Nonparametric Survival Research Articles

Survival and longevity among male Australian elite athletes: A retrospective cohort study.

To quantify and compare survival and longevity in three populations of male Australian elite athletes, relative to sex- and age-matched referents from the general population. Retrospective cohort study. This study comprised a census of three populations of male Australian elite athletes (i.e. N = 10,502 Australian Football League players, N = 803 national team rugby union players, and N = 1527 Olympic athletes) who debuted from 1-JAN-1921 to 31-DEC-2023. Nonparametric relative survival analysis was conducted using the Kaplan-Meier estimator to generate survival curves, the Pohar Perme method and a log-rank type test to estimate and compare net survival, and the life years difference measure to estimate longevity gains or losses. Olympic athletes had a continuous gain in net survival post-debut, with an average life years difference of 3.40 (95 % confidence interval 2.30 to 4.49) years at the maximum follow-up time. Australian football and rugby union players initially had modest gains in net survival post-debut, followed by a gradual decline in net survival, with average life years differences of 0.31 (95 % confidence interval -0.03 to 0.65) and -0.67 (95 % confidence interval -1.87 to 0.53) years at the maximum follow-up time, respectively. Net survival was significantly different across the three cohorts (test statistic 21.8; degrees of freedom = 2; p < 0.001). Expected survival benefits were offset in elite male Australian football and rugby union players. Further research examining cause-specific mortality is warranted to elucidate the underlying reasons for the observed lack of expected survival benefit in Australian football and rugby union players.

Multiarterial Grafting in Redo CABG: Type of Arterial Conduit and Patient Sex Determine Benefit

To evaluate whether multiarterial grafting provides incremental benefit above single arterial grafting in isolated redo CABG. From 1/1980 to 7/2020, 6559 adults underwent 6693 isolated CABG reoperations. Patients undergoing multiarterial grafting were propensity-score matched with those undergoing single arterial grafting, with or without additional vein grafts, yielding 2005 well-matched pairs. Endpoints were in-hospital postoperative complications, hospital mortality, and long-term mortality. Median follow-up was 10 years with 25% followed >17 years. Multivariable multiphase hazard modeling and nonparametric random survival forests for survival were used to identify patients for whom multiarterial grafting was most beneficial. Among propensity-matched patients, postoperative complications for multiarterial versus single arterial grafting were any reoperation, 50 (2.5%) versus 65 (3.2%); renal failure, 73 (3.6%) versus 55 (2.7%); stroke, 44 (2.2%) versus 38 (1.9%); and deep sternal infection, 36 (1.8%) versus 25 (1.2%). Hospital mortality was 1.7% (n=35) versus 2.8% (n=56) (P=.03). Comparing multiarterial to single arterial grafting, survival at 1 and 3 years was 95% versus 94% and 92% versus 88%, and at 5, 15, and 20 years, survival was 87%, 49%, and 31% versus 82%, 42%, and 25%. Better survival after multiarterial grafting was confined to males with 2 patent internal thoracic artery grafts (P<.0001). Redo CABG with multiarterial grafting can be performed with lower in-hospital mortality and similar major morbidity to single arterial grafting. It is associated with better long-term survival, particularly in males when 2 internal thoracic artery grafts are used.

External Validation of a Vision Prognostication Tool for Vision Loss after Plaque Brachytherapy

Introduction: Vision loss remains a significant concern for patients undergoing radiation treatment for uveal melanoma. Herein, we identify predictors of vision loss after plaque brachytherapy at our academic medical center and, using this independent data set, evaluate a published vision prognostication tool. Methods: A retrospective chart review identified 252 posterior uveal melanoma patients treated with iodine-125 plaque brachytherapy from 2013 to 2018. Interval-censored nonparametric survival analysis estimated risk for moderate (&lt;20/50) and severe (&lt;20/200) vision loss at 3 and 5 years post-brachytherapy. Factors affecting vision loss were evaluated using interval-censored proportional hazards regression models. The prognostication tool was evaluated using a receiver operating curve and comparing predicted risk to observed risk. Results: Increased age, tumor height, and radiation dose to optic disc were associated with vision loss for patients with baseline VA ≥20/50. For those with baseline VA ≥20/200, risk factors included dose to optic disc, female sex, and baseline vision worse than 20/50. The vision prognostication tool demonstrated a moderate ability to differentiate patients with 3-year vision &lt;20/200, with an area under the curve of 0.75. Conclusion: The vision prognostication tool was predictive for vision loss but generally overestimated risk, possibly reflecting evolving practice patterns including anti-VEGF use for radiation retinopathy.

Neoadjuvant therapy versus upfront surgery in resectable pancreatic cancer: reconstructed patient-level meta-analysis of randomized clinical trials.

Neoadjuvant treatment has shown promising results in patients with borderline resectable pancreatic ductal adenocarcinoma. The potential benefits of neoadjuvant treatment on long-term overall survival in patients with resectable pancreatic ductal adenocarcinoma have not yet been established. The aim of this study was to compare long-term overall survival of patients with resectable pancreatic ductal adenocarcinoma based on whether they received neoadjuvant treatment or underwent upfront surgery. A systematic review including randomized clinical trials on the overall survival outcomes between neoadjuvant treatment and upfront surgery in patients with resectable pancreatic ductal adenocarcinoma was conducted up to 1 August 2023 from PubMed, MEDLINE and Web of Science databases. Patient-level survival data was extracted and reconstructed from available Kaplan-Meier curves. A frequentist one-stage meta-analysis was employed, using Cox-based models and a non-parametric method (restricted mean survival time), to assess the difference in overall survival between groups. A Bayesian meta-analysis was also conducted. Five randomized clinical trials comprising 625 patients were included. Among patients with resectable pancreatic ductal adenocarcinoma, neoadjuvant treatment was not significantly associated with a reduction in the hazard of death compared with upfront surgery (shared frailty HR 0.88, 95% c.i. 0.72 to 1.08, P = 0.223); this result was consistent in the non-parametric restricted mean survival time model (+2.41 months, 95% c.i. -1.22 to 6.04, P < 0.194), in the sensitivity analysis that excluded randomized clinical trials with a high risk of bias (shared frailty HR 0.91 (95% c.i. 0.72 to 1.15; P = 0.424)) and in the Bayesian analysis with a posterior shared frailty HR of 0.86 (95% c.i. 0.70 to 1.05). Neoadjuvant treatment does not demonstrate a survival advantage over upfront surgery for patients with resectable pancreatic ductal adenocarcinoma.

Uncovering merchants’ willingness to wait in on-demand food delivery markets

While traditional on-demand food delivery services help restaurants reach more customers and enable doorstep deliveries, they also come with drawbacks, such as high commission fees and limited control over the delivery process. White-label food delivery services have emerged as an alternative, ready-to-use platform for restaurants to arrange delivery for customer orders received through their applications or websites, without the constraints imposed by traditional on-demand food delivery platforms or the need to develop an in-house delivery operation. Although several studies have investigated consumer behavior when using traditional on-demand food delivery services, there is limited research on merchants’ behavior when adopting white-label food delivery services. In this research, we develop a non-parametric survival model to estimate merchants’ willingness to wait when using white-label food delivery services and examine how various factors, such as delivery fees, the number of placed orders, and average waiting time, affect merchants’ willingness to wait, drawing on a dataset of both delivered and canceled orders from a crowd-sourcing delivery platform in Singapore. The empirical results show that merchants’ willingness to wait has a non-linear relationship with their average waiting time; it initially increases and then decreases with average waiting time. Moreover, the relationship between merchants’ willingness to wait during the pick-up stage and their average waiting times in the matching stage follows a similar non-linear trend. Merchants who have experienced lengthy waiting times on average in the matching stage tend to be less patient in the pick-up stage. This research sheds light on the stage-specific dynamics of merchants waiting behavior in white-label delivery service and provides insights for delivery platforms to optimize their operational strategies and enhance user experiences.

Survival of Children With Critical Congenital Heart Defects in the National Birth Defects Prevention Study.

Critical congenital heart defects (CCHDs) are associated with considerable morbidity and mortality. This study estimated survival of children with nonsyndromic CCHDs and evaluated relationships between exposures of interest and survival by CCHD severity (univentricular or biventricular function). This analysis included 4380 infants with CCHDs (cases) born during 1999-2011 and enrolled in the National Birth Defects Prevention Study, a multisite, population-based case-control study of major birth defects. Cases were linked to state death files. Nonparametric Kaplan-Meier survival functions were used to estimate 1- and 5-year survival probabilities overall and by severity group (univentricular/biventricular) stratified by demographic and clinical exposure variables of interest. The log-rank test was used to determine whether stratified survival curves were equivalent. Survival and 95% confidence intervals (CIs) were also estimated using Cox proportional hazards modeling adjusted for maternal age, education, race/ethnicity, study site, and birth year. One- and five-year survival rates were 85.8% (CI 84.7-86.8) and 83.7% (CI 82.5-84.9), respectively. Univentricular 5-year survival was lower than biventricular case survival [65.3% (CI 61.7-68.5) vs. 89.0% (CI 87.8-90.1; p < 0.001)]. Clinical factors (e.g. preterm birth, low birthweight, and complex/multiple defects) were associated with lower survival in each severity group. Sociodemographic factors (non-Hispanic Black race/ethnicity, <high school education, smoking, and lower household income) were only associated with survival among biventricular cases. Mortality among children with CCHDs occurred primarily in the first year of life. Survival was lower for those with univentricular defects, and social determinants of health were most important in predicting survival for those with biventricular defects.

Nonparametric conditional survival function estimation and plug-in bandwidth selection with multiple covariates

Nonparametric conditional survival function estimation and plug-in bandwidth selection with multiple covariates

Leisure Activity, Childhood Starvation, and Risk of Incident Diabetes in Older Adults: Evidence From a Chinese Nationwide Cohort Study

ObjectiveThis study explores the impact of leisure activity and the association between childhood starvation and the risk of diabetes in older Chinese adults. DesignProspective cohort study based on the Chinese Longitudinal Healthy Longevity Study (CLHLS), a nationwide cohort study in China. Setting and participantsA total of 4637 older adults aged ≥65 years, all with documented diabetes history, experiences of childhood starvation, and participation in leisure activities were recruited. MethodsChildhood starvation exposure was assessed via self-reported responses from a structured questionnaire. The leisure activities were measured by 9 distinctive components and categorized into 3 distinct categories: productive activity, recreational activity, and sedentary activity. Diabetes status was determined by self-reported, physician-diagnosed cases during the follow-up period. Nonparametric survival models were employed for analysis. ResultsOver an average follow-up period of 4.3 years, 215 of 4637 participants (4.6%) reported a confirmed diagnosis of diabetes. Nonparametric survival models showed that those reporting childhood starvation had a higher risk of late-life diabetes [hazard ratio (HR) 1.72, 95% CI 1.21-2.44]. Engaging in productive activity (HR 0.90, 95% CI 0.83-0.99) and recreational activity (HR 0.88, 95% CI 0.77-1.00) was linked with a reduced risk of late-life diabetes. Sedentary activity did not show a significant effect. Further analysis highlighted the interactions effects of leisure activities on diabetes risk across different demographic and historical exposure subgroups. Conclusions and implicationsEngaging in productive and recreational leisure activities was inversely associated with the risk of diabetes in older adults who experienced childhood starvation. Promoting such activities could be beneficial in mitigating long-term diabetes risk related to early-life nutritional deficiencies.

Outcomes of patients with advanced solid tumors who discontinued immune-checkpoint inhibitors: a systematic review and meta-analysis

The outcome of patients with metastatic tumors who discontinued immune checkpoint inhibitors (ICIs) not for progressive disease (PD) has been poorly explored. We performed a meta-analysis of all studies reporting the clinical outcome of patients who discontinued ICIs for reasons other than PD. We searched PubMed, Embase and Scopus databases, from the inception of each database to December 2023, for clinical trials (randomized or not) and observational studies assessing PD-(L)1 and CTLA-4 inhibitors in patients with metastatic solid tumors who discontinued treatment for reasons other than PD. Each study had to provide swimmer plots or Kaplan-Meier survival curves enabling the reconstruction of individual patient-level data on progression-free survival (PFS) following the discontinuation of immunotherapy. The primary endpoint was PFS from the date of treatment discontinuation overall and according to tumor histotype, type of treatment and reason of discontinuation. The Combersure's method was used to estimate meta-analytical non-parametric summary survival curves assuming random effects at study level. Thirty-six studies (2180 patients) were included. The pooled median PFS (mPFS) was 24.7 months (95% CI, 18.8-30.6) and the PFS-rate at 12, 24, and 36 months was respectively 69.8% (95% CI, 63.1-77.3), 51.0% (95% CI, 43.4-59.8) and 34.0% (95% CI, 27.0-42.9). Univariable analysis showed that the mPFS was significantly longer for patients with melanoma (43.0 months), as compared with non-small cell lung cancer (NSCLC, 13.5 months) and renal cell carcinoma (RCC, 10.0 months; between-strata comparison test p-value < 0.001); for patients treated with anti-PD-(L)1+anti-CTLA-4 as compared with anti-PD-(L)1 monotherapy (44.6 versus 19.9 months; p-value <0.001), and in NSCLC when the reason of treatment discontinuation was elective as compared with toxicity onset (19.6 versus 4.8 months; p-value=0.003). The multivariable analysis confirmed these differences. The long-term outcome of patients who stopped ICIs for reasons other than PD was substantially affected by clinicopathological features: PFS after treatment discontinuation was longer in patients with melanoma, and/or treated with anti-PD-(L)1+anti-CTLA-4, and shorter in patients with RCC or in those patients with NSCLC who stopped treatment for toxicity onset. The Italian Ministry of University and Research (PRIN 2022Y7HHNW).

Phase I trial of talimogene laherparepvec for the treatment of peritoneal surface malignancies (TEMPO).

2668 Background: Peritoneal surface malignancies (PSM) constitute a group of difficult-to-treat metastatic diseases that respond poorly to standard chemotherapy. Talimogene Laherparepvec (T-VEC), the first FDA-approved oncolytic virus (OV), demonstrates potential in treating various malignancies by effectively initiating crucial immune responses. Yet, the utility of intraperitoneal (ip) T-VEC for unresectable PSM is still unexplored. Methods: TEMPO (NCT03663712) was a nonrandomized, open-label, multicenter phase I trial involving 18 patients with unresectable PSM originating from the appendix (n=14), ovary (n=3), or small bowel (n=1). The trial aimed to determine the recommended Phase 2 dose (RP2D) and maximum tolerated dose (MTD) of ip T-VEC, as well as to evaluate its safety and tolerability. An exploratory endpoint was the Restricted Mean Survival Time (RMST), the estimated non-parametric mean survival time up until the longest patient follow-up time (i.e. maximum of the longest follow-up, and generally an underestimate of the actual mean survival time especially in small samples). Patients received an initial dose of 4x106 plaque-forming units (pfu) of ip T-VEC, followed by their assigned dose every two weeks for up to four additional doses, using a standard ‘3+3’ dose escalation scheme. The doses tested were 4x106, 4x107, and 4x108 pfu of ip T-VEC. Safety assessments were conducted regularly. Results: The RP2D is 4x108 pfu of T-VEC. The trial reported no dose-limiting toxicities, thus the MTD was not reached. Treatment-Related Adverse Events (TRAEs) were recorded and categorized by the maximum dose received (Table).One patient treated with 4x106 pfu had grade 3 abdominal pain and another grade 3 fatigue. No severe TRAEs occurred with 4x107pfu. One patient showed a grade 3 neutrophil count decrease at 4x108 pfu. The RMST, calculated with a truncation time of 5.5 months, varied across the doses. Patients receiving 4x106 pfu had an RMST of 3.1 (95% CI= 1.550-4.733) months, while those treated with 4x107 and 4x108 pfu had RMSTs of 4.2 (95% CI= 2.492-5.884) and 4.6 (95% CI= 3.551-5.722) months, respectively. Conclusions: We conclude that ip T-VEC is safe in patients with unresectable PSM. TEMPO successfully identified 4x108 pfu as the RP2D for TVEC. Notably, the RMST increased with increasing doses of TL. These results are promising, considering the limited benefit of conventional chemotherapy in treating PSM. This study positions T-VEC as a promising candidate for combination immunotherapies for PSM that amplifies the bystander immune response induced by OVs. Clinical trial information: NCT03663712 . [Table: see text]

#2214 Incidence, risk factors and outcome of calciphylaxis in Iceland

Abstract Background and Aims Calciphylaxis is a rare and serious medical condition that primarily affects patients with end-stage kidney disease (ESKD). The disorder is characterized by calcium deposition in blood vessels and soft tissues, resulting in skin ulceration and tissue necrosis. The aim of the study was to investigate the incidence, risk factors, and outcomes of calciphylaxis in Iceland. Method This was a retrospective case-control study that included individuals diagnosed with calciphylaxis in 1990-2022. Patients were identified by searching electronic medical records (EMR) for the ICD-10 diagnosis codes E83.5 and L94.2, SNOMED histology codes T01000-T03900 in combination with M55400, M55500 or M54000-M54200, documented sodium thiosulfate (STS) therapy, and by consultation with practicing nephrologists. Data on medical history, medication use, and survival were obtained from EMR. A control group of ESKD patients was matched (1:1) based on age, sex, and duration of dialysis. Total annual incidence was calculated per 1,000,000 population of Iceland each year, and for ESKD patients per 10,000 dialysis patients at the end of each year. Poisson regression was used to estimate changes in incidence. Groups were compared using non-parametric tests and survival estimated by Kaplan-Meier method and Cox regression. Results Twenty-one cases of calciphylaxis were identified, of which 18 (86%) had ESKD. The average incidence was 2/1,000,000/year. For ESKD patients, the median (interquartile range, IQR) age was 69 (66-75) years, and 50% were women. Three patients (17%) had not yet initiated dialysis at the time of diagnosis and the median duration of dialysis was 0.12 (IQR 0.0-2.8) years. The average incidence of calciphylaxis was 75 per 10,000 dialysis patients/year, increasing by 9.0% annually (Fig. 1). No significant differences between calciphylaxis cases and controls were observed for comorbid diseases or other risk factors. The median body mass index of cases was 29.5 (IQR 27.1-33.7) compared with 27.6 (IQR 24.3-30.6) kg/m2 in controls (p=0.20), and 50% of cases and 22% of controls (p=0.16) were on warfarin treatment at diagnosis. Fifteen patients received treatment with sodium thiosulfate (STS). Three of 6 patients received STS therapy before 2016 and all 12 patients after 2016. One-year mortality of ESKD patients with calciphylaxis was 44%. The median survival was 134 days in 1990-2016, increasing to 644 days after 2016. Survival difference between cases and controls was significant (p=0.04) by log rank test, while adjusted Cox regression did not show a statistically significant association between calciphylaxis and death (HR 2.0 (95% CI, 0.67-6.1)). Conclusion The incidence of calciphylaxis in Iceland seems to be increasing and is high compared to incidence rates reported in other countries. Survival, although poor, appears to have improved in recent years. The small sample size is a limitation and further research will require collaborative efforts to generate large patient samples.

Improved nonparametric survival prediction using CoxPH, Random Survival Forest & DeepHit Neural Network

In recent times, time-to-event data such as time to failure or death is routinely collected alongside high-throughput covariates. These high-dimensional bioinformatics data often challenge classical survival models, which are either infeasible to fit or produce low prediction accuracy due to overfitting. To address this issue, the focus has shifted towards introducing a novel approaches for feature selection and survival prediction. In this article, we propose a new hybrid feature selection approach that handles high-dimensional bioinformatics datasets for improved survival prediction. This study explores the efficacy of four distinct variable selection techniques: LASSO, RSF-vs, SCAD, and CoxBoost, in the context of non-parametric biomedical survival prediction. Leveraging these methods, we conducted comprehensive variable selection processes. Subsequently, survival analysis models—specifically CoxPH, RSF, and DeepHit NN—were employed to construct predictive models based on the selected variables. Furthermore, we introduce a novel approach wherein only variables consistently selected by a majority of the aforementioned feature selection techniques are considered. This innovative strategy, referred to as the proposed method, aims to enhance the reliability and robustness of variable selection, subsequently improving the predictive performance of the survival analysis models. To evaluate the effectiveness of the proposed method, we compare the performance of the proposed approach with the existing LASSO, RSF-vs, SCAD, and CoxBoost techniques using various performance metrics including integrated brier score (IBS), concordance index (C-Index) and integrated absolute error (IAE) for numerous high-dimensional survival datasets. The real data applications reveal that the proposed method outperforms the competing methods in terms of survival prediction accuracy.

Comparison of oncological and functional outcomes in Lower-limb osteosarcoma pediatric patients: a large single-center retrospective cohort study.

Treating pediatric osteosarcoma in long bones is challenging due to skeletal immaturity, which restricts the generalizability of insights derived from adult patients. Are there disparities in outcomes? How should surgical protocols be tailored for children of varying ages? What are the specific postoperative complications? A large single-center retrospective cohort study of 345 patients under 14 years old with lower-limb osteosarcoma treated in our department since 2000 was conducted to address these inquiries. A retrospective analysis of 345 pediatric patients with lower-limb osteosarcoma admitted to our department between 2000 and 2019 was conducted. Clinical and functional outcomes were compared based on age groups, surgical methods, type of prosthesis, and primary tumor location. Patients were divided into the low-age group (≤10y old) and the high-age group (>10y old). Overall survival rate (OS), progressionfree survival rate (PFS), and prosthesis survival rate were assessed using Kaplan-Meier curves, nonparametric survival analysis (log-rank test), and Univariate cox regression were used for comparison. The incidence of complications, local relapse rate (LRR), metastasis rate, final limb-salvage, and amputation rate, and Musculoskeletal Tumor Society (MSTS) score of different independent groups were further evaluated using χ2 test or Fisher's exact test, and t -test was employed to evaluate the measurement data. The average age of the patients was 11.10±2.32 years ranging from 4 to 14y, with an average follow-up duration of 48.17 months. The 5, 10, and 15-year OS rates were 50.3%, 43.8%, and 37.9%, respectively. The progression-free survival rate was 44.8% at 5 years and 41.1% at 10 years. The final limb salvage rate was 61.45%, while the final amputation rate was 38.55%. The low-age group had a higher amputation rate compared with the high-age group (48.00% vs. 33.18%, P =0.009). The overall LRR was 9.28%, and the incidence of metastasis was 28.99%. The LRR of the limb-salvage group was higher than the amputation group ( P =0.004). The low-age group experienced more prosthesis-related complications than the high-age group ( P =0.001). The most common prosthesis-related complication in the low-age group was soft-tissue failure, while the periprosthetic infection was most frequent in the high-age group. The high-age group had a higher cumulative prosthesis survival compared with the low-age group ( P =0.0097). Modular prosthesis showed better MSTS scores and higher cumulative prosthetic survival than expandable prosthesis in pediatric patients ( P <0.05). Limb preservation in pediatric patients becomes increasingly efficacious with advancing age, while consideration of amputation is warranted for younger patients. The prevailing postoperative complications associated with prosthesis encompass soft tissue failure and periprosthetic infection. Younger patients diagnosed with lower limb osteosarcoma exhibit a heightened amputation rate and a greater incidence of prosthesis-related complications.

Predicting Long-Term Childhood Survival of Newborns with Congenital Heart Defects: A Population-Based, Prospective Cohort Study (EPICARD).

Backgroud: Congenital heart defects (CHDs) are the most frequent group of major congenital anomalies, accounting for almost 1% of all births. They comprise a very heterogeneous group of birth defects in terms of their severity, clinical management, epidemiology, and embryologic origins. Taking this heterogeneity into account is an important imperative to provide reliable prognostic information to patients and their caregivers, as well as to compare results between centers or to assess alternative diagnostic and treatment strategies. The Anatomic and Clinical Classification of CHD (ACC-CHD) aims to facilitate both the CHD coding process and data analysis in clinical and epidemiological studies. The objectives of the study were to (1) Describe the long-term childhood survival of newborns with CHD, and (2) Develop and validate predictive models of infant mortality based on the ACC-CHD. Methods: This study wasbased on data from a population-based, prospective cohort study: Epidemiological Study of Children with Congenital Heart Defects (EPICARD). The final study population comprised 1881 newborns with CHDs after excluding cases that were associated with chromosomal and other anomalies. Statistical analysis included non-parametric survival analysis and flexible parametric survival models. The predictive performance of models was assessed by Harrell's C index and the Royston-Sauerbrei RD2, with internal validation by bootstrap. Results: The overall 8-year survival rate for newborns with isolated CHDs was 0.96 [0.93-0.95]. There was a substantial difference between the survival rate of the categories of ACC-CHD. The highest and lowest 8-year survival rates were 0.995 [0.989-0.997] and 0.34 [0.21-0.50] for "interatrial communication abnormalities and ventricular septal defects" and "functionally univentricular heart", respectively. Model discrimination, as measured by Harrell's C, was 87% and 89% for the model with ACC-CHD alone and the full model, which included other known predictors of infant mortality, respectively. The predictive performance, as measured by RD2, was 45% and 50% for the ACC-CHD alone and the full model. These measures were essentially the same after internal validation by bootstrap. Conclusions: The ACC-CHD classification provided the basis of a highly discriminant survival model with good predictive ability for the 8-year survival of newborns with CHDs. Prediction of individual outcomes remains an important clinical and statistical challenge.

Risk factors affecting patients survival with colorectal cancer in Morocco: survival analysis using an interpretable machine learning approach

The aim of our study was to assess the overall survival rates for colorectal cancer at 3 years and to identify associated strong prognostic factors among patients in Morocco through an interpretable machine learning approach. This approach is based on a fully non-parametric survival random forest (RSF), incorporating variable importance and partial dependence effects. The data was povided from a retrospective study of 343 patients diagnosed and followed at Hassan II University Hospital. Covariate selection was performed using the variable importance based on permutation and partial dependence plots were displayed to explore in depth the relationship between the estimated partial effect of a given predictor and survival rates. The predictive performance was measured by two metrics, the Concordance Index (C-index) and the Brier Score (BS). Overall survival rates at 1, 2 and 3 years were, respectively, 87% (SE = 0.02; CI-95% 0.84–0.91), 77% (SE = 0.02; CI-95% 0.73–0.82) and 60% (SE = 0.03; CI-95% 0.54–0.66). In the Cox model after adjustment for all covariates, sex, tumor differentiation had no significant effect on prognosis, but rather tumor site had a significant effect. The variable importance obtained from RSF strengthens that surgery, stage, insurance, residency, and age were the most important prognostic factors. The discriminative capacity of the Cox PH and RSF was, respectively, 0.771 and 0.798 for the C-index while the accuracy of the Cox PH and RSF was, respectively, 0.257 and 0.207 for the BS. This shows that RSF had both better discriminative capacity and predictive accuracy. Our results show that patients who are older than 70, living in rural areas, without health insurance, at a distant stage and who have not had surgery constitute a subgroup of patients with poor prognosis.

OP34 Risk of Disease Recurrence and Re-resections in Crohn's Disease Patients Undergoing Primary Bowel Resection: A Population-Based Study

Abstract Background The rate of having a resection for patients with Crohn's disease (CD) have decreased over decades, while the rates of re-resections seem to have been stable around 1/3 in historic cohorts. Re-resection rates might be influenced by targeted therapies. We aimed to investigate the re-resection rates and risk of recurrence in a contemporary cohort and the effect of medical treatment on these parameters. Methods This population-based cohort included all CD patients undergoing primary intestinal resection between 2010 and 2020 in Eastern Denmark, with a background population of 2,730,000 (46% of the Danish population). Individual clinical characteristics, medication, surgical procedures, and complications, as well as imaging, and endoscopy results were collected. Disease recurrence was defined as a colonoscopy with SES-CD≥3, Rutgeerts score ≥2i, inflammation or stenosis on imaging (MRI, CTA, or IUS), fecal calprotectin ≥ 250 mg/kg, starting corticosteroids after resection or re-resection due to disease activity. We characterized the cohort using nonparametric statistics (median, IQR, percentages) and survival analysis. A Cox regression analysis with propensity score incorporating Montreal classification, age, gender, smoking, and types of resection (colon, ileocecal, or small bowel) was conducted to assess the effect of initiating prophylactic biologic treatment within the first year from resection. Results A total of 631 patients had a primary resection due to disease activity, with a median follow-up from time from diagnosis of 118 months (IQR: 69-170) (Table 1). Prior to the first resection 337 (53%) patients received immunomodulators and 249 (39%) biologics; the same numbers post-surgery were 314 (50%) and 264 (42%), respectively. A total of 256 (41%) patients were resected within two years from diagnosis while this proportion increased to 424 (67%) and 533 (84 %) after 5 and 10 years, respectively. Re-resection rates due to disease activity 5 and 10 years after primary resection were 5% and 10%. The median time from primary resection to disease recurrence was 11,3 months (IQR: 4.7-24.8) (Figure 1). Median time from first to second and second to third resection were 37 months (IQR: 15-64) and 38 months (IQR: 26-55), respectively. We found no significant difference regarding prophylactic biologic therapy within the first year of resection (n 45) based on recurrence or re-resection compared to patients not starting biologics within the first year (HR 0.40, 95%CI(0.12-1.34), p=0.14). Conclusion Most patients undergoing primary resection face post-surgery disease recurrence, yet 1 in 4 remain relapse-free for decades. Despite extensive biologic therapy, 15% still require further resections due to disease activity.

Following up with Forest Inheritors: A Survival Analysis of Recently Inherited and Recently Sold Non-Industrial Forest Land in the State of Washington, USA

A growing body of literature shows that the transfer of forest land from one owner to another in the US is associated with events such as parcelization of forest land and/or the subsequent conversion of the land away from forestry land use. For individuals and families who own forest land, a key driver of ownership transfer is the eventual or actual mortality of forest owners themselves. In the State of Washington as well as the US nationally, studies on forest legacy planning reveal that most forest owners want their children or grandchildren to own their forest properties after they no longer own the forest. In contrast, the same surveys also show that a majority of US forest owners acquired their forest land by purchasing the land from a non-relative. We utilize the Washington State Forestland Database to conduct a non-parametric survival analysis of how long recently inherited forest properties remain fully owned by the new owners (i.e. there is no subsequent sale associated with the property) compared to forest land that was sold. Results show that inherited parcels have a significantly lower survival probability as measured by remaining solely within the ownership of the new owner relative to parcels that were sold within an 8-year period. This study quantifies how the mode of forest transfer influences the duration of the following ownership lifecycle and indicates that stewardship efforts should be tailored for owners who have recently inherited forest lands.

Analysis of the survival of technology-based companies linked to business incubators

ABSTRACT Technology-Based Enterprises (TBE) contribute to the economic development of nations through the diffusion of knowledge and innovation, thus being important for the economic growth and competitiveness of territo-ries and countries. Studies point out the difficulties in creating EBTs as well as the high mortality rate. Thus, this study aimed at analysing the survival of incubated and graduated EBTs linked to business incubators in the state of Minas Gerais, Brazil, considering the explanatory variables at the company, incubator, and municipality levels. Non-parametric and semi-parametric survival analysis statistical methods were used to, first, perform a descriptive analysis of survival characteristics and, second, evaluate the relationship of explanatory variables with survival time. A sample of 603 companies was used, obtained from the InovaData-MG platform, with a record of being active or inactive from January 1990 to December 2019, according to the Brazil-ian IRS website. The results pointed out that when disengaging from incuba-tors, graduated companies are subject to a higher risk of failure, and that in-cubated ventures that also went through an acceleration programme have a higher chance of survival. Failure rates showed a non-linear relationship with firm age. Firm-specific, microeconomic, and industry specifications have important implications for the survival of the studied EBTs. This study used a new data set and a statistical technique not usually employed, which allowed us to offer an important contribution to the prediction of the mortality of EBTs, helping the formulation of more assertive public policies for the growth and development of these companies.

Serum Calretinin and Genetic Variability as a Prognostic and Predictive Factor in Malignant Mesothelioma.

Calretinin is a promising diagnostic biomarker for malignant mesothelioma (MM), but less is known about its prognostic role. Our aim was to evaluate the association between serum calretinin concentration or genetic factors and the survival or outcome of cisplatin-based chemotherapy in MM. Our study included 265 MM patients. Serum calretinin concentration was determined using ELISA. Patients were genotyped for seven polymorphisms in CALB2, E2F2, MIR335, NRF1, and SEPTIN7 using competitive allele-specific PCR. Nonparametric tests, logistic regression, and survival analysis were used for statistical analysis. Higher serum calretinin concentration was associated with shorter progression-free (PFS) (HR = 1.18 (1.02-1.37), p = 0.023) and overall survival (OS) (HR = 1.20 (1.03-1.41), p = 0.023), but the association was not significant after adjusting for clinical factors (HR = 1.05 (0.85-1.31), p = 0.653 and HR = 1.06 (0.84-1.34), p = 0.613, respectively). SEPTIN7 rs3801339 and MIR335 rs3807348 were associated with survival even after adjustment (HR = 1.76 (1.17-2.64), p = 0.007 and HR = 0.65 (0.45-0.95), p = 0.028, respectively). Calretinin concentration was higher in patients who progressed after treatment with cisplatin-based chemotherapy (1.68 vs. 0.45 ng/mL, p = 0.001). Calretinin concentration above 0.89 ng/mL was associated with shorter PFS and OS from the start of chemotherapy (HR = 1.88 (1.28-2.77), p = 0.001 and HR = 1.91 (1.22-2.97), p = 0.004, respectively), even after adjusting for clinical factors (p < 0.05). MIR335 rs3807348 was associated with a better response to chemotherapy (OR = 2.69 (1.17-6.18), p = 0.020). We showed that serum calretinin is associated with survival and chemotherapy treatment outcomes in MM and could serve as a predictive biomarker.

KAPLAN-MEIER AND NELSON-AALEN ESTIMATORS FOR CREDIT SCORING

Financial institutions use credit scoring analysis to predict the probability that a customer will default. In this paper, we determine the probability of default using nonparametric survival analysis that are Kaplan-Meier and Nelson-Aalen. The analysis is based on survival function curves, cumulative hazard function curves, mean survival time, and standard error of estimators. Based on the curves of survival function for both Kaplan Meier and Nelson Aalen estimators relatively the same. Based on the curves of cumulative hazard function, mean survival time, and standard error the Nelson-Aalen estimators are slightly higher than Kaplan-Meier.