Related Experiment Video
Updated: May 7, 2026
Pretargeted Radioimmunotherapy Based on the Inverse Electron Demand Diels-Alder Reaction
Published on: January 29, 2019
Predicting pathologic tumor response to chemoradiotherapy with histogram distances characterizing longitudinal
Shan Tan1, Hao Zhang, Yongxue Zhang
1Key Laboratory of Image Processing and Intelligent Control of Ministry of Education of China, School of Automation, Huazhong University of Science and Technology, Wuhan 430074, China and Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, Maryland 21201.
New histogram distance features derived from fluorine-18 fluorodeoxyglucose (18F-FDG) positron-emission tomography (PET) scans can predict esophageal cancer response to chemoradiotherapy (CRT). These novel features capture longitudinal changes in tumor uptake, outperforming traditional PET metrics.
Area of Science:
- Nuclear Medicine
- Radiotherapy Oncology
- Medical Imaging Analysis
Background:
- Predicting pathologic tumor response to neoadjuvant chemoradiotherapy (CRT) is crucial for optimizing esophageal cancer treatment.
- Traditional positron-emission tomography (PET) metrics, such as standardized uptake value (SUV) and total lesion glycolysis (TLG), have limitations in accurately assessing treatment response.
- Longitudinal changes in fluorine-18 fluorodeoxyglucose ((18)F-FDG) uptake distribution within tumors may offer more sensitive indicators of treatment efficacy.
Purpose of the Study:
- To propose and evaluate a novel set of (18)F-FDG PET features based on histogram distances for predicting pathologic tumor response to neoadjuvant CRT in esophageal cancer patients.
- To compare the predictive performance of these histogram distance features against traditional PET response measures and texture features.
Main Methods:
- Twenty esophageal cancer patients undergoing neoadjuvant CRT followed by surgery were included.
- Pre- and post-CRT (18)F-FDG PET/CT scans were acquired, registered, and tumor regions delineated.
- Histograms of (18)F-FDG uptake were extracted from pre- and post-treatment tumor volumes. Nineteen histogram distance features quantifying longitudinal changes were analyzed using receiver operating characteristic (ROC) analysis and Mann-Whitney U test.
Main Results:
- Seven bin-to-bin histogram distances and seven crossbin histogram distances demonstrated superior predictive ability compared to maximum SUV (AUC=0.70) and TLG (AUC=0.80).
- Crossbin histogram distance features, particularly quadratic-chi distance (AUC=0.89), diffusion distance (AUC=0.88), and Kolmogorov-Smirnov distance (AUC=0.88), showed high predictive accuracy.
- These novel histogram-based features exhibited slightly higher prediction accuracy than Haralick texture features derived from post-CRT PET images.
Conclusions:
- Longitudinal analysis of (18)F-FDG uptake distribution using histogram distances is a valuable tool for predicting esophageal cancer response to CRT.
- These histogram distance features offer a promising, non-invasive method to assess treatment efficacy and guide clinical decision-making in esophageal cancer management.
