Long-term Survival Model Research Articles

Sampling with censored data: a practical guide

ABSTRACT In this review, we present a simple guide for researchers to obtain pseudo-random samples with censored data. We focus our attention on the most common types of censored data, such as type I, type II, and random censoring. We discussed the necessary steps to sample pseudo-random values from long-term survival models where an additional cure fraction is informed. For illustrative purposes, these techniques are applied in the Weibull distribution. The algorithms and codes in R are presented, enabling the reproducibility of our study. Finally, we developed an R package that encapsulates these methodologies, providing researchers with practical tools for implementation.

Predictive Models for Long Term Survival of AML Patients Treated with Venetoclax and Azacitidine or 7+3 Based on Post Treatment Events and Responses: Retrospective Cohort Study.

The treatment of acute myeloid leukemia (AML) in older or unfit patients typically involves a regimen of venetoclax plus azacitidine (ven/aza). Toxicity and treatment responses are highly variable following treatment initiation and clinical decision-making continually evolves in response to these as treatment progresses. To improve clinical decision support (CDS) following treatment initiation, predictive models based on evolving and dynamic toxicities, disease responses, and other features should be developed. This study aims to generate machine learning (ML)-based predictive models that incorporate individual predictors of overall survival (OS) for patients with AML, based on clinical events occurring after the initiation of ven/aza or 7+3 regimen. Data from 221 patients with AML, who received either the ven/aza (n=101 patients) or 7+3 regimen (n=120 patients) as their initial induction therapy, were retrospectively analyzed. We performed stratified univariate and multivariate analyses to quantify the association between toxicities, hospital events, and short-term disease responses and OS for the 7+3 and ven/aza subgroups separately. We compared the estimates of confounders to assess potential effect modifications by treatment. 17 ML-based predictive models were developed. The optimal predictive models were selected based on their predictability and discriminability using cross-validation. Uncertainty in the estimation was assessed through bootstrapping. The cumulative incidence of posttreatment toxicities varies between the ven/aza and 7+3 regimen. A variety of laboratory features and clinical events during the first 30 days were differentially associated with OS for the two treatments. An initial transfer to intensive care unit (ICU) worsened OS for 7+3 patients (aHR 1.18, 95% CI 1.10-1.28), while ICU readmission adversely affected OS for those on ven/aza (aHR 1.24, 95% CI 1.12-1.37). At the initial follow-up, achieving a morphologic leukemia free state (MLFS) did not affect OS for ven/aza (aHR 0.99, 95% CI 0.94-1.05), but worsened OS following 7+3 (aHR 1.16, 95% CI 1.01-1.31) compared to that of complete remission (CR). Having blasts over 5% at the initial follow-up negatively impacted OS for both 7+3 (P<.001) and ven/aza (P<.001) treated patients. A best response of CR and CR with incomplete recovery (CRi) was superior to MLFS and refractory disease after ven/aza (P<.001), whereas for 7+3, CR was superior to CRi, MLFS, and refractory disease (P<.001), indicating unequal outcomes. Treatment-specific predictive models, trained on 120 7+3 and 101 ven/aza patients using over 114 features, achieved survival AUCs over 0.70. Our findings indicate that toxicities, clinical events, and responses evolve differently in patients receiving ven/aza compared with that of 7+3 regimen. ML-based predictive models were shown to be a feasible strategy for CDS in both forms of AML treatment. If validated with larger and more diverse data sets, these findings could offer valuable insights for developing AML-CDS tools that leverage posttreatment clinical data.

A New Mixture Model With Cure Rate Applied to Breast Cancer Data.

We introduce a new modelling for long-term survival models, assuming that the number of competing causes follows a mixture of Poisson and the Birnbaum-Saunders distribution. In this context, we present some statistical properties of our model and demonstrate that the promotion time model emerges as a limiting case. We delve into detailed discussions of specific models within this class. Notably, we examine the expected number of competing causes, which depends on covariates. This allows for direct modeling of the cure rate as a function of covariates. We present an Expectation-Maximization (EM) algorithm for parameter estimation, to discuss the estimation via maximum likelihood (ML) and provide insights into parameter inference for this model. Additionally, we outline sufficient conditions for ensuring the consistency and asymptotic normal distribution of ML estimators. To evaluate the performance of our estimation method, we conduct a Monte Carlo simulation to provide asymptotic properties and a power study of LR test by contrasting our methodology against the promotion time model. To demonstrate the practical applicability of our model, we apply it to a real medical dataset from a population-based study of incidence of breast cancer in São Paulo, Brazil. Our results illustrate that the proposed model can outperform traditional approaches in terms of model fitting, highlighting its potential utility in real-world scenarios.

A personalized prognostic model for long-term survival in patients with intrahepatic cholangiocarcinoma: a retrospective cohort study.

This study aimed to determine the optimal cutoff points for age and tumor size of patients with intrahepatic cholangiocarcinoma (ICC) and to establish and verify a predictive nomogram of overall survival at 1, 3, and 5 years. From the SEER (Surveillance, Epidemiology, and End Results) database, 1,325 ICC patients were selected and randomly divided into training and testing cohorts at a 7:3 ratio. Using the X-tile software, age and tumor size were classified into 3 subgroups: ≤61, 62-74, and ≥75 years and ≤35, 36-55, and ≥56 mm. Subsequently, univariate and multivariate Cox regression analyses were performed using the R software in the training cohort to determine independent risk factors, compile the prediction nomogram, and verify it with the testing cohort findings. The C-indexes of the new prediction nomograms in the training and testing cohorts were 0.738 (95% confidence interval [CI], 0.718-0.758) and 0.750 (95% CI, 0.72-0.78), respectively. Furthermore, the areas under the 1-, 3-, and 5-year receiver operating characteristic (ROC) curves based on the nomogram were 0.792, 0.853, and 0.838, respectively, higher than the ROC based on the 7th and 8th editions of the American Joint Cancer Commission (AJCC) staging system. This study established and verified a prognostic nomogram that improved the accuracy of the 1-, 3-, and 5-year survival predictions for ICC patients, compared with that based on the 7th and 8th editions of the AJCC staging system, and can help clinicians make personalized survival predictions.

Prognostic Significance of Proteomics-Discovered Circulating Inflammatory Biomarkers in Patients With Pulmonary Arterial Hypertension.

Pulmonary arterial hypertension (PAH) ultimately leads to right ventricular failure and premature death. The identification of circulating biomarkers with prognostic utility is considered a priority. As chronic inflammation is recognized as key pathogenic driver, we sought to identify inflammation-related circulating proteins that add incremental value to current risk stratification models for long-term survival in patients with PAH. Plasma levels of 384 inflammatory proteins were measured with the proximity extension assay technology in patients with PAH (n=60) and controls with normal hemodynamics (n=28). Among these, 51 analytes were significantly overexpressed in the plasma of patients with PAH compared with controls. Cox proportional hazard analyses and C-statistics were performed to assess the prognostic value and the incremental prognostic value of differentially expressed proteins. A panel of 6 proteins (CRIM1 [cysteine rich transmembrane bone morphogenetic protein regulator 1], HGF [hepatocyte growth factor], FSTL3 [follistatin-like 3], PLAUR [plasminogen activator, urokinase receptor], CLSTN2 [calsyntenin 2], SPON1 [spondin 1]) were independently associated with death/lung transplantation at the time of PAH diagnosis after adjustment for the 2015 European Society of Cardiology/European Respiratory Society guidelines, the REVEAL (Registry to Evaluate Early and Long-Term PAH Disease Management) 2.0 risk scores, and the refined 4-strata risk assessment. CRIM1, PLAUR, FSTL3, and SPON1 showed incremental prognostic value on top of the predictive models. As determined by Western blot, FSTL3 and SPON1 were significantly upregulated in the right ventricle of patients with PAH and animal models (monocrotaline-injected and pulmonary artery banding-subjected rats). In addition to revealing new actors likely involved in cardiopulmonary remodeling in PAH, our screening identified promising circulating biomarkers to improve risk prediction in PAH, which should be externally confirmed.

Non-proportional hazards model with a PVF frailty term: application with a melanoma dataset

Survival data analysis often uses the Cox proportional hazards (PH) model. This model is widely applied due to its straightforward interpretation of the hazard ratio under the assumption that the hazard rates for two subjects remain constant over time. However, in several randomized clinical trials with long-term survival data comparing two new treatments, it is frequently observed that Kaplan-Meier plots exhibit crossing survival curves. This violation of the PH assumption of the Cox PH model can not be applied to evaluate the treatment's effect on survival. This paper introduces a novel long-term survival model with non-PH that incorporates a frailty term into the hazard function. This model allows us to examine the effect of prognostic factors on survival and quantify the degree of unobservable heterogeneity. The model parameters are estimated using the maximum likelihood estimation procedure, and we evaluate the performance of the proposed models through simulation studies. Additionally, we demonstrate the applicability of our approach by fitting the models to a real skin cancer dataset.

Development of a prognostic model for long-term survival of young patients with bladder cancer: a retrospective analysis of the SEER Database

ObjectivesThis study aims to present the clinical characteristics of young patients with bladder cancer (YBCa), evaluate related risk factors and construct a nomogram based on data acquired from the Surveillance,...

Designing an Implementable Clinical Prediction Model for Near-Term Mortality and Long-Term Survival in Patients on Maintenance Hemodialysis

The life expectancy of patients treated with maintenance hemodialysis (MHD) is heterogeneous. Knowledge of life-expectancy may focus care decisions on near-term versus long-term goals. The current tools are limited and focus on near-term mortality. Here, we develop and assess potential utility for predicting near-term mortality and long-term survival on MHD. Predictive modeling study. 42,351 patients contributing 997,381 patient months over 11 years, abstracted from the electronic health record (EHR) system of midsize, nonprofit dialysis providers. Demographics, laboratory results, vital signs, and service utilization data available within dialysis EHR. For each patient month, we ascertained death within the next 6 months (ie, near-term mortality) and survival over more than 5 years during receipt of MHD or after kidney transplantation (ie, long-term survival). We used least absolute shrinkage and selection operator logistic regression and gradient-boosting machines to predict each outcome. We compared these to time-to-event models spanning both time horizons. We explored the performance of decision rules at different cut points. All models achieved an area under the receiver operator characteristic curve of≥0.80 and optimal calibration metrics in the test set. The long-term survival models had significantly better performance than the near-term mortality models. The time-to-event models performed similarly to binary models. Applying different cut points spanning from the 1st to 90th percentile of the predictions, a positive predictive value (PPV) of 54% could be achieved for near-term mortality, but with poor sensitivity of 6%. A PPV of 71% could be achieved for long-term survival with a sensitivity of 67%. The retrospective models would need to be prospectively validated before they could be appropriately used as clinical decision aids. A model built with readily available clinical variables to support easy implementation can predict clinically important life expectancy thresholds and shows promise as a clinical decision support tool for patients on MHD. Predicting long-term survival has better decision rule performance than predicting near-term mortality. Clinical prediction models (CPMs) are not widely used for patients undergoing maintenance hemodialysis (MHD). Although a variety of CPMs have been reported in the literature, many of these were not well-designed to be easily implementable. We consider the performance of an implementable CPM for both near-term mortality and long-term survival for patients undergoing MHD. Both near-term and long-term models have similar predictive performance, but the long-term models have greater clinical utility. We further consider how the differential performance of predicting over different time horizons may be used to impact clinical decision making. Although predictive modeling is not regularly used for MHD patients, such tools may help promote individualized care planning and foster shared decision making.

Association of Racial Residential Segregation With Long-Term Outcomes and Readmissions After Out-of-Hospital Cardiac Arrest Among Medicare Beneficiaries.

Background The national impact of racial residential segregation on out-of-hospital cardiac arrest outcomes after initial resuscitation remains poorly understood. We sought to characterize the association between measures of racial and economic residential segregation at the ZIP code level and long-term survival and readmissions after out-of-hospital cardiac arrest among Medicare beneficiaries. Methods and Results In this retrospective cohort study, using Medicare claims data, our primary predictor was the index of concentration at the extremes, a measure of racial and economic segregation. The primary outcomes were death up to 3 years and readmissions. We estimated hazard ratios (HRs) across all 3 types of index of concentration at the extremes measures for each outcome while adjusting for beneficiary demographics, treating hospital characteristics, and index hospital procedures. In fully adjusted models for long-term survival, we found a decreased hazard of death and risk of readmission for beneficiaries residing in the more segregated White communities and higher-income ZIP codes compared with the more segregated Black communities and lower-income ZIP codes across all 3 indices of concentration at the extremes measures (race: HR, 0.87 [95% CI, 0.81-0.93]; income: HR, 0.75 [95% CI, 0.69-0.78]; and race+income: HR, 0.77 [95% CI, 0.72-0.82]). Conclusions We found a decreased hazard of death and risk for readmission for those residing in the more segregated White communities and higher-income ZIP codes compared with the more segregated Black communities and lower-income ZIP codes when using validated measures of racial and economic segregation. Although causal pathways and mechanisms remain unclear, disparities in outcomes after out-of-hospital cardiac arrest are associated with the structural components of race and wealth and persist up to 3 years after discharge.

Multi-omics indicators of long-term survival benefits after immune checkpoint inhibitor therapy

Molecular indicators of long-term survival (LTS) in response to immune-checkpoint inhibitor (ICI) treatment have the potential to provide both mechanistic and therapeutic insights. In this study, we construct predictive models of LTS following ICI therapy based on data from 158 clinical trials involving 21,023 patients of 25 cancer types with available 1-year overall survival (OS) rates. We present evidence for the use of 1-year OS rate as a surrogate for LTS. Based on these and corresponding TCGA multi-omics data, total neoantigen, metabolism score, CD8+ Tcell, and MHC_score were identified as predictive biomarkers. These were integrated into a Gaussian process regression model that estimates "long-term survival predictive score of immunotherapy" (iLSPS). We found that iLSPS outperformed the predictive capabilities of individual biomarkers and successfully predicted LTS of patient groups with melanoma and lung cancer. Our study explores the feasibility of modeling LTS based on multi-omics indicators and machine-learning methods.

A web-based prediction model for long-term cancer-specific survival of middle-aged patients with early-stage gastric cancer: a multi-institutional retrospective study.

This study constructed and validated a prognostic model to evaluate long-term cancer-specific survival (CSS) in middle-aged patients with early gastric cancer (EGC). We extracted clinicopathological data from relevant patients between 2004 and 2015 from Surveillance, Epidemiology, and End Results (SEER) database, and randomly divided the patients into a training group (N = 688) and a validation group (N = 292). In addition, 102 Chinese patients were enrolled for external validation. Univariate and multivariate Cox regression models were used to screen for independent prognostic factors, and a nomogram was constructed to predict CSS. We used the concordance index (C-index), calibration curve, receiver operating characteristic (ROC) curve, and decision curve analysis (DCA) to evaluate the predictive performance of the model. Univariate and multivariate COX regression analyses showed that tumor location, differentiation grade, N stage, chemotherapy, and number of regional nodes examined were independent risk factors for prognosis, and these factors were used to construct the nomogram. The C-index of the model in the training cohort, internal validation cohort, and external validation cohort was 0.749 (95% CI 0.699-0.798), 0.744 (95% CI 0.671-0.818), and 0.807 (95% CI 0.721-0.893), respectively. The calibration curve showed that the model had an excellent fit. The DCA curve showed that the model had good predictive performance and practical clinical value. This study developed and validated a new nomogram to predict CSS in middle-aged patients with EGC. The prediction model has unique and practical value and can help doctors carry out individualized treatment and judge prognosis.

Reoperative aortic root replacement for prosthetic aortic valve endocarditis: impact of aortic graft.

Existing aortic graft complicates the surgical management of prosthetic valve endocarditis (PVE); yet, its impact has not been well studied. We compared outcomes of patients with prior aortic valve replacement (AVR) versus aortic surgery plus AVR, who underwent reoperative aortic root replacement (ARR) for PVE of the aortic valve. All patients who underwent reoperative ARR for PVE between 2004 and 2021 from 2 aortic centres were included. Two groups were formed based on the presence/absence of aortic graft: prior aortic surgery (AO) and prior AVR (AV) alone. Inverse propensity treatment weighting matched the groups. The Kaplan-Meier method was used to analyse long-term survival, and Fine and Gray model was used to compare the cumulative incidence of reoperation. A total of 130 patients were included (AO n = 59; AV n = 71). After matching, AO patients had increased stroke incidence (12.4% vs 0.9%) and renal failure requiring dialysis (11.5% vs 2.5%). In-hospital mortality was comparable (21.5% AO and 18.6% AV). Survival over 5 years was 68.9% (56.6-83.8%) in AO and 62.7% (48.1-81.7%) in AV (P = 0.70). The cumulative incidence of reoperation was similar [AO 6.3% (0.0-13.2%) vs AV 6.1% (0.0-15.1%), P = 0.69]. Reoperative ARRs for prosthetic valve/graft endocarditis are high-risk procedures. AO patients had higher incidence of postoperative morbidity versus AV patients. For all patients surviving operative intervention, survival and reoperation rates over 5 years were comparable between groups.

Predicting Long-term Disease-free Survival After Resection of Pancreatic Ductal Adenocarcinoma: A Nationwide Cohort Study.

To develop a prediction model for long-term (≥5years) disease-free survival (DFS) after the resection of pancreatic ductal adenocarcinoma (PDAC). Despite high recurrence rates, ~10% of patients have long-term DFS after PDAC resection. A model to predict long-term DFS may aid individualized prognostication and shared decision-making. This nationwide cohort study included all consecutive patients who underwent PDAC resection in the Netherlands (2014-2016). The best-performing prognostic model was selected by Cox-proportional hazard analysis and Akaike's Information Criterion, presented by hazard ratios (HRs) with 95% confidence intervals (CIs). Internal validation was performed, and discrimination and calibration indices were assessed. In all, 836 patients with a median follow-up of 67 months (interquartile range 51-79) were analyzed. Long-term DFS was seen in 118 patients (14%). Factors predictive of long-term DFS were low preoperative carbohydrate antigen 19-9 (logarithmic; HR 1.21; 95% CI 1.10-1.32), no vascular resection (HR 1.33; 95% CI 1.12-1.58), T1 or T2 tumor stage (HR 1.52; 95% CI 1.14-2.04, and HR 1.17; 95% CI 0.98-1.39, respectively), well/moderate tumor differentiation (HR 1.44; 95% CI 1.22-1.68), absence of perineural and lymphovascular invasion (HR 1.42; 95% CI 1.11-1.81 and HR 1.14; 95% CI 0.96-1.36, respectively), N0 or N1 nodal status (HR 1.92; 95% CI 1.54-2.40, and HR 1.33; 95% CI 1.11-1.60, respectively), R0 resection margin status (HR 1.25; 95% CI 1.07-1.46), no major complications (HR 1.14; 95% CI 0.97-1.35) and adjuvant chemotherapy (HR 1.74; 95% CI 1.47-2.06). Moderate performance (concordance index 0.68) with adequate calibration (slope 0.99) was achieved. The developed prediction model, readily available at www.pancreascalculator.com, can be used to estimate the probability of long-term DFS after resection of pancreatic ductal adenocarcinoma.

Adaptation of local businesses to the new era during the COVID-19 and the Russia-Ukraine war: case of Çeşme

ABSTRACT Turkey’s economic impact has been severe due to the lack of international tourists and the Covid-19 pandemic. Specifically, the tourist destination Çeşme has suffered economic fluctuations due to decreased international visitors, the Russia-Ukraine war, and insufficient domestic tourists to fill the gap. The study aims to identify deficiencies that help stakeholders manage tourism development amid Covid-19 and the Russia-Ukraine war. The data were collected through semi-structured interviews (n = 34) with local business owners operating for at least 25 years. The study provides insights into the recent tourism development during the pandemic and presents a high level of resilience among local businesses. The business owners perceive the current pandemic and the Russia-Ukraine war situation differently psychologically. Most resilient business owners look for business opportunities by adapting their business models for long-term survival despite the pandemic and the Russia-Ukraine war affecting international tourism.

Multivariable prediction model for both 90-day mortality and long-term survival for individual patients with perihilar cholangiocarcinoma: does the predicted survival justify the surgical risk?

The risk of death after surgery for perihilar cholangiocarcinoma is high; nearly one in every five patients dies within 90 days after surgery. When the oncological benefit is limited, a high-risk resection may not be justified. This retrospective cohort study aimed to create two preoperative prognostic models to predict 90-day mortality and overall survival (OS) after major liver resection for perihilar cholangiocarcinoma. Separate models were built with factors known before surgery using multivariable regression analysis for 90-day mortality and OS. Patients were categorized in three groups: favourable profile for surgical resection (90-day mortality rate below 10 per cent and predicted OS more than 3 years), unfavourable profile (90-day mortality rate above 25 per cent and/or predicted OS below 1.5 years), and an intermediate group. A total of 1673 patients were included. Independent risk factors for both 90-day mortality and OS included ASA grade III-IV, large tumour diameter, and right-sided hepatectomy. Additional risk factors for 90-day mortality were advanced age and preoperative cholangitis; those for long-term OS were high BMI, preoperative jaundice, Bismuth IV, and hepatic artery involvement. In total, 294 patients (17.6 per cent) had a favourable risk profile for surgery (90-day mortality rate 5.8 per cent and median OS 42 months), 271 patients (16.2 per cent) an unfavourable risk profile (90-day mortality rate 26.8 per cent and median OS 16 months), and 1108 patients (66.2 per cent) an intermediate risk profile (90-day mortality rate 12.5 per cent and median OS 27 months). Preoperative risk models for 90-day mortality and OS can help identify patients with resectable perihilar cholangiocarcinoma who are unlikely to benefit from surgical resection. Tailored shared decision-making is particularly essential for the large intermediate group.

A nomogram predictive model for long-term survival in spontaneous intracerebral hemorrhage patients without cerebral herniation at admission

Stratification of spontaneous intracerebral hemorrhage (sICH) patients without cerebral herniation at admission, to determine the subgroups may be suffered from poor outcomes or benefit from surgery, is important for following treatment decision. The aim of this study was to establish and verify a de novo nomogram predictive model for long-term survival in sICH patients without cerebral herniation at admission. This study recruited sICH patients from our prospectively maintained ICH patient database (RIS-MIS-ICH, ClinicalTrials.gov Identifier: NCT03862729) between January 2015 and October 2019. All eligible patients were randomly classified into a training cohort and a validation cohort according to the ratio of 7:3. The baseline variables and long-term survival outcomes were collected. And the long-term survival information of all the enrolled sICH patients, including the occurrence of death and overall survival. Follow-up time was defined as the time from the onset to death of the patient or the last clinical visit. The nomogram predictive model was established based on the independent risk factors at admission for long-term survival after hemorrhage. The concordance index (C-index) and ROC curve were used to evaluate the accuracy of the predictive model. Discrimination and calibration were used to validate the nomogram in both the training cohort and the validation cohort. A total of 692 eligible sICH patients were enrolled. During the average follow-up time of 41.77 ± 0.85 months, a total of 178 (25.7%) patients died. The Cox Proportional Hazard Models showed that age (HR 1.055, 95% CI 1.038–1.071, P < 0.001), Glasgow Coma Scale (GCS) at admission (HR 2.496, 95% CI 2.014–3.093, P < 0.001) and hydrocephalus caused by intraventricular hemorrhage (IVH) (HR 1.955, 95% CI 1.362–2.806, P < 0.001) were independent risk factors. The C index of the admission model was 0.76 and 0.78 in the training cohort and validation cohort, respectively. In the ROC analysis, the AUC was 0.80 (95% CI 0.75–0.85) in the training cohort and was 0.80 (95% CI 0.72–0.88) in the validation cohort. SICH patients with admission nomogram scores greater than 87.75 were at high risk of short survival time. For sICH patients without cerebral herniation at admission, our de novo nomogram model based on age, GCS and hydrocephalus on CT may be useful to stratify the long-term survival outcomes and provide suggestions for treatment decision-making.

Prognostic Factors for the Long-Term Survival after Hematopoietic Stem Cell Transplantation in Patients with Hodgkin Lymphoma.

This study investigated the possible prognostic factors for the long-term survival (Cure Rate) of Hodgkin Lymphoma patients who underwent HSCT. This retrospective cohort study analyzed 116 Patients diagnosed with Hodgkin Lymphoma who received autologous hematopoietic stem cell transplantation (Auto-HSCT) between the years 2007 and 2014 and followed up until 2017. The information regarding patients' survival had been collected using phone calls, and their pre-transplant information was available in the archived documents. Prognostic effects were investigated using long-term survival models. Patients with obesity had five times higher odds of long-term survival (cure) than the others (P=0.06). Also, the recurrence experience after HSCT negatively impacted the curing potential by 78% (P=0.05). Also, with 32 years as the change point, patients younger than 32 had 76% fewer odds of surviving long-term (P=0.03), and Poor transfused stem cell dose of CD34+ (<0.16 × 106 cells/ml) reduced the odds of long-term survival by 92% (P=0.01). According to the statistical models used in this study, obesity can increase the curing potential of Hodgkin lymphoma after transplantation. Meanwhile, aging, poor transfused CD34+ cells, and recurrence after HSCT were associated with lower survival following HSCT.

The AUGIS Survival Predictor: Prediction of Long-Term and Conditional Survival After Esophagectomy Using Random Survival Forests.

The aim of this study was to develop a predictive model for overall survival after esophagectomy using pre/postoperative clinical data and machine learning. For patients with esophageal cancer, accurately predicting long-term survival after esophagectomy is challenging. This study investigated survival prediction after esophagectomy using a RandomSurvival Forest (RSF) model derived from routine data from a large, well-curated, national dataset. Patients diagnosed with esophageal adenocarcinoma or squamous cell carcinoma between 2012 and 2018 in England and Wales who underwent an esophagectomy were included. Prediction models for overall survival were developed using the RSF method and Cox regression from 41 patient and disease characteristics. Calibration and discrimination (time-dependent area under the curve) were validated internally using bootstrap resampling. The study analyzed 6399 patients, with 2625 deaths during follow-up. Median follow-up was 41 months. Overall survival was 47.1% at 5 years. The final RSF model included 14 variables and had excellent discrimination with a 5-year time-dependent area under the receiver operator curve of 83.9% [95% confidence interval (CI) 82.6%-84.9%], compared to 82.3% (95% CI 81.1%-83.3%) for the Cox model. The most important variables were lymph node involvement, pT stage, circumferential resection margin involvement (tumor at < 1 mm from cut edge) and age. There was a wide range of survival estimates even within TNM staging groups, with quintiles of prediction within Stage 3b ranging from 12.2% to 44.7% survival at 5 years. An RSF model for long-term survival after esophagectomy exhibited excellent discrimination and well-calibrated predictions. At a patient level, it provides more accuracy than TNM staging alone and could help in the delivery of tailored treatment and follow-up.

A novel intrapericardial pulsatile device for individualized, biventricular circulatory support without direct blood contact

ObjectiveDue to severely limited donor heart availability, durable mechanical circulatory support remains the only treatment option for many patients with end-stage heart failure. However, treatment complexity persists due to its univentricular support modality and continuous contact with blood. We investigated the function and safety of reBEAT (AdjuCor GmbH), a novel, minimal invasive mechanical circulatory support device that completely avoids blood contact and provides pulsatile, biventricular support. MethodsFor each animal tested, an accurately sized cardiac implant was manufactured from computed tomography scan analyses. The implant consists of a cardiac sleeve with three inflatable cushions, 6 epicardial electrodes and driveline connecting to an electro-pneumatic, extracorporeal portable driver. Continuous epicardial electrocardiogram signal analysis allows for systolic and diastolic synchronization of biventricular mechanical support. In 7 pigs (weight, 50-80 kg), data were analyzed acutely (under beta-blockade, n = 5) and in a 30-day long-term survival model (n = 2). Acquisition of intracardiac pressures and aortic and pulmonary flow data were used to determine left ventricle and right ventricle stroke work and stroke volume, respectively. ResultsEach implant was successfully positioned around the ventricles. Automatic algorithm electrocardiogram signal annotations resulted in precise, real-time mechanical support synchronization with each cardiac cycle. Consequently, progressive improvements in cardiac hemodynamic parameters in acute animals were achieved. Long-term survival demonstrated safe device integration, and clear and stable electrocardiogram signal detection over time. ConclusionsThe present study demonstrates biventricular cardiac support with reBEAT. Various demonstrated features are essential for realistic translation into the clinical setting, including safe implantation, anatomical fit, safe device–tissue integration, and real-time electrocardiogram synchronized mechanical support, result in effective device function and long-term safety.

321 RANDOM SURVIVAL FOREST FOR PERSONALISED PROGNOSTICATION AFTER ESOPHAGECTOMY

Abstract For patients with esophageal cancer, producing accurate prediction models for long-term survival after esophagectomy has proved challenging. We investigated whether Random Survival Forests (RSF), a machine learning method, could produce an accurate prognostic model for overall survival after esophagectomy. Methods The study used data from the 'National Oesophago-Gastric Cancer Audit' (NOGCA) and included patients diagnosed with esophageal adenocarcinoma or squamous cell carcinoma between 2012 and 2018 in England and Wales and who underwent a curative esophagectomy with adequate lymphadenectomy (15 lymph nodes) and survived to discharge (n = 6838). Missing data was handled using multiple imputation. 15 variables were selected for inclusion using Random Forest variable importance and used to train the final model. The same variables with non-linearity transformations were used to develop a traditional Cox regression model for comparison. Results Median survival was 50 months. The final RSF model had good discrimination on internal validation with a C-index of 0.7627 (0.7625–0.7629), exceeding the cox model 0.7539 (0.7541–0.7537). At 3 years post-surgery, overall survival was 56.2%. The RSF yielded a mean predicted survival of 59.3% (IQR 33.3–87.1%) with good calibration (Figure 1) compared to 57.4% (38.4%–79.8%) for the cox model. The most influential variables were lymph node involvement and pT/ypT stage, however other variables including neoadjuvant treatment completion and surgical complications were also important. Decision curve analysis was undertaken which also showed an increased net benefit with the RSF model. Conclusion A Random Forest survival model provided better performance in predicting survival after curative esophagectomy. This will allow more personalised predictions to be delivered clinicians and patients. An online web app is provided at https://uoscancer.shinyapps.io/NOGCA/