Prediction of Surgical Outcome by Modeling Based on Risk Factors of Morbidity After Pulmonary Resection for Lung Cancer in Older Adults

KAMLN: A Knowledge-aware Multi-label Network for Lung Cancer Complication Prediction

Surgical resection is now the only curative approach for early stage lung cancer patients. However, postoperative complications pose a significant threat to the health and life of patients. The current complication prediction methods usually ignore the potential causal relationships between different complications, which may limit their predictive performances. To exploit this knowledge, we propose a knowledge-aware multi-label network (KAMLN) for complication prediction. In this approach, we first construct a knowledge graph to describe the potential causal relationships between different complications. Then, we design a neural network based on this knowledge graph to learn the relationships between different complications to achieve better predictive performances. Experiments using 593 lung cancer patients' data show that the KAMLN achieves a micro-AUC value of 0.664±0.100, which is better than the baseline methods. The SHAP analysis indicates lymph node dissection has a significant impact on multiple complications. Based on the experimental results, the proposed KAMLN can effectively utilize prior knowledge between different complications to achieve more accurate and fine-grained complication prediction.

A Hierarchy-driven Multi-label Network with Label Constraints for Post-operative Complication Prediction of Lung Cancer. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2023;2023:1-6. [PMID: 38083421 DOI: 10.1109/embc40787.2023.10339943] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Lung cancer is one of the most dangerous cancers all over the world. Surgical resection remains the only potentially curative option for patients with lung cancer. However, this invasive treatment often causes various complications, which seriously endanger patient health. In this study, we proposed a novel multi-label network, namely a hierarchy-driven multi-label network with label constraints (HDMN-LC), to predict the risk of complications of lung cancer patients. In this method, we first divided all complications into pulmonary and cardiovascular complication groups and employed the hierarchical cluster algorithm to analyze the hierarchies between these complications. After that, we employed the hierarchies to drive the network architecture design so that related complications could share more hidden features. And then, we combined all complications and employed an auxiliary task to predict whether any complications would occur to impose the bottom layer to learn general features. Finally, we proposed a regularization term to constrain the relationship between specific and combined complication labels to improve performance. We conducted extensive experiments on real clinical data of 593 patients. Experimental results indicate that the proposed method outperforms the single-label, multi-label baseline methods, with an average AUC value of 0.653. And the results also prove the effectiveness of hierarchy-driven network architecture and label constraints. We conclude that the proposed method can predict complications for lung cancer patients more effectively than the baseline methods.Clinical relevance-This study presents a novel multi-label network that can more accurately predict the risk of specific postoperative complications for lung cancer patients. The method can help clinicians identify high-risk patients more accurately and timely so that interventions can be implemented in advance to ensure patient safety.

Association between averaged intraoperative nociceptive response index and postoperative complications after lung resection surgery

OBJECTIVES

Since postoperative complications, defined as Clavien-Dindo grade ≥II, correlate with long-term survival after lung resection surgery in patients with primary lung cancer, identification of intraoperative risk factors for postoperative complications is crucial for better perioperative management. In the present study, we investigated the possible association between intraoperative variables for use in anaesthetic management and Clavien-Dindo grade ≥II.

METHODS

In this multi-institutional observational study, consecutive adult patients undergoing video-assisted thoracic surgery for primary lung cancer under general anaesthesia from March 2019 to April 2021 were enrolled. All patients were divided into 2 groups with Clavien-Dindo grade <II and ≥II. Uni- and multivariable analyses were performed to identify intraoperative risk factors.

RESULTS

After univariable analysis between patients with Clavien-Dindo grade <II (n = 415) and ≥II (n = 121), multivariable analysis revealed higher averaged nociceptive response (NR) index during surgery (mean NR), male sex, lower body mass index, longer duration of surgery, higher blood loss and lower urine volume, as independent risk factors for postoperative complications. In sensitivity analysis, based on the cut-off value of mean NR for postoperative complications, all patients were divided into high and low mean NR groups. The incidence of postoperative complications was significantly higher in patients with high mean NR (n = 332) than in patients with low mean NR (n = 204; P < 0.001).

CONCLUSIONS

Higher mean NR, as intraoperative variables for use in anaesthetic management, is associated with the higher incidence of postoperative complications after primary lung cancer surgery.

The Effect of Major and Minor Complications After Lung Surgery on Length of Stay and Readmission

The effect of post-operative adverse events (AEs) on patient outcomes such as length of stay (LOS) and readmissions to hospital is not completely understood. This study examined the severity of AEs from a high-volume thoracic surgery center and its effect on the patient postoperative LOS and readmissions to hospital. This study includes patients who underwent an elective lung resection between September 2018 and January 2020. The AEs were grouped as no AEs, 1 or more minor AEs, and 1 or more major AEs. The effects of the AEs on patient LOS and readmissions were examined using a survival analysis and logistic regression, respectively, while adjusting for the other demographic or clinical variables. Among 488 patients who underwent lung surgery, (Wedge resection [n = 100], Segmentectomy [n = 51], Lobectomy [n = 310], Bilobectomy [n = 10], or Pneumonectomy [n = 17]) for either primary (n = 440) or secondary (n = 48) lung cancers, 179 (36.7%) patients had no AEs, 264 (54.1%) patients had 1 or more minor AEs, and 45 (9.2%) patients had 1 or more major AEs. Overall, the median of LOS was 3 days which varied significantly between AE groups; 2, 4, and 8 days among the no, minor, and major AE groups, respectively. In addition, type of surgery, renal disease (urinary tract infection [UTI], urinary retention, or acute kidney injury), and ASA (American Society of Anesthesiology) score were significant predictors of LOS. Finally, 58 (11.9%) patients were readmitted. Readmission was significantly associated with AE group (P = 0.016). No other variable could significantly predict patient readmission. Overall, postoperative AEs significantly affect the postoperative LOS and readmission rates.

Perioperative Factors for Predicting the Need for Postoperative Intensive Care after Major Lung Resection

Postoperative management after major lung surgery is critical. This study evaluates risk factors for predicting mandatory intensive care unit (ICU) admission immediately after major lung resection. We retrospectively reviewed patients for whom the surgeon requested an ICU bed before major lung resection surgery. Patients were classified into three groups. Univariable and multivariable logistic regression analyses were performed, and a clinical nomogram was constructed. Among 340 patients, 269, 50, and 21 were classified into the no need for ICU, mandatory ICU admission, and late-onset complication groups, respectively. Predictive postoperative diffusion capacity of the lung for carbon monoxide (47.2 (interquartile range (IQR) 43.3–65.7)% versus vs. 67.8 (57.1–79.7)%; p = 0.003, odds ratio (OR) 0.969, 95% confidence interval (CI) 0.95–0.99), intraoperative blood loss (400.00 (250.00–775.00) mL vs. 100.00 (50.00–250.00) mL; p = 0.040, OR 1.001, 95% CI 1.000–1.002), and open thoracotomy (p = 0.030, OR 2.794, 95% CI 1.11–7.07) were significant predictors for mandatory ICU admission. The risk estimation nomogram demonstrated good accuracy in estimating the risk of mandatory ICU admission (concordance index 83.53%). In order to predict the need for intensive care after major lung resection, preoperative and intraoperative factors need to be considered.

Long-term impact of complications after lung resections in non-small cell lung cancer

Background

Postoperative complications after lung resection are common and fatal. The immediate effects of postoperative complications are related to poor prognosis; however, the long-term effects have not been assessed. Thus, this investigation aimed to clarify the long-term effects of postoperative complications among patients with resected non-small cell lung cancer (NSCLC).

Methods

This retrospective cohort study included 345 patients with resected NSCLC from a single institution. We used the Clavien-Dindo classification to classify postoperative complications. Postoperative complications were defined as complications with a Clavien-Dindo grade of ≥2. The Kaplan-Meier method was used to evaluate survival. Prognostic factors were analyzed using a Cox proportional hazard model.

Results

There were 110 patients with postoperative complications (31.9%). The 5-year overall survival (OS), recurrence-free survival (RFS), and cause-specific survival (CSS) rates were significantly lower in patients with complications than in those without complications [OS: 66.1%, 95% confidence interval (CI): 55.4-74.8% vs. 78.0%, 95% CI: 71.8-83.1%, P=0.001; RFS: 48.8%, 95% CI: 38.1-58.7% vs. 70.8%, 95% CI: 64.2-76.4%, P<0.001; CSS: 82.7%, 95% CI: 72.8-89.3% vs. 88.2%, 95% CI: 82.8-92.0%, P=0.005]. The 5-year OS was lower in the pulmonary complication group than in the other complication group (58.1%, 95% CI: 40.0-72.4% vs. 70.5%, 95% CI: 56.6-80.6%, P=0.033). Postoperative complications were indicated as a poor prognostic factor for OS (hazard ratio, 1.67; 95% CI: 1.11-2.53; P=0.002).

Conclusions

Postoperative complications were associated with unfavorable OS because of the worse prognosis of postoperative pulmonary complications.