Automated Image Quantitation for Cancer Response Assessment and Clinical Trial Endpoints

Cancer imaging is moving from a purely diagnostic tool to a source of measurable, reproducible data that can guide treatment decisions and trial outcomes. Artificial intelligence now allows researchers to extract quantitative information directly from medical images, turning scans into structured, analyzable data rather than static pictures for visual review.Automated image quantitation, including tumor segmentation, radiomic feature extraction, and functional imaging metrics from PET, DCE-MRI, and DWI, offers a non-invasive way to measure tumor burden, track treatment response, and predict outcomes over time. These tools have the potential to replace or supplement manual, observer-dependent assessments like RECIST with faster, more consistent, and more sensitive measurements, and to serve directly as endpoints in clinical trials.Yet moving automated quantitation from research settings into clinical response assessment and trial use remains difficult. Variability across scanners and institutions, a lack of standardized acquisition and analysis protocols, limited prospective validation, and unclear regulatory pathways all slow adoption. This Research Topic explores how AI-driven quantitation of medical images can be developed, validated, and applied as a reliable measure of cancer response and a trial endpoint.Our goal is to show how automated image quantitation can move from algorithm development to accepted use as a response assessment tool and a clinical trial endpoint. We aim to address the challenges that stand in the way of this transition, including reproducibility across sites and scanners, interpretability of automated measures, and the evidence needed for regulatory qualification.By bringing together medical physicists, radiologists, oncologists, and AI researchers, we connect computational innovation with the practical needs of response assessment in the clinic and in trials. We focus on how automated quantitation of tumor size, shape, texture, and function can improve response evaluation, patient stratification, and therapeutic decision-making.This Research Topic welcomes submissions on the development, validation, and clinical or trial application of automated image quantitation in oncology. We accept experimental, computational, and translational studies across all cancer imaging modalities.We welcome Original Research articles, Reviews, and Methodological papers. Topics of interest include:o Automated tumor segmentation and volumetric quantitationo Radiomic and texture-based feature extraction for treatment responseo Quantitative functional and molecular imaging (PET, DCE-MRI, DWI) for response monitoringo Automated alternatives or complements to RECIST and other response criteriao Use of automated image quantitation as clinical trial endpoints or stratification toolso Reproducibility, standardization, and multi-site/multi-scanner generalizability of quantitation methodso Regulatory and qualification pathways for automated imaging endpointsThrough these papers, we aim to build a resource that helps move automated image quantitation from algorithm to accepted clinical and trial practice.Please note that manuscripts consisting solely of bioinformatics or computational analysis of public omics databases that are not supplemented by relevant functional validation (clinical cohort or biological validation in vitro or in vivo) are out of scope for this Research Topic. Topic editors: Yuming Jiang, Sandeep Kumar Mishra, Harshan Ravi Source: https://www.frontiersin.org/research-topics/84327/automated-image-quantitation-for-cancer-response-assessment-and-clinical-trial-endpoints