Construction and Verification of Nomogram Model for Lung Adenocarcinoma With ≤ 5 Bone-Only Metastases Basing on Hematology Markers

OBJECTIVES: This retrospective study investigated prognostic factors in advanced lung adenocarcinoma (LUAD) with one to five bone-only metastasis (BOM) and developed a nomogram model to estimate patient survival. METHODS: We investigated patients with advanced LUAD with one to five bone-only metastasis at the initial diagnosis and diagnosed between 2013 and 2019 in two hospitals. A formula named Risk-H was constructed using hematological variables screened by LASSO-Cox regression analysis in the internal set and verified by the external set. Two nomogram models were developed by clinical variables selected by LASSO-Cox regression analysis with or without Risk-H in the internal set. The concordance index (C-index), calibration curves, time-dependent receiver operating characteristic (ROC) analysis, area under the curve (AUC), and decision curve analysis (DCA) were formulated to verify nomogram models. The primary endpoint was overall survival. RESULTS: We finally included 125 and 69 patients, respectively, in the internal and external sets for analysis. The following were significant hematology prognostic factors and were included in the Risk-H formula: alkaline phosphatase and albumin, leukocyte. Four clinical factors, including loss of weight, sensitive mutation status, T and N stage, with or without Risk-H were used to establish nomogram models. C-index, calibration curves, ROC analysis, AUC, and DCA showed the addition of hematological data improved the predictive accuracy of survival. CONCLUSIONS: Pretreatment peripheral blood indexes may be a meaningful serum biomarker for prognosis in LUAD. The addition of Risk-H to the nomogram model could serve as a more economical, powerful, and practical method to predict survival for LUAD patients with one to five BOM.