Abstract

<h3>Purpose/Objective(s)</h3> Brain metastases are frequent complications for lung cancer patients. The aim of this paper was to explore whether there were alterations in brain glucose uptake and brain texture in lung cancer patients, as well as to investigate possible associations between brain alterations and lung cancers via the state-of-the-art total-body PET/CT. <h3>Materials/Methods</h3> Forty-four subjects including 27 healthy controls and 17 lung cancer patients were enrolled in this study. Lung cancer patients were tested for serum lung cancer markers including carcinoembryonic antigen (CEA), neuron specific enolase (NSE), pro-gastrin-releasing peptide (ProGRP), cytokeratin-19-fragment (CYFRA21-1) and squamous cell carcinoma (SCC). 2-deoxy-2-[fluorine-18] fluoro-D-glucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) images were acquired using the uEXPLORER PET/CT system. Standardized uptake value normalized by lean body mass (SUL) was calculated as indicators of brain glucose uptake. Differences in brain SUL maps between lung cancer patients and healthy controls were assessed. Correlation analysis was conducted between aberrant brain glucose uptake, SUL values of cancer lesions, and serum concentrations of lung cancer markers. In addition, we performed PET/CT radiomics to investigate whether radiomic features from the regions with aberrant glucose uptake were predictive in distinguishing lung cancer from healthy controls. Specifically, we extracted 86 radiomic features from each brain regions using an open-source python package open-source software, including the intensity features and texture features. Recursive feature elimination was used to eliminate redundant features. The selected features were used by a linear support vector machine (SVM) classifier. <h3>Results</h3> Compared with healthy controls, lung cancer patients showed increased SUL values in the right middle temporal gyrus (MTG), left brainstem, bilateral lentiform nucleus and right putamen. Correlation analysis showed that the serum level of CYFRA21-1 was negatively correlated with SUL values of the right MTG, left brainstem and right putamen. Radiomic features extracted from the above four brain regions could accurately classify lung cancer patients and healthy controls with an accuracy of 100%. Furthermore, radiomic features extracted from brain regions with aberrant glucose uptake were more predictive than radiomic features from the entire brain. <h3>Conclusion</h3> Lung cancer patients experienced altered brain glucose uptake and changes in brain textures compared with healthy controls. These alterations may reflect possible lung-brain interactions and potential brain metastasis of lung cancer.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call