Incorporation of whole-body metabolic tumor burden into current prognostic models for non-small cell lung cancer patients with spine metastasis

Numerous prognostic models are utilized for surgical decision and prognostication in metastatic spine tumors. However, these models often fail to consider the whole-body tumor burden into account, which may be crucial for the prognosis of metastatic cancers. A potential surrogate marker for tumor burden, whole-body metabolic tumor burden (wMTB), can be calculated from total lesion glycolysis (TLG) obtained from 18F-Fludeoxyglucose positive emission tomography (18F-FDG PET) images. We aimed to improve prognostic power of current models by incorporating wMTB for non-small cell lung cancer (NSCLC) patients with spine metastases. Retrospective analysis using a review of electrical medical records and survival data. In this study, we included 74 NSCLC patients with image proven spine metastases. Increase in Integrated Discrimination Improvement (IDI) index after incorporation of wMTB into prognostic scores. Enrolled patients’ baseline data, cancer characteristics and survival status were retrospectively collected. Five widely used prognostic scores (Tomita, Katagiri, Tokuhashi, Global Spine Tumor Study Group [GSTSG], New England Spine Metastasis Score [NESMS]), and TLG indexes were calculated for all patients. The relationships among survival time, prognostic models and TLG values were analyzed. Improvement of prognostic power was validated by incorporating significant TLG index into significant current models. Among current prognostic models, Tomita (EGFR wild-type), Katagiri, GSTSG and Tokuhashi were significantly related to patient survival. Among TLG indexes, LogTLG3 was significantly related to survival. Incorporation of LogTLG3 into significant prognostic models resulted in positive IDI index until 3 years in all models. This study showed that incorporation of wMTB improved prognostic power of current prognostic models of metastatic spine tumors.