Reconstruction-aware radiomics for risk stratification of subsolid pulmonary nodules: technical evidence, clinical limits, and translation priorities.
- 1Department of Radiology, Zigong First People's Hospital, Zigong, Sichuan, China.
Abstract
Subsolid pulmonary nodules (SSNs), including pure ground-glass nodules (pGGNs) and part-solid nodules (PSNs), span transient inflammatory abnormalities and adenocarcinoma-spectrum lesions with widely different management implications. This critical narrative review treats CT reconstruction as part of the quantitative biomarker definition and asks whether targeted or high-spatial-resolution reconstruction provides stable, task-specific information beyond standardized thin-section CT. Evidence was synthesized across direct SSN reconstruction-domain studies, technical robustness studies, and clinical prediction studies. Direct imaging studies support improved depiction of selected fine structures, while same-scan GGN evidence demonstrates substantial pixel-size dependence of radiomic features and model composition. However, most clinical models were not evaluated in paired reconstruction domains, external validation with calibration remains uncommon, and no prospective study demonstrating management or patient benefit from reconstruction-aware radiomics in SSNs was identified in the literature reviewed. Persistent pGGNs provide a biological and technical rationale for investigating incremental quantitative information, but greater benefit than in PSNs has not been demonstrated. In PSNs, any model must add value beyond solid-component measurements, consolidation-to-tumor ratio, morphology, and longitudinal growth. Translation therefore requires reconstruction and segmentation robustness testing, strong conventional comparators, external calibration, prespecified thresholds, and prospective decision-linked evaluation. Until such benefit and safety are demonstrated, reconstruction-aware radiomics should remain investigational or confined to governed, domain-checked decision-support pathways.
Related
- Impact of deep learning image reconstruction on CT radiomics reproducibility in patients with pulmonary nodules
- Cancer-Associated Parasitic Neglected Tropical Diseases: Precision Diagnostics, Risk Stratification, and Implementation Priorities
- Radiomics-based machine learning models decipher heterogeneity of anoikis-related molecular subtypes for prognostic stratification in head and neck squamous cell carcinoma
- Nomogram models integrating multiregion radiomics, quantitative computed tomography, and clinical indicators for distinguishing connective tissue disease-associated interstitial lung disease from idiopathic pulmonary fibrosis and improving the prediction of gender-age-physiology stage
- Multimodal preoperative risk stratification of pancreatic mass lesions using clinical variables, imaging radiomics and plasma cfDNA methylation: protocol for an ambispective multicentre diagnostic prediction model study
BibTeX
@article{liu2026reconstructionaware,
title = {{Reconstruction-aware radiomics for risk stratification of subsolid pulmonary nodules: technical evidence, clinical limits, and translation priorities.}},
author = {Yujian Liu and Yu Peng and Feng Chen and Jie Huang and Jie Yang and Jianquan Zhong},
journal = {Frontiers in medicine},
year = {2026},
month = oct,
doi = {10.3389/fmed.2026.1955926},
url = {https://europepmc.org/article/MED/42903588},
note = {CC BY},
}
RIS
TY - JOUR
TI - Reconstruction-aware radiomics for risk stratification of subsolid pulmonary nodules: technical evidence, clinical limits, and translation priorities.
AU - Yujian Liu
AU - Yu Peng
AU - Feng Chen
AU - Jie Huang
AU - Jie Yang
AU - Jianquan Zhong
PY - 2026
DA - 2026/10/06
JO - Frontiers in medicine
DO - 10.3389/fmed.2026.1955926
UR - https://europepmc.org/article/MED/42903588
AB - Subsolid pulmonary nodules (SSNs), including pure ground-glass nodules (pGGNs) and part-solid nodules (PSNs), span transient inflammatory abnormalities and adenocarcinoma-spectrum lesions with widely different management implications. This critical narrative review treats CT reconstruction as part of the quantitative biomarker definition and asks whether targeted or high-spatial-resolution reconstruction provides stable, task-specific information beyond standardized thin-section CT. Evidence was synthesized across direct SSN reconstruction-domain studies, technical robustness studies, and clinical prediction studies. Direct imaging studies support improved depiction of selected fine structures, while same-scan GGN evidence demonstrates substantial pixel-size dependence of radiomic features and model composition. However, most clinical models were not evaluated in paired reconstruction domains, external validation with calibration remains uncommon, and no prospective study demonstrating management or patient benefit from reconstruction-aware radiomics in SSNs was identified in the literature reviewed. Persistent pGGNs provide a biological and technical rationale for investigating incremental quantitative information, but greater benefit than in PSNs has not been demonstrated. In PSNs, any model must add value beyond solid-component measurements, consolidation-to-tumor ratio, morphology, and longitudinal growth. Translation therefore requires reconstruction and segmentation robustness testing, strong conventional comparators, external calibration, prespecified thresholds, and prospective decision-linked evaluation. Until such benefit and safety are demonstrated, reconstruction-aware radiomics should remain investigational or confined to governed, domain-checked decision-support pathways.
SN - 2296-858X
ER -