Artificial Intelligence Enabled Imaging Biomarkers for Distinguishing Pseudoprogression from True Progression after Cellular Immunotherapy in Glioblastoma

Pseudoprogression, transient radiographic worsening caused by treatment-related inflammation rather than tumor growth, is a well-recognized pitfall after chemoradiotherapy for glioblastoma and is becoming more relevant as checkpoint inhibitors and cellular immunotherapies such as chimeric antigen receptor (CAR) T cells enter clinical testing. Conventional structural magnetic resonance imaging (MRI) cannot reliably separate pseudoprogression from true progression, exposing patients to premature treatment discontinuation or unnecessary intensification. Quantitative, artificial-intelligence-based imaging analysis (radiomics, and its molecular extension radiogenomics) is an emerging candidate solution. In this narrative review we synthesize the evidence on radiomics and radiogenomics for differentiating pseudoprogression from true progression in glioblastoma, with emphasis on the immunotherapy and cellular-therapy context. We summarize the pathobiology and reported frequency of pseudoprogression after standard and immune-based therapies, outline the limitations of current response criteria (RANO, iRANO, RANO 2.0), and review representative studies of texture analysis, hand-crafted radiomics, deep-learning radiomics, and radiogenomics. Although several studies report promising discrimination (reported area-under-the-curve values of approximately 0.88-0.92 in different cohorts), the evidence is limited by small sample sizes, retrospective designs, inconsistent definitions and protocols, and infrequent external validation. Progress will require prospective multicenter cohorts with standardized acquisition and segmentation, adherence to TRIPOD and CLAIM reporting standards, transparent and interpretable models, and integration with clinical, molecular, and liquid-biopsy data. A dedicated evidence base specific to cellular immunotherapy is urgently needed.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-18
DOI
https://doi.org/10.21203/rs.3.rs-11295181/v1
Primary Topic
Radiomics and Machine Learning in Medical Imaging
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Artificial Intelligence Enabled Imaging Biomarkers for Distinguishing Pseudoprogression from True Progression after Cellular Immunotherapy in Glioblastoma

Chunlei He, Yafeng Wang, xiaotong zhao, wubing zhou
Research Square
Radiomics and Machine Learning in Medical Imaging
preprint

Artificial Intelligence Enabled Imaging Biomarkers for Distinguishing Pseudoprogression from True Progression after Cellular Immunotherapy in Glioblastoma

Chunlei He, Yafeng Wang, xiaotong zhao, wubing zhou
preprint en

Abstract

Pseudoprogression, transient radiographic worsening caused by treatment-related inflammation rather than tumor growth, is a well-recognized pitfall after chemoradiotherapy for glioblastoma and is becoming more relevant as checkpoint inhibitors and cellular immunotherapies such as chimeric antigen receptor (CAR) T cells enter clinical testing. Conventional structural magnetic resonance imaging (MRI) cannot reliably separate pseudoprogression from true progression, exposing patients to premature treatment discontinuation or unnecessary intensification. Quantitative, artificial-intelligence-based imaging analysis (radiomics, and its molecular extension radiogenomics) is an emerging candidate solution. In this narrative review we synthesize the evidence on radiomics and radiogenomics for differentiating pseudoprogression from true progression in glioblastoma, with emphasis on the immunotherapy and cellular-therapy context. We summarize the pathobiology and reported frequency of pseudoprogression after standard and immune-based therapies, outline the limitations of current response criteria (RANO, iRANO, RANO 2.0), and review representative studies of texture analysis, hand-crafted radiomics, deep-learning radiomics, and radiogenomics. Although several studies report promising discrimination (reported area-under-the-curve values of approximately 0.88-0.92 in different cohorts), the evidence is limited by small sample sizes, retrospective designs, inconsistent definitions and protocols, and infrequent external validation. Progress will require prospective multicenter cohorts with standardized acquisition and segmentation, adherence to TRIPOD and CLAIM reporting standards, transparent and interpretable models, and integration with clinical, molecular, and liquid-biopsy data. A dedicated evidence base specific to cellular immunotherapy is urgently needed.

Research Square
The Sixth People's Hospital of Shenyang (CN)
Good health and well-being
Radiomics and Machine Learning in Medical Imaging
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon