Predicting Intensive Care Unit Length of Stay After Acute Type A Aortic Dissection Surgery Using Machine Learning

Predictors, prognosis and costs of prolonged intensive care unit stay after surgery for type A aortic dissection

BACKGROUND

The outcomes after prolonged treatment in the intensive care unit (ICU) after surgery for Stanford type A aortic dissection (TAAD) have not been previously investigated.

METHODS

This analysis included 3538 patients from a multicenter study who underwent surgery for acute TAAD and were admitted to the cardiac surgical ICU.

RESULTS

The mean length of stay in the cardiac surgical ICU was 9.9±9.5 days. The mean overall costs of treatment in the cardiac surgical ICU 24086±32084 €. In-hospital mortality was 14.8% and 5-year mortality was 30.5%. Adjusted analyses showed that prolonged ICU stay was associated with significantly lower risk of in-hospital mortality (adjusted OR 0.971, 95%CI 0.959-0.982), and of five-year mortality (adjusted OR 0.970, 95%CI 0.962-0.977), respectively. Propensity score matching analysis yielded 870 pairs of patients with short ICU stay (2-5 days) and long ICU stay (>5 days) with balanced baseline, operative and postoperative variables. Patients with prolonged ICU stay (>5 days) had significantly lower in-hospital mortality (8.9% vs. 17.4%, <0.001) and 5-year mortality (28.2% vs. 30.7%, P=0.007) compared to patients with short ICU-stay (2-5 days).

CONCLUSIONS

Prolonged ICU stay was common after surgery for acute TAAD. However, when adjusted for multiple baseline and operative variables as well as adverse postoperative events and the cluster effect of hospitals, it was associated with favorable survival up to 5 years after surgery.

TMEM52B Isoforms P18 and P20 Differentially Promote the Oncogenesis and Metastasis of Nasopharyngeal Carcinoma

Transmembrane protein 52B (TMEM52B), a newly identified tumor-related gene, has been reported to regulate various tumors, yet its role in nasopharyngeal carcinoma (NPC) remains unclear. Transcriptomic analysis of NPC cell lines reveals frequent overexpression of TMEM52B, and immunohistochemical results show that TMEM52B is associated with advanced tumor stage, recurrence, and decreased survival time. Depleting TMEM52B inhibits the proliferation, migration, invasion, and oncogenesis of NPC cells in vivo. TMEM52B encodes two isoforms, TMEM52B-P18 and TMEM52B-P20, differing in their N-terminals. While both isoforms exhibit similar pro-oncogenic roles and contribute to drug resistance in NPC, TMEM52B-P20 differentially promotes metastasis. This functional discrepancy may be attributed to their distinct subcellular localization; TMEM52B-P18 is confined to the cytoplasm, while TMEM52B-P20 is found both at the cell membrane and in the cytoplasm. Mechanistically, cytoplasmic TMEM52B enhances AKT phosphorylation by interacting with phosphoglycerate kinase 1 (PGK1), fostering NPC growth and metastasis. Meanwhile, membrane-localized TMEM52B-P20 promotes E-cadherin ubiquitination and degradation by facilitating its interaction with the E3 ubiquitin ligase NEDD4, further driving NPC metastasis. In conclusion, the TMEM52B-P18 and TMEM52B-P20 isoforms promote the metastasis of NPC cells through different mechanisms. Drugs targeting these TMEM52B isoforms may offer therapeutic benefits to cancer patients with varying degrees of metastasis.

A comparison of single and double arterial cannulation for cardiopulmonary bypass for acute type A aortic surgery: A single center, retrospective observational study

Background: Acute type A aortic dissection (ATAAD) is a cardiovascular emergency and has high mortality and morbidity. We retrospectively compared the effects on outcomes of single arterial cannulation via axillary artery (AAC) with double arterial cannulation via axillary and femoral artery (DAC) in patients who underwent cardiopulmonary bypass (CPB) for ATAAD.Methods: Between January 2017 and May 2021, four hundred 29 patients who underwent aortic arch repair with circulatory arrest for ATAAD were divided into AAC group (n = 283) and DAC group (n = 146). The propensity score-matched (PSM) analysis were performed to compare the characteristics and outcomes of the groups.Results: After PSM (n = 137 in each), the DAC group had a longer duration of CPB (229 vs 244, p = 0.011), aortic cross-clamp time (121 vs 149, p < 0.001), durations of Intensive Care Unit (ICU) stay (7 vs 8, p = 0.014) and hospital stay (19 vs 25, p < 0.001) compared with AAC group. The incidences of dialysis (21% vs. 31%, p = 0.073), postoperative stroke (9% vs 15%, p = 0.143), ECMO support (2% vs 7%, p = 0.077), in-hospital mortality (7% vs 14%, p = 0.071) and follow-up mortality (10% vs 19%, p = 0.059) showed no significant difference between two groups. Multivariate logistic regression analysis showed postoperative ECMO (OR: 16.69, 95% CI: 1.78-156.29; p = 0.014) or stroke (OR: 11.34, 95% CI: 2.64-48.72; p < 0.001) were associated with in-hospital mortality. Univariate Cox regression results showed stroke history (OR: 4.61, 95% CI: 1.90-11.16; p = 0.001), aortic valvuloplasty (OR: 0.21, 95% CI: 0.07-0.59; p = 0.003), postoperative ALT day1 (OR: 1.00, 95% CI: 1.00-1.00; p = 0.008), ECMO (OR: 16.30, 95% CI: 4.78-55.61; p < 0.001), tracheotomy (OR: 3.78, 95% CI: 1.08-13.20; p = 0.037), postoperative stroke (OR: 4.61, 95% CI: 1.90-11.16; p < 0.001) and re-exploration for bleeding (OR: 3.52, 95% CI: 1.01-12.27; p = 0.048) were associated to follow-up mortality.Conclusions: For surgical treatment of ATAAD with CPB when compared to double axillary and femoral artery, single axillary cannulation was associated with shorter durations of CPB and ACC as well as ICU and hospital stays but no with significant difference in mortality.

Personalised preinterventional risk stratification of mortality, length of stay and hospitalisation costs in transcatheter aortic valve implantation using a machine learning algorithm: a pilot trial

INTRODUCTION

Risk stratification based on Euroscore II (ESII) is used in some centres to assist decisions to perform transcatheter aortic valve implant (TAVI) procedures. ESII is a generic, non-TAVI-specific metric, and its performance fades for mortality at follow-up longer than 30 days. We investigated if a TAVI-specific predictive model could achieve improved predictive preinterventional accuracy of 1-year mortality compared with ESII.

PATIENTS AND METHODS

In this prospective pilot study, 284 participants with severe symptomatic aortic valve stenosis who underwent TAVI were enrolled. Standard clinical metrics (American Society of Anesthesiology (ASA), New York Heart Association and ESII) and patient-reported outcome measures (EuroQol-5 Dimension-Visual Analogue Scale, Kansas City Cardiomyopathy Questionnaire and Clinical Frailty Scale (CFS)) were assessed 1 day before TAVI. Using these data, we tested predictive models (logistic regression and decision tree algorithm (DTA)) with 1-year mortality as the dependent variable.

RESULTS

Logistic regression yielded the best prediction, with ASA and CFS as the strongest predictors of 1-year mortality. Our logistic regression model score showed significantly better prediction accuracy than ESII (area under the curve=0.659 vs 0.800; p=0.002). By translating our results to a DTA, cut-off score values regarding 1-year mortality risk emerged for low, intermediate and high risk. Treatment costs and length of stay (LoS) significantly increased in high-risk patients.

CONCLUSIONS AND SIGNIFICANCE

A novel TAVI-specific model predicts 1-year mortality, LoS and costs after TAVI using simple, established, transparent and inexpensive metrics before implantation. Based on this preliminary evidence, TAVI team members and patients can make informed decisions based on a few key metrics. Validation of this score in larger patient cohorts is needed.

Risk factors for prolonged postoperative ICU stay in the patients with Stanford type A aortic dissection

OBJECTIVE

To investigate the independent risk factors for postoperative prolonged ICU stay in patients with Stanford type A aortic dissection (TAAD) and assess the clinical outcomes of prolonged ICU stay.

METHOD

The clinical data of 100 patients with TAAD admitted to the Department of Cardiovascular Surgery, First Affiliated Hospital of Anhui Medical University from December 2018 to September 2022 were retrospectively collected and analyzed. Patients were divided into two groups, based on the postoperative ICU stay (7 days as the threshold), regular ICU stay group (< 7 days) and prolonged ICU stay group (≥ 7 days). First, preoperative and intraoperative materials were collected for univariate analysis. Then, the significant variables after univariate analysis were analyzed using logistic regression, and the final independent risk factors for prolonged ICU stay were determined. Meanwhile, the postoperative clinical outcomes were analyzed with the aim of assessing the clinical outcomes due to prolonged ICU stay.

RESULTS

There were 65 and 35 patients in the regular ICU stay group and the prolonged ICU stay group, respectively. In accordance with the result of univariate analysis in the two groups, emergency surgery (χ2 = 13.598; P < 0.001), preoperative urea nitrogen (t = 3.006; P = 0.004), cardiopulmonary bypass (CPB) time (t = 2.671; P = 0.001) and surgery time (t = 2.630; P = 0.010) were significant. All significant variates were analyzed through logistic regression, and it was found that emergency surgery (OR = 0.192; 95% CI: 0.065-0.561), preoperative urea nitrogen (OR = 0.775; 95% CI: 0.634-0.947) and cardiopulmonary time (OR = 0.988; 95% CI: 0.979-0.998) were independent risk factors for prolonged postoperative ICU stay. The Receiver Operating Characteristic (ROC) curves of these three factors were also effective in predicting postoperative prolonged ICU stay (Emergency surgery, AUC = 0.308, 95% CI: 0.201-0.415; Preoperative urea nitrogen, AUC = 0.288, 95% CI: 0.185-0.392; cardiopulmonary time, AUC = 0.340, 95% CI: 0.223-0.457). Moreover, compared with a single factor, the predictive value of combined factors was more significant (AUC = 0.810, 95% CI: 0.722-0.897). For the comparison of postoperative data in the two groups,, compared with the regular ICU stay group, the incidence of adverse events in the prolonged ICU stay group increased significantly, including limb disability of limbs (χ2 = 22.182; P < 0.001), severe organ injury (χ2 = 23.077; P < 0.001), tracheotomy (χ2 = 17.582; P < 0.001), reintubation (χ2 = 28.020; P < 0.001), 72 h tracheal extubation after surgery (χ2 = 29.335; P < 0.001), 12 h consciousness recovery after surgery (χ2 = 18.445; P < 0.001), ICU re-entering (χ2 = 9.496; P = 0.002) and irregular discharging (χ2 = 24.969; P < 0.001).

CONCLUSION

Emergency surgery, preoperative urea nitrogen, and CPB time are risk factors for postoperative prolonged ICU stay after TAAD surgery. Furthermore, prolonged ICU stay is associated with worse clinical outcomes. Hence, a reasonable strategy should be adopted proactively focusing on the risk factors to shorten ICU stays and improve clinical outcomes.

Prediction of the development of contrast‑induced nephropathy following percutaneous coronary artery intervention by machine learning

Contrast-induced nephropathy (CIN) is associated with increased mortality and morbidity in patients with coronary artery disease undergoing elective percutaneous coronary intervention(PCI). We developed a machine learning-based risk stratification model to predict contrast-induced nephropathy after PCI. A study retrospectively enrolling 240 patients eligible for PCI from December 2017 to May 2020 was performed. CIN was defined as a rise in serum creatinine levels ≥0.5 mg/dL or ≥25% from baseline within 72 h after surgery. Eight machine learning methods were performed based on clinical variables. Shapley Additive exPlanation values were also used to interpret the best-performing prediction models. Development of CIN was found in 37 patients(16.5%) after PCI. There were 11 significant predictors of CIN, including uric acid, peripheral vascular disease, cystatin C, creatine kinase-MB, haemoglobin, N-terminal pro-brain natriuretic peptide, age, diabetes, systemic immune-inflammatory index, total protein, and low-density lipoprotein. Regarding the efficacy of the machine learning model that accurately predicted CIN, SVM exhibited the most outstanding AUC value of 0.784. The SHAP and radar plots were used to illustrate the positive and negative effects of the 11 features attributed to the SVM. Machine learning models have the potential to identify the risk of CIN for elective PCI patients.

Myocardial Priority Promotes Cardiovascular Recovery for Acute Type A Aortic Dissection Combined with Coronary Artery Disease Undergoing Aortic Arch Surgery

The optimal surgical strategy for acute type A aortic dissection (ATAAD) with coronary artery disease (CAD) remains unclear. The goal of this study was to investigate the cardiovascular protective effects of the myocardial priority (MP) strategy or traditional selective cerebral perfusion (SCP) in ATAAD with CAD. A total of 214 adults were analyzed retrospectively, of which 80 underwent the MP strategy intraoperatively. Seventy-nine pairs were propensity-score-matched and divided into SCP and MP groups. The follow-up period ranged from 6 to 36 months. The MP group had a significantly shorter myocardial ischemic time, higher perfusion flow, higher radial artery pressure, and lower incidence of NIRS decrease >20% of the base value, but a longer lower limb circulatory arrest and bypass time than the SCP group. Although similar adverse cardiac and cerebrovascular events were observed in both groups, a shorter posthospital stay, less blood loss and transfusion, higher postoperative hemoglobin, lower creatinine, and higher PaO2/FiO2 were observed in the MP group. Subgroup analysis showed that when the TIMI Risk Score was <4, the MP group had a lower incidence of low cardiac output and lower postoperative cTnI level. The follow-up patients had similar morbidities between the two groups. The novel MP strategy is associated with a shortened myocardial ischemic time, better maintained perfusion of vital organs, and postoperative recovery after surgery for ATAAD combined with non-severe CAD.

Circulating biomarker-based risk stratifications individualize arch repair strategy of acute Type A aortic dissection via the XGBoosting algorithm

Aims

The incremental usefulness of circulating biomarkers from different pathological pathways for predicting mortality has not been evaluated in acute Type A aortic dissection (ATAAD) patients. We aim to develop a risk prediction model and investigate the impact of arch repair strategy on mortality based on distinct risk stratifications.

Methods and results

A total of 3771 ATAAD patients who underwent aortic surgery retrospectively included were randomly divided into training and testing cohorts at a ratio of 7:3 for the development and validation of the risk model based on multiple circulating biomarkers and conventional clinical factors. Extreme gradient boosting was used to generate the risk models. Subgroup analyses were performed by risk stratifications (low vs. middle-high risk) and arch repair strategies (proximal vs. extensive arch repair). Addition of multiple biomarkers to a model with conventional factors fitted an ABC risk model consisting of platelet-leucocyte ratio, mean arterial pressure, albumin, age, creatinine, creatine kinase-MB, haemoglobin, lactate, left ventricular end-diastolic dimension, urea nitrogen, and aspartate aminotransferase, with adequate discrimination ability {area under the receiver operating characteristic curve (AUROC): 0.930 [95% confidence interval (CI) 0.906-0.954] and 0.954, 95% CI (0.930-0.977) in the derivation and validation cohort, respectively}. Compared with proximal arch repair, the extensive repair was associated with similar mortality risk among patients at low risk [odds ratio (OR) 1.838, 95% CI (0.559-6.038); P = 0.316], but associated with higher mortality risk among patients at middle-high risk [OR 2.007, 95% CI (1.460-2.757); P < 0.0001].

Conclusion

In ATAAD patients, the simultaneous addition of circulating biomarkers of inflammatory, cardiac, hepatic, renal, and metabolic abnormalities substantially improved risk stratification and individualized arch repair strategy.

Comparison of Four Machine Learning Techniques for Prediction of Intensive Care Unit Length of Stay in Heart Transplantation Patients

Background

Post-operative heart transplantation patients often require admission to an intensive care unit (ICU). Early prediction of the ICU length of stay (ICU-LOS) of these patients is of great significance and can guide treatment while reducing the mortality rate among patients. However, conventional linear models have tended to perform worse than non-linear models.

Materials and Methods

We collected the clinical data of 365 patients from Wuhan Union Hospital who underwent heart transplantation surgery between April 2017 and August 2020. The patients were randomly divided into training data (N = 256) and test data (N = 109) groups. 84 clinical features were collected for each patient. Features were validated using the Least Absolute Shrinkage and Selection Operator (LASSO) regression’s fivefold cross-validation method. We obtained Shapley Additive explanations (SHAP) values by executing package “shap” to interpret model predictions. Four machine learning models and logistic regression algorithms were developed. The area under the receiver operating characteristic curve (AUC-ROC) was used to compare the prediction performance of different models. Finally, for the convenience of clinicians, an online web-server was established and can be freely accessed via the website https://wuhanunion.shinyapps.io/PredictICUStay/.

Results

In this study, 365 consecutive patients undergoing heart transplantation surgery for moderate (NYHA grade 3) or severe (NYHA grade 4) heart failure were collected in Wuhan Union Hospital from 2017 to 2020. The median age of the recipient patients was 47.2 years, while the median age of the donors was 35.58 years. 330 (90.4%) of the donor patients were men, and the average surgery duration was 260.06 min. Among this cohort, 47 (12.9%) had renal complications, 25 (6.8%) had hepatic complications, 11 (3%) had undergone chest re-exploration and 19 (5.2%) had undergone extracorporeal membrane oxygenation (ECMO). The following six important clinical features were selected using LASSO regression, and according to the result of SHAP, the rank of importance was (1) the use of extracorporeal membrane oxygenation (ECMO); (2) donor age; (3) the use of an intra-aortic balloon pump (IABP); (4) length of surgery; (5) high creatinine (Cr); and (6) the use of continuous renal replacement therapy (CRRT). The eXtreme Gradient Boosting (XGBoost) algorithm presented significantly better predictive performance (AUC-ROC = 0.88) than other models [Accuracy: 0.87; sensitivity: 0.98; specificity: 0.51; positive predictive value (PPV): 0.86; negative predictive value (NPV): 0.93].

Conclusion

Using the XGBoost classifier with heart transplantation patients can provide an accurate prediction of ICU-LOS, which will not only improve the accuracy of clinical decision-making but also contribute to the allocation and management of medical resources; it is also a real-world example of precision medicine in hospitals.

Personalized Preoperative Prediction of the Length of Hospital Stay after TAVI Using a Dedicated Decision Tree Algorithm

Background: The aim of this study was to identify pre-operative parameters able to predict length of stay (LoS) based on clinical data and patient-reported outcome measures (PROMs) from a scorecard database in patients with significant aortic stenosis who underwent TAVI (transfemoral aortic valve implantation). Methods: 302 participants (51.7% males, age range 78.2−84.2 years.) were prospectively recruited. After computing the median LoS value (=6 days, range = 5−8 days), we implemented a decision tree algorithm by setting dichotomized values at median LoS as the dependent variable and assessed baseline clinical variables and PROMs (Clinical Frailty Scale (CFS), EuroQol-5 Dimension-5 Levels (EQ-5D) and Kansas City Cardiomyopathy Questionnaire (KCCQ)) as potential predictors. Results: Among clinical parameters, only peripheral arterial disease (p = 0.029, HR = 1.826) and glomerular filtration rate (GFR, cut-off < 33 mL/min/1.73 m2, p = 0.003, HR = 2.252) were predictive of LoS. Additionally, two PROMs (CFS; cut-off = 3, p < 0.001, HR = 1.324 and KCCQ; cut-off = 30, p = 0.003, HR = 2.274) were strong predictors. Further, a risk score for LoS (RS_LoS) was calculated based on these predictors. Patients with RS_LoS = 0 had a median LoS of 5 days; patients RS_LoS ≥ 3 had a median LoS of 8 days. Conclusions: based on the pre-operative values of the above four predictors, a personalized prediction of LoS after TAVI can be achieved.