Accelerated Failure Time Survival Research Articles

Extent of MGMT promoter methylation modifies the effect of temozolomide on overall survival in patients with glioblastoma: a regional cohort study.

Background MGMT methylation in glioblastoma predicts response to temozolomide but dichotomizing methylation status may mask the true prognostic value of quantitative MGMT methylation. This study evaluated whether extent of MGMT methylation interacts with the effect of temozolomide on overall survival.MethodsWe included consecutive glioblastoma patients aged ≥16 years diagnosed (April 2012–May 2020) at a neuro-oncology center. All patients had quantitative MGMT methylation measured using pyrosequencing. Those with MGMT methylated tumors were stratified into high and low methylation groups based on a cut-off using Youden index on 2-year survival. Our accelerated failure time survival models included extent of MGMT methylation, age, postoperative Karnofsky performance score, extent of resection, temozolomide regimen, and radiotherapy.ResultsThere were 414 patients. Optimal cut-off point using Youden index was 25.9% MGMT methylation. The number of patients in the unmethylated, low and high methylation groups was 223 (53.9%), 81 (19.6%), and 110 (26.6%), respectively. In the adjusted model, high (hazard ratio [HR] 0.60, 95% confidence intervals [CI] 0.46–0.79, P = 0.005) and low (HR 0.67, 95% CI 0.50–0.89, P < 0.001) methylation groups had better survival compared to unmethylated group. There was no evidence for interaction between MGMT methylation and completed temozolomide regimen (interaction term for low methylation P = 0.097; high methylation P = 0.071). This suggests no strong effect of MGMT status on survival in patients completing temozolomide regimen. In patients not completing the temozolomide regimen, higher MGMT methylation predicted better survival (interaction terms P < 0.001).ConclusionsQuantitative MGMT methylation may provide additional prognostic value. This is important when assessing clinical and research therapies.

Duration of Postdiarrheal Enteric Pathogen Carriage in Young Children in Low-resource Settings.

BackgroundProlonged enteropathogen shedding after diarrhea complicates the identification of etiology in subsequent episodes and is an important driver of pathogen transmission. A standardized approach has not been applied to estimate the duration of shedding for a wide range of pathogens.MethodsWe used a multisite birth cohort of children 0–24 months of age from whom diarrheal and monthly nondiarrheal stools were previously tested by quantitative polymerase chain reaction for 29 enteropathogens. We modeled the probability of detection of the etiologic pathogen before and after diarrhea using a log-normal accelerated failure time survival model and estimated the median duration of pathogen carriage as well as differences in subclinical pathogen carriage 60 days after diarrhea onset in comparison to a prediarrhea baseline.ResultsWe analyzed 3247 etiologic episodes of diarrhea for the 9 pathogens with the highest attributable burdens of diarrhea. The median duration of postdiarrheal carriage varied widely by pathogen, from about 1 week for rotavirus (median, 8.1 days [95% confidence interval {CI}, 6.2–9.6]) to >1 month for Cryptosporidium (39.5 days [95% CI, 30.6–49.0]). The largest increases in subclinical pathogen carriage before and after diarrhea were seen for Cryptosporidium (prevalence difference between 30 days prior and 60 days after diarrhea onset, 0.30 [95% CI, .23–.39]) and Shigella (prevalence difference, 0.21 [95% CI, .16–.27]).ConclusionsPostdiarrheal shedding was widely variable between pathogens, with strikingly prolonged shedding seen for Cryptosporidium and Shigella. Targeted antimicrobial therapy and vaccination for these pathogens may have a relatively large impact on transmission.

Development of nonlinear probabilistic S-N curves using survival analysis techniques with application to steel bridges

This paper presents the application of survival analysis techniques, widely used in medical research, to the development of nonlinear probabilistic S-N curves for fatigue including the effects of other covariates. The number of cycles to failure is treated as a fictitious “time-to-event” parameter. Parametric survival analyses were performed using fatigue test data for steel bridges. The loglogistic accelerated failure time survival model was determined to be best suited for steel bridges. Results indicate that the current linear S-N curves for bridges do not provide a consistent probability of failure. Survival analysis can be a powerful tool for probabilistic fatigue analyses.

Bus arrival time prediction and measure of uncertainties using survival models

This study uses survival models to estimate the arrival time of buses at the downstream stops and intersections. Both accelerated failure time (AFT) and Cox regression based hazard models were considered in this study. Two different types of events: (i) buses arriving at bus stops and (ii) buses arriving at signalised intersections were included for measuring arrival times. Weibull and log-logistic distribution models were fitted for obtaining the arrival times against both the events separately. Various other factors such as distance, speed, bus stop dwell time, passenger count, gradient of the road, intersection length and signal details which included green time, red time, cycle length and so on were considered as explanatory variables. The proposed study was tested on a study corridor of length 59.48 km in the Mumbai arterial roads using public transport (buses). The results reveal that arrival times predicted using the developed models provided smaller uncertainties for 70% of the prediction and reduced prediction variation by 10%. The mean absolute percentage error value obtained for the AFT survival models was 10.04. Overall, the AFT model approach appears to be a promising method compared to Cox regression to predict bus arrival times and the associated uncertainties.

Interplay between Socioeconomic Markers and Polygenic Predisposition on Timing of Dementia Diagnosis.

OBJECTIVESIdentifying the interplay between socioeconomic markers (education and financial resources) and polygenetic predisposition influencing the time of dementia and the diagnosis of clinical Alzheimerʼs disease (AD) dementia is of central relevance for preventive strategies.DESIGNProspective cohort design.SETTINGThe English Longitudinal Study of Aging is a household survey data set of a representative sample.PARTICIPANTSA total of 7,039 individuals aged 50 years and older participated in the study. Of these, 320 (4.6%) were diagnosed with dementia over the 10‐year follow‐up.MEASUREMENTSPolygenic score (PGS) for Alzheimerʼs disease (AD‐PGS) was calculated using summary statistics from the International Genomics of Alzheimerʼs Project. An accelerated failure time survival model was used to investigate interactions between AD‐PGS and socioeconomic markers on the timing of dementia and clinical AD dementia diagnosis.RESULTSA one standard deviation increase in AD‐PGS was associated with an accelerated time to dementia diagnosis by 4.8 months. The presence of the apolipoprotein E gene (APOE‐ε4) was associated with an earlier dementia onset by approximately 24.9 months, whereas intermediate and low levels of wealth were associated with an accelerated time to dementia diagnosis by 12.0 months and 18.7 months, respectively. A multiplicative interaction between AD‐PGS and years of completed schooling in decelerating the time to clinical AD dementia by 3.0 months suggests educational attainment may serve as a protective mechanism against AD diagnosis among older people with a higher polygenic risk. Interaction between AD‐PGS and lower wealth accelerated the time to clinical AD dementia diagnosis by 21.1 to 24.1 months.CONCLUSIONSocioeconomic markers influence the time to dementia and clinical AD dementia diagnosis, particularly in those with a higher polygenic predisposition. J Am Geriatr Soc 68:1529‐1536, 2020.

Further Increasing Fisher's Information for Parameters of Association in Accelerated Failure Time Models via Double Extreme Ranks

Double Extreme Ranked Set Sampling (DERSS) was first introduced by Samawi (2002) as a modification to the well-known Ranked Set Sampling (RSS) and Extreme Ranked Set Sampling (ERSS). In this article, we provide a modification to DERSS scheme with ranking based on an easy-to-rank baseline auxiliary variable known to be associated with survival time. We show that using the modified DERSS improves the performance of the Accelerated failure time (AFT) survival model and provides a more efficient estimator of the hazard ratio than that based on their counter parts simple random sample (SRS), RSS and ERSS, which results in reducing the sample size required and hence the total cost of the study. Our theoretical and simulation studies show the superiority of using the modified DERSS for AFT survival models compared with using SRS, RSS and ERSS. A numerical example based on Worcester Heart Attack Study is presented to illustrate the implementation of the DERSS.

Temporal Effects of Financial Globalization on Income Inequality

Through parametric accelerated failure time survival analysis on a panel of 120 countries for 1970-2018, the study finds empirical evidence that time to upsurge in income inequality declines after a country liberalizes investment across borders. This occurs with a significant non-monotonic hazard rate of upsurge, arguably via the channels of capital-skill complementarity and wage inequality. This effect is robust to distributional assumptions, alternative measurements of financial globalization and income inequality, sample restriction to pre-crisis period and residual heteroskedasticity. Moreover, it is reasonably generalizable across different economic arrangements due to employing diverse panel and handling issues of imputation and cross-country standardization in contemporary income inequality data.

Gastric adenocarcinoma in young adult patients: patterns of care and survival in the United States.

Evidence regarding gastric cancer patients < 40years old is limited. This study examines young adults with gastric adenocarcinoma in the National Cancer Database to describe demographics and treatment practices, and to develop a nomogram to predict survival. The database was queried for adult patients diagnosed with gastric adenocarcinoma from 2004 to 2013. Patients were stratified into two age groups: <40 and ≥ 40years. The database was analyzed to compare demographics, clinical characteristics, and treatments used for each group. Differences in survival were assessed using Kaplan-Meier curves and log-rank test. For adults < 40years old, an accelerated failure time survival model was fitted for overall survival and a descriptive nomogram was constructed. Of 70,084 patients included, 2615 (4%) were < 40years old and 67,469 (96%) were ≥ 40years. Compared to older patients, adults < 40years old were more likely to be female (46 vs. 35%), non-white (31 vs. 23%), Hispanic (32 vs. 11%), from the northeast (36 vs. 23%), and to present with stage IV disease (59 vs. 42%) and bone metastases (36 vs. 21%; p < 0.001 for all). The nomogram showed clinical stage as the strongest predictor of overall survival, followed by treatment, grade, race, Charlson-Deyo comorbidity score, and sex. Young adults with gastric adenocarcinoma are more likely to be Hispanic, female, from the northeast, and to present with metastases. Despite these differences, clinical stage, treatment, and tumor grade are most predictive of overall survival for young adult patients.

Development of thoracic injury risk functions for the THOR ATD

The Test Device for Human Occupant Restraint (THOR) 50th percentile male anthropomorphic test device (ATD) aims to improve the ability to predict the risk of chest injury to restrained automobile occupants by measuring dynamic chest deflection at multiple locations. This research aimed to describe the methods for developing a thoracic injury risk function (IRF) using the multi-point chest deflection metrics from the 50th percentile male THOR Metric ATD with the SD-3 shoulder and associating to post-mortem human subjects (PMHS) outcomes that were matched on identical frontal and frontal-oblique impact sled testing conditions. Several deflection metrics were assessed as potential predictor variables for AIS 3+ injury risk, including a combined metric, called PC Score, which was generated from a principal component analysis. A parametric survival analysis (specifically, accelerated failure time (AFT) with Weibull distribution) was assessed in the development of the IRF. Model fit was assessed using various modeling diagnostics, including the area under the receiver operating characteristic curve (AUC). Models based on resultant deflection consistently exhibited improved fit compared to models based on x-axis deflection or chord deflection. Risk functions for the THOR PC Score and Cmax (maximum resultant deflection) were qualitatively equivalent, producing AUCs of 0.857 and 0.861, respectively. Adjusting for the potential confounding effects of age, AFT survival models with Cmax or PC Score as the primary deflection metric resulted in the THOR injury risk models with the best combination of biomechanical appropriateness, potential utility and model fit, and may be recommended as injury predictors.

Low-grade and high-grade endometrial stromal sarcoma: A National Cancer Database study

ObjectiveTo provide refined prognostic information from large cohorts of women with low-grade or high-grade endometrial stromal sarcoma (ESS). MethodsWe performed an observational retrospective cohort analysis of women diagnosed with low-grade or high-grade ESS from the 1998–2013 National Cancer Database. Kaplan-Meier and multivariable accelerated failure time survival analyses were performed to identify prognostic factors after multiple imputation of missing data. Recursive partitioning methods were used to rank prognostic factors in high-grade ESS. Matched cohort analyses were performed to hypothesis-test effects of adjuvant treatments. ResultsWe identified 2414 and 1383 women with low-grade or high-grade ESS, respectively. Women with high-grade ESS had markedly decreased survival compared to women with low-grade ESS (five-year survival (95% CI): 32.6 (30.1–35.3%) versus 90.5% (89.3–91.8%), P<0.001). Among women with high-grade ESS, median survival (95% CI) was only 19.9 (17.1–22.1) months. Increased age and tumor size were associated with decreased survival in low-grade ESS. In high-grade ESS, additional negative prognostic factors were distant or nodal metastasis, omission of lymphadenectomy, and pathologically-positive surgical margins (all P<0.001). Use of adjuvant chemotherapy (time ratio (TR) (95% CI): 1.36 (1.17–1.58), P<0.001) and radiotherapy (TR (95% CI): 1.57 (1.32–1.87), P<0.001) were associated with increased survival for high-grade ESS. ConclusionThe contrasting excellent versus poor prognosis of low-grade versus high-grade ESS, respectively, was confirmed. The best treatment of high-grade ESS is early and complete surgical resection including lymphadenectomy. Adjuvant chemotherapy and radiotherapy may increase survival of women with high-grade ESS.

Prognosis and treatment of uterine leiomyosarcoma: A National Cancer Database study

ObjectiveTo determine overall survival and factors associated with survival of women with uterine leiomyosarcoma. MethodsWe performed an observational cohort study of women with uterine leiomyosarcoma (n=7455) from the 1998–2013 National Cancer Database. Kaplan-Meier and multivariable accelerated failure time survival analyses were performed to investigate predictors of survival. Sensitivity and matched cohort analyses were performed to evaluate the roles of oophorectomy, lymphadenectomy, and chemotherapy in early leiomyosarcoma and chemotherapy in metastatic leiomyosarcoma. ResultsMedian (interquartile range) age at diagnosis was 54 (48–63) years. Older age, higher comorbidity, black race, higher stage or grade, larger tumor size, lymph node involvement, metastasis at diagnosis, positive surgical margin, adjuvant chemotherapy, and brachytherapy were independently associated with decreased survival by unmatched cohort analyses. Private insurance was associated with increased survival. By matched cohort analyses, omitting oophorectomy was not associated with survival among women≤51years old at diagnosis (event time ratio (ETR) (95% CI) 1.06 (0.90–1.25), P=0.48). Omitting lymphadenectomy was not associated with survival (ETR (95% CI) 1.02 (0.94–1.10), P=0.60). Among women with stage I leiomyosarcoma, adjuvant chemotherapy was not associated with increased survival (ETR (95% CI) 0.91 (0.78–1.05), P=0.18). Chemotherapy was associated with increased survival of women with metastatic leiomyosarcoma (median survival (95% CI) 19.4 (16.4–23.0) versus 10.9 (7.7–14.3) months, ETR (95% CI) 1.66 (1.46–1.90), P<0.001). ConclusionEarly and complete resection is the best-evidenced treatment for uterine leiomyosarcoma. Oophorectomy and lymphadenectomy may be safely omitted for clinically uterus-confined leiomyosarcoma. Chemotherapy increases survival of women with metastatic leiomyosarcoma.

Prognostic impact of tumour burden measured by quantitative real-time PCR from sentinel lymph nodes of melanoma patients: data from 10-year follow-up.

Sentinel node (SN) biopsy is regarded as standard of care for patients (pts) with cutaneous melanoma ≥1.0mm of thickness. In the recent AJCC classification, findings in the SN are simply classified as positive or negative. In our analyses, we were interested whether quantitative real-time PCR (qRT-PCR) is able to predict disease-free survival (DFS) and overall survival (OS) depending on tumour burden in the SN. One hundred and forty-five pts were analysed using qRT-PCR for tyrosinase. Results were analysed using accelerated failure time survival model and cox proportional hazards models using the R statistics framework. Forty-one pts (28%) were positive according to qRT-PCR. In total, 12 of 41 pts showed tumour deposits in the SN using S100 and/or HMB-45-labelled immunohistochemistry as well. One patient had micrometastases detected by immunohistochemistry staining but failed in the qRT-PCR. After 10years of follow-up, 34 patients recurred and 27 patients died. Significant differences for DFS and OS were detected for sex, increasing tumour thickness, ulceration of the primary tumour, and metastatic spread in the SN determined by histology as well as qRT-PCR. Quantitative analyses showed a logarithmic correlation between tumour burden and prognosis. However, as multivariate analyses reveal qRT-PCR was not superior compared to classical histology or immunohistology.

Using survival models to estimate bus travel times and associated uncertainties

Transit agencies often provide travelers with point estimates of bus travel times to downstream stops to improve the perceived reliability of bus transit systems. Prediction models that can estimate both point estimates and the level of uncertainty associated with these estimates (e.g., travel time variance) might help to further improve reliability by tempering user expectations. In this paper, accelerated failure time survival models are proposed to provide such simultaneous predictions. Data from a headway-based bus route serving the Pennsylvania State University-University Park campus were used to estimate bus travel times using the proposed survival model and traditional linear regression frameworks for comparison. Overall, the accuracy of point estimates from the two approaches, measured using the root-mean-squared errors (RMSEs) and mean absolute errors (MAEs), was similar. This suggests that both methods predict travel times equally well. However, the survival models were found to more accurately describe the uncertainty associated with the predictions. Furthermore, survival model estimates were found to have smaller uncertainties on average, especially when predicted travel times were small. Tests for transferability over time suggested that the models did not over-fit the dataset and validated the predictive ability of models established with historical data. Overall, the survival model approach appears to be a promising method to predict both expected bus travel times and the uncertainty associated with these travel times.

Modelling total duration of traffic incidents including incident detection and recovery time

Traffic incidents are key contributors to non-recurrent congestion, potentially generating significant delay. Factors that influence the duration of incidents are important to understand so that effective mitigation strategies can be implemented. To identify and quantify the effects of influential factors, a methodology for studying total incident duration based on historical data from an ‘integrated database’ is proposed. Incident duration models are developed using a selected freeway segment in the Southeast Queensland, Australia network. The models include incident detection and recovery time as components of incident duration. A hazard-based duration modelling approach is applied to model incident duration as a function of a variety of factors that influence traffic incident duration. Parametric accelerated failure time survival models are developed to capture heterogeneity as a function of explanatory variables, with both fixed and random parameters specifications. The analysis reveals that factors affecting incident duration include incident characteristics (severity, type, injury, medical requirements, etc.), infrastructure characteristics (roadway shoulder availability), time of day, and traffic characteristics. The results indicate that event type durations are uniquely different, thus requiring different responses to effectively clear them. Furthermore, the results highlight the presence of unobserved incident duration heterogeneity as captured by the random parameter models, suggesting that additional factors need to be considered in future modelling efforts.