In oncology clinical trials, on-treatment biopsy samples are taken to confirm the mode of action of new molecules, among other reasons. Yet, the time point of sample collection is typically scheduled according to 'Expert Best Guess'. We have developed an approach integrating digital pathology and mathematical modelling to provide clinical teams with quantitative information to support this decision. Using digitised biopsies from an ongoing clinical trial as the input to an agent-based mathematical model, we have quantitatively optimised and validated the model demonstrating that it accurately recapitulates observed biopsy samples. Furthermore, the validated model can be used to predict the dynamics of simulated biopsies, with applications from protocol design for phase 1–2 studies to the conception of combination therapies, to personalised healthcare.

在肿瘤学临床试验中，治疗中的活检样本被用来确认新分子的作用方式，以及其他原因。然而，样本采集的时间点通常是根据“专家最佳猜测”安排的。我们开发了一种整合数字病理学和数学建模的方法，为临床团队提供定量信息以支持这一决定。使用来自正在进行的临床试验的数字化活检作为基于代理的数学模型的输入，我们对模型进行了定量优化和验证，证明它准确地概括了观察到的活检样本。此外，经过验证的模型可用于预测模拟活检的动态，从 1-2 期研究的方案设计到联合疗法的概念，