Abstract
BACKGROUND: Bone metastasis considerably undermines the prognosis of advanced primary liver cancer patients. Though its impact is well-recognized, the clinical field still lacks robust predictive models that can accurately forecast patient outcomes and aid in treatment effectiveness evaluation. Addressing this gap is paramount for improving patient management and survival. MATERIALS AND METHODS: We conducted an extensive analysis using data from the SEER database (2010-2020). COX regression analysis was applied to identify prognostic factors for primary liver cancer with bone metastasis (PLCBM). Nomograms were developed and validated to predict survival outcomes in PLCBM patients. Additionally, propensity score matching and Kaplan-Meier survival analyses lent additional insight by dissecting the survival advantage conferred by various treatment strategies. RESULTS: A total of 470 patients with PLCBM were included in our study. The median overall survival (OS) and cancer-specific survival (CSS) for these patients were both 5 months. We unveiled several independent prognosticators for OS and CSS, spanning demographic to therapeutic parameters like marital status, cancer grade, histological type, and treatments received. This discovery enabled the formulation of two novel nomograms-now verified to eclipse the predictive prowess of the traditional TNM staging system regarding discrimination and clinical utility. Additionally, propensity score matching analysis showed the effectiveness of surgeries, radiotherapy, and chemotherapy in improving OS and CSS outcomes for PLCBM patients. CONCLUSIONS: Our investigation stands out by introducing pioneering nomograms for prognostic evaluation in PLCBM, a leap forward compared to existing tools. Far exceeding mere academic exercise, these nomograms hold immense clinical value, serving as a foundation for nuanced risk stratification systems and delivering dynamic, interactive guides, allowing healthcare professionals and patients to assess individual bone metastasis survival probabilities and personalize treatment selection.