Follow up after endoscopic resection in submucosal invasive colorectal cancers


Corresponding: Hiroaki Ikematsu, Department of Gastroenterology and GI Oncology, National Cancer Center Hospital East, Kashiwanoha 6-5-1, Kashiwa, Chiba 277-8577, Japan. Email:


Submucosal invasive colorectal cancers (SM-CRC) have approximately a 10% chance of lymph node metastasis, which requires surgical resection including lymph node dissection for curative treatment. It is important to optimally survey patients after curative resection for SM-CRC in order to detect early recurrence. In the present report, we principally show the long-term outcomes after follow up of SM-CRC resected endoscopically based on a report of the literature and our experience in Japan. The long-term outcomes of low-risk SM-CRC endoscopically resected alone or high-risk SM-CRC with additional surgical resection with lymph node dissection are excellent. However, the risk of local recurrence of endoscopic resection alone in patients with high-risk submucosal invasive cancer was significantly higher in rectal cancer as compared to similar colonic cancer. Patients with submucosal rectal cancer showing high-risk pathological features are, therefore, strongly recommended to undergo additional treatment. We consider that longer follow up is required for patients with SM-CRC because recurrence occurred relatively later in SM-CRC compared to advanced colorectal cancer.


Lymph node metastasis (LNM) occurs in approximately 6–12% of patients with submucosal invasive colorectal cancers (SM-CRC).[1-4] According to the Paris classification and Japanese clinical guidelines, a SM-CRC with negative vertical margins, well- or moderately differentiated adenocarcinoma, with no evidence of vascular or lymphatic invasion and an invasion depth of <1000 μm has a low risk for LNM and local recurrence.[5-7] SM-CRC lesions that are positive for any of these risk factors are classified as being at high risk for LNM. Based on these guidelines, it has been recommended that endoscopic treatment only would suffice for patients with low-risk SM-CRC. However, if any high-risk findings are observed during histological examination of the resected specimen, additional curative surgical resection with lymph node dissection is recommended.[7] There are, however, some cases of high-risk SM-CRC that are followed up by endoscopic resection only due to various patient-related factors including the patient's comorbid state or wishes. The long-term outcomes for this patient group is, however, still unknown. The present report will discuss the long-term outcomes after endoscopic and surgical resection for all SM-CRC.

Japanese Clinical Guidelines

Historically, factors such as the depth of submucosal invasion,[5] histological type (i.e. poorly differentiated adenocarcinoma, signet-ring cell carcinoma, and mucinous carcinoma), the presence of a poorly differentiated area and muconodules at the site of deepest invasion, budding,[8] and vascular invasion have been reported to be risk factors for regional LMN in SM-CRC.[8, 9] According to the recently updated Japanese guidelines by Watanabe et al., SM-CRC are classified as lesions at high risk for LNM if any of the following findings are observed during histological examination of the resected specimen: (i) depth of SM invasion ≥1000 μm; (ii) lymphovascular invasion positive; (iii) poorly differentiated adenocarcinoma, signet-ring cell carcinoma, or mucinous carcinoma; and (iv) Grade 2/3 budding at the site of deepest invasion.[7] If any of the high-risk findings are observed during histological examination of the endoscopic resected specimen, surgical resection with lymph node dissection is considered as an additional treatment. However, it was noted that approximately 90% of patients with SM-CRC with a depth of invasion of ≥1000 μm did not have LNM. It is, therefore, important to determine whether additional treatment is indicated after sufficiently considering other factors in addition to the depth of SM invasion, such as whether other risk factors for lymph node metastasis are present, the physical and social background of the patient, as well as the patient's wishes.

Endoscopic diagnosis of patients with SM-CRC

Endoscopic diagnosis is important in selecting the optimal treatment strategy for tackling SM-CRC. In SM-CRC with a high-risk factor of LMN, generally only the depth of invasion can be ascertained endoscopically. Kudo et al. initially reported the efficacy of invasion depth diagnosis using a pit pattern assessment with the magnifying colonoscope. They reported that the non-structural pit pattern V could be visualized in both intramucosal and submucosal cancers with 65% (128/195) of the lesions corresponding to invasive cancer infiltrating into the submucosa.[10] The clinical classification of crypt patterns was further proposed by Fujii et al. to discriminate between m-sm1 and sm2 or beyond.[11] Matsuda et al. from the same group reported that sensitivity, specificity and diagnostic accuracy of the invasive pattern to differentiate m-ca or sm1 (<1000 μm) from sm2–3 (≥1000 μm) were 85.6%, 99.4%, and 98.8%, respectively.[12] Therefore, it may be possible to gauge the appropriate treatment strategy in real time using magnification colonoscopy.

Endoscopic submucosal dissection (ESD) has been used to successfully resect large colorectal lesions en bloc.[13] In the case when a lesion cannot be assessed confidently, it may be possible to determine treatment strategy after endoscopic resection is carried out to enable the pathologist to ascertain the features described above. This strategy is particularly important in rectal lesions, as rectal surgery is certainly more invasive than colonic surgery.

Outcomes of patients with SM-CRC

The long-term outcomes of endoscopic resection alone, without lymph node dissection, in patients with high-risk submucosal invasive colorectal cancer remain unknown. We examined the long-term outcomes after endoscopic resection for SM-CRC patients and found that the Japanese guidelines enabled appropriate decision-making.[14] Endoscopic resection is adequate for low-risk submucosal invasive colorectal cancer, with additional curative surgical resection recommended for high-risk submucosal invasive colorectal cancer. Oka et al. reported that the validity of endoscopic resection without additional surgical resection for cases with low-risk SM-CRC was proven.[15] We reported that the risk of local recurrence was significantly higher in high-risk submucosal rectal cancer patients treated by endoscopic resection only (against guidelines: patients who were unfit or refused further surgery) as compared to similar submucosal invasive colonic cancers[16] (Figs 1-3). Therefore, at present, patients with submucosal rectal cancer with lesions showing high-risk pathological features are strongly recommended to have additional surgical resection along with lymph node dissection. Kobayashi et al. reported that the recurrence rate after surgical curative resection for node-negative T1 colorectal cancer was very low.[17]

Figure 1.

(A) Sessile lesion Is, 8 mm, identified in the lower rectum 1 cm from Herrmann's line. (B) Part of the lesion has a depressed area. (C) Chromoscopy with indigocarmine shows a definite depression. (D) Magnification with crystal violet staining demonstrates an invasive pattern in the depressed area. Based on these findings, the tumor was diagnosed as early colon cancer with deep submucosal invasion and surgical resection was recommended. (E) The patient rejected surgical treatment. Therefore, the lesion was resected endoscopically (endoscopic mucosal resection). (F) Negative tumor margins were confirmed endoscopically.

Figure 2.

Histopathological examination of the resected specimen demonstrates well-differentiated adenocarcinoma invading the submucosal layer (sm deep; 2000 μm). Lymphatic invasion was positive.

Figure 3.

(A) Submucosal tumor-like recurrence was found 2 cm oral from the original endoscopic mucosal resection scar at follow-up colonoscopy 5 years later. (B) Computed tomography confirmed a mass beyond the rectal wall in the same site.

Follow up of patients with SM-CRC

It is important to optimally survey patients after curative resection for SM-CRC in order to detect early recurrence. Oka et al. reported that recurrence was observed within 3 years after endoscopic resection only.[15] Studies evaluating recurrences after curative resection of colorectal cancer including advanced cancer in Europe and the USA have shown that approximately 50% and 70% of recurrences are detected within 1 year and 2 years after surgery, respectively.[18, 19] This increases to almost 100% within 5 years after surgery.[19] In contrast, we have previously reported that in patients with early colorectal cancer, approximately 10% of recurrences were detected within 1 year after resection, approximately 35% within 2 years, and approximately 15% of the recurrences after 5 years after resection.[16] In looking at long-term outcomes according to the endoscopic treatment strategy (ESD vs endoscopic mucosal resection [EMR]),[20] local recurrence occurred in 2.1% (3/145) in the ESD group compared to 14.5% (33/228) in the EMR group. Two endoscopic piecemeal mucosal resection (EPMR) patients required surgery because of invasive recurrence despite a histological diagnosis of intramucosal cancer at treatment.


Patients with SM-CRC treated by endoscopic resection, surgical resection, or both, based on the Japanese clinical guidelines, had good long-term clinical outcomes. The risk of local recurrence was, however, higher in patients with high-risk submucosal rectal cancer treated with endoscopic resection only against guidelines as compared with similar patients with submucosal colon cancer.

The treatment guidelines are determined based on the risk of LNM. Several studies have reported the pathological and prognostic factors that can predict LNM in SM-CRC patients treated by colorectal resection and lymph node dissection.[6, 7, 17, 21-24] There are, however, few reports on the risk factors for recurrent lesions. The long-term outcome after endoscopic resection of SM-CRC is completely unknown. We hence investigated long-term outcomes and reported that the risk of local recurrence was higher in high-risk SM-CRC patients treated by endoscopic resection only against guidelines.[14, 16] We found that submucosal rectal cancer demonstrated a higher risk of local recurrence. We calculated the hazard ratios for death and recurrence in high-risk SM-CRC patients treated by endoscopic resection only for the following variables: risk stratification, age, sex, resection method, location, morphology, tumor size, differentiation, depth, lymphatic invasion, and vascular invasion by Cox regression analysis to detect a risk factor for local recurrence. It revealed that the only significant risk factor for local recurrence was the location of the lesion (hazard ratio, 6.73; 95% confidence interval, 1.04–43.43, P = 0.045).

Based on these results, we strongly recommend additional curative surgery for patients with high-risk submucosal rectal cancer to prevent local metastasis. The treatment strategy should be determined only after endoscopic resection is carried out and the pathology evaluated.

In contrast, adaptation of the endoscopic resectable lesion may be extended for patients with high-risk submucosal colon cancer. We consider that further long-term outcome evaluations, as well as further investigation of the risk factors and detailed pathological evaluations are required.

Patients with deep SM-CRC who reject surgery may be difficult to manage. Some Western centers have advocated chemoradiotherapy for these patients. There are also some reports of the usefulness of adjuvant chemoradiotherapy for advanced-stage rectal cancers.[25-27] Unfortunately, data are still lacking regarding submucosal rectal cancers and more studies are eagerly awaited.

In the past, we have reported that SM-CRC takes a longer time to recur than does advanced cancer. Approximately 15% of recurrences were detected after 5 years after resection.[16] Therefore, it may be worthwhile to follow these patients for at least 5 years or even longer after endoscopic resection. However, as the number of recurrent cases in that study was small, additional data may be necessary to clarify the actual surveillance strategy. Nonetheless, it is recommended that physical examinations, blood tests including carcinoembryonic antigen level, chest radiography examinations, and computed tomography of the abdomen and pelvis are done postoperatively every 6 months for the first 3 years and every 12 months thereafter. A total colonoscopy should also be carried out yearly. The optimal surveillance schedule after curative endoscopic or surgical resection for SM-CRC should be validated in further studies. It is postulated that the surveillance strategy for rectal and colonic lesions may be different considering these various factors.

In conclusion, patients with SM-CRC treated by endoscopic or surgical resection, or both, based on the Japanese clinical guidelines, had good long-term clinical outcomes. The risk of local recurrence for high-risk rectal SM-CRC resected by endoscopic resection alone is significantly higher than that for the colon. In patients with high-risk rectal lesions, additional therapy is therefore strongly recommended.

Conflict of Interests

Authors declare no conflict of interests for this article.