Successful Salvage Underwater Endoscopic Mucosal Resection for Early Colorectal Cancer with Unclear Horizontal and Vertical Margin after Cold Snare Polypectomy: Two Case Reports Authors: Fumihiko Nakamura, Kentaro Ishii, Masahiko Kuroda, Yukako Nemoto Published: 2026 ABSTRACT We sometimes encounter a case of early colorectal cancer with positive or unidentified margin after cold snare polypectomy (CSP). Case 1 was a 70-year-old woman who underwent total colonoscopy at our center. An 8-mm lesion classified as 0-IIa+IIc in the Paris classification was detected in the descending colon and removed with CSP. Case 2 was a 60-year-old man who underwent total colonoscopy at our center. A 6-mm lesion classified as 0-Is in the Paris classification was detected in the rectum and removed with CSP. Both lesions resulted in well-differentiated adenocarcinoma with unidentified horizontal and vertical margins pathologically. We safely performed underwater endoscopic mucosal resection (UEMR) for the post-CSP scar because residual cancer could not be ruled out completely. Pathological results showed that there was no residual cancer in either case. We report two cases of successful salvage UEMR for early colorectal cancer with unidentified horizontal and vertical margins after CSP. Salvage UEMR after CSP in the early stage may facilitate the exploration of residual cancer or the resection of residual cancer. Introduction Cold snare polypectomy (CSP) is increasingly being performed for small colorectal polyps less than 10 mm in size. This procedure is fundamentally indicated for adenomatous lesions, and not carcinoma[1]. However, there are rare cases where lesions resected with CSP are pathologically found to be early colorectal cancer with unclear or positive margins. In such cases, the exact pathological depth of invasion becomes unknown. At present, there are no clear clinical guidelines indicating whether observation is sufficient or additional treatment is recommended for such cases. The decision on how to treat these cases tends to depend on each institution. We report two cases of successful salvage underwater endoscopic mucosal resection (UEMR) for early colorectal cancer with unclear horizontal and vertical margins after CSP. Case Presentation Case 1 A 70-year-old woman underwent total colonoscopy screening at our hospital. A flat elevated lesion with slight redness and central depression (0-IIa+IIc in the Paris classification), which was 8 mm in diameter, was detected in the proximal side of the descending colon (Fig. 1a, b). Magnified narrow band imaging (NBI) revealed an irregular vessel and surface pattern in the depressed area of the lesion, and this area was classified as Type 2B in Japan NBI expert team classification, which indicated Tis cancer[2] (Fig. 1c). Endoscopic mucosal resection (EMR) was preferred for this lesion. Initial endoscopic treatment of case 1 at our center. a. A flat elevated lesion was detected in the proximal side of the descending colon. This lesion has a reddish surface with fold convergence. b. NBI shows an irregular depressed area. This lesion was classified as 0-IIa+IIc in the Paris classification and was 8 mm in size. c. Magnified NBI shows Type 2B (irregular vessel and surface pattern) in the JNET classification, suggesting Tis cancer. d. The lesion was removed with CSP. The muscularis mucosa remained (green arrow), although en-bloc resection was achieved. e. The tumor was diagnosed as well-differentiated adenocarcinoma. However, the lateral and vertical margins are indistinct (hematoxylin and eosin staining, blue box: ×40, red box: ×200). NBI, narrow band imaging; JNET, Japan NBI expert team; CSP, cold snare polypectomy However, a trainee removed this lesion with CSP (Fig. 1d). A pathological examination revealed well-differentiated adenocarcinoma. In addition, the pathological depth of invasion and horizontal and vertical margins were difficult to evaluate (Fig. 1e). A quick surveillance colonoscopy was performed 1 month later because there was a possibility of missing the post-CSP scar. This scar was identified and a point on the concentric circle of the scar was marked with a tattoo. No obvious signs of residual lesion were endoscopically found at the scar (Fig. 2a, b). However, the embedded or microscopic presence of residual cancer under the mucosa could not be ruled out completely. We performed diagnostic and salvage UEMR for this post-CSP scar. We used the Snare Master (Olympus Corp., Tokyo, Japan). However, this snare was slippery because of the stiffness of the scar. Subsequently, using a bipolar snare (Dragonare; Xeon Medical Corp., Tokyo, Japan), the scar area was able to be strangulated with the snare, and en-bloc resection was performed without any complications (Fig. 2c, d). Additional endoscopic treatment (1 month after initial treatment) and pathological findings of case 1. a. A scar after polypectomy was detected in the descending colon. b. Residual lesion cannot be seen endoscopically in magnified NBI. c, d. Additional UEMR was performed with simultaneous en-bloc resection. e. A pathological examination showed no evidence of residual cancer or a tumor. Additionally, remarkable fibrosis in the submucosal layer can be seen over a broad range in section no. 2 (hematoxylin and eosin staining, ×40). The submucosal layer was resected with adequate depth. NBI, narrow band imaging; UEMR, underwater endoscopic mucosal resection A pathological examination showed inflamed colonic mucosa after polypectomy, confirming adequate depth of submucosal resection, with no evidence of a residual tumor (Fig. 2e). Deep sections of the pathology specimen were additionally prepared and evaluated, but no residual findings were identified. Case 2 A 60-year-old man underwent total colonoscopy to investigate the cause of a positive fecal occult blood test at our center. A 6-mm lesion classified as 0-Is type in the Paris classification was detected in the lower and upper rectal boundary. Although this lesion did not have a depressed area, it appeared to have an expanded appearance. Magnified investigation with chromoendoscopy or NBI was not performed, but the possibility of Tis cancer was suspected (Fig. 3a). EMR was planned, and local submucosal injection was performed. However, an inexperienced assistant overtightened the snare at the base of the lesion, leading to an electrocautery-free resection instead of the intended EMR. Consequently, the procedure was equivalent to a CSP (Fig. 3b). Then, Additional hot forceps polypectomy was performed and several clips were applied. A pathological examination revealed well-differentiated adenocarcinoma. The resected specimen experienced fragmentation during its retrieval through the endoscopic channel. The pathological depth of invasion and horizontal and vertical margins were difficult to evaluate (Fig. 3c). Initial endoscopic treatment of case 2 at our center. a. A small Is type polyp, 6 mm in size with a white spot on the surrounding mucosa, was detected in the boundary between the upper rectum and the lower rectum. b. The lesion was resected. c. The resected specimen was divided into many parts. The tumor was diagnosed as well-differentiated adenocarcinoma. The lateral and vertical margins are indistinct (hematoxylin and eosin staining, blue box: ×40, red box: ×200). Diagnostic and salvage endoscopic treatment was performed 20 days later because of the possibility of embedded or microscopic residual cancer. We identified a post-CSP scar, and there were no obvious signs of residual cancer in the scar endoscopically at that time. The clips were removed with forceps. UEMR was performed with the Dragonare snare, and the scar was excised en-bloc without any complications (Fig. 4a-d). A pathological examination showed regenerative changes, confirming adequate depth of submucosal resection, with no residual tumor in the scar (Fig. 4e). Deep sections of the pathology specimen were additionally prepared and evaluated as in Case 1, but no residual findings were identified. Additional endoscopic treatment (20 days after initial treatment) and pathological findings of case 2. a. A scar after polypectomy with several clips was observed in the same location as the rectum. b–d. After removal of the clips, additional UEMR was performed with simultaneous en-bloc resection. e. The resected specimen was divided into eight sections. A panoramic view of section no. 5 is shown. A pathological examination showed no evidence of residual cancer. Remarkable fibrosis in the submucosal layer is visible over a broad range (hematoxylin and eosin staining, ×40). The submucosal layer was resected with adequate depth. UEMR, underwater endoscopic mucosal resection According to these salvage treatments, both lesions in cases 1 and 2 were pathologically indicated as Tis cancer without a residual lesion. However, we will continue more cautious and intensive long-term surveillance. Discussion CSP is effective and safe for small colorectal adenomatous polyp[3]. However, this procedure often fails to remove the muscularis mucosae and a portion of the submucosal layer, leading to unidentified or positive margins[4,5]. Consequently, the invasion depth and curability of adenocarcinoma in our cases could not be confirmed pathologically. These outcomes resulted from the application of CSP for cancerous lesions. In Case 1, this was stemmed from the insufficient recognition of magnified endoscopic findings. In Case 2, although magnified endoscopy was omitted to shorten the procedure time, EMR was planned based on suspected Tis cancer. However, an unintentional electrocautery-free resection occurred due to inadequate coordination with the assistant. These cases highlight the necessity of performing an accurate endoscopic diagnosis by systematically combining normal observation with white light imaging and magnified endoscopy with NBI or chromoendoscopy[2,6]. Furthermore, endoscopists are required to possess management skills to ensure precise procedures by sharing the procedural intent with their assistants. There are currently no clear treatment guidelines for Tis cancer with unclear or positive margins. Inoki et al. reported that a 14.6% residual lesion was observed among borderline resection cases of CSP for high-grade dysplasia or T1a cancer with an indefinite margin, as in our cases (increasing significantly to 50.0% in cases with a positive margin)[7]. In contrast, the reported residual rate for adenomas with unclear or positive margins was 1.9%[8]. Meanwhile, there have been reports of cases where the lesions developed as submucosal cancer with lymphovascular invasion or advanced cancer after CSP[7,9]. Furthermore, if the residual cancer develops submucosally, the submucosal protrusion gradually becomes more noticeable, while changes in the surface structure are difficult to detect even with endoscopic imaging diagnosis[7]. For these reasons, meticulous care must be taken during follow-up, and it might be desirable to thoroughly resect the post-CSP scar, including the surrounding mucosa and submucosa in cases with unclear or positive vertical margin. In these cases, performing additional colonoscopy and placing a tattoo around the scar as soon as possible for identifying the scar in the future are important. A post-CSP scar is small and the identification rate was reported to be less than 50% in surveillance colonoscopy[7]. In the present cases, the scars clearly detected because surveillance colonoscopy was performed shortly after the initial procedure: approximately 1 month later in case 1 and 20 days later in case 2. The risk of residual lesions requires repeated colonoscopy. This imposes a significant physical and psychological burden on patients, including not only the burden of preparation and the colonoscopy itself, but also anxiety about residual lesions. Taking these factors into consideration, diagnostic surveillance and early salvage treatment would be recommended. UEMR is useful for residual lesions on the scars. In a study on residual laterally spreading tumor treated with endoscopic piecemeal mucosal resection, Kim et al. found that UEMR, compared with conventional EMR, achieved significantly higher en-bloc resection rates (47.2% vs. 15.9%) and endoscopic complete removal rates (88.9% vs. 31.8%)[10]. They also reported that UEMR for scarring of residual lesions maintained a strong safety profile, with no perforation or delayed bleeding. In the underwater immersion setting, the scarred mucosa and submucosa involute toward the lumen, making them easily graspable by the snare, whereas the muscularis propria remains annular, thereby ensuring safety. In contrast, when local injections are performed on scar lesions during conventional EMR, only the submucosal layer immediately surrounding the scar is elevated and snaring becomes difficult[10]. Similarly, endoscopic submucosal dissection (ESD) for scar lesions also requires trained techniques and is associated with increased procedure time and higher perforation risk, largely due to the difficulty in identifying the appropriate dissection layer within the fibrotic submucosa[11]. Our cases showed that salvage UEMR enabled safe and adequately deep additional resection of the scar. Although salvage UEMR for residual lesions after endoscopic resection has been covered in many studies, to the best of our knowledge, there have been no reports of early diagnostic and salvage intervention with UEMR for early colorectal cancer with unidentified or positive horizontal and vertical margins after CSP. Consequently, this procedure indicated that these two lesions were Tis cancer without remnant after CSP. Salvage UEMR for a post-CSP scar, especially in early colorectal cancer with an unknown or positive vertical margin, allows endoscopists to infer whether the lesion remains and to perform additional resection safely even if the lesion is remaining or embedded at the cellular level under the scar. A more thorough histological information of the lesion obtained by salvage UEMR may provide the patients and physicians confidence in the appropriateness of a future endoscopic treatment strategy. Taking these factors into account, our reported advantages of early introduction of UEMR are summarized as follows: 1) Early intervention for reliable localization: Post-CSP scars are small and hardly detectable. By performing early-stage UEMR, the lesion site can be precisely localized, ensuring that the area at risk for residual cancer is accurately addressed before the scar becomes unidentifiable. 2) Appropriate snaring and resection for fibrotic scars: Unlike EMR or ESD, which often fail to lift the scarred area, UEMR utilizes buoyancy to make the fibrotic area protrude, thereby facilitating a deeper and more secure snaring safely, ensuring an adequate vertical margin. 3) Definitive diagnosis through deep resection: Salvage UEMR can resect the submucosa deeply, allowing for a thorough histological assessment. Evaluating these deep sections can be a valuable indicator for ruling out microscopic or embedded residual cancer, providing a more reliable basis for future surveillance. However, our report has several limitations. 1) This report details a short-term outcome based on the two salvage UEMR cases. Therefore, we acknowledge that a longer follow-up period is required to definitively conclude the absence of true residual cancer. 2) Our cases were limited only to Tis cancer; thus, we cannot definitively state that prophylactic diagnostic and salvage UEMR, as performed in our cases, is the decisively appropriate prophylactic additional treatment for all early colorectal cancers, including T1a cancer. 3) The pathological diagnosis after salvage UEMR indicated that the initial lesion was Tis cancer. However, this remains an inference, as the possibility of T1a cancer cannot be completely excluded owing to the nature of the initial CSP and subsequent UEMR. The assessment of critical pathological factors such as lymphovascular invasion and tumor budding would have been insufficient in our cases if the original lesion were T1a cancer, leaving a residual risk of metastasis. Considering these limitations, long-term regular surveillance for our cases needs to be continued rather than adopting the standard surveillance strategy for completely resected Tis cancer[12]. This surveillance primarily involves colonoscopy, but it can also include tumor marker measurement via blood samples and CT scan if clinically indicated. Simultaneously, we hope to validate the long-term effectiveness of our salvage UEMR approach through the further accumulation of cases in the future. Conclusion We report two patients in whom additional UEMR was performed as a diagnostic and salvage procedure for early colorectal cancer with unidentified horizontal and vertical margins after CSP. The present report also provides an important learning experience, particularly for trainees. Author contributions Fumihiko Nakamura was responsible for the endoscopic diagnosis and treatment and designed and drafted the manuscript. Kentaro Ishii and Yukako Nemoto were responsible for the revision of the manuscript. Masahiko Kuroda made the pathological diagnosis. All authors approved the final manuscript. Competing interests The authors declare that they have no competing interests. Ethics approval This case reports were approved by the Institutional Review Board at Kohseichuo General Hospital (Approval Code: 2025-04; 19 June 2025). Informed consent Written informed consent was obtained from the patients for publication of this case report and the accompanying images.