Survival Model Predictive Accuracy and ROC Curves

By Patrick J. Heagerty, Yingye Zheng. Biometrics, 2005

1,389 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/7c529617cf5579f8c9683299ec33c57c17451e94

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Time-to-Event Prediction with Neural Networks and Cox Regression (2019, background): “This is likely a result of its interpretability, as it has a close relationship to classiﬁcation accuracy (Ishwaran et al., 2008), and ROC AUC (Heagerty and Zheng, 2005).”
A unified inference procedure for a class of measures to assess improvement in risk prediction systems with survival data (2012, methodology): “Recently, the time-specific ROC curve methods have been modified to deal with censored event time data [4, 5].”
A Multi-Task Learning Formulation for Survival Analysis (2016, methodology): “Due to the presence of censoring in the data, the standard evaluation metrics for regression such as root of mean square error and R(2) are not suitable for measuring the performance in survival analysis [9].”
Interstitial lung disease in systemic sclerosis: a simple staging system. (2008, methodology): “Optimal HRCT extent and FVC thresholds, identified against mortality using proportional hazards analysis, were compared with alternative thresholds using a modified time-dependent receiver operating characteristic curve (ROC) analysis ( 14 ) (see online supplement).”
Predictive validity performance indicators in violence risk assessment: a methodological primer. (2013, background): “Alternative models have been recently proposed to take stochastic information into account (Chambless & Diao, 2006; Heagerty & Zheng, 2005), though these have not made their way into the mainstream yet.”
Assessing the performance of prediction models: a framework for some traditional and novel measures (2010, methodology): “Furthermore, time-dependent c statistics have been proposed.(6,38) In addition to the c statistic, the discrimination slope can be used as a simple measure for how well subjects with and without the outcome are separated.”
Estimating and comparing time‐dependent areas under receiver operating characteristic curves for censored event times with competing risks (2013, background): “[3, 4] introduced the time-dependent ROC curves and proposed several definitions of cases and controls.”
Estimation of time‐dependent area under the ROC curve for long‐term risk prediction (2006, background): “Hagerty and Zheng [21] contrast the above sensitivity=speci city parameters, called ‘Cumulative=Dynamic’, with alternative time-dependent parameters called ‘Incident=Static’, where now, in our notation, at time tm sens(tm; K)=P(Zi¿K |Di(tm)=1; Di(tm−1)=0) and spec(tm; K)=P(Zi6K |Di(t∗)=0) for some x”
Whole slide images based cancer survival prediction using attention guided deep multiple instance learning networks (2020, methodology): “To evaluate the performances in survival prediction, we take the concordance index (C-index) and area under curve (AUC) as our evaluation metrics [49].”
Machine Learning for Survival Analysis: A Survey (2017, methodology): “In order to evaluate the performance during a follow-up period, Heagerty and Zheng deﬁned the C-index for a ﬁxed follow-up time period (0 , t ∗ ) as the weighted average of AUC values at all possible observation time points [Heagerty and Zheng 2005].”
Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges (2016, background): “Overall, the above considerations stay the same, though usually the choice of loss functions will change for different outcomes.(39,47)”
Dynamic Predictions and Prospective Accuracy in Joint Models for Longitudinal and Time‐to‐Event Data (2011, methodology): “In this last part we follow similar arguments to Zheng and Heagerty (2007); Antolini, Boracchi, and Biganzoli (2005); and Heagerty and Zheng (2005), but suitably adapted to the joint modeling framework.”