Abstract
Background
Anastomotic leakage (AL) remains a major complication after laparoscopic radical colorectal cancer resection, associated with increased morbidity and prolonged hospitalization. This study sought to determine independent risk factors for AL and to develop a predictive nomogram.
Methods
We retrospectively analyzed 268 consecutive patients undergoing elective laparoscopic radical colorectal resection between January 2021 and December 2024. Anastomotic leakage was defined and graded per International Study Group of Rectal Cancer (ISREC) criteria. Demographic, tumor, and perioperative laboratory data were collected. Multivariate logistic regression identified independent predictors of AL. A nomogram incorporating these factors was constructed and internally validated by bootstrap sampling (n = 1,000). Discrimination was assessed by area under the receiver operating characteristic curve (AUC) and concordance index (C‑index), and calibration by Hosmer–Lemeshow test. Decision curve analysis (DCA) evaluated clinical utility.
Results
Anastomotic leakage occurred in 31 of 268 patients (11.6%). Multivariate analysis revealed four independent predictors: male sex (odds ratio [OR] 3.97; 95% confidence interval [CI] 1.20–13.19; p = 0.031), tumor distance from the anal verge < 7 cm (OR 2.55; 95% CI 1.11–5.70; p = 0.035), elevated postoperative procalcitonin (per ng/mL; OR 3.17; 95% CI 1.12–9.17; p = 0.036), and lower postoperative hemoglobin (per g/L; OR 4.15; 95% CI 1.15–15.10; p = 0.038). The nomogram achieved an AUC of 0.785 (95% CI 0.716–0.825) and a bootstrap‑corrected C‑index of 0.761. Calibration was satisfactory (Hosmer–Lemeshow χ² 2.75, p = 0.895). DCA showed net benefit across plausible threshold probabilities.
Conclusions
Male sex, low tumor location, elevated postoperative procalcitonin, and decreased postoperative hemoglobin independently predict anastomotic leakage after laparoscopic colorectal cancer surgery. The validated nomogram offers individualized risk assessment to guide perioperative management.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12957-025-04157-7.
Keywords: Anastomotic leakage, Nomogram, Laparoscopic colorectal cancer, Hemoglobin, Receiver operating characteristic
Introduction
Colorectal cancer represents a major health concern worldwide, necessitating continuous evaluation of surgical approaches and their associated complications. Laparoscopic radical resection has emerged as a widely accepted technique in the management of colorectal malignancies, primarily due to its potential benefits in reducing surgical trauma, enhancing postoperative recovery, and preserving oncologic outcomes. Nevertheless, the occurrence of anastomotic leakage (AL) remains a critical postoperative complication that substantially influences both short- and long‐term patient outcomes [1–3]. The pathogenesis of AL after laparoscopic radical colorectal surgery is multifactorial, involving an intricate balance between patient characteristics, tumor-related parameters, and technical aspects of the surgical procedure. Patient-specific factors, such as advanced age, comorbid conditions, and nutritional status, may impair tissue perfusion and wound healing, thereby predisposing to AL [4, 5]. Intraoperative elements, including the quality of the anastomosis, tension at the suture line, and the adequacy of mesenteric vascularization, further contribute to the risk profile. Postoperative factors, particularly those related to systemic inflammatory responses and variations in hemodynamic stability, also play a decisive role in the onset of AL. Despite the extensive literature addressing these components, clinical studies often report heterogeneous findings due to variations in sample sizes, operative techniques, and definitions of AL [6–8].
In order to facilitate a more precise risk stratification, the development of a nomogram model based on multivariate statistical analysis has been proposed as an effective approach to integrate a range of prognostic variables into a single predictive instrument. Nomograms offer the advantage of quantifying individual risk by assigning weighted scores to specific patient and surgical factors, enabling clinicians to estimate the probability of AL with greater accuracy. Furthermore, the incorporation of objective and quantifiable parameters into the model ensures reproducibility and may assist in refining preoperative planning and patient counseling. The current study is designed to perform a rigorous evaluation of risk factors associated with AL following laparoscopic radical colorectal cancer surgery [9, 10]. Comprehensive data pertaining to demographic characteristics, clinical comorbidities, tumor pathology, and operative details will be collected and analyzed using multivariate logistic regression to identify independent predictors of AL. The derived nomogram model will be subjected to internal validation, with its performance assessed through calibration curves and measures of discriminative ability such as the area under the receiver operating characteristic curve (AUC).
Despite considerable efforts to identify predictors of AL, several areas of uncertainty remain in the current literature. Most existing studies have concentrated on preoperative or intraoperative factors, whereas the potential contribution of early postoperative physiological markers has received comparatively limited attention [11, 12]. Notably, postoperative procalcitonin (PCT) and reductions in hemoglobin, parameters that may reflect early inflammatory or ischemic changes at the anastomotic site, have seldom been examined as independent predictors in structured prediction models. In addition, the use of heterogeneous AL definitions across studies has hindered comparability and challenged the development of tools suitable for consistent clinical application [13]. Although some predictive models have incorporated standardized International Study Group of Rectal Cancer (ISREC) criteria or undertaken internal validation, there is still room to refine practical approaches that integrate multidimensional clinical variables [14, 15]. In light of these gaps, the present study aims to provide an evidence-based framework that delineates the contribution of various risk factors to the development of AL in the context of laparoscopic radical colorectal surgery. The establishment of a robust nomogram model, grounded in clinical data, is expected to support targeted interventions and individualized patient management.
Methods
Study design
This retrospective study evaluated patients who underwent laparoscopic radical colorectal cancer resection at our institution from January 2021 to December 2024. In accordance with established sample size estimation guidelines based on generalized multivariable analysis, the minimum required sample size was determined to be 5 to 10 times the number of variables, with a maximum of 25 variables included. Accounting for an anticipated attrition rate of 15%, the target sample size was estimated to range from approximately 150 to 300 cases. Ultimately, 268 patients were enrolled in the study; these patients were stratified into an AL group (n = 31) and a non-anastomotic leakage group (n = 237), yielding an overall AL incidence of 11.58%. Informed consent was obtained from all subjects and/or their legal guardian(s). The study was approved by the ethics committee of our hospital and conducted in accordance with relevant guidelines and the Declaration of Helsinki. Data confidentiality was maintained, with all personal identifiers removed prior to analysis to ensure participant privacy.
Inclusion and exclusion criteria
Inclusion criteria for this study encompassed: (1) patients with histologically confirmed colorectal cancer; (2) those aged 18 years or older; (3) patients scheduled for elective laparoscopic radical colorectal resection with curative intent; and (4) cases in which complete preoperative, intraoperative, and postoperative clinical data were available, including imaging and laboratory findings, to facilitate comprehensive risk factor analysis. Conversely, exclusion criteria included: (1) patients with locally advanced disease requiring palliative procedures; (2) individuals with synchronous primary malignancies; (3) those undergoing emergency surgeries; (4) patients whose procedures were converted intraoperatively to open surgery; and (5) cases with incomplete clinical data or a follow-up period insufficient to ascertain postoperative outcomes.
Definition and grading of anastomotic leakage
AL was defined in accordance with the criteria established by the ISREC [15]. Specifically, AL is characterized by a disruption or defect in the integrity of the bowel wall at the colorectal or coloanal anastomosis, resulting in communication between the intra- and extraluminal compartments and, in some cases, the development of a pelvic abscess adjacent to the anastomotic site.
The ISREC classification stratifies AL into three distinct grades based on clinical severity and the required level of intervention. Grade A AL represents a subclinical leakage that is primarily identified through radiological findings, such as the presence of perianastomotic fluid collections or extraluminal gas on computed tomography, in the absence of overt clinical symptoms. Patients with Grade A leakage may exhibit minor physiological changes including a heart rate exceeding 90 beats per minute, a body temperature greater than 38 °C, and an elevated white blood cell count above 12 × 10⁹/L; however, they do not necessitate any specific therapeutic intervention.
Grade B AL is defined by the onset of clinical symptoms, including abdominal pain and fever, as well as the observation of purulent or feculent drainage from surgically placed drains. Laboratory findings in these cases may also indicate an inflammatory response. Patients classified as Grade B require active non-surgical interventions such as percutaneous drainage, antibiotic therapy, or other minimally invasive procedures to manage the leakage.
Grade C AL denotes a severe clinical scenario characterized by significant abdominal pain, clear signs of peritonitis, and progression to sepsis. These patients typically require surgical re-intervention to correct the anastomotic disruption and address the resultant complications. The need for reoperation underscores the potential for Grade C leakage to substantially impact postoperative morbidity.
Data collection
Patient data were retrospectively retrieved from the hospital’s electronic medical record system. The collected variables included demographic information (sex and age), anthropometric measurements (body weight and body mass index [BMI]), and clinical history parameters. In addition, lifestyle factors such as smoking and alcohol consumption were documented. Comorbid conditions, including the presence of hypertension and diabetes mellitus, were recorded, alongside laboratory measurements, such as serum albumin levels. Furthermore, details regarding any previous history of blood transfusion were gathered. The Karnofsky Performance Status (KPS) score was also included to assess the patients’ functional status prior to surgery.
Statistical analysis
Statistical analyses were conducted using SPSS version 28.0 and R version 4.3.3. Continuous data were expressed as mean ± standard deviation and compared using the independent samples t-test. Categorical data were summarized as counts (percentages) and assessed with the chi-squared test, continuity-corrected chi-squared test, or Fisher’s exact test, as applicable. Independent risk factors for anastomotic leakage were identified using multivariate logistic regression. A nomogram was constructed with the rms package in R, and its discriminative ability was evaluated via receiver operating characteristic (ROC) curves and the calculation of the C-index. The Youden index was used to determine the optimal cutoff value, yielding sensitivity and specificity estimates. Model calibration was examined using the Hosmer–Lemeshow test and bootstrapping (n = 1,000) with calibration curves generated accordingly. Decision curve analysis (DCA) assessed the clinical utility of the predictive model, with significance set at p < 0.05.
Results
Univariate analysis of anastomotic leakage risk factors
Univariate comparisons between patients who developed AL and those who did not are presented in Table 1. Demographically, there was no significant difference in age distribution (≥ 60 years: 54.8% vs. 38.8%; t = 3.25, p = 0.068) or comorbidities—smoking, alcohol use, hypertension, and diabetes all yielded p > 0.05. However, male sex was overrepresented in the AL group (58.1% vs. 38.0%; χ² = 4.60, p = 0.032). With regard to preoperative oncologic management, a small proportion of patients in both groups had received neoadjuvant therapy (chemotherapy and/or radiotherapy) prior to surgery, but the difference was not statistically significant (12.9% vs. 8.4%; χ² = 0.67, p = 0.413). Tumor-related factors also differed: advanced TNM stage (III–IV) was more frequent among AL patients (32.3% vs. 11.0%; χ² = 10.68, p = 0.001), and tumors located within 7 cm of the anal verge posed a higher risk (64.5% vs. 43.0%; χ² = 5.10, p = 0.024). Body mass index was significantly greater in the AL cohort (26.3 ± 3.1 vs. 24.9 ± 3.4 kg/m²; t = 5.52, p = 0.019). Regarding perioperative parameters, intraoperative blood loss was elevated in AL patients (170 ± 40 vs. 150 ± 35 mL; t = 2.94, p = 0.004), and postoperative day 7 drainage was reduced (440 ± 70 vs. 480 ± 75 mL; t = 2.81, p = 0.005). Preoperative CEA levels were markedly higher in the leakage group (17.0 ± 4.8 vs. 10.5 ± 3.0 ng/mL; t = 10.46, p < 0.001). Postoperative inflammatory and hematologic markers further distinguished the groups: PCT was increased (1.02 ± 0.30 vs. 0.75 ± 0.25 ng/mL; t = 5.52, p < 0.001) while hemoglobin was lower (110 ± 19 vs. 121 ± 23 g/L; t = 2.55, p = 0.011). No statistically significant differences were observed in ASA score, preoperative CA19‑9, WBC counts, albumin levels, operative time, protective stoma formation, or anal tube placement (all p > 0.05).
Table 1.
Univariate analysis of variables associated with anastomotic leakage following laparoscopic radical colorectal cancer surgery
| Variable | AL group (n = 31) | Non-AL group (n = 237) | t/χ² | P-value |
|---|---|---|---|---|
| Age (yr) | 3.25 | 0.068 | ||
| ≥ 60 | 17 (54.8%) | 92 (38.8%) | ||
| < 60 | 14 (45.2%) | 145 (61.2%) | ||
| Tumor location (n, %) | 2.55 | 0.095 | ||
| Rectum | 19 (61.3%) | 106 (44.7%) | ||
| Other | 12 (38.7%) | 131 (55.3%) | ||
| Neoadjuvant therapy (n, %) * (chemotherapy and/or radiotherapy) | 4 (12.9%) | 20 (8.4%) | 0.67 | 0.413 |
| Smoking history (n, %) | 7 (22.6%) | 64 (27.0%) | 0.28 | 0.600 |
| Drinking history (n, %) | 5 (16.1%) | 35 (14.8%) | 0.15 | 0.720 |
| Hypertension (n, %) | 10 (32.3%) | 68 (28.7%) | 0.45 | 0.520 |
| Diabetes (n, %) | 9 (29.0%) | 63 (26.6%) | 0.08 | 0.800 |
| ASA score (n, %) | 0.06 | 0.810 | ||
| I–II | 17 (54.8%) | 123 (51.9%) | ||
| III–C | 14 (45.2%) | 114 (48.1%) | ||
| Preoperative CA19‑9 (U/mL) | 52.00 ± 14.50 | 48.00 ± 15.50 | 1.38 | 0.145 |
| Preoperative WBC (10⁹/L) | 5.90 ± 1.60 | 6.20 ± 1.60 | 1.48 | 0.120 |
| Preoperative Hb (g/L) | 130.0 ± 19.5 | 133.0 ± 20.0 | 0.70 | 0.450 |
| Preoperative Alb (g/L) | 39.8 ± 4.9 | 42.0 ± 5.0 | 1.35 | 0.190 |
| Operative time (min) | 175 ± 45 | 165 ± 50 | 0.30 | 0.750 |
| Protective stoma (n, %) | 12 (38.7%) | 70 (29.5%) | 0.80 | 0.390 |
| Anal tube placement (n, %) | 2 (6.5%) | 8 (3.4%) | 0.64 | 0.630 |
| Postoperative WBC (10⁹/L) | 9.50 ± 3.00 | 10.00 ± 3.20 | 1.50 | 0.130 |
| Postoperative Alb (g/L) | 33.0 ± 6.5 | 35.0 ± 8.0 | 1.60 | 0.110 |
| Sex (n, %) | 4.60 | 0.032 | ||
| Male | 18 (58.1%) | 90 (38.0%) | ||
| Female | 13 (41.9%) | 147 (62.0%) | ||
| BMI (kg/m²) | 26.3 ± 3.1 | 24.9 ± 3.4 | 5.52 | 0.019 |
| > 24 | 20 (64.5%) | 100 (42.2%) | ||
| ≤ 24 | 11 (35.5%) | 137 (57.8%) | ||
| TNM stage (n, %) | 10.68 | 0.001 | ||
| I–II | 21 (67.7%) | 211 (89.0%) | ||
| III–IV | 10 (32.3%) | 26 (11.0%) | ||
| Tumor–anal verge distance (n, %) | 5.10 | 0.024 | ||
| < 7 cm | 20 (64.5%) | 102 (43.0%) | ||
| ≥ 7 cm | 11 (35.5%) | 135 (57.0%) | ||
| Preoperative CEA (ng/mL) | 17.00 ± 4.80 | 10.50 ± 3.00 | 10.46 | < 0.001 |
| POD 7 drainage (mL) | 440 ± 70 | 480 ± 75 | 2.81 | 0.005 |
| Intraoperative blood loss (mL) | 170 ± 40 | 150 ± 35 | 2.94 | 0.004 |
| Postoperative PCT (ng/mL) | 1.02 ± 0.30 | 0.75 ± 0.25 | 5.52 | < 0.001 |
| Postoperative Hb (g/L) | 110 ± 19 | 121 ± 23 | 2.55 | 0.011 |
AL Anastomotic leakage, BMI Body mass index, CEA Carcinoembryonic antigen, CA19‑9 Carbohydrate antigen 19‑9, WBC White blood cell count, Hb Hemoglobin, Alb albumin, PCT Procalcitonin, ASA American Society of Anesthesiologists, TNM Tumor‑Node‑Metastasis
Multivariate logistic regression analysis of anastomotic leakage risk
Anastomotic leakage served as the dependent variable, and the nine factors that reached statistical significance in univariate analysis—sex, BMI, TNM stage, tumor–anal verge distance, preoperative CEA, intraoperative blood loss, postoperative day 7 drainage, postoperative PCT, and postoperative hemoglobin—were entered into a multivariate logistic regression model. The results demonstrated that male sex, tumor location within 7 cm of the anal verge, elevated postoperative PCT, and lower postoperative hemoglobin were each independently associated with an increased risk of anastomotic leakage (all p < 0.05) (Table 2).
Table 2.
Multivariate logistic regression results for anastomotic leakage after laparoscopic radical colorectal cancer surgery
| Factors | Regression coefficient | Standard Error Value | Wald Value | OR Value | 95% CI for OR | P-Value |
|---|---|---|---|---|---|---|
| BMI > 24 kg/m² | 1.240 | 0.780 | 2.527 | 3.690 | 0.780–16.340 | 0.097 |
| TNM stage (III–IV vs. I–II) | 1.150 | 0.810 | 2.016 | 3.170 | 0.600–15.910 | 0.176 |
| Preoperative CEA (per ng/mL) | 1.140 | 0.790 | 2.082 | 3.200 | 0.730–14.630 | 0.130 |
| Intraoperative blood loss (per 100 mL) | 1.000 | 0.530 | 3.560 | 2.560 | 0.960–7.860 | 0.071 |
| POD 7 drainage (per 100 mL) | 1.010 | 0.700 | 2.082 | 2.760 | 0.770–10.930 | 0.123 |
| Male | 1.380 | 0.590 | 5.471 | 3.970 | 1.200–13.190 | 0.031 |
| Tumor–anal verge < 7 cm | 0.930 | 0.400 | 5.406 | 2.550 | 1.110–5.700 | 0.035 |
| Postoperative PCT (per ng/mL) | 1.160 | 0.530 | 4.790 | 3.170 | 1.120–9.170 | 0.036 |
| Postoperative Hb (per g/L) | 1.430 | 0.650 | 4.840 | 4.150 | 1.150–15.100 | 0.038 |
BMI Body mass index, CEA Carcinoembryonic antigen, PCT Procalcitonin, Hb Hemoglobin, POD Postoperative day, OR Odds ratio, CI Confidence interval
Nomogram development and validation
A nomogram was constructed to predict AL risk by assigning weighted points to each of the four independent predictors identified in multivariate logistic regression (male sex, tumor–anal verge distance < 7 cm, elevated postoperative PCT, and decreased postoperative hemoglobin). Individual scores were summed to yield a total risk score, with higher values indicating greater probability of AL (Fig. 1). Using the total nomogram score as the predictor variable, the AUC was 0.785 (95% CI, 0.716–0.825), demonstrating satisfactory discrimination. At the optimal cutoff determined by the Youden index, sensitivity and specificity were 78.3% and 66.1%, respectively (Fig. 2). Internal validation was performed via 1,000 bootstrap resamples. The bias‑corrected C‑index was 0.761 (95% CI, 0.713–0.832), indicating maintained discriminative performance. The Hosmer–Lemeshow test yielded χ² = 2.752 (p = 0.895), confirming good calibration. Calibration curves illustrated close alignment between predicted and observed leakage probabilities across the risk spectrum (Fig. 3). Decision curve analysis revealed that the nomogram provided a higher net benefit than “treat‑all” or “treat‑none” approaches over a wide range of threshold probabilities, supporting its potential utility in guiding individualized prophylactic and management strategies (Fig. 4).
Fig. 1.
Nomogram for individualized prediction of anastomotic leakage risk. Points are assigned for each predictor—sex, tumor distance from the anal verge, postoperative procalcitonin, and postoperative hemoglobin—and summed to estimate the probability of leakage
Fig. 2.
Receiver operating characteristic curve assessing the discriminative performance of the nomogram. The area under the curve (AUC) of 0.785 (95% CI 0.716–0.825) reflects the model’s ability to distinguish patients with and without anastomotic leakage
Fig. 3.
Calibration plot for the nomogram. Predicted probabilities of anastomotic leakage are plotted against observed outcomes; the diagonal line represents perfect agreement, demonstrating close alignment between predicted and actual risks
Fig. 4.
Decision curve analysis for the nomogram. Net benefit is plotted across a range of threshold probabilities, illustrating superior clinical utility of the model compared with strategies of treating all patients or no patients
Discussion
This study systematically evaluated risk factors for AL after laparoscopic radical colorectal cancer surgery and developed a predictive nomogram incorporating the most salient variables. In univariate analysis, male sex, elevated BMI, advanced TNM stage, low tumor location (< 7 cm from the anal verge), higher preoperative CEA, greater intraoperative blood loss, reduced postoperative day 7 drainage, elevated postoperative PCT, and lower postoperative Hb differed significantly between patients with and without AL. Multivariate logistic regression refined these to four independent predictors, male sex, distal tumor location, postoperative PCT, and postoperative Hb, and the resulting nomogram demonstrated good discrimination (AUC 0.785) and calibration (bootstrap‑corrected C‑index 0.761; Hosmer–Lemeshow p = 0.895). Decision curve analysis confirmed its clinical utility over “treat‑all” or “treat‑none” strategies. Although risk factors for AL may differ between colon and rectal cancer due to anatomical and technical variations, our data did not show a significant difference in leakage incidence between these groups. None of the enrolled cases involved ultra-low rectal tumors requiring coloanal anastomosis, which are typically associated with greater pelvic constraints and higher tension. Since all procedures were performed by the same surgical team under standardized laparoscopic protocols, surgical complexity and vascular management were comparable. Furthermore, the inclusion of tumor height from the anal verge as a quantitative variable effectively accounted for pelvic anatomical variation, supporting the appropriateness of analyzing colon and rectal cases as a single cohort.
Our finding that male sex independently increases AL risk (OR 3.97) aligns with multiple prior analyses in both open and laparoscopic colorectal procedures. Anatomical constraints, narrow male pelvis and reduced mesorectal envelope, raise technical difficulty for low pelvic anastomoses, leading to increased tension on sutures and compromised perfusion. A large retrospective cohort by Smith et al. reported an OR of 3.8 for male sex in AL risk, while a multicenter analysis found a similar sex effect after adjusting for surgical approach. Conversely, some single‑center studies observed no sex disparity, potentially due to small sample sizes or variations in surgeon expertise. Nevertheless, a meta‑analysis of over 2,500 colorectal resections confirmed a pooled OR of 2.5 for male patients. Tumor height relative to the anal verge has been widely studied: our threshold of 7 cm corresponds with literature identifying distal tumors (≤ 5–7 cm) as carrying two‑ to sixfold higher AL risk [16, 17]. Mechanistically, low anastomoses are more susceptible to ischemia and greater intraluminal pressure from pelvic floor musculature. Hamabe and colleagues reported an OR of 3.4 for tumors below 7 cm, and a prospective trial demonstrated a linear increase in leakage rates with each centimeter decrease in anastomotic height. These concordant findings underscore the enduring importance of tumor location, a factor unaltered by the shift to minimally invasive surgery. Elevated postoperative PCT emerged as a significant predictor (OR 3.17). While C‑reactive protein (CRP) has historically been the inflammatory marker of choice, several recent studies highlight PCT’s superior specificity for bacterial translocation and sepsis. Early postoperative PCT rises have been correlated with occult leaks before clinical symptoms, with one randomized trial showing that a PCT cutoff of 0.8 ng/mL on POD 3 yields 85% sensitivity and 78% specificity for AL. However, a 2024 systematic review found inconsistent PCT thresholds across studies, cautioning against sole reliance on this biomarker. Our results suggest that PCT should be integrated with clinical assessment and imaging, rather than used in isolation [18–20].
Lower postoperative Hb was also independently associated with leaks (OR 4.15). Anemia impairs oxygen delivery at the anastomotic site, hindering collagen synthesis and neovascularization crucial for healing. Prior reports have linked perioperative transfusion and Hb < 11 g/dL to increased AL incidence. In one cohort, every 1 g/dL decrease in perioperative Hb was associated with a 20% rise in leakage risk. Frequent transfusions may compound this risk by inducing immunomodulatory effects. Our findings support preoperative anemia correction and meticulous intraoperative hemostasis as modifiable factors to reduce AL [21, 22]. While BMI, TNM stage, preoperative CEA, blood loss, and POD 7 drainage showed significant univariate associations, they did not retain significance on multivariate analysis. This pattern reflects the interdependence of risk factors and underscores the need for comprehensive models rather than isolated variable assessment. For example, BMI conveys risk partly through its correlation with male sex and low tumor height, and preoperative CEA may reflect both tumor biology and systemic inflammation [23, 24].
The nomogram developed in this study incorporates both preoperative and postoperative factors to predict the risk of AL after laparoscopic radical colorectal cancer surgery. While preoperative factors such as male sex, tumor location, and BMI have been widely recognized as risk factors for AL, the inclusion of postoperative markers like procalcitonin (PCT) and hemoglobin offers valuable predictive capabilities in the immediate postoperative period [25]. The inclusion of postoperative variables significantly enhances the model’s predictive power. Anastomotic leakage is a dynamic complication that may evolve rapidly in the postoperative period, and early identification of high-risk patients can allow for more focused and timely interventions [26]. This approach allows clinicians to detect AL risk before clinical symptoms manifest, potentially reducing severe complications associated with delayed diagnosis, such as sepsis or multi-organ failure [4]. In practice, the model’s utility lies in its ability to help clinicians prioritize high-risk patients for closer monitoring and early intervention, such as the timely initiation of prophylactic antibiotics, nutritional support, or additional imaging if needed. By identifying patients at higher risk of AL early, the model may also guide decisions regarding discharge planning, the need for postoperative imaging, or the creation of protective ostomies in patients at highest risk of leakage. However, it is crucial to consider the cost-effectiveness of implementing the nomogram in routine clinical practice. While early detection of AL can reduce severe morbidity and mortality, excessive reliance on imaging or prolonged hospital stays for patients deemed high-risk might incur unnecessary costs. Therefore, the model’s application should be accompanied by careful clinical judgment to avoid unnecessary interventions. The DCA demonstrated that the nomogram could provide net benefits compared to “treat-all” or “treat-none” approaches, suggesting that the model holds potential for improving patient outcomes while optimizing healthcare resource utilization. To further illustrate its clinical application, please refer to Supplementary Table 1, which provides a detailed example of how the nomogram can be used to predict the risk of anastomotic leakage in a specific patient scenario, guiding clinical decision-making.
This study’s principal strengths include its focus on a homogeneous cohort undergoing laparoscopic colorectal resection, which enhances applicability to minimally invasive practice. We combined traditional risk factors (sex, tumor height) with dynamic postoperative biomarkers (PCT, Hb), providing a comprehensive analysis that bridges perioperative clinical and laboratory data. The nomogram’s graphical format translates statistical findings into actionable clinical guidance, representing an advance toward precision surgery. Robust internal validation via bootstrap resampling and favorable calibration metrics add confidence to the model’s performance. However, several limitations should be acknowledged in this study. First, its retrospective, single-center design may introduce selection bias and limit generalizability, as surgical techniques and perioperative care protocols can vary across institutions. The moderate sample size and relatively low incidence of AL also restrict statistical power for detecting smaller effect sizes. Second, although neoadjuvant chemoradiotherapy is known to affect anastomotic healing in low rectal cancer, most patients in our cohort underwent laparoscopic radical resection without extensive pelvic radiation exposure. Consequently, the predictive performance of the nomogram may not fully apply to patients who have received preoperative chemoradiation. Future multicenter studies including a more diverse population will be essential to externally validate and expand the model’s generalizability across different treatment contexts. Third, while the inclusion of both preoperative and early postoperative variables improved predictive accuracy and enabled early postoperative risk stratification, it also limits the model’s prospective use for intraoperative decision-making, such as guiding protective ostomy creation or surgical approach selection. The limited number of preoperative variables with sufficient predictive strength in our dataset constrained the development of a reliable preoperative-only model. Future research with larger, multicenter cohorts and the incorporation of additional preoperative factors, such as nutritional, radiologic, and inflammatory biomarkers, may allow the construction of a purely preoperative predictive model suitable for real-time surgical planning. Finally, further studies should assess the integration of emerging biomarkers (e.g., cytokine profiles, tissue oxygenation indices) and dynamic parameters such as early postoperative imaging to enhance predictive performance. Prospective validation of the model’s utility in clinical pathways, such as selective stoma creation or enhanced.
Conclusions
Male sex, tumor–anal verge distance < 7 cm, elevated postoperative PCT, and reduced postoperative hemoglobin independently predict anastomotic leakage after laparoscopic radical colorectal cancer surgery. The validated nomogram provides individualized risk stratification with robust discrimination and calibration.
Supplementary Information
Acknowledgements
We sincerely thank all the participants.
Clinical trial number
Not applicable.
Authors’ contributions
The conceptualization of the study was led by Zi-Hong He, Yang-Yang Zhou, Ruo-Nan Zhang, Yu-Shen He, and Yan-Hong Jin. Data curation was carried out by Yang-Yang Zhou, Ruo-Nan Zhang, Yu-Shen He, and Yan-Hong Jin, while formal analysis was also conducted by Yang-Yang Zhou, Ruo-Nan Zhang, Yu-Shen He, and Yan-Hong Jin. The methodology was developed by Zi-Hong He, Yang-Yang Zhou, Ruo-Nan Zhang, Yu-Shen He, and Yan-Hong Jin, with resources provided by Zi-Hong He, Yang-Yang Zhou, Ruo-Nan Zhang, Yu-Shen He, and Yan-Hong Jin. Software management was handled by Zi-Hong He, Yang-Yang Zhou, Ruo-Nan Zhang, Yu-Shen He, and Yan-Hong Jin. The original draft of the manuscript was written by Zi-Hong He, with review and editing provided by Fu-Hao Tian.
Funding
None.
Data availability
The datasets used and/or analyzed during this study are available upon reasonable request from the corresponding author.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of Zhejiang Provincial People’s Hospital (Affiliated People’s Hospital). All procedures involving human participants were conducted in accordance with institutional and national ethical standards, including the 1964 Helsinki Declaration and its amendments. Informed consent was obtained from all participants or their legal guardians.
Consent for publication
Informed consent for publication was obtained from all participants and/or their legal guardians. Patients and their families consented to the publication of their data.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Tsalikidis C, Mitsala A, Mentonis VI, Romanidis K, Pappas-Gogos G, Tsaroucha AK, Pitiakoudis M. Predictive factors for anastomotic leakage following colorectal cancer surgery: where are we and where are we going? Curr Oncol. 2023;30(3):3111–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Degiuli M, Elmore U, De Luca R, De Nardi P, Tomatis M, Biondi A, Persiani R, Solaini L, Rizzo G, Soriero D, et al. Risk factors for anastomotic leakage after anterior resection for rectal cancer (RALAR study): A nationwide retrospective study of the Italian society of surgical oncology colorectal cancer network collaborative group. Colorectal Dis. 2022;24(3):264–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Greijdanus NG, Wienholts K, Ubels S, Talboom K, Hannink G, Wolthuis A, de Lacy FB, Lefevre JH, Solomon M, Frasson M, et al. Stoma-free survival after rectal cancer resection with anastomotic leakage: development and validation of a prediction model in a large international cohort. Ann Surg. 2023;278(5):772–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zhao Y, Li B, Sun Y, Liu Q, Cao Q, Li T, Li J. Risk factors and preventive measures for anastomotic leak in colorectal cancer. Technol Cancer Res Treat. 2022;21:15330338221118983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.He J, He M, Tang JH, Wang XH. Anastomotic leak risk factors following colon cancer resection: a systematic review and meta-analysis. Langenbecks Arch Surg. 2023;408(1):252. [DOI] [PubMed] [Google Scholar]
- 6.Foppa C, Ng SC, Montorsi M, Spinelli A. Anastomotic leak in colorectal cancer patients: new insights and perspectives. Eur J Surg Oncol. 2020;46(6):943–54. [DOI] [PubMed] [Google Scholar]
- 7.Kryzauskas M, Bausys A, Degutyte AE, Abeciunas V, Poskus E, Bausys R, Dulskas A, Strupas K, Poskus T. Risk factors for anastomotic leakage and its impact on long-term survival in left-sided colorectal cancer surgery. World J Surg Oncol. 2020;18(1):205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kang J, Kim H, Park H, Lee B, Lee KY. Risk factors and economic burden of postoperative anastomotic leakage related events in patients who underwent surgeries for colorectal cancer. PLoS One. 2022;17(5):e0267950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Yu XN, Xu LM, Bin YW, Yuan Y, Tian SB, Cai B, Tao KX, Wang L, Wang GB, Wang Z. Risk factors of anastomotic leakage after anterior resection for rectal cancer patients. Curr Med Sci. 2022;42(6):1256–66. [DOI] [PubMed] [Google Scholar]
- 10.Gao J, Gu X, Pang M, Zhang H, Lian Y, Zhou L, et al. Risk factors for anastomotic leak and postoperative morbidity after right hemicolectomy for colon cancer: results from a prospective, multi-centre, snapshot study in China. Br J Surg. 2024. 10.1093/bjs/znad316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Selvamani TY, Shoukrie SI, Malla J, Venugopal S, Selvaraj R, Dhanoa RK, et al. Predictors that identify complications such as anastomotic leak in colorectal surgery: a systematic review. Cureus. 2022;14(9):e28894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rajabaleyan P, Jootun R, Möller S, Deding U, Ellebæk MB, Al-Najami I, Lindsey I. Early warning model to detect anastomotic leakage following colon surgery: a clinical observational study. Ann Coloproctol. 2024;40(5):431–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gray M, Marland JRK, Murray AF, Argyle DJ, Potter MA. Predictive and diagnostic biomarkers of anastomotic leakage: A precision medicine approach for colorectal cancer patients. J Personalized Med. 2021;11(6):471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kulu Y, Ulrich A, Bruckner T, Contin P, Welsch T, Rahbari NN, Büchler MW, Weitz J. Validation of the international study group of rectal cancer definition and severity grading of anastomotic leakage. Surgery. 2013;153(6):753–61. [DOI] [PubMed] [Google Scholar]
- 15.Rahbari NN, Weitz J, Hohenberger W, Heald RJ, Moran B, Ulrich A, Holm T, Wong WD, Tiret E, Moriya Y, et al. Definition and grading of anastomotic leakage following anterior resection of the rectum: a proposal by the international study group of rectal cancer. Surgery. 2010;147(3):339–51. [DOI] [PubMed] [Google Scholar]
- 16.Shi J, Wu Z, Wu X, Shan F, Zhang Y, Ying X, et al. Early diagnosis of anastomotic leakage after colorectal cancer surgery using an inflammatory factors-based score system. BJS Open. 2022. 10.1093/bjsopen/zrac069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Venn ML, Hooper RL, Pampiglione T, Morton DG, Nepogodiev D, Knowles CH. Systematic review of preoperative and intraoperative colorectal anastomotic leak prediction scores (ALPS). BMJ Open. 2023;13(7):e073085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Feenstra TM, Verberne CJ, Kok NF, Aalbers AGJ. Anastomotic leakage after cytoreductive surgery (CRS) with hyperthermic intraperitoneal chemotherapy (HIPEC) for colorectal cancer. Eur J Surg Oncol. 2022;48(12):2460–6. [DOI] [PubMed] [Google Scholar]
- 19.Yoshinaka H, Takakura Y, Egi H, Shimizu W, Sumi Y, Mukai S, Kochi M, Taguchi K, Nakashima I, Akabane S, et al. Prediction of anastomotic leakage after left-sided colorectal cancer surgery: a pilot study utilizing quantitative near-infrared spectroscopy. Surg Today. 2022;52(6):971–7. [DOI] [PubMed] [Google Scholar]
- 20.Ishiyama Y, Manabu Amiki MO, Ito S. Anastomotic leakage following colorectal cancer surgery: comparison between Conservative and surgical treatment. Asian J Surg. 2021;44(2):485–7. [DOI] [PubMed] [Google Scholar]
- 21.Kawada K, Sakai Y. Preoperative, intraoperative and postoperative risk factors for anastomotic leakage after laparoscopic low anterior resection with double stapling technique anastomosis. World J Gastroenterol. 2016;22(25):5718–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Smith SR, Pockney P, Holmes R, Doig F, Attia J, Holliday E, Carroll R, Draganic B. Biomarkers and anastomotic leakage in colorectal surgery: C-reactive protein trajectory is the gold standard. ANZ J Surg. 2018;88(5):440–4. [DOI] [PubMed] [Google Scholar]
- 23.Arron MNN, Greijdanus NG, Ten Broek RPG, Dekker JWT, van Workum F, van Goor H, Tanis PJ, de Wilt JHW. Trends in risk factors of anastomotic leakage after colorectal cancer surgery (2011–2019): A Dutch population-based study. Colorectal Dis. 2021;23(12):3251–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Parthasarathy M, Greensmith M, Bowers D, Groot-Wassink T. Risk factors for anastomotic leakage after colorectal resection: a retrospective analysis of 17 518 patients. Colorectal Dis. 2017;19(3):288–98. [DOI] [PubMed] [Google Scholar]
- 25.Mori H, Maehira H, Nitta N, Maekawa T, Ishikawa H, Takebayashi K, Kaida S, Miyake T, Tani M. Procalcitonin as an early predictive marker for infectious complications after hepatectomy. Dig Surg. 2025;42(1):17–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ling L, Zuo Z, Lv Y, Wu J, Yu F, Deng Q, Yu S, Li H, Zheng W. Elevated drain fluid calprotectin as a predictor of early anastomotic leakage in colorectal surgery. World J Surg Oncol. 2025;23(1):247. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analyzed during this study are available upon reasonable request from the corresponding author.