CT radiomics-based model for predicting TMB and immunotherapy response in non-small cell lung cancer

Jiexiao Wang ^1,2, Jialiang Wang ^1,2, Xiang Huang ^2,3

¹School of Basic Medical Sciences, Anhui Medical University, Hefei, Anhui, China.
²Anhui Province Key Laboratory of Medical Physics and Technology, Institute of Health and Medical Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui, China.
³Hefei Cancer Hospital, Chinese Academy of Sciences, Hefei, Anhui, China.
⁴University of Science and Technology of China, Hefei, Anhui, China.
⁵Anhui Province Key Laboratory of Medical Physics and Technology, Institute of Health and Medical Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui, China. wsj@cmpt.ac.cn.
⁶School of Basic Medical Sciences, Anhui Medical University, Hefei, Anhui, China. bhong@hmfl.ac.cn.
⁷Anhui Province Key Laboratory of Medical Physics and Technology, Institute of Health and Medical Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui, China. bhong@hmfl.ac.cn.
⁸Hefei Cancer Hospital, Chinese Academy of Sciences, Hefei, Anhui, China. bhong@hmfl.ac.cn.
⁹School of Basic Medical Sciences, Anhui Medical University, Hefei, Anhui, China. wanghz@hfcas.ac.cn.
¹⁰Anhui Province Key Laboratory of Medical Physics and Technology, Institute of Health and Medical Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, Anhui, China. wanghz@hfcas.ac.cn.
¹¹Hefei Cancer Hospital, Chinese Academy of Sciences, Hefei, Anhui, China. wanghz@hfcas.ac.cn.

⁰School of Basic Medical Sciences, Anhui Medical University, Hefei, Anhui, China.

Bmc Medical Imaging +

|

February 15, 2024

View abstract on PubMed

Summary

This study developed a radiomics model using CT scans to predict tumor mutational burden (TMB) in non-small cell lung cancer (NSCLC). The model accurately identifies patients likely to respond to immunotherapy, improving treatment precision.

Area Of Science

Radiology
Oncology
Medical Imaging Analysis

Background

Tumor mutational burden (TMB) is a key biomarker for predicting immunotherapy response in non-small cell lung cancer (NSCLC).
Radiomics offers advanced quantitative analysis of medical imaging features.
Accurate TMB prediction is crucial for personalized immunotherapy in NSCLC.

Purpose Of The Study

To develop and validate a radiomic model for predicting TMB levels in NSCLC patients using CT imaging.
To assess the model's ability to predict immunotherapy response based on CT-derived radiomic features.
To evaluate the clinical utility of radiomics for TMB classification and treatment selection in NSCLC.

Main Methods

Retrospective analysis of pre-operative chest CT images from 127 NSCLC patients.
Extraction of 1037 radiomic features using 3D-Slicer software.
Construction and validation of a TMB prediction model using LASSO regression, multiple logistic regression, ROC curves, and calibration curves on training and external datasets.

Main Results

Three radiomic features (Flatness, Autocorrelation, Minimum) were identified as significant predictors of TMB level.
The radiomic model achieved an AUC of 0.816 in the training dataset for TMB prediction.
External validation demonstrated AUCs of 0.775 and 0.762 for TMB prediction and immunotherapy efficacy prediction, respectively.

Conclusions

A CT-based radiomic model effectively predicts TMB levels in NSCLC.
This approach offers a cost-effective method for improving TMB classification.
The model supports precise immunotherapy treatment selection for NSCLC patients.

Keywords:

Computed tomography Immunotherapy Non-small cell lung cancer Radiomics Tumor mutation burden