Abstract
PURPOSE: Prognostication of lung cancer patients remains challenging clinically. This study aims to address this problem by investigating the utility of bone glucose metabolism networks as a prognostic biomarker of lung cancer patients. The rationale for targeting bone metabolic networks specifically comes from the long-recognised role of bone in innate immunity and the recent appreciation of the key role bones play in regulating whole-body glucose metabolism. METHODS: This study is a retrospective analysis of data from the multi-centre trial ACRIN 6668 on non-small cell lung carcinoma (NSCLC). Conventional [(18)F]Fluorodeoxyglucose (FDG) Positron Emission Tomography (PET) standardised uptake value (SUV) analysis of lung lesions and bones as well as network analysis of bones were carried out in 34 stage IIIB non-operable NSCLC patients before and after chemoradiation. RESULTS: Conventional lung tumour SUV peak analysis of PET data cannot predict NSCLC patient survival (p = 0.23-0.35), while network analysis of bone glucose metabolism pre-chemoradiation can significantly predict patient survival (p = 0.0003) and identify three distinct clusters of survivors (short-term, long-term and mix-term). Chemoradiation treatment results in homogenisation of the innate survivors' clusters into adapted mix-term clusters and loss of predictive value of bone glucose metabolism networks (p = 0.77). CONCLUSIONS: Innate pre-treatment skeletal glucose metabolism networks of NSCLC patients can predict patient survival in contrast with adapted post-treatment networks, supporting the role of bones in both innate immunity and regulation of systemic glucose metabolism. Newly proposed network analysis performs better than conventional SUV PET analysis, and should be considered as a prognostication tool in the management of NSCLC patients.