Methods for a similarity measure for clinical attributes based on survival data analysis

Comparison of Synthetic Data Generation Techniques for Control Group Survival Data in Oncology Clinical Trials: Simulation Study

Background

Synthetic patient data (SPD) generation for survival analysis in oncology trials holds significant potential for accelerating clinical development. Various machine learning methods, including classification and regression trees (CART), random forest (RF), Bayesian network (BN), and conditional tabular generative adversarial network (CTGAN), have been used for this purpose, but their performance in reflecting actual patient survival data remains under investigation.

Objective

The aim of this study was to determine the most suitable SPD generation method for oncology trials, specifically focusing on both progression-free survival (PFS) and overall survival (OS), which are the primary evaluation end points in oncology trials. To achieve this goal, we conducted a comparative simulation of 4 generation methods, including CART, RF, BN, and the CTGAN, and the performance of each method was evaluated.

Methods

Using multiple clinical trial data sets, 1000 data sets were generated by using each method for each clinical trial data set and evaluated as follows: (1) median survival time (MST) of PFS and OS; (2) hazard ratio distance (HRD), which indicates the similarity between the actual survival function and a synthetic survival function; and (3) visual analysis of Kaplan-Meier (KM) plots. Each method's ability to mimic the statistical properties of real patient data was evaluated from these multiple angles.

Results

In most simulation cases, CART demonstrated the high percentages of MSTs for synthetic data falling within the 95% CI range of the MST of the actual data. These percentages ranged from 88.8% to 98.0% for PFS and from 60.8% to 96.1% for OS. In the evaluation of HRD, CART revealed that HRD values were concentrated at approximately 0.9. Conversely, for the other methods, no consistent trend was observed for either PFS or OS. CART demonstrated better similarity than RF, in that CART caused overfitting and RF (a kind of ensemble learning approach) prevented it. In SPD generation, the statistical properties close to the actual data should be the focus, not a well-generalized prediction model. Both the BN and CTGAN methods cannot accurately reflect the statistical properties of the actual data because small data sets are not suitable.

Conclusions

As a method for generating SPD for survival data from small data sets, such as clinical trial data, CART demonstrated to be the most effective method compared to RF, BN, and CTGAN. Additionally, it is possible to improve CART-based generation methods by incorporating feature engineering and other methods in future work.

Multistate models of developmental toxicity: Application to valproic acid-induced malformations in the zebrafish embryo

For the determination of acute toxicity of chemicals in zebrafish (Danio rerio) embryos, the OECD test guideline 236, relative to the Fish Embryo Toxicity Test (FET), stipulates a dose-response analysis of four lethal core endpoints and a quantitative characterization of abnormalities including their time-dependency. Routinely, the data are analyzed at the different observation times separately. However, observations at a given time strongly depend on the previous effects and should be analyzed jointly with them. To solve this problem, we developed multistate models for occurrence of developmental malformations and live events in zebrafish embryos exposed to eight concentrations of valproic acid (VPA) the first five days of life. Observations were recorded daily per embryo. We statistically infer on model structure and parameters using a numerical Bayesian framework. Hatching probability rate changed with time and we compared five forms of its time-dependence; a constant rate, a piecewise constant rate with a fixed hatching time at 48 h post fertilization, a piecewise constant rate with a variable hatching time, as well as a Hill and Gaussian form. A piecewise constant function of time adequately described the hatching data. The other transition rates were conditioned on the embryo body concentration of VPA, obtained using a physiologically-based pharmacokinetic model. VPA impacted mostly the malformation probability rate in hatched and non-hatched embryos. Malformation reversion probability rates were lowered by VPA. Direct mortality was low at the concentrations tested, but increased linearly with internal concentration. The model makes full use of data and gives a finer grain analysis of the teratogenic effects of VPA in zebrafish than the OECD-prescribed approach. We discuss the use of the model for obtaining toxicological reference values suitable for inter-species extrapolation. A general result is that complex multistate models can be efficiently evaluated numerically.