Predictive value of postoperative C-reactive protein-to-albumin ratio in anastomotic leakage after esophagectomy

Automated machine learning model for predicting anastomotic strictures after esophageal cancer surgery: a retrospective cohort study

BACKGROUND

Anastomotic strictures (AS) frequently occurs in patients following esophageal cancer surgery, significantly affecting their long-term quality of life. This study aims to develop a machine learning model to predict high-risk AS, enabling early intervention and precise management.

METHODS

A total of 1549 patients underwent radical esophageal cancer surgery and were split into a training set (1084) and a validation set (465). Adaptive Synthetic Sampling (ADASYN) handled class imbalance, while Boruta and Least Absolute Shrinkage and Selection Operator (LASSO) with cross-validation refined key features. High-correlation features (r > 0.8) were assessed using variance inflation factors (VIFs) and clinical relevance. Machine learning models were trained and evaluated using area under curve (AUC), accuracy, sensitivity, specificity, calibration curves, and decision curve analysis (DCA). Shapley Additive exPlanations (SHAP) analysis improved model interpretability.

RESULTS

Seven critical variables were finalized, including anastomotic leakage (AL), neoadjuvant therapy (NCRT), suture method (SM), endoscopic assistance (EA), white blood cell count (WBC), albumin (Alb), and Suture site (SS). The Gradient Boosting Machine (GBM) model achieved the highest AUC, with 0.886 in the training set and 0.872 in the validation set. Shapley Additive Explanations (SHAP) analysis indicated that AL, SM, and NCRT were the most significant variables for model predictions.

CONCLUSION

The GBM machine learning model constructed in this study can effectively identify high-risk patients for AS following esophageal cancer surgery, offering strong support for earlier postoperative detection and precise clinical management.

Biomarkers associated with anastomotic leakage after esophagectomy: a systematic review

PURPOSE

Anastomotic leakage (AL) is one of the most important complications that occurs after upper gastrointestinal surgery, registering rates of 20-30% after esophagectomy. The role of systemic inflammatory biomarkers to predict anastomotic leaks is controversial and needs systematization.

METHODS

A systematic review based on the PRISMA guidelines criteria was performed. PubMed, Scopus, and Embase were queried using MESH Terms and All Fields key words to identify studies investigating a range of immune-inflammatory factors in predicting AL.

RESULTS

Twenty-four studies were included in this review. The total number of included patients was 5903, ranging in each study from 42 to 612. The included studies reported patients that underwent different techniques of esophagectomy (Ivor Lewis, McKeown, Orringer or thoracoabdominal esophagectomy) and 23 out of 24 studies included patients that underwent neoadjuvant treatment. While different biomarkers at different timepoints were analyzed, most studies have indicated postoperative biomarkers, between day 3 and day 5 to reach statistical significance.

CONCLUSIONS

Systemic inflammatory biomarkers represent potential risk stratification and predicting tools for AL after esophageal surgery, but more studies need to be conducted to validate their clinical utility.

Early postoperative systemic inflammatory response as predictor of anastomotic leakage after esophagectomy: a systematic review and meta-analysis

BACKGROUND AND PURPOSE

Postesophagectomy anastomotic leakage occurs in up to 16% of patients and is the main cause of morbidity and mortality. The leak severity is determined by the extent of contamination and the degree of sepsis, both of which are related to the time from onset to treatment. Early prediction based on inflammatory biomarkers such as C-reactive protein (CRP) levels, white blood cell counts, albumin levels, and combined Noble-Underwood (NUn) scores can guide early management. This review aimed to determine the diagnostic accuracy of these biomarkers.

METHODS

This study was designed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and registered in the PROSPERO (International Prospective Register of Systematic Reviews) database. Two reviewers independently conducted searches across PubMed, MEDLINE, Web of Science, and Embase. Sources of bias were assessed, and a meta-analysis was performed.

RESULTS

Data from 5348 patients were analyzed, and 13% experienced leakage. The diagnostic accuracy of the serum biomarkers was analyzed, and pooled cutoff values were identified. CRP levels were found to have good diagnostic accuracy on days 2 to 5. The best discrimination was identified on day 2 for a cutoff value < 222 mg/L (area under the curve = 0.824, sensitivity = 81%, specificity = 88%, positive predictive value = 38.6%, and negative predictive value = 98%). A NUn score of >10 on day 4 correlated with poor diagnostic accuracy.

CONCLUSION

The NUn score failed to achieve adequate accuracy. CRP seems to be the only valuable biomarker and is a negative predictor of postesophagectomy leakage. Patients with a CRP concentration of <222 mg/L on day 2 are unlikely to develop a leak, and patients can safely proceed through their enhanced recovery after surgery protocol. Patients with a CRP concentration of <127 mg/L on day 5 can be safely discharged when clinically possible.

Systemic Inflammatory Response and the Noble and Underwood (NUn) Score as Early Predictors of Anastomotic Leakage after Esophageal Reconstructive Surgery

Anastomotic leakage (AL) remains the main cause of post-esophagectomy morbidity and mortality. Early detection can avoid sepsis and reduce morbidity and mortality. This study evaluates the diagnostic accuracy of the Nun score and its components as early detectors of AL. This single-center observational cohort study included all esophagectomies from 2010 to 2020. C-reactive protein (CRP), albumin (Alb), and white cell count (WCC) were analyzed and NUn scores were calculated. The area under the curve statistic (AUC) was used to assess their predictive accuracy. A total of 74 of the 668 patients (11%) developed an AL. CRP and the NUn-score proved to be good diagnostic accuracy tests on postoperative day (POD) 2 (CRP AUC: 0.859; NUn score AUC: 0.869) and POD 4 (CRP AUC: 0.924; NUn score AUC: 0.948). A 182 mg/L CRP cut-off on POD 4 yielded a 87% sensitivity, 88% specificity, a negative predictive value (NPV) of 98%, and a positive predictive value (PPV) of 47.7%. A NUn score cut-off > 10 resulted in 92% sensitivity, 95% specificity, 99% NPV, and 68% PPV. Albumin and WCC have limited value in the detection of post-esophagectomy AL. Elevated CRP and a high NUn score on POD 4 provide high accuracy in predicting AL after esophageal cancer surgery. Their high negative predictive value allows to select patients who can safely proceed with enhanced recovery protocols.

Routine Radiologic Assessment for Anastomotic Leak Is Not Necessary in Asymptomatic Patients After Esophagectomy for Esophageal Cancer

BACKGROUND

Anastomotic leaks (AL) are a major source of post-esophagectomy morbidity and patients are often initially asymptomatic. Debate exists on timing and utility of imaging to detect AL post-esophagectomy. We sought to evaluate the efficacy and timing of radiographic AL evaluation in esophageal cancer patients post-esophagectomy.

METHODS

A retrospective database of esophageal cancer patients who underwent esophagectomy at a single institution from 2004 to 2020 was used to determine the utilization, timing, and sensitivity of radiologic testing for AL post-esophagectomy.

RESULTS

Seventy-six patients were identified of which 37 (49%) had a cervical anastomosis. Sixty-four (84%) underwent 71 "asymptomatic radiographic leak tests" (ARLT), 7 of which had 2 different tests, including: 41 fluoroscopic esophagrams (58%), 18 CT-esophagrams (25%), and 12 upper GI studies (17%). Seventeen patients (22%) developed clinical signs of AL (hemodynamic instability, leukocytosis) and underwent "symptomatic radiographic leak tests" (SRLT) with fluoroscopic esophagram (n = 9, 12%), CT-esophagram (n = 7, 9%), or upper GI study (n = 1, 1%). ARLT and SRLT were positive in 2/64 (3%) and 17/17 (100%) patients, respectively, for 19 total ALs (25%). Among the 17 SRLT( +) patients, 1 was also ARLT( +), 13 were initially ARLT( -), and 3 were not evaluated by ARLT. The median postoperative day for ARLT and SRLT was 4.0 (IQR 3.0-5.5) and 9.0 days (IQR 6.0-13.0), respectively, with a statistically significant difference (p < 0.005). The sensitivity and specificity of ARLT for detecting AL were 13.3% and 100.0%, respectively.

CONCLUSIONS

Based on the low ARLT sensitivity, routine use of imaging to detect asymptomatic ALs post-esophagectomy may be limited. Symptomatic ALs were often present in a delayed fashion, even after initial negative imaging.