Augmenting Insufficiently Accruing Oncology Clinical Trials Using Generative Models: Validation Study

Journal of Medical Internet Research, Год журнала: 2025, Номер 27, С. e66821 - e66821

Опубликована: Март 5, 2025

Insufficient patient accrual is a major challenge in clinical trials and can result underpowered studies, as well exposing study participants to toxicity additional costs, with limited scientific benefit. Real-world data provide external controls, but insufficient affects all arms of study, not just controls. Studies that used generative models simulate more patients were the scenarios considered, replicability criteria, number models, evaluated. This aimed perform comprehensive evaluation on extent be compensate for trials. We performed retrospective analysis using 10 datasets from 9 fully accrued, completed, published cancer For each trial, we removed latest recruited (from 10% 50%), trained model remaining patients, simulated replace ones augment available data. then replicated this augmented dataset determine if findings remained same. Four different evaluated: sequential synthesis decision trees, Bayesian network, adversarial variational autoencoder. These compared sampling replacement (ie, bootstrap) simple alternative. Replication analyses 4 metrics: agreement, estimate standardized difference, CI overlap. Sequential replication metrics removal up 40% last (decision agreement: 88% 100% across datasets, 100%, cannot reject difference null hypothesis: overlap: 0.8-0.92). Sampling was next most effective approach, agreement varying 78% 89% datasets. There no evidence monotonic relationship estimated effect size recruitment order these studies. suggests earlier trial systematically than those later, at least partially explaining why early effectively later trial. The fidelity generated relative training Hellinger distance high cases. an oncology few 60% target recruitment, enable simulation full had continued accruing alternative drawing conclusions study. results demonstrating potential rescue poorly trials, studies are needed confirm generalize them other diseases.

Язык: Английский

Synthetic tabular data generation in Federated Learning environments: A practical use case for Acute Myeloid Leukemia (Preprint)

Опубликована: Март 20, 2025

Data scarcity and dispersion pose significant obstacles in biomedical research, particularly when addressing rare diseases. In such scenarios, Synthetic Generation (SDG) has emerged as a promising path to mitigate the first issue. Concurrently, Federated Learning (FL) is machine learning paradigm where multiple nodes collaborate create centralized model with knowledge that distilled from data different nodes, but without need for sharing it. This research explores combination of SDG FL technologies context Acute Myeloid Leukemia, hematological disorder, evaluating their combined impact quality generated artificial datasets. To evaluate privacy- fidelity-related federating distribution scenarios numbers comparing them baseline model. A state-of-the-art Generative Adversarial Network architecture was trained considering four scenarios: (1) non-federated all available, (2) federated scenario evenly distributed among (3) unevenly randomly (imbalanced data), (4) non-IID distributions. For each fixed set node quantities (3, 5, 7, 10) considered assess its impact, evaluated attending fidelity-privacy trade-off. The computed fidelity metrics exhibited statistically deteriorations (P < 0.001) ranging 0.21% 21.23% due federation process. When experiments diverse no strong tendencies were observed, even if specific comparisons resulted significative differences. Privacy mainly maintained while obtaining maximum improvements 55.17% 26.23, although they not significant. Within scope use case this paper, act an algorithm results loss compared maintaining privacy levels. However, deterioration does significantly increase number used train models grows, though differences found comparisons. fact amount differently neither most nor metrics, similar scenarios.

Язык: Английский

An attempt to generate panoramic radiographs including jaw cysts using StyleGAN3

Dentomaxillofacial Radiology, Год журнала: 2024, Номер unknown

Опубликована: Сен. 2, 2024

The purpose of this study was to generate radiographs including dentigerous cysts by applying the latest generative adversarial network (GAN; StyleGAN3) panoramic radiography.

Язык: Английский