ViVo: A temporal modeling framework that boosts statistical power and minimizes animal usage

Abstract

Preclinical tumor studies are often limited by high variability and small sample sizes, reducing statistical power and masking treatment effects. We present an exponential framework that estimates tumor growth rates (r) independently of initial burden and introduces Tumor Growth Rate (TGR) matrices to map treatment effects across time windows. Applied to a public xenograft dataset and four additional mouse models, exponential fits achieved high agreement with raw data (median R2 typically > 0.8-0.9). When resampling matched group sizes (n=3-7), our approach consistently outperformed conventional endpoint and daily non-parametric analyses, revealing treatment effects that standard comparisons missed. The framework also predicts tumor weights for animals euthanized early, enabling their inclusion in final analyses and enhancing statistical consistency while advancing the 3Rs principles. To ensure broad adoption, we developed ViVo, and open-source web platform (gcanudo-barreras.github.io/ViVo-Platform/) that provides accessible, standardized analysis of in vivo tumor kinetics and therapeutic efficacy.