Introduction

The risk factors for synchronous lymph node metastasis (LNM) in pT1 colorectal cancer (CRC) are submucosal invasion depth (≥1000µm), lymphovascular involvement (LVI), poorly differentiated components (por), and tumor budding (BD). On the other hand, in pedunculated lesions with invasion confined to the head of the polyp, the incidence of LNM and recurrence have been reported to be very low (0-2.4% and 0%, respectively).^1,2 However, the robust evidence has been limited.

Aims & Methods

This retrospective study aimed to compare the incidence of LNM and recurrence between the head and stalk invasion groups, based on submucosal invasion assessment using Haggitt Level 2 (HL2; the junction line between normal and neoplastic epithelium), and to evaluate the impact of LNM/recurrence risk factors in pedunculated pT1 CRC.
Patients with pedunculated pT1 CRC treated by endoscopic or surgical resection in our hospital between January 2010 and December 2021 were included. Those with multiple invasive CRCs, synchronous advanced cancer in other organs, and familial adenomatous polyposis were excluded. The submucosal invasion was classified into two groups, head invasion (above HL2) and stalk invasion (beyond HL2). Clinicopathological characteristics, incidence of LMN, and recurrence rate were analyzed between patients with head and stalk invasion.

Results

The study included 66 patients with 66 lesions. The median age was 67 years, and 52 patients (78.8%) were male. The median tumor size was 20mm, and the location was as follows: 4 lesions (6.1%) in the right colon and 62 lesions (93.9%) in the left colon. There were 40 head invasion and 26 stalk invasion lesions. There were no significant differences between the two groups for LVI (17.5% vs 26.9%), por (2.5% vs 7.5%), and BD (5.0% vs 7.7%). Surgical resection was performed more frequently in the stalk invasion group (20% vs 77%, P < 0.01). Synchronous LNM was observed in 2 patients in the head invasion group (5.0%, 95% confidence interval [CI] 0.6-16.9), and 1 patient in the stalk invasion group (3.8%, 95%CI 0.1-19.6) (P = 1). Recurrence was observed in 1 patient in the head invasion group (2.5%, 95%CI 0.1-13.2), and 1 patient in the stalk invasion group (3.8%, 95%CI 0.1-19.6) (P = 1), both of them were distant metastasis. In patients without LVI, por, nor BD, there was no LMN in either group, but recurrence was observed in 1/32 patients in the head invasion group (3.1%, 95%CI 0-17.1), and 1/16 patients in the stalk invasion group (6.3%, 95%CI 0-30.3) (P = 0.61). The patient with recurrence in the head invasion group without LVI/por/BD revealed multiple liver metastases one year after endoscopic resection, and he received chemotherapy but died of colon cancer one and a half years after recurrence.

Conclusion

In pedunculated pT1 CRC, synchronous LNM and recurrence were observed in the head invasion group, with no significant difference from the stalk invasion group. Among cases without LVI/por/BD, there was no LNM but recurrence was observed in both of head and stalk invasion groups. Contrary to the previous reports, head invasion was not a completely LNM/recurrence-free condition, so endoscopic resection alone cannot be considered curative.

References

1. Matsuda T, et al. Risk of lymph node metastasis in patients with pedunculated type early invasive colorectal cancer: A retrospective multicenter study, Cancer Sci. 2022;102:1693-1697.
2. Asayama N, et al. Long-term outcomes after treatment for pedunculated-type T1 colorectal carcinoma: a multicenter retrospective cohort study, J Gastroenterol. 2016;51:702-710.

Introduction

Colorectal cancer (CRC) is one of the leading causes of cancer-related mortality and morbidity worldwide. Its tumor microenvironment is composed of a complex interplay of various monocytes, dendritic cells, macrophages and T-cells. Among them, there are various subtypes of T-cells, each of which plays a very important role and has been reported to be associated with prognosis in various cancers. In CRC, there are many reports on the association between CD8+ T-cells and Foxp3+ T-cells and prognosis. However, these are all reports mainly focused on advanced cancers. No studies on T-cells focused on early-stage CRC have been reported to date. Among early-stage CRC, submucosal deep invasive (T1b) CRC has a lymph node metastasis rate of about 10%. Although submucosal invasion depth, vascular invasion, histologic grade, or budding grade are known to be pathological risk factors of lymph node metastasis, there have been no reports evaluating the relationship between tumor microenvironment and lymph node metastasis. Therefore, in this study, we aimed to investigate the characteristics of T-cells according to the degree of CRC invasion and the association between T-cells and the risk of lymph node metastasis in T1b CRC.

Aims & Methods

Initially, 60 patients with colorectal neoplasia [20 with intramucosal neoplasia (IM group); 20 with submucosal invasive CRC (SM group); and 20 with advanced cancer (AD group)] treated at Hiroshima University Hospital from January 2011 to December 2015 were randomly selected. We examined changes in T-cells with tumor invasion. Next, 84 patients with T1b CRC who underwent initial surgical resection with lymph node dissection or additional surgical resection with lymph node dissection after endoscopic resection at Hiroshima University Hospital from March 2013 to July 2019 were selected. We examined the relationship between T-cells and lymph node metastasis in T1b CRC. For all specimens, the phenotype and number of T-cells were evaluated using the triple immunofluorescence method of CD4 (helper T-cell marker), CD8 (killer T-cell marker), and Foxp3 (regulatory T-cell marker). We counted T-cells by subtype in the center (CT) and the invasive front (IF) of the tumor and examined their association with lymph node metastasis.

Results

In the IM group, there was no significant difference in the number of any subtype between CT and IF, while in the SM and AD groups, all subtypes were more numerous in IF than CT. They were also predominantly more numerous as the degree of tumor invasion with IM, SM, and AD. The high number of Foxp3 in IF and the high ratio of Foxp3/CD4 and Foxp3/CD8 were positively correlated with lymph node metastasis.

Conclusion

High Foxp3 counts and high Foxp3/CD4 and high Foxp3/CD8 ratios in IF of T1b CRC may help predict lymph node metastasis.

Disclosure

The authors have no conflicts of interest or financial ties to disclose.

Introduction

Accurate prediction of lymph node metastasis (LNM) in submucosal invasive (T1) colorectal cancer (CRC) is essential for determining the need for additional surgery following endoscopic resection. Conventional histopathological assessments, however, are limited by interobserver variability. This study aimed to develop and externally validate a stacking-based artificial intelligence (AI) model integrating histopathological whole slide images (WSIs) and clinical parameters to predict the risk of LNM in T1 CRC.

Aims & Methods

This retrospective study included 1,056 patients with histologically confirmed T1 CRC treated at Showa Medical University Northern Yokohama Hospital. Of these, 593 cases were used for model training and 100 for internal testing. External validation was conducted on 25 cases from two independent institutions.
The proposed stacking ensemble model incorporated the Multiple Instance Self-Training (MIST) score derived from WSIs, patient sex, and tumor location (colon vs. rectum). Base learners included logistic regression, XGBoost, and random forest, with a meta-learner synthesizing their outputs to generate the final LNM prediction. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) in both internal and external cohorts.

Results

LNM was observed in 69 (11.6%) training cases, 15 (15.0%) test cases, and 2 (8.0%) external validation cases. The model achieved an AUC of 0.6965 in the internal test cohort and 0.7500 in the external validation cohort, demonstrating stable and generalizable predictive performance.
A representative case involved a 64-year-old man with a 16-mm type 0-Is lesion in the sigmoid colon resected via endoscopic mucosal resection (EMR). Histopathology showed well-differentiated tubular adenocarcinoma (tub1 > tub2), submucosal invasion depth of 3,300 µm (T1b), tumor budding grade BD1, Ly0, V1a, and negative resection margins (HM0, VM0). Despite guideline-based recommendations prompting additional surgical resection, no LNM was found (N0, 0/16), and the patient remained recurrence-free five years postoperatively—consistent with the AI model’s prediction.

Conclusion

The stacking-based AI model exhibited robust and generalizable performance in predicting LNM in T1 CRC, both internally and across external institutions. This approach offers a complementary tool to conventional pathology and may enhance individualized decision-making regarding the necessity of additional surgical intervention.

Introduction

Tumor budding is a significant prognostic value in the management of colorectal cancer (CRC). There are grades of tumor budding: 0-4 buds as low, 5-9 buds as intermediate, and more than 10 buds as high. However, the specific prognosis difference between cases with 0 buds and those with 1-4 buds, is not well-established due to a lack of comparative studies.

Aims & Methods

The aim of this study is to examine and compare the rate of lymph node (LN) metastasis and prognosis by distinguishing between 0 buds and 1-4 buds within the low-grade category (0-4 buds) of tumor budding in T1 and T2 CRC. This study retrospectively identified a total of 224 cases of T1 and T2 CRC from 2015 to 2018 across three medical institutions using medical records. The pathology, including the assessment of tumor budding, was subsequently reconfirmed, and the recurrence and survival of patients were evaluated up to December 2023.

Results

Patients in the BD1 group exhibited a higher T stage compared to BD0, accompanied by significantly increased rates of lymphovascular invasion and perineural invasion. The prevalence of LN metastasis was 14.8%. No significant difference in LN metastasis was observed between BD0 and BD1. In a multivariate analysis exploring factors associated with LN metastasis, relevant factors included lymphatic invasion, perineural invasion, and 5 buds or more. (Table 1) There were no significant distinctions in 5-year survival and progression free survival rate between the BD0 and BD1 groups (P=0.971).
Table 1. Univariate and multivariate analysis for risk factor of lymph node metastasis

Conclusion

The study confirmed that there was no significant difference in LN metastasis and prognosis between BD0 and BD1.

References

1. Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA: A Cancer Journal for Clinicians 2023;73(1):17-48.
2. Lugli A, Kirsch R, Ajioka Y, Bosman F, Cathomas G, Dawson H, et al. Recommendations for reporting tumor budding in colorectal cancer based on the International Tumor Budding Consensus Conference (ITBCC) 2016. Mod Pathol 2017;30(9):1299-311.

Introduction

Tip-in endoscopic mucosal resection (EMR) has shown a favorable resection outcome for large colorectal polyps.^{1, 2} However, reported outcomes have been limited to experts’ performance. To generalize the outcomes, trainees’ performance of Tip-in EMR is an interest, however, available data are absent.

Aims & Methods

To clarify whether Tip-in EMR can be achieved effectively and safely even by trainees, in this single-center retrospective study, clinical outcomes of Tip-in EMR between experts and trainees were compared. Inclusion was 15–25 mm non-pedunculated colorectal neoplasms resected by Tip-in EMR between January 2014 and December 2020. Exclusion was residual lesions after the previous resection. Data were collected from medical records and prospectively recorded endoscopy and pathology databases. Baseline characteristics, treatment results, histological outcomes, and surveillance outcomes were analyzed. In our institution, trainees perform Tip-in EMR under the supervision of experts, and switch to experts when experts decide to switch during the procedure, all of which were also analyzed as the trainee group. Independent procedure completion and independent en bloc resection were defined as achieving procedure completion and en bloc resection without switching, respectively. Failure of independent en bloc resection was defined as piecemeal resection or switching. Residual lesions were defined as endoscopically detected residuals that were pathologically confirmed as neoplasia. All procedures performed by trainees were grouped into three periods (period 1: 1–10 cases, period 2: 11–20 cases, and period 3: ≥21 cases), and the learning curve was assessed. A logistic regression analysis was performed to examine the risk factors of failure of independent en bloc resection.

Results

A total of 597 lesions (550 patients) were analyzed. Among them, 438 lesions (402 patients) and 159 lesions (148 patients) were resected by 6 experts and 14 trainees, respectively. The independent procedure completion was 146 cases (91.8%) in the trainees and 438 cases (100%) in the experts. The en bloc resection rates did not differ between trainees and experts (83.0% vs. 88.6%, P = .098). However, the independent en bloc resection rate was significantly lower in trainees (77.4% vs. 88.6%, P = .001). There was no significant difference in the incidence of adverse events (2.5% vs. 5.7%, P = .165). Among 380 lesions (63.6%) of 347 patients who received surveillance colonoscopy, 11 residual lesions were identified; 3 lesions (3.1%) in the trainees and 8 (3.5%) in the experts (P = 1), all of them were removed endoscopically. For the learning curve of trainees, the independent en bloc resection (OR 3.4, 95%CI 2.0–5.7, P < .001) was significantly lower and took longer procedure time (median 8.6 min vs. 5.8 min, P < .001) in period 1 than those of experts, but not in period 2 and 3. In multivariate analysis, risk factors of failure of independent en bloc resection were non-polypoid morphology (OR 3.4, 95% CI: 1.6–7.3; P = .001), positive non-lifting signs (OR 3.1, 95%CI: 1.2–8.0; P = .023), lesions with an underlying semilunar fold (OR 3.6, 95%CI: 2.0–6.3; P < .001), and trainees with experience ≤10 procedures (OR 3.6, 95%CI: 2.1-6.3, P<.001).

Conclusion

We demonstrated favorable outcomes of Tip-in EMR for 15-25 mm lesions performed by trainees. However, Tip-in EMR should be performed under the supervision of experts, especially for trainees with experience of ≤10 cases.

References

1. Imai K, Hotta K, Ito S, et al. Tip-in Endoscopic Mucosal Resection for 15- to 25-mm Colorectal Adenomas: A Single-Center, Randomized Controlled Trial (STAR Trial). Am J Gastroenterol. 2021;116:1398-1405.
2. Takada K, Hotta K, Imai K, et al. Tip-in EMR as an alternative to endoscopic submucosal dissection for 20- to 30-mm nonpedunculated colorectal neoplasms. Gastrointestinal Endosc. 2022;96:849-856.