Updated external validation of the SORG machine learning algorithms for prediction of ninety-day and one-year mortality after surgery for spinal metastasis

Introduction: The incidence of device infection is constantly increasing; requiring transvenous lead extraction (TLE). Data regarding predictors of short and long-term mortality after TLE for infection are limited. Methods: We collected data regarding 30-day and 1-year mortality of patients undergoing TLE at a university hospital between April 2004 and June 2015. Patients with less than 30-day follow up were excluded. Results: Out of total 1223 TLE procedures, 700 were performed for infectious indications. 30-day follow-up was available for 620 patients (88.6%) and 1-year follow up was available for 541 patients (77.3%). Overall 30-day mortality was 9% (4.3% for pocket infection, 12.3% for systemic infection) and 1-year mortality was 27.5% (16% for pocket infection, 35.8% for systemic infection). Patient age, end-stage renal disease, history of valve replacement, atrial fibrillation, Staphylococcus aureus infection, systemic infection, any procedural complication, elevated WBC count, low hemoglobin, need for CCU admission, need for pressor support, acute kidney injury, cardiogenic shock and need for blood transfusion were predictors of both 30-day and 1-year mortality in univariate analysis. Any retained fragment was predictor of 30-day mortality. Peripheral vascular disease and low platelet count were predictors of 1-year mortality. Patient age, history of valve replacement, need for pressor support and low hemoglobin were independent predictors of 30-day as well as 1-year mortality in multivariate analysis. End-stage renal disease, atrial fibrillation, elevated WBC count and need for CCU admission were independent predictors of 30-day mortality. Systemic infection, low platelet count and need for blood transfusion were predictors of 1-year mortality. Strongest predictor of 30-day mortality was history of valve replacement (Odds ratio 4.23) and strongest predictor of 1-year mortality was need for pressor support (Odds ratio 3.12). Conclusions: In conclusion, 30-day and 1-year mortality after device infection remains high despite successful TLE. Patient age, history of valve replacement, need for pressor support and low hemoglobin are independent predictors of both short and long-term mortality in multivariate analysis.

Artificial intelligence and spine: rise of the machines

Preoperative prognostication of short-term postoperative mortality in patients with spinal metastatic disease can improve shared decision making around end-of-life care. To (1) develop machine learning algorithms for prediction of short-term mortality and (2) deploy these models in an open access web application. The American College of Surgeons, National Surgical Quality Improvement Program was used to identify patients that underwent operative intervention for metastatic disease. Four machine learning algorithms were developed, and the algorithm with the best performance across discrimination, calibration, and overall performance was integrated into an open access web application. The 30-d mortality for the 1790 patients undergoing surgery for spinal metastatic disease was 8.49%. Preoperative factors used for prognostication were albumin, functional status, white blood cell count, hematocrit, alkaline phosphatase, spinal location (cervical, thoracic, lumbosacral), and severity of comorbid systemic disease (American Society of Anesthesiologist Class). In this population, machine learning algorithms developed to predict 30-d mortality performed well on discrimination (c-statistic), calibration (assessed by calibration slope and intercept), Brier score, and decision analysis. An open access web application was developed for the best performing model and this web application can be found here: https://sorg-apps.shinyapps.io/spinemets/. Machine learning algorithms are promising for prediction of postoperative outcomes in spinal oncology and these algorithms can be integrated into clinically useful decision tools. As the volume of data in oncology continues to grow, creation of learning systems and deployment of these systems as accessible tools may significantly enhance prognostication and management.

Patients presenting with neurological deficits and/or pain due to spinal metastasis usually require immediate or subacute surgical treatment. Nevertheless, it is unclear whether or not side effects of primary cancer location might influence postoperative complication rate. We therefore analyzed our spinal database to identify factors influencing early postoperative complications after surgery for symptomatic spinal metastases. From 2013 to 2017, 163 consecutive patients suffering from symptomatic spinal metastases were treated at our department. Early postoperative complications were defined as any postoperative event requiring additional medical or surgical treatment within 30days of spinal surgery. A multivariate regression analysis was performed to identify independent predictors for postoperative complications after surgery for spinal metastasis. Overall, 39 of 163 patients who underwent spinal surgery for spinal metastasis developed early postoperative complications throughout the treatment course (24%). Preoperative ASA score ≥ 3 (p = 0.003), preoperative C-reactive protein level > 10mg/l (p = 0.008), preoperative Karnofsky Performance Score < 60% (p = 0.03), radiation treatment within 2months of surgery (p = 0.01), presence of diabetes mellitus (p = 0.008), and preoperative complete neurological impairment (p = 0.04) were significant and independent predictors for early postoperative complications in patients with surgery for spinal metastasis. The ability to preoperatively predict postoperative complication risk is valuable to select critically ill patients at higher risk requiring special attention. Therefore, the present study identified several significant and independent risk factors for the development of early postoperative complication in patients who underwent surgery for spinal metastasis.

External validation of a predictive algorithm for in-hospital and 90-day mortality after spinal epidural abscess

Development of machine learning algorithms for prediction of mortality in spinal epidural abscess

Predictive factors of the 30-day mortality after surgery for spinal metastasis: Analysis of a nationwide database

P38. Predicting 30-day mortality after surgery for metastatic disease of the spine: the H2-FAILS score

External validation of the SORG machine learning algorithms for predicting 90-day and 1-year survival of patients with lung cancer-derived spine metastases: a recent bi-center cohort from China

Scoring systems for metastatic spine disease focus on predicting long- to medium-term mortality or a combination of perioperative morbidity and mortality. However, accurate prediction of perioperative mortality alone may be the most important factor when considering surgical intervention. We aimed to develop and evaluate a new tool, the H2-FAILS score, to predict 30-day mortality after surgery for metastatic spine disease. Using the National Surgical Quality Improvement Program database, we identified 1195 adults who underwent surgery for metastatic spine disease from 2010 to 2018. Incidence of 30-day mortality was 8.7% (n = 104). Independent predictors of 30-day mortality were used to derive the H2-FAILS score. H2-FAILS is an acronym for: Heart failure (2 points), Functional dependence, Albumin deficiency, International normalized ratio elevation, Leukocytosis, and Smoking (1 point each). Discrimination was assessed using area under the receiver operating characteristic curve (AUC). The H2-FAILS score was compared with the American Society of Anesthesiologists Physical Status Classification (ASA Class), the 5-item modified Frailty Index (mFI-5), and the New England Spinal Metastasis Score (NESMS). Internal validation was performed using bootstrapping. Alpha = 0.05. Predicted 30-day mortality was 1.8% for an H2-FAILS score of 0 and 78% for a score of 6. AUC of the H2-FAILS was 0.77 (95% confidence interval: 0.72-0.81), which was higher than the mFI-5 (AUC 0.58, p < 0.001), ASA Class (AUC 0.63, p < 0.001), and NESMS (AUC 0.70, p = 0.004). Internal validation showed an optimism-corrected AUC of 0.76. The H2-FAILS score accurately predicts 30-day mortality after surgery for spinal metastasis. Prognostic level III.

40. Surgical Site Infection in Spinal Metastasis - Risk Factor and Countermeasure

Assessing the utility of a clinical prediction score regarding 30-day morbidity and mortality following metastatic spinal surgery: the New England Spinal Metastasis Score (NESMS)

International external validation of the SORG machine learning algorithms for predicting 90-day and one-year survival of patients with spine metastases using a Taiwanese cohort

IntroductionThe aim of our study was to validate the original Charlson Comorbidity Index (1987) (CCI) and adjusted CCI (2011) as a prediction model for 30-day and 1-year mortality after hip fracture surgery. The secondary aim of this study was to verify each variable of the CCI as a factor associated with 30-day and 1-year mortality.MethodsA prospective database of two-level II trauma teaching hospitals in the Netherlands was used. The original CCI from 1987 and the adjusted CCI were calculated based on medical history. To validate the original CCI and the adjusted CCI, the CCI was plotted against the observed 30-day and 1-year mortality, and the area under the curve (AUC) was calculated.ResultsA total of 3523 patients were included in this cohort study. The mean of the original CCI in this cohort was 5.1 (SD ± 2.0) and 4.6 (SD ± 1.9) for the adjusted CCI. The AUCs of the prediction models were 0.674 and 0.696 for 30-day mortality for the original and adjusted CCIs, respectively. The AUCs for 1-year mortality were 0.705 and 0.717 for the original and adjusted CCIs, respectively.ConclusionsA higher original and adjusted CCI is associated with a higher mortality rate. The AUC was relatively low for 30-day and 1-year mortality for both the original and adjusted CCIs compared to other prediction models for hip fracture patients in our cohort. The CCI is not recommended for the prediction of 30-day and 1-year mortality in hip fracture patients.

Commentary: Development of Machine Learning Algorithms for Prediction of 30-Day Mortality After Surgery for Spinal Metastasis.