Development and validation of a LASSO-based nomogram for predicting anastomotic leakage in elderly patients after laparoscopic gastrectomy

Introduction

Gastric cancer (GC), the fifth most prevalent malignancy globally, imposes a disproportionate burden in China, accounting for 50–70% of newly diagnosed cases worldwide (1). Although global incidence rates have declined, GC remains the second most commonly diagnosed cancer and the second leading cause of cancer-related mortality in China (2). With rapid population aging, over 60% of GC cases in China now occur in individuals aged 65 years and older, with incidence peaking around 70 years of age (3). Laparoscopic gastrectomy (LG) has emerged as a preferred alternative to open surgery due to its advantages, including reduced intraoperative blood loss, shorter postoperative recovery, and superior cosmetic outcomes (4). Nevertheless, postoperative complications such as anastomotic leakage (AL), wound infection, and intestinal obstruction remain prevalent, with AL representing one of the most clinically devastating complications (5,6).

AL, characterized by disruption of the anastomotic site resulting in extraluminal contamination, typically occurs 3 to 26 days postoperatively. It is associated with mortality rates exceeding 50% and reoperation rates surpassing 60% in GC patients (7,8), underscoring the critical need for reliable perioperative tools to identify high-risk patients. While evidence regarding overall morbidity differences between elderly and younger LG patients remains conflicting (4,9), elderly populations are particularly vulnerable to AL due to physiological frailty, higher comorbidity burdens, and atypical clinical presentations that obscure timely diagnosis (10). Current diagnostic modalities, including radiographic imaging and laboratory testing, exhibit limited sensitivity in detecting early-stage AL (11). Consequently, there is an urgent demand for validated perioperative risk stratification models specifically tailored to elderly patients undergoing LG.

Current clinical risk models for predicting AL predominantly rely on conventional logistic regression methodologies (12), which frequently exhibit limitations such as overfitting, suboptimal variable selection in small cohorts, and insufficient validation within geriatric-specific populations. Furthermore, many existing frameworks incorporate variables that lack generalizability across diverse surgical contexts (13). To address these limitations, machine learning (ML) approaches, such as the least absolute shrinkage and selection operator (LASSO), demonstrate distinct advantages in high-dimensional data analysis. LASSO enables automated variable selection by penalizing redundant predictors through regularization, thereby enhancing both model interpretability and generalizability (14). When integrated with nomograms—user-friendly graphical tools that translate statistical models into actionable clinical algorithms—LASSO-derived frameworks hold significant potential to bridge the gap between computational complexity and clinical applicability (15).

Despite advancements in predictive modeling in clinical settings, no prior studies have developed or validated a LASSO-based nomogram specifically designed to predict AL in elderly patients undergoing LG. Current studies either focus on mixed-age populations or employ regression methods prone to overfitting when applied to this high-risk demographic (16,17). This oversight is clinically significant, as the aging population exhibits unique anatomical, immunological, and physiological challenges that require tailored predictive strategies (3). The present study addresses this critical gap through two primary objectives: to develop the first LASSO-regularized nomogram for perioperative AL prediction in elderly LG patients, harmonizing clinical utility (via nomogram visualization) with statistical robustness (via LASSO variable selection); and to rigorously validate its discriminative performance and calibration accuracy in clinical samples (18). By establishing an evidence-based, clinician-friendly predictive tool, this research aims to refine preoperative risk stratification, inform perioperative decision-making, and ultimately reduce the morbidity and mortality burden of AL in this vulnerable population. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-897/rc).

Methods

Patient population

Clinical data were retrospectively collected from 1,467 GC patients following LG at Shengjing Hospital of China Medical University in Shenyang city, China, between February 2019 and April 2024. The inclusion criteria were as follows: (I) age ≥65 years; (II) histopathologically confirmed diagnosis of primary gastric adenocarcinoma; (III) curative-intent R0 resection with D1/D1 + α/D1 + β/D2 lymphadenectomy, as confirmed by postoperative pathological diagnosis; (IV) missing data <10%; (V) absence of distant metastasis, adjacent organ invasion, or peritoneal dissemination; (VI) tumor-node-metastasis (TNM) staging I–III, as classified according to the 7th edition of the American Joint Committee on Cancer (AJCC) staging system; (VII) diagnosis of AL within 30 days postoperatively; and (VIII) all surgeries performed by experienced surgeons with a minimum of 200 prior standard LG procedures. The exclusion criteria were: (I) prior neoadjuvant therapy; (II) intraoperative conversion to open laparotomy; (III) concurrent multi-organ resection; (IV) postoperative intraperitoneal hyperthermic perfusion chemotherapy; (V) emergency surgery for gastric bleeding, obstruction, or perforation; (VI) palliative resection; and (VII) prolonged preoperative administration of nonsteroidal anti-inflammatory drugs or immunosuppressants. A total of 884 eligible patients were included in the final cohort. A flowchart depicting the study selection process is shown in Figure S1.

Ethics statement

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shengjing Hospital of China Medical University (No. 2024PS151K) and individual consent for this retrospective analysis was waived.

Definition of AL

AL was defined as the disruption of gastrointestinal continuity involving the leakage of luminal contents beyond the anastomotic site, affecting the gastrojejunal, jejunojejunal, or esophagojejunal anastomosis (7). In this study, AL diagnosis required fulfillment of at least one of the following criteria: (I) the presence of gastrointestinal fluids (e.g., bile, enteric contents) in abdominal drainage, concomitant with clinical signs such as fever, chills, abdominal distension, or abdominal pain; (II) leukocytosis or an elevated neutrophil ratio; (III) the detection of orally administered methylene blue in abdominal drainage; (IV) radiographic evidence of contrast extravasation at the anastomotic site on fluoroscopy imaging; or (V) the direct visualization of anastomotic dehiscence via postoperative abdominal computed tomography (CT) or endoscopic examination.

Sample size

The sample size was calculated based on the events per variable (EPV) principle for multivariate logistic regression analysis, which mandates a minimum of 10 EPV per candidate predictor variable (19). A preliminary analysis of 100 randomly selected cases estimated an AL incidence rate of 10% among elderly patients following LG (20), with up to six variables projected for inclusion in the final model. To satisfy statistical requirements, a minimum sample size of 600 patients was calculated. To ensure robustness, the final analysis included 854 patients.

Data collection

The following data variables were extracted from the electronic medical record system: (I) demographic and clinical characteristics: age, gender (male/female), body mass index (BMI, kg/m²), alcohol consumption (yes/no), smoking status (yes/no), comorbidities (diabetes, hypertension, pyloric obstruction, pulmonary insufficiency, cardiovascular disease, and renal disease), history of abdominal surgery (yes/no), and American Society of Anesthesiologists (ASA) score (I–II/III). (II) Oncologic information: tumor differentiation grade (low/medium-high), tumor location (upper, middle, lower, or multiple parts), tumor size (cm), TNM stage (I–II/III), lymphovascular invasion status (yes/no), and nerve invasion (yes/no). (III) Intraoperative records: anastomotic type (Billroth I, Billroth II, Billroth II+Braun, or Roux-en-Y), extent of surgical resection (Total, distal, or proximal), intraoperative blood loss (mL), blood transfusion (yes/no), and operation duration (minutes). (IV) Preoperative laboratory parameters (assessed immediately prior to surgery): systemic inflammatory markers: white blood cell count (WBC, ×10⁹/L), neutrophil count (×10⁹/L), monocyte count (×10⁹/L), and C-reactive protein (CRP, mg/L); nutrient and hematologic indicators: albumin (g/L), hemoglobin (g/L), lymphocyte count (10⁹/L), and platelet count (×10⁹/L). The preoperative prognostic nutritional index (PNI) value was calculated using the formula: albumin (g/L) + 5 × total lymphocyte count (10⁹/L). The preoperative neutrophil-lymphocyte ratio (NLR) value was calculated as neutrophil count (10⁹/L) divided by lymphocyte count (10⁹/L). The preoperative platelet-lymphocyte ratio (PLR) value was calculated as platelet count (10⁹/L) divided by lymphocyte count (10⁹/L). The preoperative lymphocyte-monocyte ratio (LMR) was calculated as lymphocyte count (10⁹/L) divided by monocyte count (10⁹/L).

Statistical analysis

Statistical analyses were performed using SPSS version 26.0 (IBM Corp., Chicago, USA) and R software 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were assessed for normality via the Kolmogorov-Smirnov test. Parametric data were presented as mean ± standard deviation (SD), while non-normally distributed variables were described as median with interquartile range (P25, P75). Categorical variables were summarized as frequency (percentage). Intergroup comparisons were performed using the Student’s t-test, the Mann-Whitney U test, the Chi-squared test, or Fisher’s exact test, as appropriate. Missing values constituted <10% of the dataset and were addressed using multiple imputation (21).

The cohort was randomly assigned to a training set (n=620) and a validation set (n=264) in a 7:3 ratio. First, optimal predictors with non-zero coefficients were identified using the LASSO regression method, employing 10-fold cross-validation based on both lambda (λ) min (minimum) and lambda 1 − standard error (SE) criteria. The selected factors were further analyzed using multivariate logistic regression analysis to determine odds ratio (OR) and corresponding 95% confidence interval (CI). Predictors with P<0.05 were incorporated into the final nomogram model. The area under the receiver operating characteristic (ROC) curve (AUC) was calculated to evaluate the accuracy and discriminatory ability of the prediction model, where an AUC value closer to 1 signified superior predictive performance. Calibration plots and the Hosmer-Lemeshow test (P>0.05 indicating good fit) were used to assess the goodness-of-fit of the nomogram. Decision curve analysis (DCA) was performed to assess the clinical practicability of the nomogram by analyzing net benefit across predefined probability thresholds. A two-tailed P<0.05 was considered statistically significant.

Results

Clinical characteristics of participants

This retrospective study comprised 884 elderly patients who underwent LG. The cohort included 565 males (63.9%) and 319 females (36.1%), with a mean age of 72.90±5.05 years (range, 65–89 years). Postoperative AL was identified in 120 cases (13.6%). Comparative analysis between the AL (n=120) and non-AL (n=764) groups revealed statistically significant differences in age (P<0.001), ASA (P<0.001), pyloric obstruction (P=0.01), diabetes (P<0.001), cardiovascular disease (P<0.001), intraoperative blood loss (P<0.001), blood transfusion (P=0.005), preoperative hemoglobin (P=0.02), preoperative albumin (P=0.001), preoperative CRP (P<0.001), and PNI (P<0.001) (Table 1). The cohort was randomly allocated to a training set (n=620, 70%) and a validation set (n=264, 30%), with no significant baseline differences observed between the two cohorts upon univariate analysis (all P>0.05, see Table S1).

LASSO and multivariate logistic regression analysis

Monocyte, neutrophil, lymphocyte, albumin, and platelet were excluded for LASSO and multivariate regression analysis due to their collinearity with PLR, NLR, LMR, and PNI. LASSO regression analysis with 10-fold cross-validation [λ_1−SE =0.0489, log(λ) =−3.018] identified five non-zero coefficient variables as predictors of AL: age, ASA, diabetes, blood loss, and PNI (Figure 1A,1B). Subsequent multivariable logistic regression analysis within the training set demonstrated that age (OR: 1.298, 95% CI: 1.210–1.393), ASA (OR: 2.794, 95% CI: 1.555–5.020), diabetes (OR: 4.358, 95% CI: 2.282–8.321), blood loss (OR: 2.963, 95% CI: 1.649–5.324), and PNI (OR: 0.309, 95% CI: 0.168–0.569) were independent predictors for AL with statistical significance (all P<0.01) (Figure 2).

Figure 1 Predictor selection using a LASSO regression model. (A) Coefficient path plot of LASSO regression model. Each line represents the coefficient of a predictor variable, starting from the full model (left) and shrinking to zero as λ increases (right). (B) 10-fold cross-validation plot for optimal λ selection. The left dotted vertical line indicates the λ_min and the right dotted vertical line shows the λ_1−SE. The predictors in this study were selected using the λ_1−SE criterion. λ, lambda; LASSO, least absolute shrinkage and selection operator; SE, standard error.

Figure 2 Forest plot of multivariate logistic regression analysis of the risk factors for AL in elderly patients following LG. AL, anastomotic leakage; ASA, American Society of Anesthesiologists; CI, confidence interval; LG, laparoscopic gastrectomy; OR, odds ratio; PNI, prognostic nutritional index.

Construction of the nomogram prediction model

The significant predictors identified via multivariate logistic regression analysis were incorporated into a nomogram to quantitatively estimate the risk of AL in elderly patients following LG. As illustrated in Figure 3A, the top scale of the nomogram model represented the risk score for each variable. The total score for an individual could be obtained by summing all factor scores: higher total scores correspond to an elevated probability of AL. For instance, an 88-year-old patient with ASA class III, non-diabetic status, intraoperative blood loss <200 mL, and a PNI <50 would yield a cumulative score of 257, corresponding to an AL risk probability of 0.814 (red arrow) (Figure 3B).

Figure 3 Nomogram for individualized predicting AL in elderly patients following LG. (A) Nomogram prediction model to evaluate the risk of AL based on age, ASA, diabetes, blood loss, and PNI. (B) Clinical application example: for an 88-year-old patient (ASA III, non-diabetes, blood loss <200 mL, and PNI <50), the total score [257] corresponds to an estimated AL risk of 0.814 (red arrow). Red dots represent the parameter-specific values of the patient in the example; the asterisks (***) besides each variable denote the predictive factors derived from multivariable regression analysis. AL, anastomotic leakage; ASA, American Society of Anesthesiologists; LG, laparoscopic gastrectomy; PNI, prognostic nutritional index.

Evaluation of the prediction model

AUC values were used to measure the discriminative performance of the nomogram model. The AUC values for the training cohort were 0.870 (95% CI: 0.826–0.913) and for the validation cohort were 0.890 (95% CI: 0.828–0.952), respectively (Figure 4A). These findings suggested the nomogram model had good diagnostic power in effectively differentiating AL from non-AL patients.

Figure 4 Predictive performance and clinical utility of the nomogram. (A) ROC curve demonstrating discriminative ability in training [AUC =0.870, 95% CI: 0.826–0.913; optimal threshold =0.130 (sensitivity =0.774, specificity =0.845)] and validation cohorts [AUC =0.890, 95% CI: 0.828–0.952; optimal threshold =0.226 (sensitivity =0.899, specificity =0.750)]. (B) Calibration plots comparing predicted probabilities of AL and observed outcomes. (C) DCA quantifying net benefit across threshold probabilities, affirming the model’s clinical applicability. AL, anastomotic leakage; AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; ROC, receiver operating characteristic.

Calibration of the nomogram prediction model

Calibration plots and the Hosmer-Lemeshow test were performed to evaluate the goodness-of-fit of the nomogram model. As shown in Figure 4B, the calibration curves revealed strong concordance between the predicted probability and observed outcomes in both training and validation cohorts. Additionally, the Hosmer-Lemeshow test further corroborated model fit in both the training (χ²=8.791, P=0.46) and validation cohorts (χ²=2.978, P=0.97).

Clinical validity of the nomogram prediction model

DCA curves were used to assess the net clinical benefit of the nomogram across threshold probabilities. In the training cohort (Figure 4C), the nomogram outperformed the “treat-all” (black line above the x-axis) and “treat-none” (gray curve) strategies at a wide range of threshold probabilities, yielding superior net benefit. A similar trend was observed in the validation cohort, underscoring the model’s clinical applicability (Figure 4C).

Discussion

AL following LG remains a significant challenge in postoperative management, particularly in elderly patients who exhibit heightened frailty and diminished physiological reserve (16,22). The clinical urgency to address this complication underscores the importance of developing robust risk prediction tools for the timely identification of high-risk elderly GC patients, thereby potentially mitigating adverse outcomes. In this study, we systematically developed and rigorously validated a diagnostic nomogram by incorporating readily accessible clinical variables, such as demographic characteristics, comorbidities, and laboratory indices. LASSO regression and multivariable logistic regression analyses identified five independent predictors of AL: advanced age, ASA classification ≥3, diabetes, intraoperative blood loss (≥200 mL), and preoperative PNI <50. The model demonstrated robust discriminative accuracy, excellent calibration consistency, and favorable clinical utility. This tool provides a practical clinical decision-support instrument, enabling early risk identification and guiding tailored perioperative strategies to reduce AL-associated morbidity in elderly patients undergoing LG. This integration into clinical workflows may enhance postoperative monitoring paradigms and resource allocation for high-risk individuals.

In the present study, the incidence of AL in elderly patients following LG was approximately 13.6%, closely aligning with the 15.1% rate reported by Trapani et al. in a cohort of Italian GC patients undergoing LG (23). Advanced age was identified as an independent predictor of AL in our model, likely attributable to the age-related physiological decline in immune function and diminished capacity to withstand surgical stress (24). Consistently, a meta-analysis demonstrated that elderly GC patients experienced a higher risk of postoperative AL than their younger counterparts (25). Furthermore, Maejima et al. found that advanced age independently increased the risk of AL in GC patients undergoing total gastrectomy (26). These findings are further corroborated by a nomogram model developed by Tu et al., which identified advanced age as a critical risk factor for AL in a multicenter cohort of 3,632 Chinese patients receiving gastrectomy (27). These collective results advocate for the implementation of preoperative geriatric assessments in elderly LG patients, facilitating personalized interventions such as intensive nutritional optimization, structured prehabilitation programs, and vigilant postoperative monitoring to address age-related vulnerabilities. Such measures align with nomogram-guided risk stratification to prioritize resource allocation for high-risk subgroups and improve clinical outcomes.

Among the various comorbidities, diabetes emerged as a significant predictor of postoperative AL among elderly GC patients undergoing LG. Accumulating evidence underscores that hyperglycemia disrupts anastomotic healing through multifactorial mechanisms, such as suppressed collagen synthesis, dysregulated angiogenesis, and impaired leukocyte function, collectively compromising anastomotic integrity (28). Consequently, diabetic patients exhibit a substantially elevated AL risk compared to non-diabetic cohorts following gastric resection (29). This finding aligns with prior research by Li et al., who identified diabetes as a critical prognostic variable in a predictive model for AL after rectal cancer resection (30). The association is further corroborated by a recent systematic review by Bracale et al., which highlighted diabetes as a crucial patient-related risk factor influencing AL development in a multinational cohort of 42,489 gastrectomy patients (13). To translate these pathophysiological insights into clinical practice, our LASSO-derived nomogram facilitates preoperative risk stratification of diabetic patients, thereby guiding tailored perioperative management. For high-risk individuals with diabetes, stringent glycemic optimization is recommended to attenuate hyperglycemia-induced microvascular dysfunction (31). These protocols align with suggestions proposed by Tan et al., who emphasized preoperative glycemic control for diabetic patients to improve anastomotic healing following colorectal surgery (32).

The nomogram’s identification of preoperative PNI as a protective factor underscores the pivotal role of nutritional optimization in preventing AL. Accumulating evidence confirms that preoperative nutritional status, as quantified by PNI, significantly correlates with surgical outcomes due to its direct impact on postoperative tissue repair mechanisms (17). Notably, GC patients with low PNI exhibit an elevated incidence of AL following LG, indicating that compromised nutritional reserves may critically impair anastomotic integrity (33). In this study, PNI <50 was considered a statistically significant threshold for independently predicting heightened AL risk, aligning with prior research. For instance, Yu et al. developed a nomogram model validating PNI as a predictor of AL following esophagectomy in esophageal carcinoma patients (34). Similarly, a recent LASSO regression-derived risk model reaffirmed the clinical utility of PNI in predicting AL among patients undergoing radical gastrectomy (17). By incorporating PNI into our LASSO-based nomogram, clinicians acquire a robust, clinically actionable framework to stratify high-risk patients and tailor preoperative nutritional rehabilitation strategies. This underscores the necessity of proactive, multidisciplinary interventions to address malnutrition in these high-risk cohorts (35). For example, nutritional deficits may be ameliorated through early collaboration with clinical nutrition specialists, incorporating enteral or parenteral supplementation, nutrient enrichment, or elective delays in surgery until predefined nutritional thresholds are attained (36).

Another significant predictor of AL among elderly patients following LG was found to be an ASA class III. The ASA classification is a widely utilized tool for assessing preoperative physical status and operative risk (37). Our findings suggest that higher ASA classifications correlate with poorer baseline health status, which may contribute to reduced physiological resilience, elevated operative risks, and increased vulnerability to AL. For patients stratified as high-risk via the nomogram, we recommend comprehensive preoperative optimization, including systematic comorbidity management and nutritional supplementation to mitigate surgical risk. These measures align with studies by Rencuzogullari et al. (38) and Yang et al. (39), who emphasized ASA’s utility in preoperative risk stratification for AL, particularly in elderly cohorts with diminished physiological reserves. Based on our nomogram model, clinicians can shift from a reactive to a preventive paradigm, targeting modifiable risk factors such as ASA-linked comorbidities. For high-risk patients, surgeons may prioritize techniques such as tension-free anastomoses, selective intraoperative leak testing, or reinforcement with staple-line bolster materials to strengthen anastomotic integrity (40). Cira et al. corroborate this approach, demonstrating that compromised anastomotic healing in high-risk cohorts is intrinsically linked to suboptimal preoperative health status (41).

The identification of intraoperative blood loss (≥200 mL) as a significant predictor highlights the imperative for proactive surgical strategies to mitigate its multifactorial impact on anastomotic healing. Elevated blood loss exacerbates hypoxia, systemic inflammation, immune dysregulation, and coagulation imbalances, collectively compromising the structural integrity of the anastomotic site and increasing susceptibility to leakage (33,42). Similarly, intraoperative blood loss has been associated with AL, as proposed by Li et al. (30) in their nomogram-guided hemostasis protocol for colorectal surgery. To address this, technical refinements such as vascular skeletonization, judicious application of energy devices, and staged dissection in high-risk anatomical regions may minimize hemorrhagic complications, particularly in elderly patients with fragile vasculature (43). Notably, substantial intraoperative blood loss often necessitates transfusion to maintain hemodynamic stability (44). While our multivariate analysis excluded transfusions as an independent risk factor for AL, their documented association with AL suggests indirect pathophysiological interactions, such as transfusion-related immunomodulation or volume-associated stress, which may exacerbate leakage in vulnerable cohorts (45). By incorporating this nomogram into intraoperative pathways, clinicians can dynamically adjust strategies: reducing blood loss through technical refinements, tailoring transfusion practices, and escalating surveillance for high-risk cases (43).

Practice implications

The development of a LASSO-based nomogram for predicting AL in elderly patients undergoing LG represents a significant advancement in perioperative management and personalized therapeutic strategies for this vulnerable cohort. By incorporating five readily accessible clinical variables—age, diabetes, PNI, intraoperative blood loss, and ASA—this model establishes a dynamic risk stratification framework. This enables clinicians to categorize patients into distinct risk tiers, facilitating targeted interventions aligned with individualized AL risk profiles. High-risk patients identified by the nomogram warrant a proactive, multidisciplinary perioperative approach for support (46). For example, geriatric patients with low PNI scores, indicative of malnutrition, may benefit from preoperative nutrition protocols or protein supplementation to bolster tissue repair mechanisms (47). Concurrently, stringent glycemic control in diabetic patients could mitigate metabolic stress associated with surgical trauma. The inclusion of intraoperative blood loss as a predictor underscores the imperative for meticulous surgical technique in high-risk cases. Surgeons may prioritize advanced hemostatic strategies, minimize trauma during dissection, or reinforce anastomotic sites with bolstering techniques in patients with heightened susceptibility to AL during operation (48). Furthermore, anesthesiologists may leverage ASA classification to tailor hemodynamic monitoring and fluid resuscitation protocols, optimizing perfusion to anastomotic regions. Postoperatively, the nomogram facilitates personalized surveillance strategies. High-risk patients should undergo intensive monitoring, including serial laboratory evaluations and early imaging studies to detect subclinical signs of AL. Early mobilization and enteral nutritional support following LG may accelerate recovery, particularly in patients with preoperative frailty markers, as indicated by ASA classification. Integrating this nomogram into routine clinical practice empowers clinicians to shift from standardized protocols to precision medicine approaches. Its adoption aligns with the mandate for personalized surgical care, ensuring risk-adapted strategies that address the unique physiological vulnerabilities of elderly patients, thereby optimizing recovery and long-term quality of life (49).

Limitations

There are some limitations in the current study. First, the cross-sectional design restricts the ability to infer causal relationships between the identified predictors and the incidence of AL. Longitudinal studies are essential to determine whether risk factors directly contribute to AL or merely correlate with underlying physiological vulnerabilities. Second, the geographic and demographic constraints of the dataset, collected solely from Shenyang City and limited to elderly patients (≥65 years), may introduce selection bias. The exclusion of younger patients and those undergoing alternative surgical approaches (e.g., open or robotic gastrectomy) limits the generalizability of the nomogram. To address this, multicenter, prospective studies incorporating diverse geographic populations, varied operative techniques, and extended follow-up periods are critical to validate and refine the model’s applicability. Third, the current analysis excludes clinically pertinent variables such as preoperative radiotherapy, antibiotic use, or multi-organ resection, which may confound AL risk. Future iterations of the nomogram should incorporate these parameters to enhance predictive precision, especially in complex cases involving multimodal therapies. Finally, external validation remains imperative to verify the nomogram’s reliability across institutions with differing surgical protocols, patient demographics, and postoperative care standards.

Conclusions

In conclusion, this study developed and validated a LASSO-based nomogram to predict the risk of AL in elderly patients undergoing LG. By systematically integrating five predictors—age, diabetes, PNI, blood loss, and ASA—the model demonstrated strong discriminative power, reliable calibration, and significant clinical utility across training and validation cohorts. Notably, this nomogram represents a practical, evidence-based tool for identifying high-risk elderly patients who require tailored perioperative strategies. Our work provides a foundation for future research to advance personalized risk prediction in geriatric surgical oncology.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-897/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-897/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-897/prf

Funding: None.

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-897/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shengjing Hospital of China Medical University (No. 2024PS151K) and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. He F, Wang S, Zheng R, et al. Trends of gastric cancer burdens attributable to risk factors in China from 2000 to 2050. Lancet Reg Health West Pac 2024;44:101003. [Crossref] [PubMed]
2. He Y, Wang Y, Luan F, et al. Chinese and global burdens of gastric cancer from 1990 to 2019. Cancer Med 2021;10:3461-73. [Crossref] [PubMed]
3. Saif MW, Makrilia N, Zalonis A, et al. Gastric cancer in the elderly: an overview. Eur J Surg Oncol 2010;36:709-17. [Crossref] [PubMed]
4. Shimada S, Sawada N, Oae S, et al. Safety and curability of laparoscopic gastrectomy in elderly patients with gastric cancer. Surg Endosc 2018;32:4277-83. [Crossref] [PubMed]
5. Liao G, Wang Z, Zhang W, et al. Comparison of the short-term outcomes between totally laparoscopic total gastrectomy and laparoscopic-assisted total gastrectomy for gastric cancer: a meta-analysis. Medicine (Baltimore) 2020;99:e19225. [Crossref] [PubMed]
6. Sakamoto T, Fujiogi M, Matsui H, et al. Short-Term Outcomes of Laparoscopic and Open Total Gastrectomy for Gastric Cancer: A Nationwide Retrospective Cohort Analysis. Ann Surg Oncol 2020;27:518-26. [Crossref] [PubMed]
7. Baiocchi GL, Giacopuzzi S, Marrelli D, et al. International consensus on a complications list after gastrectomy for cancer. Gastric Cancer 2019;22:172-89. [Crossref] [PubMed]
8. Seicean RI, Puscasu D, Gheorghiu A, et al. Anastomotic Leakage after Gastrectomy for Gastric Cancer. J Gastrointestin Liver Dis 2023;32:526-35. [Crossref] [PubMed]
9. Mohri Y, Yasuda H, Ohi M, et al. Short- and long-term outcomes of laparoscopic gastrectomy in elderly patients with gastric cancer. Surg Endosc 2015;29:1627-35. [Crossref] [PubMed]
10. Kim EJ, Seo KW, Yoon KY. Laparoscopy-assisted distal gastrectomy for early gastric cancer in the elderly. J Gastric Cancer 2012;12:232-6. [Crossref] [PubMed]
11. Berlth F, Wichmann D, Fusco S, et al. Anastomotic leakage following surgical resection in the upper gastrointestinal tract. Chirurgie (Heidelb) 2024;95:871-877. [Crossref] [PubMed]
12. Shi J, Wu Z, Wu X, et al. Early Diagnosis of Anastomotic Leakage After Gastric Cancer Surgery Via Analysis of Inflammatory Factors in Abdominal Drainage. Ann Surg Oncol 2022;29:1230-41. [Crossref] [PubMed]
13. Bracale U, Peltrini R, De Luca M, et al. Predictive Factors for Anastomotic Leakage after Laparoscopic and Open Total Gastrectomy: A Systematic Review. J Clin Med 2022;11:5022. [Crossref] [PubMed]
14. Ali H, Shahzad M, Sarfraz S, et al. Application and impact of Lasso regression in gastroenterology: A systematic review. Indian J Gastroenterol 2023;42:780-90. [Crossref] [PubMed]
15. Houqiong J, Yuli Y, Yahang L, et al. LASSO-based nomogram predicts the risk factors of low anterior resection syndrome for middle and low rectal cancer underwent robotic surgery. Surg Endosc 2024;38:3378-87. [Crossref] [PubMed]
16. Wakahara T, Ueno N, Maeda T, et al. Impact of Gastric Cancer Surgery in Elderly Patients. Oncology 2018;94:79-84. [Crossref] [PubMed]
17. Zhao T, Li L, Wang Y, et al. Prognostic nutritional index combined with carcinoembryonic antigen and carbohydrate antigen 242 for early prediction of anastomotic leakage after radical gastrectomy for gastric cancer. Am J Transl Res 2023;15:4668-77. [PubMed]
18. Wang H, Ding Y, Zhuang M, et al. Application and progress of nomograms in gastric cancer. Front Med (Lausanne) 2025;12:1510742. [Crossref] [PubMed]
19. Peduzzi P, Concato J, Kemper E, et al. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373-9. [Crossref] [PubMed]
20. Jung HS, Park YK, Ryu SY, et al. Laparoscopic Total Gastrectomy in Elderly Patients (≥70 Years) with Gastric Carcinoma: A Retrospective Study. J Gastric Cancer 2015;15:176-82. [Crossref] [PubMed]
21. Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009;338:b2393. [Crossref] [PubMed]
22. Miao X, Guo Y, Ding L, et al. A dynamic online nomogram for predicting the heterogeneity trajectories of frailty among elderly gastric cancer survivors. Int J Nurs Stud 2024;153:104716. [Crossref] [PubMed]
23. Trapani R, Rausei S, Reddavid R, et al. Risk factors for esophago-jejunal anastomosis leakage after total gastrectomy for cancer. A multicenter retrospective study of the Italian research group for gastric cancer. Eur J Surg Oncol 2020;46:2243-7. [Crossref] [PubMed]
24. Pan Y, Chen K, Yu WH, et al. Laparoscopic gastrectomy for elderly patients with gastric cancer: A systematic review with meta-analysis. Medicine (Baltimore) 2018;97:e0007. [Crossref] [PubMed]
25. Merga ZC, Lee JS, Gong CS. Outcomes of Gastrectomy for Gastric Cancer in Patients Aged >80 Years: A Systematic Literature Review and Meta-Analysis. J Gastric Cancer 2023;23:428-50. [Crossref] [PubMed]
26. Maejima K, Taniai N, Yoshida H. Risk Factors for Esophagojejunal Anastomotic Leakage in Gastric Cancer Patients after Total Gastrectomy. J Nippon Med Sch 2023;90:64-8. [Crossref] [PubMed]
27. Tu RH, Lin JX, Zheng CH, et al. Development of a nomogram for predicting the risk of anastomotic leakage after a gastrectomy for gastric cancer. Eur J Surg Oncol 2017;43:485-92. [Crossref] [PubMed]
28. Zhong Y, Sun R, Li W, et al. Risk factors for esophageal anastomotic stricture after esophagectomy: a meta-analysis. BMC Cancer 2024;24:872. [Crossref] [PubMed]
29. Lin X, Li J, Chen W, et al. Diabetes and risk of anastomotic leakage after gastrointestinal surgery. J Surg Res 2015;196:294-301. [Crossref] [PubMed]
30. Li R, Zhou J, Zhao S, et al. Prediction model of anastomotic leakage after anterior resection for rectal cancer-based on nomogram and multivariate analysis with 1995 patients. Int J Colorectal Dis 2023;38:139. [Crossref] [PubMed]
31. Jan YK, Kelhofer N, Tu T, et al. Diagnosis, Pathophysiology and Management of Microvascular Dysfunction in Diabetes Mellitus. Diagnostics (Basel) 2024;14:2830. [Crossref] [PubMed]
32. Tan DJH, Yaow CYL, Mok HT, et al. The influence of diabetes on postoperative complications following colorectal surgery. Tech Coloproctol 2021;25:267-78. [Crossref] [PubMed]
33. Oshi M, Kunisaki C, Miyamoto H, et al. Risk Factors for Anastomotic Leakage of Esophagojejunostomy after Laparoscopy-Assisted Total Gastrectomy for Gastric Cancer. Dig Surg 2018;35:28-34. [Crossref] [PubMed]
34. Yu WQ, Gao HJ, Shi GD, et al. Development and validation of a nomogram to predict anastomotic leakage after esophagectomy for esophageal carcinoma. J Thorac Dis 2021;13:3549-65. [Crossref] [PubMed]
35. Chen XW, Guo XC, Cheng F. Impact of nutritional support on immunity, nutrition, inflammation, and outcomes in elderly gastric cancer patients after surgery. World J Gastrointest Surg 2024;16:2175-82. [Crossref] [PubMed]
36. Zhao XN, Lu J, He HY, et al. Clinical significance of preoperative nutritional status in elderly gastric cancer patients undergoing radical gastrectomy: A single-center retrospective study. World J Gastrointest Surg 2024;16:2211-20. [Crossref] [PubMed]
37. Doyle DJ, Hendrix JM, Garmon EH. American Society of Anesthesiologists Classification. Treasure Island (FL): StatPearls Publishing; 2023.
38. Rencuzogullari A, Benlice C, Valente M, et al. Predictors of Anastomotic Leak in Elderly Patients After Colectomy: Nomogram-Based Assessment From the American College of Surgeons National Surgical Quality Program Procedure-Targeted Cohort. Dis Colon Rectum 2017;60:527-36. [Crossref] [PubMed]
39. Yang L, Zhang P, Yang W, et al. Development and Validation of a Novel Nomogram Model for Early Diagnosis of Anastomotic Leakage After Laparoscopic Colorectal Cancer Surgery. Surg Infect (Larchmt) 2024; Epub ahead of print. [Crossref] [PubMed]
40. Reischl S, Wilhelm D, Friess H, et al. Innovative approaches for induction of gastrointestinal anastomotic healing: an update on experimental and clinical aspects. Langenbecks Arch Surg 2021;406:971-80. [Crossref] [PubMed]
41. Cira K, Wilhelm D, Neumann PA. Modern approaches and technologies to prevent anastomotic leakage in the gastrointestinal tract. Chirurgie (Heidelb) 2024;95:895-900. [Crossref] [PubMed]
42. Koyanagi K, Ozawa S, Oguma J, et al. Blood flow speed of the gastric conduit assessed by indocyanine green fluorescence: New predictive evaluation of anastomotic leakage after esophagectomy. Medicine (Baltimore) 2016;95:e4386. [Crossref] [PubMed]
43. Deguchi Y, Fukagawa T, Morita S, et al. Identification of risk factors for esophagojejunal anastomotic leakage after gastric surgery. World J Surg 2012;36:1617-22. [Crossref] [PubMed]
44. Kamei T, Kitayama J, Yamashita H, et al. Intraoperative blood loss is a critical risk factor for peritoneal recurrence after curative resection of advanced gastric cancer. World J Surg 2009;33:1240-6. [Crossref] [PubMed]
45. Kim SH, Son SY, Park YS, et al. Risk Factors for Anastomotic Leakage: A Retrospective Cohort Study in a Single Gastric Surgical Unit. J Gastric Cancer 2015;15:167-75. [Crossref] [PubMed]
46. Ushimaru Y, Nagano S, Kawabata R, et al. Enhancing surgical outcomes in elderly gastric cancer patients: the role of comprehensive preoperative assessment and support. World J Surg Oncol 2024;22:136. [Crossref] [PubMed]
47. Nomura E, Seki T, Yatabe K, et al. Study of the therapeutic strategy to improve survival outcomes from the perspective of perioperative conditions in elderly gastric cancer patients: a propensity score-matched analysis. World J Surg Oncol 2024;22:197. [Crossref] [PubMed]
48. Sun Q, Yang Z, Xu R, et al. Smart responsive staple for dynamic promotion of anastomotic stoma healing. Bioact Mater 2024;39:630-42. [Crossref] [PubMed]
49. Leonard ME, Williamson AJH, Weiss R, et al. A Framework to Optimize Primary Care of Older Surgical Patients: A Qualitative Study of Geriatricians. JAMA Netw Open 2025;8:e2456787. [Crossref] [PubMed]