Siamese Survival Analysis with Competing Risks

Abstract

Survival analysis in the presence of multiple possible adverse events, i.e., competing risks, is a pervasive problem in many industries (healthcare, finance, etc.). Since only one event is typically observed, the incidence of an event of interest is often obscured by other related competing events. This nonidentifiability, or inability to estimate true cause-specific survival curves from empirical data, further complicates competing risk survival analysis. We introduce Siamese Survival Prognosis Network (SSPN), a novel deep learning architecture for estimating personalized risk scores in the presence of competing risks. SSPN circumvents the nonidentifiability problem by avoiding the estimation of cause-specific survival curves and instead determines pairwise concordant time-dependent risks, where longer event times are assigned lower risks. Furthermore, SSPN is able to directly optimize an approximation to the C-discrimination index, rather than relying on well-known metrics which are unable to capture the unique requirements of survival analysis with competing risks.

A review on statistical and machine learning competing risks methods

“…It is also important to provide details on how such data were generated or processed, as well as a clear description of any ad hoc choices made (e.g., inclusion/exclusion criteria). For instance, the Surveillance, Epidemiology, and End Results (SEER) Program data sets have been employed to showcase several CR methods (e.g., Alaa & van der Schaar, 2017; Bellot & van der Schaar, 2018a, 2018b; Nemchenko et al., 2018; Zhang & Zhou, 2018). However, detailed information on how the data set used was preprocessed is usually not provided (in some cases, authors do not even provide the full list of covariates used in the analysis).…”

“…For example, to perform feature selection, a penalized multinomial logistic regression (e.g., as implemented in glmnet) could be employed. Other approaches specifically developed for discrete-time CR data include Siamese Survival Prognosis Network (Nemchenko et al, 2018) and DeepHit (Lee et al, 2018), both using neural networks. Another recent approach, by Sparapani et al (2020), is based on Bayesian additive regression trees (BART, Hill et al, 2020).…”

Competing risk between in-hospital mortality and recovery: An application of DeepHit on COVID-19 clinical data

“…The cmprsk package in R is used to t the Fine-Gray model. We used the materials provided by Nemchenko et al 7 to t the model and nd the performance metric.…”

“…6 This approach is inadequate as the occurrence of an event of interest is often obscured by other competing events. 7 Fine-Gray model is one of the survival techniques which has looked into this competing risk aspect. Hence, it is considered as a benchmark in competing risk analysis.…”

Competing risk between in-hospital mortality and recovery: An application of DeepHit on COVID-19 clinical data

“…But incorporating this into classical survival models is not trivial. 7 Most of the competing risk models are based on cause-specific cumulative incidence function(CIF). Cause-specific CIF gives the probability that the event k occurs on or before time t for a patient with covariates x. Fine-Gray model is one of the survival models which has looked into the competing risk aspect.…”