Related Experiment Video
Updated: Aug 15, 2026
Using Micro-computed Tomography for the Assessment of Tumor Development and Follow-up of Response to Treatment in a Mouse Model of Lung Cancer
Published on: May 20, 2016
Radiomics-Based AI for Predicting Neoadjuvant Immunochemotherapy Pathological Response in Non-Small Cell Lung Cancer:
Ziqi Jiang1, Yuan Xu1, Shuyu Jia2
1Department of Thoracic Surgery, Peking Union Medical College Hospital, Chinese Academy of Medical Science, No.1 Shuaifuyuan, Wangfujing, Beijing, Dongcheng District, 100730, China, 86 13621021237.
Radiomics-based artificial intelligence (AI) significantly improves prediction of treatment response in non-small cell lung cancer (NSCLC) compared to traditional methods. This AI approach offers higher accuracy for assessing pathological complete response (pCR) and major pathological response (MPR) after neoadjuvant therapy.
Area of Science:
- Medical Imaging and Artificial Intelligence
- Oncology
- Radiomics
Background:
- Non-small cell lung cancer (NSCLC) is a leading cause of cancer mortality globally.
- Early prediction of response to neoadjuvant therapy is crucial for effective patient management.
- Current methods for assessing treatment response in NSCLC have limitations.
Purpose of the Study:
- To evaluate the diagnostic performance of radiomics-based AI in predicting pathological complete response (pCR) and major pathological response (MPR) after neoadjuvant immunochemotherapy in NSCLC.
- To compare the accuracy of AI models against traditional radiological criteria for response prediction.
- To assess the potential of AI in improving preoperative response assessment for NSCLC patients.
Main Methods:
- Systematic literature search across major databases (PubMed, Embase, Cochrane, Web of Science) up to October 2025.
- Inclusion of studies using CT or PET/CT-based AI models for predicting pCR or MPR in NSCLC.
- Meta-analysis of sensitivity, specificity, and AUC using a bivariate random effects model, with methodological quality appraisal using PROBAST+AI.
Main Results:
- Analysis of 23 studies with 2004 patients revealed AI models significantly outperformed traditional criteria for predicting pCR (AUC 0.85 vs. not specified) and MPR (AUC 0.88 vs. 0.65).
- AI models achieved pooled sensitivity of 0.77 and specificity of 0.79 for pCR, and sensitivity of 0.80 and specificity of 0.83 for MPR.
- Positron emission tomography/CT-based AI models showed higher specificity for MPR compared to CT-based models (0.95 vs. 0.80).
Conclusions:
- Radiomics-based AI demonstrates high diagnostic accuracy and superior sensitivity for preoperative response assessment in NSCLC.
- AI shows significant potential to enhance the prediction of pathological response to neoadjuvant therapy.
- Further large-scale, prospective, multicenter trials are needed to validate AI findings for clinical implementation.