Serum nutritional predictive biomarkers and risk assessment for anastomotic leakage after laparoscopic surgery in rectal cancer patients

Abstract

BACKGROUND

Anastomotic leakage (AL) is one of the severest complications after laparoscopic surgery for middle/low rectal cancer, significantly impacting patient outcomes. Identifying reliable predictive factors for AL remains a clinical challenge. Serum nutritional biomarkers have been implicated in surgical outcomes but are underexplored as predictive tools for AL in this setting. Our study hypothesizes that preoperative serum levels of prealbumin (PA), albumin (ALB), and transferrin (TRF), along with surgical factors, can accurately predict AL risk.

AIM

To determine the predictive value of preoperative serum nutritional biomarkers for rectal cancer AL following laparoscopic surgery.

METHODS

In the retrospective cohort study carried out at a tertiary cancer center, we examined 560 individuals who underwent laparoscopic procedures for rectal cancer from 2018 to 2022. Preoperative serum levels of PA, ALB, and TRF were measured. We employed multivariate logistic regression to determine the independent risk factors for AL, and a predictive model was constructed and evaluated using receiver operating characteristic curve analysis.

RESULTS

AL occurred in 11.96% of cases, affecting 67 out of 560 patients. Multivariate analysis identified PA, ALB, and TRF as the independent risk factor, each with an odds ratio of 2.621 [95% confidence interval (CI): 1.582-3.812, P = 0.012], 3.982 (95%CI: 1.927-4.887, P = 0.024), and 2.109 (95%CI: 1.162-2.981, P = 0.031), respectively. Tumor location (< 7 cm from anal verge) and intraoperative bleeding ≥ 300 mL also increased AL risk. The predictive model demonstrated an excellent accuracy, achieving an area under the receiver operating characteristic curve of 0.942, a sensitivity of 0.844, and a specificity of 0.922, demonstrating an excellent ability to discriminate.

CONCLUSION

Preoperative serum nutritional biomarkers, combined with surgical factors, reliably predict anastomotic leakage risk after rectal cancer surgery, highlighting their importance in preoperative assessment.

Key Words: Rectal cancer; Laparoscopic operation; Anastomotic leakage; Albumin; Prealbumin; Transferrin

Core Tip: This study establishes a robust predictive model for anastomotic leakage in middle/low rectal cancer patients undergoing laparoscopic surgery, leveraging preoperative serum levels of prealbumin, albumin, and transferrin. Demonstrating a high precision with an area under the receiver operating characteristic curve of 0.942, this model underscores the critical role of nutritional status in preoperative assessments. By integrating these serum nutritional biomarkers, our approach enhances the prediction of postoperative complications and aids in refining pre-surgical strategies to enhance patient results.

INTRODUCTION

Colorectal cancer (CRC) ranks among the most prevalent malignancies globally, posing a significant public health threat. According to World Health Organization reports, CRC is the third most common cancer worldwide and the second leading cause of cancer-related deaths[1,2]. Annually, more than 1.9 million new cases of CRC are diagnosed, leading to nearly 935000 deaths[3]. CRC not only imposes a substantial burden on patients both physically and psychologically but also exerts considerable economic pressure on the global health system. With advancements in medical technology, laparoscopic surgery has become an essential approach for treating middle and low rectal cancer. Compared to traditional open surgery, laparoscopic surgery provides multiple benefits, including less trauma, quicker recovery, shorter hospitalization, and fewer postoperative complications[4]. A series of randomized controlled trials and other studies have verified that laparoscopic surgery for patients with middle/low rectal cancer yields favorable oncological outcomes comparable to open surgery, and may even show better prognoses in some aspects[5]. However, despite the development of laparoscopic techniques enhancing the safety and efficacy of surgery, the overall survival rate for rectal cancer is still influenced by postoperative complications, especially anastomotic leakage (AL).

AL is defined as the leakage of luminal contents from a surgical join through an unnatural passage into the abdominal cavity following abdominal surgery, especially following rectal cancer resection. The incidence of AL in laparoscopic surgery for rectal cancer is reported to vary between 3% to 15%, depending on surgical skill, individual patient differences, and the quality of postoperative management[6]. Although this incidence rate is relatively low, the clinical consequences of AL can be extremely severe, including increased risk of death, prolonged hospital stays, and significantly increased medical costs. From a clinical perspective, AL not only elevates the mortality risk for patients but also can result in a range of severe complications such as intra-abdominal infection, peritonitis, and even organ failure[7-9]. Therefore, even among rectal cancer patients who achieve tumor eradication, AL can significantly decrease the quality of life. Lots of researches have indicated that patients with AL exhibit lower overall and disease-free survival rates compared to those without AL, in addition to extended durations of hospital stay. On average, the hospital stay for patients with AL is more than twice as long as for those without this complication, not only increasing the burden on patients but also imposing significant economic stress on the healthcare system. The occurrence of AL can augment the treatment costs by nearly $30000[10]. In light of the serious consequences of AL, medical research continues to explore strategies for its prevention and management. This includes optimizing surgical techniques, improving the accuracy of preoperative assessments, and implementing more personalized measures in postoperative care. Despite these efforts, AL remains one of the most challenging complications in laparoscopic surgery for rectal cancer.

Serum nutritional biomarkers such as albumin (ALB), prealbumin (PA), and transferrin (TRF) are widely utilized to assess the nutritional status of patients and its effect on surgical outcomes[11]. These biomarkers are clinically significant due to their sensitivity and specificity for the early diagnosis of malnutrition and monitoring of treatment response. PA[12], a protein with a short half-life, reflects short-term changes in nutritional status, while ALB[13] and TRF[14] indicate mid- to long-term nutritional reserves. Studies have shown that low levels of these serum nutritional biomarkers are associated with lots of adverse clinical outcomes, including an increased risk of postoperative complications, extended hospital stays, and elevated mortality rates[15]. In the context of surgical operations, especially abdominal surgeries like resection for rectal cancer, nutritional status is crucial for patients’ postoperative recovery. Malnutrition not only affects wound healing but also lowers immune response, increases the risk of infection, and thereby promotes the occurrence of postoperative complications, including AL[16]. AL not only prolongs the recovery process but may also lead to the need for intensive care, reoperation, or even death. Therefore, preoperative assessment of serum nutritional biomarkers enables physicians to identify patients with malnutrition, providing them with personalized nutritional support to optimize preoperative status and reduce the risk of postoperative complications. Recent research further confirms the value of serum nutritional biomarkers in predicting surgical outcomes. For instance, rectal cancer patients with low ALB levels have a significantly increased risk of postoperative infection[17-19]; another study highlighted that levels of PA and TRF are closely related to the speed of postoperative recovery[20-22]. These findings underscore the importance of incorporating serum nutritional biomarker assessment into surgical treatment plans. Despite the recognized importance of serum nutritional biomarkers in assessing the influence of preoperative nutritional status on surgical outcomes, their application in predicting the likelihood of AL following laparoscopic surgery for middle and low rectal cancer remains significantly underexplored. Existing research primarily focuses on general statistics of AL incidence or explores the relationship between malnutrition and surgical complications in non-specific types of surgeries, with fewer studies specifically addressing laparoscopic rectal cancer surgery and its prognostic significance of specific nutritional biomarkers for AL risk[23,24]. Furthermore, even where related research exists, it often pertains only to a single nutritional marker, such as ALB, with limited investigation into other potentially predictive biomarkers like PA and TRF[25]. This limitation reduces the possibility of a comprehensive assessment of nutritional status, potentially impacting the accurate assessment of patients’ preoperative risk and corresponding intervention measures.

Here, we conducted a retrospective analysis of 560 patients receiving laparoscopic radical procedures for middle and low rectal cancer, including 67 cases with postoperative AL and 493 without. By comparing the serum nutritional biomarker levels between these two groups, we aim to reveal the relationship between these biomarkers and the risk of AL. Furthermore, logistic regression analysis was used to determine the independent risk factor influencing the occurrence of AL and construct a predictive model to assess the AL risk after laparoscopic surgery in the middle/low rectal cancer patient.

MATERIALS AND METHODS

General information of patients

This study included 560 patients with middle/low rectal cancer who received laparoscopic radical surgery at the Affiliated Cancer Hospital of Xinjiang Medical University from December 2018 to December 2022. Inclusion criteria were: (1) Patients diagnosed with middle/low rectal cancer; (2) Age over 18 years; (3) Patients undergoing laparoscopic rectal cancer radical surgery; and (4) Patients with complete preoperative data and examinations. Exclusion criteria comprised: (1) Patients with a prior history of distant metastasis; (2) Patients who underwent radiotherapy or chemotherapy prior to surgery; (3) Patients with other major complications before surgery; and (4) Patients whose procedure was converted to open surgery. This study was approved by the Ethics Committee of the Affiliated Cancer Hospital of Xinjiang Medical University (Approval No. G-2021005), and all included patients were thoroughly briefed and signed consent forms.

The collected clinical data of patients mainly included basic information [gender, age, body mass index (BMI), comorbid conditions like diabetes and hypertension, smoking or drinking history], surgery-related factors (preoperative intestinal obstruction, preventive stoma, surgical time, intraoperative bleeding, and neoadjuvant treatment), tumor-related factors [tumor size, tumor location, and tumor node metastasis (TNM) stage], and preoperative hematological indicators [white blood cell (WBC), platelet (PLT), hemoglobin (Hb), C-reactive protein (CRP), PA, ALB, and TRF]. Smoking history was defined as a binary variable, indicating whether a patient had ever smoked at any point in their life, without quantifying the exposure in terms of pack-years or other indices. Preoperative hematological values were measured from fasting peripheral venous blood taken within 24 hours before surgery to accurately assess the nutritional status of middle/low rectal cancer patients undergoing laparoscopic radical procedures. Tumor location was determined by measuring the distance from the edge of the anus to the bottom of the tumor. Tumors were categorized into two groups: Those located less than 7 cm from the anal verge and those located 7 cm or more from the anal verge. This categorization was used to evaluate the impact of tumor location on the risk of AL.

Diagnosis and grade of AL

We categorized patients into the AL group (n = 67) and the non-AL group (n = 493) based on the occurrence of AL within 2 weeks postoperatively. Diagnosis of AL was based on[26]: (1) Clinical presentation: Postoperative unexplained abdominal pain and distension, unexplained persistent fever, with or without signs of peritoneal irritation; (2) Drainage characteristics: A significant increase in the volume of pelvic drainage fluid, which may contain gas, pus, feces, etc.; (3) Digital rectal examination: The leakage site at the anastomosis can be palpated; and (4) Imaging examination: Extravasation of contrast agent, discontinuity of the intestinal tract, air around the anastomosis; (5) Laboratory tests: Significant increase or higher than normal values in WBC count, neutrophil count, percentage of neutrophils, and CRP levels. Based on the specific severity of AL[27], it can be classified into: (1) Grade A (n = 30, minor leakage requiring no additional intervention); (2) Grade B (n = 26, requiring active intervention but not reoperation); and (3) Grade C (n = 11, severe cases requiring reoperation for treatment).

Statistical analysis

We performed statistical analysis by using SPSS 26.0 (IBM Corp, Armonk, NY, United States). All variable data which is continuous are presented as mean ± SD. Comparisons among three or more groups were conducted using the χ² test, and the independent samples t-test was used to compare two groups. The risk factors for postoperative AL were analyzed using univariate and multivariate logistic regression models. Receiver operating characteristic (ROC) curves were used to analyze the predictive value of serum nutritional indicators for postoperative AL. The predictive model was built in R platform, and the “rms” package created nomograms[28,29]. The evaluation of serum nutritional indicators and predictive models were completed through comparing sensitivity, specificity, and the area under the ROC curve (AUC). AUC was compared using the DeLong test. The calibration of the predictive model was diagnosed and visualized using calibration analysis from the “rms” package. A P value of less than 0.05 was considered statistically significant.

RESULTS

Clinical and pathological characteristics of AL and non-AL patients

A total of 560 patients undergoing laparoscopic surgery for the middle and low rectal cancer were included, with 67 patients experiencing postoperative AL, resulting in an incidence rate of 11.96%. Table 1 compares the clinical and pathological characteristics between AL patients (n = 67) and non-AL patients (n = 493). Statistical analysis revealed no significant differences between the two groups regarding gender (P = 0.488) and age (P = 0.180), indicating these factors are not primary influencers of AL occurrence. A significant correlation was found between smoking history and the occurrence of AL (P = 0.002), suggesting smoking may increase the risk of postoperative AL. However, patients with comorbidities such as hypertension and diabetes did not show significant statistical differences (P = 0.140 and P = 0.958, respectively), implying these conditions have a limited impact on AL. The proportion of patients with preoperative intestinal obstruction in the AL group was much higher (P = 0.027), suggesting the presence of preoperative intestinal obstruction may require more cautious surgical strategies. Surgical time and intraoperative bleeding were also significant factors affecting the occurrence of AL, with surgery lasting over 180 minutes (P < 0.001) and intraoperative blood loss over 300 mL (P = 0.041) significantly increasing the risk of AL. Other clinical parameters, such as BMI, drinking history, neoadjuvant treatment, tumor size, TNM stage, tumor location, WBC, PLT, and CRP, revealed no statistical differences between AL and non-AL groups, indicating these factors have a minor effect on AL occurrence.

Table 1 Correlation of anastomotic leakage with clinical and laboratorial parameters in middle and low rectal cancer patients undergoing laparoscopic operation.

Characteristics	AL (n = 67)	Non-AL (n = 493)	χ2	P value
Gender
Male	37	250	0.481	0.488
Female	30	243
Age
< 60	19	181	1.794	0.180
≥ 60	48	312
BMI (kg/m2)
< 24	57	395	0.930	0.335
≥ 24	10	98
Hypertension
Yes	21	114	2.178	0.140
No	46	379
Diabetes
Yes	12	87	0.003	0.958
No	55	406
Smoking history
Yes	19	68	9.536	0.002
No	48	425
Drinking history
Yes	23	178	0.081	0.776
No	44	315
Preoperative intestinal obstruction
Yes	11	40	4.914	0.027
No	56	453
Preventive stoma
Yes	7	32	1.425	0.233
No	60	461
Surgical time (minutes)
< 180	39	417	27.133	< 0.001
≥ 180	28	76
Intraoperative bleeding (mL)
< 300	58	461	4.189	0.041
≥ 300	9	32
Neoadjuvant treatment
Yes	5	43	0.119	0.730
No	62	450
Tumor size (cm)
< 5	61	439	0.246	0.620
≥ 5	6	54
TNM stage
I	50	397	1.454	0.483
II	12	72
III	5	24
Tumor location (cm)
< 7	30	159	4.138	0.042
≥ 7	37	334
WBC (× 109/L)
< 10	50	395	1.091	0.296
≥ 10	17	98
Hb (g/L)
< 130 (male) or 120 (female)	8	56	0.033	0.855
≥ 130 (male) or 120 (female)	59	437
PLT (× 109/L)
< 400	57	381	2.102	0.147
≥ 400	10	112
CRP (mg/L)
< 10	48	385	1.400	0.237
≥ 10	19	108

AL: Anastomotic leakage; BMI: Body mass index; TNM: Tumor node metastasis; WBC: White blood cell; Hb: Hemoglobin; PLT: Platelet; CRP: C-reactive protein.

Comparison of preoperative serum nutritional indicator levels between AL and non-AL patients

We compared preoperative serum nutritional indicators between AL (n = 67) and non-AL (n = 493) middle/rectal cancer patients after laparoscopic surgery. Results in Table 2 show that levels of PA, ALB, and TRF were significantly lower in the AL group, at 245.74 ± 55.18 mg/L vs 331.83 ± 67.12 mg/L, 32.19 ± 4.87 g/L vs 38.74 ± 5.62 g/L, and 2.24 ± 0.32 g/L vs 2.89 ± 0.54 g/L, respectively. T-tests confirmed these differences were statistically significant (P < 0.001), indicating a significant association between preoperative nutritional status and the occurrence of AL.

Table 2 Comparison of preoperative serum nutritional factors between anastomotic leakage patients and non-anastomotic leakage patients.

Group	n	Serum nutritional factors (mean ± SD)
		PA (mg/L)	ALB (g/L)	TRF (g/L)
AL	67	245.74 ± 55.18	32.19 ± 4.87	2.24 ± 0.32
Non-AL	493	331.83 ± 67.12	38.74 ± 5.62	2.89 ± 0.54
t		10.045	9.086	9.621
P value		< 0.001	< 0.001	< 0.001

AL: Anastomotic leakage; PA: Prealbumin; ALB: Albumin; TRF: Transferrin.

Comparison of preoperative serum nutritional indicator levels among different grades of AL patients

In the study comparing preoperative serum nutritional indicators across different severity grades of AL in patients, we found that levels of PA, ALB, and TRF significantly decreased as the severity of AL increased, as shown in Table 3. The average levels of PA, ALB, and TRF for Grade A (minor leakage) patients were 312.04 mg/L, 36.81 g/L, and 2.76 g/L, respectively, while those for Grade C (severe leakage) patients dropped to 207.18 mg/L, 22.60 g/L, and 1.98 g/L. Statistical analysis showed these differences were highly significant (P < 0.001), revealing a significant correlation between the status of preoperative nutrition and the severity of AL, emphasizing the importance of preoperative nutritional assessment in predicting AL risk.

Table 3 Comparison of preoperative serum nutritional factors among different grades of anastomotic leakage patients.

Grade	n	Serum nutritional factors (mean ± SD)
		PA (mg/L)	ALB (g/L)	TRF (g/L)
Grade A	30	312.04 ± 56.61	36.81 ± 5.29	2.76 ± 0.41
Grade B	26	243.19 ± 54.28	30.72 ± 4.18	2.27 ± 0.34
Grade C	11	207.18 ± 43.66	22.60 ± 3.76	1.98 ± 0.27
F		19.801	39.323	23.173
P value		< 0.001	< 0.001	< 0.001

PA: Prealbumin; ALB: Albumin; TRF: Transferrin.

Risk factor analysis for AL after laparoscopic surgery for middle and low rectal cancer

In this study, univariate and multivariate logistic regression analyses were conducted to explore the risk factors for AL. As shown in Table 4, univariate analysis identified six factors significantly associated with an increased risk of AL. Specifically, the odds ratios (ORs) for PA, ALB, TRF, smoking history, preoperative intestinal obstruction, intraoperative bleeding ≥ 300 mL, and tumor location < 7 cm were 1.721 [95% confidence interval (CI): 1.271-2.198, P = 0.021], 1.594 (95%CI: 1.165-2.063, P = 0.017), 1.982 (95%CI: 1.366-2.598, P = 0.030), 1.369 (95%CI: 1.064-1.517, P = 0.032), 2.083 (95%CI: 1.695-3.206, P = 0.028), 1.436 (95%CI: 1.102-1.763, P = 0.021), and 1.498 (95%CI: 1.073-2.019, P = 0.012), respectively. However, after multivariate logistic regression analysis, only PA, ALB, TRF, intraoperative bleeding ≥ 300 mL, and tumor location < 7 cm were significantly associated with an increased risk of AL, with adjusted OR values of 2.621 (95%CI: 1.582-3.812, P = 0.012), 3.982 (95%CI: 1.927-4.887, P = 0.024), 2.109 (95%CI: 1.162-2.981, P = 0.031), 4.182 (95%CI: 2.108-5.482, P = 0.009), and 3.124 (95%CI: 1.779-4.215, P = 0.016), respectively. Notably, the significance of smoking history and preoperative intestinal obstruction was not maintained in the multivariate analysis, suggesting their effects may be modulated by other variables. Additionally, factors such as gender, age, BMI, hypertension, diabetes, drinking history, preventive stoma, surgical time, neoadjuvant treatment, tumor size, TNM stage, WBC, PLT, and CRP showed no significant correlation with the risk of AL in the univariate analysis and were not included or showed no significance in the multivariate analysis.

Table 4 Univariate and multivariate logistic regression analysis of risk factors for anastomotic leakage.

Factors	Reference	Univariate			Multivariate
		OR	95%CI	P value	OR	95%CI	P value
PA	Continuous variable	1.721	1.271-2.198	0.021	2.621	1.582-3.812	0.012
ALB	Continuous variable	1.594	1.165-2.063	0.017	3.982	1.927-4.887	0.024
TRF	Continuous variable	1.982	1.366-2.598	0.030	2.109	1.162-2.981	0.031
Gender	0 = male, 1 = female	0.982	0.764-1.175	0.281	/	/	/
Age	0 = < 60, 1 = ≥ 60	1.021	0.835-1.284	0.473	/	/	/
BMI	0 = < 24 kg/m2, 1 = ≥ 24 kg/m2	0.894	0.712-1.092	0.501	/	/	/
Hypertension	0 = No, 1 = Yes	1.123	0.902-1.351	0.124	/	/	/
Diabetes	0 = No, 1 = Yes	1.076	0.863-1.376	0.374	/	/	/
Smoking history	0 = No, 1 = Yes	1.369	1.064-1.517	0.032	1.170	1.086-1.375	0.088
Drinking history	0 = No, 1 = Yes	0.994	0.704-1.328	0.584	/	/	/
Preoperative intestinal obstruction	0 = No, 1 = Yes	2.083	1.695-3.206	0.028	0.927	0.786-1.159	0.617
Preventive stoma	0 = No, 1 = Yes	1.046	0.903-1.254	0.438	/	/	/
Surgical time	0 = < 180 minutes, 1 = ≥ 180 minutes	1.098	0.916-1.328	0.330	/	/	/
Intraoperative bleeding	0 = < 300 mL, 1 = ≥ 300 mL	1.436	1.102-1.763	0.021	4.182	2.108-5.482	0.009
Neoadjuvant treatment	0 = No, 1 = Yes	1.121	0.921-1.328	0.287	/	/	/
Tumor size	0 = < 5 cm, 1 = ≥ 5 cm	0.899	0.711-1.197	0.319	/	/	/
TNM stage	0 = I, 1 = II or III	0.876	0.702-1.354	0.437	/	/	/
Tumor location	0 = ≥ 7 cm, 1 = < 7 cm	1.498	1.073-2.019	0.012	3.124	1.779-4.215	0.016
WBC	0 = < 10 × 109/L, 1 = ≥ 10 × 109/L	0.926	0.804-1.154	0.572	/	/	/
PLT	0 = < 400 × 109/L, 1 = ≥ 400 × 109/L	0.957	0.726-1.368	0.439	/	/	/
CRP	0 = < 10 mg/L, 1 = ≥ 10 mg/L	1.051	0.804-1.328	0.265	/	/	/

OR: Odds ratio; CI: Confidence interval; PA: Prealbumin; ALB: Albumin; TRF: Transferrin; BMI: Body mass index; TNM: Tumor node metastasis; WBC: White blood cell; Hb: Hemoglobin; PLT: Platelet; CRP: C-reactive protein.

The diagnostic levels of the three preoperative serum nutritional biomarkers (PA, ALB, TRF) for AL occurrence, either independently or combined, were explored. ROC curve analysis showed that the optimal cutoff value for preoperative serum PA in predicting AL after laparoscopic surgery for middle and low rectal cancer was 288.50 mg/L, with an AUC of 0.790 (95%CI: 0.702-0.879), and corresponding sensitivity and specificity of 70.00% and 71.43%, respectively. The optimal cutoff value for preoperative serum ALB was 33.30 g/L, with an AUC of 0.812 (95%CI: 0.722-0.902), and corresponding sensitivity and specificity of 71.74% and 72.09%, respectively. The optimal cutoff value for preoperative serum TRF was 2.550 g/L, with an AUC of 0.793 (95%CI: 0.702-0.884), and corresponding sensitivity and specificity of 70.21% and 74.51%, respectively. The combined prediction of preoperative serum nutritional indicators (PA + ALB + TRF) for AL occurrence showed an AUC of 0.864 (95%CI: 0.800-0.929), with corresponding sensitivity and specificity of 81.03% and 77.97%, respectively. The value of combining preoperative serum PA, ALB, TRF for predicting postoperative AL exceeds that of any single indicator, which is shown in Figure 1.

Figure 1 Evaluation of the predictive performance of receiver operating characteristic curves for preoperative serum prealbumin, albumin and transferrin independently or in combination for anastomotic leakage. PA: Prealbumin; ALB: Albumin; TRF: Transferrin; TPR: True positive rate; FPR: False positive rate.

Predictive model for AL after laparoscopic surgery for middle and low rectal cancer based on preoperative serum nutritional biomarkers

Based on the five independent risk factors for AL identified through multivariate logistic regression analysis (PA, ALB, TRF, intraoperative bleeding, and tumor location), we constructed a predictive model for rectal cancer AL risk following laparoscopic surgery, as shown in Figure 2A. The model demonstrated good predictive performance, established by the ROC curve analysis. The AUC of the model reached 0.942 (95%CI: 0.913-0.971), with a sensitivity of 0.844 and a specificity of 0.922, indicating high discriminative ability (Figure 2B). The calibration performance of the model was assessed through calibration plots. The calibration plot (Figure 2C) shows the calibration curve closely aligning with the 45-degree diagonal line, with a concordance index of 0.838 (95%CI: 0.780-0.897, P < 0.001), indicating a high degree of consistency between predicted probabilities and observed rates of AL. Furthermore, the Hosmer-Lemeshow test yielded a P value of 0.328, suggesting no significant discrepancy between model predictions and actual observations, thereby validating the model’s good calibration.

Figure 2 Construction and evaluation of the nomogram of anastomotic leakage risk after laparoscopic radical surgery for middle and low rectal cancer based on preoperative serum nutritional markers. A: The nomogram; B: Receiver operating characteristic curve; C: Calibration. TPR: True positive rate; FPR: False positive rate; AUC: Area under the receiver operating characteristic curve; CI: Confidence interval.

DISCUSSION

In this study, we conducted a statistical analysis on the incidence of AL following laparoscopic surgery in 560 middle/low rectal cancer patients. Although our study is retrospective, efforts were made to ensure group homogeneity by applying consistent inclusion criteria and standardizing surgical procedures across facilities. Additional analyses confirmed that demographic and clinical characteristics were comparable between AL and non-AL groups, with some differences observed in smoking history and intraoperative factors. No significant clustering of AL cases was found among specific surgeons or facilities. Out of 560 patients, 67 patients experienced AL postoperatively, marking an incidence rate of 11.96%. This rate is comparable to the previous report[30] in a cohort of 5398 rectal cancer patients (10.20%) and by Li et al[31] in 497 rectal cancer patients (10.26%), yet it is lower than the incidence reported by Peltrini et al[32] in 367 rectal cancer patients (17.4%). The incidence of AL in our study (11.96%) is comparable to the range reported in similar studies. Factors such as tumor location, intraoperative blood loss, and patient comorbidities contribute to this rate, aligning with established risk profiles for AL in laparoscopic rectal surgery. These findings suggest that the laparoscopic approach to rectal surgery does not elevate the risk of AL postoperatively in comparison to open surgery. Variations in these results may also be attributable to different inclusion criteria and patient populations. AL following radical resection for rectal cancer arises from numerous factors such as characteristics, tumor status, as well as surgical factors. Our results indicate that the tumor location, the volume of surgical blood loss, and preoperative hematological nutritional biomarkers have an influence on the occurrence of AL.

The influence of tumor location on the incidence of AL in middle and low rectal cancer has reached a consensus both domestically and internationally[33], and the risk of developing AL increases as the tumor’s proximity to the anal margin decreases[34-38]. Our study arrived at the same conclusion. This could be due to the larger wound created during resection when the tumor is nearer to the anus. Additionally, the damage to tissues and blood vessels caused by electrocoagulation during surgery can lead to exudation and bleeding, thereby reducing the blood supply to the anastomotic site and increasing the risk of postoperative AL[39]. Hb is associated with perfusion and oxygenation at the edges of the anastomosis, which are critical factors for the healing of the anastomosis. Hence, postoperative anemia due to excessive intraoperative blood loss has been described as a risk factor for leakage[40]. Multiple studies have found[41-43] that Hb levels below 110 g/L increase the AL risk after rectal cancer surgery, which can be attributed to the decreased capacity to transport oxygen to tissues and the subsequent risk of ischemia[44,45]. Both surgical blood loss and blood transfusion are independently associated with an increased risk of anastomotic failure[46]. Blood loss may lead to ischemia at the anastomotic site, impairing healing. Transfusion may induce immunosuppression, thereby increasing the risk of infectious diseases around the anastomosis[47,48].

From a pathophysiological perspective, malnutrition can increase the risk of AL through various mechanisms, including impaired wound healing, diminished immune response, and elevated infection risk. Thus, nutritional status not only indirectly affects the speed of postoperative recovery and long-term prognosis but also directly impacts the occurrence of postoperative complications. PA and ALB are among the most thoroughly investigated nutritional biomarkers in the realm of AL. Over 35% of CRC patients exhibit low serum PA levels preoperatively, with low serum PA identified as a predictor of postoperative recurrence[49] and an independent risk factor for prolonged hospital stay[50]. Numerous studies have established a direct correlation between preoperative ALB levels and the risk of postoperative AL[51-53], aligning with our findings that low preoperative serum PA and ALB levels are independent risk factors for rectal AL following cancer surgery. This association may be attributed to reduced perioperative nutritional status leading to compromised immune function and increased risk of infection and its propagation[54]. TRF, a primary iron-binding protein in human plasma, plays a critical role in transporting iron absorbed from the gastrointestinal tract and iron released from the breakdown of red blood cells, serving as an important indicator for nutritional assessment. TRF may mediate nutrition and tumor prognosis through iron ions. A cohort study by Herrinton et al[55] suggested that low TRF saturation is associated with an elevated risk of colon and rectal cancer in men. Sawayama et al[56] found that low preoperative serum TRF levels in 501 patients undergoing surgery for stages I-III CRC were inversely correlated with recurrence-free survival, overall survival, and cancer-specific survival. Additionally, TRF saturation can serve as an indicator of the efficacy of chemotherapy in CRC. Ochiai et al[57] reported that an increase in blood iron saturation 48 hours after administering a standard first-line chemotherapy regimen for advanced CRC was associated with a shorter median survival time. Patients with colorectal liver metastases undergoing partial hepatectomy who have serum TRF levels below 190 mg/dL have a poorer prognosis[58], suggesting that serum TRF levels could be a potential prognostic indicator for adverse outcomes in CRC patients with liver metastases. Preoperative serum TRF levels have also been predictive of complications in other diseases. Patients with low preoperative serum TRF levels face an increased risk of pneumonia following esophageal cancer resection[59] and delayed spontaneous closure of gastrointestinal cutaneous fistulas[60]. However, until now, no studies have identified a relationship between TRF and the risk of AL after rectal cancer surgery. This study is the first to identify low preoperative serum TRF level as a risk factor for AL.

Beyond identifying the diagnostic potential of preoperative serum nutritional biomarkers for the risk of AL, our study meticulously developed and validated a predictive model based on these markers to forecast the risk of AL in patients with low and mid-rectal cancer undergoing laparoscopic surgery. The cornerstone of our discovery is the model’s robust performance, highlighted by an AUC of 0.942, demonstrating exceptional sensitivity and specificity. The ROC-AUC analysis was chosen not only due to its favorable results but because it is the most appropriate tool for evaluating the discriminative ability of our predictive model in the clinical context. Additionally, supplementary statistical evaluations (Hosmer-Lemeshow and NRI) confirmed the robustness of our findings. This achievement underscores the model’s capability in accurately differentiating between high-risk and low-risk patients for AL. Incorporating preoperative serum nutritional biomarkers (PA, ALB, and TRF) along with intraoperative blood loss and tumor location as predictive factors not only enhanced the model’s predictive accuracy but also aligned with the evolving understanding of AL’s multifactorial etiology. The clinical consequences of AL, ranging from prolonged hospital stays to increased morbidity and even mortality, pose a significant challenge in colorectal surgery. The significance of AL affects not only immediate postoperative outcomes but also long-term survival and quality of life. Hence, the identification of reliable predictors of AL is not just an academic endeavor but a clinical necessity. By enabling the early identification of individuals at high risk, such a predictive model offers avenues for preoperative interventions. Particularly, our model underscores the critical role of nutritional status, proving the inclusion of serum nutritional biomarkers highlights a potentially modifiable risk factor. Given the strong performance of this model, we have integrated it into our clinical practice to aid in preoperative decision-making. Specifically, for patients identified as high-risk for AL by our predictive model, we engage in detailed discussions with the surgical team and the patient during the preoperative phase. These discussions not only cover the potential risks of AL but also explore the appropriate strategies to mitigate these risks. We now more frequently recommend the use of a protective stoma during curative resection of rectal cancer for these high-risk patients. A protective stoma significantly reduces the risk of severe complications associated with AL, such as intra-abdominal infections, prolonged hospital stays, and the need for reoperation. By applying the results of this predictive model to surgical strategy, we are able to provide more personalized treatment plans for high-risk patients, thereby improving their overall postoperative outcomes. This strategy has been shown to effectively reduce surgical complications, aligning with current best practices and literature. We believe that this approach helps to optimize the treatment process for rectal cancer patients, reduce the incidence of postoperative AL, and ultimately improve the long-term quality of life for these patients. We recommend further studies to validate the applicability of this model across diverse patient populations and to explore additional potential predictive factors to further enhance the precision of surgical decision-making.

CONCLUSION

In this retrospective analysis, we leveraged clinical data from 560 middle/low rectal cancer patients undergoing laparoscopic surgery, focusing on preoperative serum nutritional biomarkers. We developed a predictive model that identifies patients at increased risk of developing AL based on PA, ALB, and TRF, achieving an AUC of 0.942. Highlighting the significant impact of nutritional status on AL risk, our study not only deepens the understanding of factors influencing surgical outcomes but also emphasizes the essential role of integrating nutritional assessments into preoperative planning. This strategy is poised to significantly improve patient care by enhancing postoperative recovery and potentially affecting long-term outcomes in rectal surgery.