia.

Hot avulsion is another alternative method to eliminate non-lifting neoplasia. Similar to CAST, hot avulsion is not technically challenging. The non-lifting neoplasia is grasped with a hot biopsy forceps, gently tented towards the lumen, and resected by applying short bursts of an electrosurgical energy mode that contains a component of cutting current. In comparison with CAST, hot avulsion has shown comparable recurrence rates (15 % and 10.3 %) and complication profiles [5] [6]. However, the CAST technique has theoretical advantages over hot avulsion. As stated by the authors, the histology specimen is not distorted by the cautery effect when applying CAST. In addition, CAST uses only soft coagulation current, without any cutting current, and potentially mitigates against deep mural injury, whereas with hot avulsion, which incorporates cutting current, its tissue effect is less controllable than soft coagulation, so it is theoretically more likely to cause a deep mural injury.

From an endoscopic skill perspective, CAST is a rather easy procedure. Moreover, the endoscopic equipment used for CAST is widely available. It can be performed with a single device, the snare used for tissue resection, so it has the potential to save time and money. These are important factors allowing its broad utilization by any endoscopic practice. Other techniques for treating non-lifting colorectal neoplasia, like full-thickness endoscopic resection, are still in their infancy and require additional devices, time, and more technical proficiency than CAST.

With regard to the generalizability of this study, we need to acknowledge that not all endoscopists are equal. The investigators reported < 5 % of invasive cancer, even in the NNL cohort, which emphasizes the importance of patient selection via accurate lesion assessment using enhanced imaging modalities (e. g. narrow-band imaging) prior to attempting any resection technique. Applying CAST to lesions that have an increased likelihood of harboring submucosal invasive cancer, such as Paris 0-IIc lesions with Kudo Vn pit pattern, places the patient in unnecessary jeopardy, given the risk of perforation and the likelihood of an inadequate outcome from an oncologic perspective.

Therefore, in our opinion, CAST is a very promising technique that, for obvious reasons, should not be “CAST away,” but it needs to be applied selectively and only for appropriate colorectal lesions. CAST will be best suited for endoscopists with adequate experience in differentiating benign from malignant non-lifting lesions.