Personalizing patient risk of a life-altering event: An application of machine learning to hemiarch surgery.

OBJECTIVE: The study objective was to assess a machine learning model's ability to predict the occurrence of life-altering events in hemiarch surgery and determine contributing patient characteristics and intraoperative factors. METHODS: In total, 602 patients who underwent hemiarch replacement at a high-volume aortic center from 2009 to 2022 were included. Patients were randomly divided into training (80%) and testing (20%) sets with various eXtreme gradient boosting candidate models constructed to predict the risk of experiencing life-altering events, including stroke, mortality, or new renal replacement therapy requirement. A total of 64 input parameters from the index hospitalization were identified, including 24 demographic characteristics as well as 8 preoperative and 32 intraoperative variables. A SHapley Additive exPlanation beeswarm plot was generated to identify and interpret the impact of individual features on the predictions of the final model. RESULTS: A life-altering event was noted in 15% (90/602) of patients who underwent hemiarch replacement, including urgent/emergency cases and dissections. The final eXtreme Gradient Boosting model demonstrated a cross-validation accuracy of 88% on the testing set and was well calibrated as evidenced by a low Brier score of 0.12. The best performing model achieved an area under the receiver operating characteristic curve of 0.76 and an area under the precision recall curve of 0.55. The SHapley Additive exPlanation beeswarm plot provided insights into key features that significantly influenced model prediction. CONCLUSIONS: Machine learning demonstrated superior accuracy in predicting hemiarch patients who would experience a life-altering event. This model may help to guide patients and clinicians in stratifying risk on an individual basis, which may in turn influence clinical decision-making.

Spinal cord protection for thoracoabdominal aortic aneurysm repair: from bench to bedside.

This keynote lecture and corresponding presentation discuss the anatomy and pathophysiology surrounding spinal cord injury in aortic surgery. This article will discuss risk factors and mechanisms for spinal cord injury, including loss of direct and collateral spinal cord perfusion and ischemia-reperfusion injury. This review will examine these elements in both the laboratory and clinical setting, in addition to other neuroprotective strategies applied in clinical practice. Addressing spinal cord injury requires an integrated and considerate approach to simultaneously optimize spinal cord blood flow, promote collateralization and improve ischemic tolerance. Given the catastrophic clinical consequences for both the patient and their caregivers, continuing to investigate and examine spinal cord injury is of the utmost importance.

Comparison Study of Iliac Branch Endoprosthesis when Used on and off Label.

BACKGROUND: The aim of this study is to compare how instructions for use (IFU) affected perioperative and intermediate term outcomes for common iliac artery aneurysms (CIAA) treated with the Gore Excluder iliac branch endoprosthesis (IBE). METHODS: A retrospective analysis was performed of all patients treated at two affiliated academic centers from September 2016 to May 2020. Outcomes were compared between IFU and nonIFU IBE cases. Criteria for nonIFU included: (1) use with a nonGore aortic endoprosthesis (n = 10), (2) isolated IBE (n = 3), and (3) requiring nondedicated covered stents for additional extension into a more suitable landing zone in the ipsilateral internal iliac artery or one of its branches (n = 11). Perioperative and intermediate term data were collected for both groups. The primary end points were free from the major adverse event (MAE) at 30 days and primary effectiveness at 1 year. RESULTS: A total of 51 CIAA (39 patients) were treated with an IBE. Overall, 15 patients were treated under IFU and 24 under nonIFU. The IFU group mean age was older (72 vs. 67 years, P = 0.03), and males (97%) were primarily treated. Comorbidities were similar except nonIFU had more patients with previous endovascular abdominal aortic aneurysm repair on presentation (0 vs. 4 cases, P = 0.04). Procedure (178 vs. 264 min, P = 0.02) and fluoroscopy (52 vs. 74 min, P = 0.04) times were longer in the nonIFU group. Technical success was 100% for both groups, and there was no difference in device related reintervention at 30 days (0 vs. 1, P = 0.44). There was no MAE in either group at 30 days. Intervention for any endoleak was similar between the groups (2 vs. 3, P = 0.94). Percent CIAA sac regression was similar between the groups (19% vs. 18%, P = 0.21). There was no difference for primary effectiveness at 1 year (93% vs. 92%, P = 0.85). There was one death per group at one year not related to an aortic or iliac cause. CONCLUSIONS: In properly selected patients with complex anatomy, IBE can be used with nondedicated aortic and internal iliac components with good early term outcomes.

Survival following lung transplantation: A population-based nested case-control study.

BACKGROUND: Lung transplantation is the mainstay of treatment for patients with end-stage respiratory failure. This study sought to evaluate survival following transplantation compared to the general population and quantify standardized mortality ratios (SMRs) using a nested case-control study design. METHODS: Control subjects were nonhospitalized inhabitants of the United States identified through the National Longitudinal Mortality Study. Case subjects were adults who underwent lung transplantation between 1990 and 2007 and identified through the Organ Procurement and Transplantation Network. Propensity-matching (5:1, nearest neighbor, caliper = 0.1) was utilized to identify suitable control subjects based on age, sex, race, and location of residency. The primary study endpoint was 10-year survival. RESULTS: About 14,977 lung transplant recipients were matched to 74,885 nonhospitalized US residents. The 10-year survival rate of lung transplant recipients was 28% (95% confidence interval [CI] = 27%-29%). The population expected mortality rate was 19 deaths/100 person-years while the observed ratio was 104 deaths/100 person-years (SMR = 5.39, 95% CI = 5.35-5.43). The largest discrepancies between observed and expected mortality rates were in females (SMR = 5.97), Hispanic (SMR = 10.70), and single lung recipients (SMR = 5.92). SMRs declined over time (1990-1995 = 5.79, 1996-2000 = 5.64, and 2001-2007 = 5.10). Standardized mortality peaks in the first year after transplant and decreases steadily over time. CONCLUSIONS: Lung transplant recipients experience a fivefold higher SMR compared to the nonhospitalized population. Long-term mortality rates have experienced consistent decline over time.

Appraisal of Donation After Circulatory Death: How Far Could We Expand the Heart Donor Pool?

BACKGROUND: the incidence of organ donation after circulatory death (DCD) is increasing; however, heart use has lagged behind other solid organs. Ex vivo perfusion devices are under United States Food and Drug Administration review for use in DCD heart recovery. This study sought to measure the potential increase in the donor pool if DCD heart donation becomes widely adopted. METHODS: DCD donor data were obtained from Organ Procurement and Transplantation Network database. Selection criteria included donor age 18 to 49 years, donors meeting Maastricht III criteria, warm ischemia time ≤30 minutes, and donation between 2015 and 2020. Exclusion criteria were coronary disease, prior myocardial infarction, ejection fraction <0.50, significant valve disease, bacteremia, pulmonary capillary wedge pressure >15 mm Hg, and history of HIV/hepatitis C virus infections. RESULTS: There were 12 813 DCD donors during this period, of which 3528 met study criteria, and 70 hearts (2%) were transplanted. The use of DCD hearts would represent an additional 48 heart transplants per month, which corresponds to a 21% (3458 of 16 521) increase across the country. Median warm ischemia was 23 minutes, with no difference between hearts that were or were not transplanted (23 vs 22.5 minutes, P = .97). The frequency with which other organs were successfully transplanted was kidney, 92%; liver, 44%; lung, 7%; intestine, 0%; and pancreas, 2%. CONCLUSIONS: Wide adoption of DCD heart transplantation could yield a substantial increase in the donor pool size, with approximately 580 additional organs being available each year across the United States. This would represent the largest increase in the donor pool in the modern era of heart transplantation.

Measuring the effect of the COVID-19 pandemic on solid organ transplantation.

BACKGROUND: The COVID-19 pandemic has uniquely affected the United States. We hypothesize that transplantation would be uniquely affected. METHODS: In this population-based cohort study, adult transplantation data were examined as time series data. Autoregressive-integrated-moving-average models of transplantation rates were developed using data from 1990 to 2019 to forecast the 2020 expected rates in a theoretical scenario if the pandemic did not occur to generate observed-to-expected (O/E) ratios. RESULTS: 32,594 transplants were expected in 2020, and only 30,566 occurred (O/E 0.94, CI 0.88-0.99). 58,152 waitlist registrations were expected and 50,241 occurred (O/E 0.86, CI 0.80-0.94). O/E ratios of transplants were kidney 0.92 (0.86-0.98), liver 0.96 (0.89-1.04), heart 1.05 (0.91-1.23), and lung 0.92 (0.82-1.04). O/E ratios of registrations were kidney 0.84 (0.77-0.93), liver 0.95 (0.86-1.06), heart 0.99 (0.85-1.18), and lung 0.80 (0.70-0.94). CONCLUSIONS: The COVID-19 pandemic was associated with a significant deficit in transplantation. The impact was strongest in kidney transplantation and waitlist registration.