Qian J, Tao YL, Zheng SS. Risk modeling of delayed postoperative bleeding after endoscopic submucosal dissection for early colorectal cancer and precancerous lesions. World J Gastrointest Surg 2025; 17(9): 107326 [DOI: 10.4240/wjgs.v17.i9.107326]
Corresponding Author of This Article
Shu-Sen Zheng, MD, Department of Hepatobiliary and Pancreatic Surgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, No. 79 Qingchun Road, Hangzhou 310003, Zhejiang Province, China. shusenzheng@zju.edu.cn
Research Domain of This Article
Gastroenterology & Hepatology
Article-Type of This Article
Retrospective Study
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Jun Qian, Shu-Sen Zheng, Department of Hepatobiliary and Pancreatic Surgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou 310003, Zhejiang Province, China
Jun Qian, Department of Colorectal Surgery, Zhejiang Cancer Hospital, Hangzhou 310022, Zhejiang Province, China
Ya-Li Tao, Endoscopy Center, Zhejiang Cancer Hospital, Hangzhou 310000, Zhejiang Province, China
Shu-Sen Zheng, National Health Commission Key Laboratory of Combined Multi-Organ Transplantation, Hangzhou 310003, Zhejiang Province, China
Author contributions: Qian J designed the research and wrote the first manuscript; Qian J, Tao YL and Zheng SS conceived the research and analyzed the data; Qian J conducted the analysis and provided guidance for the research; and all authors reviewed and approved the final manuscript.
Institutional review board statement: This study was approved by the Medical Ethics Committee of Zhejiang Cancer Hospital.
Informed consent statement: Patients were not required to give informed consent to the study because the analysis used anonymous clinical data that were obtained after each patient agreed to treatment by written consent.
Conflict-of-interest statement: All the authors report no relevant conflicts of interest for this article.
Data sharing statement: No additional data are available.
Open Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Shu-Sen Zheng, MD, Department of Hepatobiliary and Pancreatic Surgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, No. 79 Qingchun Road, Hangzhou 310003, Zhejiang Province, China. shusenzheng@zju.edu.cn
Received: May 14, 2025 Revised: June 9, 2025 Accepted: July 25, 2025 Published online: September 27, 2025 Processing time: 133 Days and 1 Hours
Abstract
BACKGROUND
As a minimally invasive technique, endoscopic submucosal dissection (ESD) is widely used in treating early colorectal cancer (ECRC) and precancerous lesions (PCLs). However, a common postoperative complication - delayed postoperative bleeding (DPOB) - can significantly hinder patient recovery.
AIM
To build and validate a predictive model for assessing post-ESD DPOB risk in ECRC and PCL patients, utilizing logistic regression methodology.
METHODS
A retrospective review was conducted on ECRC/PCL 302 patients who received ESD at our hospital between July 2021 and July 2024. The cohort was stratified based on the incidence of DPOB following ESD, forming DPOB and non-DPOB groups. Through allocation, they were further allocated into model and validation cohorts. Clinical variables from both cohorts were collated and subjected to univariate analysis to determine potential factors associated with post-ESD DPOB. Subsequently, we constructed a predictive model for DPOB risk employing logistic regression analysis. Model performance assessment used receiver operating characteristic curves in both the training and validation cohorts, with internal validation accomplished via 10-fold cross-validation.
RESULTS
The occurrence rate of DPOB was 9.93%. Univariate analysis revealed that the number of lesions, lesion size, lesion location, degree of submucosal fibrosis, and intraoperative bleeding were significantly associated with DPOB. Binary logistic regression analysis identified the number of lesions, lesion size, lesion location, and degree of submucosal fibrosis as independent DPOB determinants. A nomogram that was developed to quantify the DPOB risk exhibited that an increment in the total score corresponded to an increased risk. The model achieved area under the curve values of 0.831 and 0.821 in the model and validation groups, respectively, with P values of 0.853 and 0.203 in the Hosmer-Lemeshow test. The model demonstrated robust discriminative performance, with an average area under the curve of 0.795 (95% confidence interval: 0.702-0.887) in 10-fold cross-validation.
CONCLUSION
Collectively, the presence of multiple lesions, lesion size of ≥ 3 cm, lesion localization in the rectum, and severe fibrosis are significant independent predictors of DPOB in patients undergoing surgery for ECRC or PCLs. The proposed risk prediction model, which integrates these factors, demonstrates excellent predictive accuracy and clinical utility, thereby providing a valuable tool for risk stratification and postoperative management in this patient population.
Core Tip: Currently, research on risk factors for delayed postoperative bleeding (DPOB) in patients undergoing endoscopic submucosal dissection for early colorectal cancer or precancerous lesions remains limited. This investigation set out to evaluate post-endoscopic submucosal dissection DPOB risk in these patients and to support more informed clinical management by identifying key influencing factors. Multiple lesions, lesions ≥ 3 cm in diameter, rectal lesion location, and marked fibrosis emerged as key independent predictors of DPOB. For patients exhibiting these risk factors, enhanced postoperative surveillance and tailored management strategies are therefore advised to improve their outcomes.
Citation: Qian J, Tao YL, Zheng SS. Risk modeling of delayed postoperative bleeding after endoscopic submucosal dissection for early colorectal cancer and precancerous lesions. World J Gastrointest Surg 2025; 17(9): 107326
Ranking as the third most commonly diagnosed malignancy worldwide, colorectal cancer (CRC) is responsible for the second greatest cancer mortality burden, posing a grave threat to human life and health[1]. Attributed to the abnormal growth of glandular epithelial cells in the colon or rectal tissues, this disease is etiologically linked to a combination of genetic predisposition, environmental exposures, and modifiable lifestyle factors like sedentary behavior, obesity, and tobacco use[2,3]. Early CRC (ECRC) often presents with vague and nonspecific clinical manifestations. As the disease progresses, patients may develop symptoms such as altered bowel patterns (e.g., intermittent diarrhea and constipation), hemafecia, and abdominal discomfort, which severely impair their quality of life[4,5]. Recent epidemiological data highlight the global impact of CRC, with 2020 data documenting over 147950 new diagnoses (including 104610 colon and 43340 rectal cancers) in the United States alone and nearly 1 million CRC-attributable fatalities worldwide[6,7]. Endoscopic submucosal dissection (ESD) has emerged as a fundamental therapeutic approach for ECRC and precancerous lesions (PCLs), offering significant clinical benefits compared to conventional surgical interventions. These advantages include minimized tissue trauma, enhanced treatment effectiveness, and decreased rates of postsurgical complications[8,9]. However, the procedure’s technical complexity within the intricate colorectal anatomy creates specific challenges, particularly the risk of delayed postoperative bleeding (DPOB), a serious complication that compromises patient recovery and treatment success[10,11]. For optimal postoperative management of ECRC and PCLs treated with ESD, the identification of DPOB risk factors becomes imperative, as this knowledge holds substantial clinical relevance for refining therapeutic protocols and improving patient outcomes.
Current evidence regarding predictors of post-ESD DPOB in these patients remains insufficient. The present investigation employs logistic regression methodology to establish a clinically applicable scoring system for DPOB risk assessment following ESD procedures in ECRC and PCL cases, offering enhanced predictive capability for this postoperative complication.
MATERIALS AND METHODS
Case selection
Eligibility criteria: Patients must fulfill ECRC or PCL diagnostic criteria[12]. Preoperative assessments using magnifying chromoendoscopy or endoscopic ultrasound (EUS) confirmed that the depth of tumor invasion was confined to the mucosal or superficial submucosal layer. EUS findings that demonstrate a clear demarcation between the lesion and the muscularis propria were considered acceptable for inclusion in cases where preoperative assessment could not definitively determine whether the invasion depth extended to the deep submucosal layer.
Exclusion criteria: Patients were excluded if preoperative EUS or pelvic computerized tomography imaging indicated deep mucosal layer invasion or extracolonic extension as well as evidence of lymph node metastasis. Additional exclusion criteria were postoperative pathological confirmation of non-epithelial tumors (e.g., stromal tumors, neuroendocrine tumors, leiomyomas, granular cell tumors), diagnosis of other non-neoplastic pathological changes, or identification of a type V mucosal pit pattern (irregular VA or unstructured VN types) on magnifying chromoendoscopy[13]. Further, cases that involve significant intraoperative hemorrhage or perforation that require conversion to surgical intervention, a history of colorectal malignancy surgery or radiotherapy/chemotherapy, and incomplete clinical, endoscopic, or pathological records were excluded.
The Zhejiang Cancer Hospital ethics committee approved this study. After rigorous inclusion and exclusion criteria application, this study enrolled 302 patients diagnosed with ECRC or PCLs and treated at Zhejiang Cancer Hospital, from July 2021 to July 2024. Participants were stratified based on the occurrence of DPOB into the non-DPOB (n = 272) and DPOB groups (n = 30). DPOB was a hemorrhage originating from the resection site occurring 12-14 days after ESD, accompanied by hematochezia and a reduced hemoglobin of > 20 g/L[14].
Intervention protocol
All patients underwent ESD. Comprehensive preoperative preparations, including electrocardiogram, blood routine examination, and pelvic-abdominal computerized tomography imaging, were performed. The intervention commenced with endoscopic identification of the target lesion, assisted by chromoendoscopic techniques to enhance pathological margin delineation. An adequate elevation of the lesion was then achieved through submucosal injection of an adrenergic-containing solution combined with indigo carmine and sodium hyaluronate, thus facilitating sufficient tissue separation. After that, the submucosal microvasculature was managed through a three-step process: Precision coagulation using bipolar electrocautery, vessel transection with an electrosurgical device, and focused thermoablation of bleeding points. Following this, a mucosal incision was deliberately placed at a 5-mm peripheral margin from the lesion boundaries, after which submucosal dissection was performed to facilitate en-bloc resection. The resection site underwent a comprehensive endoscopic evaluation to confirm complete tumor excision and to detect potential hemorrhagic events. When active bleeding was visualized, prompt hemostatic measures were initiated to achieve effective control. Upon procedure completion, the endoscope was carefully removed with concurrent aspiration of residual intraluminal gas and fecal contents. Patients were subsequently maintained on a complete bowel rest status for optimal wound recovery. Intravenous antibiotic prophylaxis was administered to minimize infection risks, with close and continuous clinical surveillance implemented to promptly identify any signs of procedure-related morbidity or clinical deterioration.
Data collection and outcome measures
Patient demographics and clinical information, such as gender, age, comorbidities (e.g., hypertension, diabetes, chronic obstructive pulmonary disease), smoking/alcohol consumption status, lesion number/size/Location, submucosal fibrosis extent, and intraoperative bleeding, were systematically obtained. These variables were analyzed employing univariate and multivariate statistical methods to determine potential risk factors associated with DPOB in patients undergoing treatment for ECRC or PCLs. The degree of submucosal fibrosis was assessed and categorized based on standardized criteria[15] as no fibrosis (F0), which is characterized by the presence of a blue transparent layer after the submucosal injection of sodium hyaluronate and indigo carmine; mild fibrosis (F1), which is the appearance of white, muscular-like structures within the blue transparent layer; severe fibrosis (F2), which is identified by the exclusive presence of white, muscular-like structures in the submucosal layer, with the absence of a blue transparent layer. Intraoperative bleeding was rigorously defined as any bleeding event that requires endoscopic hemostatic intervention, including the use of argon plasma coagulation, hemostatic forceps, or clipping devices to achieve effective hemostasis[16].
Statistical analysis
Statistical Package for the Social Sciences version 22.0 or GraphPad Prism was used for data analysis. Categorical variables were expressed as frequencies and percentages [n (%)]. Comparative analyses of categorical variables were conducted using the χ² test or Fisher’s exact test, as appropriate. To maintain consistent event rates between the modeling and validation groups considering the relatively low DPOB incidence (9.93%), this research used simple random sampling to partition the data in a 2:1 ratio, resulting in 201 and 101 patients in the modeling and validation cohorts, respectively. This method improves the accuracy and reliability of model assessment by maintaining balanced event rates between groups in internal validation. Variables determined as statistically significant in the univariate analysis were subsequently included in a multivariate logistic regression model to identify independent risk factors that contribute to DPOB in patients with ECRC and PCLs. Multicollinearity among the predictor variables was assessed using the variance inflation factor, with a variance inflation factor value of < 5 indicating no significant multicollinearity. To ensure the robustness and generalizability of the logistic regression model, a 10-fold cross-validation process was applied. The dataset was randomly categorized into 10 equally sized subsets. The model was iteratively refitted to each of the 10 training sets that comprise 90% of the data. The area under the curve (AUC) was calculated for the remaining 10% of the data in each iteration, and the mean AUC was computed. A P value of less than 0.05 indicated statistical significance.
RESULTS
Univariate analysis of risk factors for DPOB in patients with ECRC and PCLs
According to data derived from the univariate analysis, variables including the number of lesions, lesion size, lesion location, degree of submucosal fibrosis, and intraoperative bleeding, exhibited a close correlation with DPOB in patients undergoing surgery for ECRC or PCLs (P < 0.05). Conversely, factors, such as gender, age, comorbidities (hypertension, diabetes, and chronic obstructive pulmonary disease), smoking status, and alcohol consumption, demonstrated no significant association with delayed bleeding (P > 0.05). Table 1 summarizes the detailed results.
Table 1 Univariate analysis of risk factors for delayed postoperative bleeding in patients with early colorectal cancer and precancerous lesions, n (%).
Multivariate analysis of risk factors for DPOB in patients with ECRC and PCLs
Variables that demonstrate statistical significance in the univariate analysis, namely the number of lesions, lesion size, lesion location, degree of submucosal fibrosis, and intraoperative bleeding, were incorporated as independent variables, with delayed bleeding considered as the dependent variable. Among the variables, multicollinearity diagnostics confirmed the absence of significant collinearity. Logistic regression analysis identified the number of lesions, lesion size, lesion location, and degree of submucosal fibrosis as independent risk factors for DPOB (P < 0.05). Conversely, intraoperative bleeding did not independently correlate with DPOB (P > 0.05). A predictive nomogram was subsequently developed by integrating these factors, with each variable assigned a weighted score. The cumulative score was utilized to estimate the risk of delayed bleeding, wherein higher total scores corresponded to an increased possibility of delayed bleeding. Tables 2 and 3 and Figure 1 present detailed results, including variable assignments and scoring.
Validation of the nomogram model for predicting DPOB in patients with ECRC and PCLs
Model performance was assessed through the AUCs. The AUC of the modeling group was 0.831 [95% confidence interval (CI): 0.717-0.944], with the Hosmer–Lemeshow test demonstrating that the predicted probability did not deviate significantly from the actual risk distribution (χ2 = 1.975, P = 0.853). The AUC of the validation group was 0.821 (95%CI: 0.664-0.979), and the Hosmer–Lemeshow test yielded a P value of 0.203 (χ2 = 8.510). The 10-fold cross-validation of the model generated a mean AUC of 0.795 (95%CI: 0.702-0.887), exhibiting its robust discriminative ability across diverse subsets of the training dataset. Figure 2 and Tables 4 and 5 present detailed results.
Figure 2 Receiver operating characteristic curves in the modeling and validation groups.
A: Receiver operating characteristic curve of the modeling group; B: Receiver operating characteristic curve of the validation group. AUC: Area under the curve.
Table 4 Receiver operating characteristic analysis data for the predictive performance of the risk model for delayed postoperative bleeding.
This study conducted comprehensive univariate and multivariate analyses to investigate the risk factors associated with DPOB in patients with ECRC and PCLs. The observed incidence of DPOB in our cohort was 9.93%, which is closely congruent with the 10.30% reported by Yu et al[17].
Our univariate analysis identified several significant predictors, including the number of lesions, lesion size, lesion location, degree of submucosal fibrosis, and intraoperative bleeding. Multivariate analysis further confirmed these independent determinants of DPOB occurrence: Lesion multiplicity, lesions exceeding 3 cm in diameter, rectal lesion location, and severe fibrosis. These clinical correlations are mechanistically supported by several factors: (1) Procedural complexity escalates substantially with multiple lesions, as each requires individual resection. Simultaneous hemostatic demands across multiple sites may compromise procedural accuracy, frequently resulting in inadequate hemostasis and elevated DPOB risk[17]; (2) Lesions with greater diameter demonstrate increased procedural difficulty because of their more extensive vascular involvement, which impairs both complete resection and effective hemostasis. Additionally, the greater surface area requires more extensive coagulation, where any insufficiency permits delayed bleeding manifestations[18-20]; (3) Rectal lesions present unique vascular challenges due to the region’s abundant blood supply with larger arterial and venous diameters, especially in distal rectal locations. This anatomical region benefits from dual perfusion via both the perianal and inferior rectal arteries, creating technical difficulties in achieving complete hemostasis during electrocautery. Inadequate coagulation in this region causes delayed bleeding[21]. Furthermore, the rectum’s physiological role in stool storage subjects the surgical site to mechanical stress during defecation, including pressure and peristaltic contractions, which precipitate vascular rupture and subsequent bleeding[22]; and (4) Submucosal injection is crucial for lesion elevation during ESD. Severe fibrosis in the submucosal layer complicates this process, causing inadequate lesion lifting and unclear dissection planes. Further, fibrotic tissue interferes with effective electrocoagulation, obscuring the surgical field and increasing the likelihood of missing exposed vessels, both of which contribute to DPOB development[23,24].
Previous studies have revealed the influencing factors of DPOB in patients with ECRC and PCLs postoperatively. Cai et al[25] identified the maximum lesion diameter of ≥ 3.0 cm in patients with ECRC as an independent predictor of delayed bleeding after ESD, which is consistent with the results of our study. Similarly, Zhu et al[26] reported that advanced age, female sex, low-grade intraepithelial neoplasia, hypertension, lesion size of ≥ 4 cm, operative duration of ≥ 120 minutes, and the absence of hemostatic clip application were significant risk factors for delayed bleeding after ESD in patients with gastric PCLs, which corroborates and extends the present results. Xu et al[27] identified that risk factors for delayed bleeding after endoscopic resection of colorectal tumors included colorectal tumors located in the proximal colon, a history of antithrombotic medication use, high-grade intraepithelial neoplasia or early cancer, and severe submucosal fibrosis, providing complementary evidence to our findings. Further, Terasaki et al[28] demonstrated that lesion location in the cecum is a crucial independent risk factor for delayed bleeding after ESD for colorectal tumors, which is congruent with our results. The risk prediction model integrating the number of lesions, lesion size, lesion location, and degree of submucosal fibrosis demonstrated robust predictive performance for DPOB in patients undergoing surgery for ECRC and PCLs. The modeling group achieved an AUC of 0.831, whereas the validation group reached an AUC of 0.821. Furthermore, the 10-fold cross-validation generated an average AUC of 0.795, confirming the model’s excellent robustness and generalization performance.
Extensive research has examined DPOB risk factors across diverse clinical contexts. A study by Park et al[29] determined younger age, aspirin use, polyp size, and immediate post-procedural bleeding as significant predictors of post-polypectomy delayed bleeding. Similarly, Aizawa et al[30] revealed that larger polyp size (> 5 mm), clip application, and antithrombotic therapy were significant predictors of delayed bleeding after cold snare polypectomy in patients receiving antithrombotic treatment. Furthermore, Park et al[31] indicated that prophylactic endoscopic clipping may provide protective benefits against delayed bleeding after endoscopic resection of ampullary tumors, although the results did not reach statistical significance. This underscores the necessity for ongoing research to determine and validate effective preventive measures for delayed bleeding in high-risk patient populations.
CONCLUSION
In conclusion, multiple lesions, lesion size of ≥ 3 cm, lesion localization in the rectum, and severe fibrosis are independent risk factors for DPOB in patients with ECRC and PCLs. Patients who present with these clinical characteristics should be closely monitored and managed with targeted interventions to reduce the risk of delayed bleeding. Future studies are recommended to focus on expanding sample sizes to further identify the potential effect of intraoperative bleeding and to refine risk stratification models, thereby ultimately improving clinical outcomes for this patient population.
Footnotes
Provenance and peer review: Unsolicited article; Externally peer reviewed.
Peer-review model: Single blind
Specialty type: Gastroenterology and hepatology
Country of origin: China
Peer-review report’s classification
Scientific Quality: Grade B
Novelty: Grade C
Creativity or Innovation: Grade C
Scientific Significance: Grade B
P-Reviewer: Yang CFJ S-Editor: Bai Y L-Editor: A P-Editor: Xu ZH
Hamada K, Uedo N, Tonai Y, Arao M, Suzuki S, Iwatsubo T, Kato M, Shichijo S, Yamasaki Y, Matsuura N, Nakahira H, Kanesaka T, Yamamoto S, Akasaka T, Hanaoka N, Takeuchi Y, Higashino K, Ishihara R, Okada H, Iishi H, Fukui K, Shimokawa T. Efficacy of vonoprazan in prevention of bleeding from endoscopic submucosal dissection-induced gastric ulcers: a prospective randomized phase II study.J Gastroenterol. 2019;54:122-130.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 38][Cited by in RCA: 39][Article Influence: 6.5][Reference Citation Analysis (0)]
Gomella PT, Solomon J, Ahdoot M, Gurram S, Lebastchi AH, Levy E, Krishnasamy V, Kassin MT, Chang R, Wood BJ, Linehan WM, Ball MW. Timing, incidence and management of delayed bleeding after partial nephrectomy in patients at risk for recurrent, bilateral, multifocal renal tumors.Urol Oncol. 2024;42:222.e1-222.e7.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 2][Cited by in RCA: 5][Article Influence: 5.0][Reference Citation Analysis (0)]
Wang L, Liu ZQ, Zhang JY, Chen SY, Zhong YS, Zhang YQ, Chen WF, Ma LL, Qin WZ, Hu JW, Cai MY, Yao LQ, Li QL, Zhou PH. Management of delayed bleeding of upper gastrointestinal endoscopic submucosal tunneling procedures: a retrospective single-center analysis and brief meta-analysis.J Gastroenterol Hepatol. 2023;38:2174-2184.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 3][Reference Citation Analysis (0)]
