Nomogram based on homogeneous and heterogeneous associated factors for predicting bone metastases in patients with different histological types of lung cancer.

ABSTRACT

BACKGROUND: The purpose of the present study was to characterize the prevalence, associated factors, and to construct a nomogram for predicting bone metastasis (BM) with different histological types of lung cancer. PATIENTS AND METHODS: This study was a descriptive study that basing on the invasive lung cancer patients diagnosed between 2010 and 2014 in Surveillance, Epidemiology, and End Results program. A total of 125,652 adult patients were retrieved. Logistic regression analysis was conducted to investigate homogeneous and heterogeneous factors for BM occurrence. Nomogram was constructed to predict the risk for developing BM and the performance was evaluated by the receiver operating characteristics curve (ROC) and the calibration curve. The overall survival of the patients with BM was analyzed using the Kaplan-Meier method and the survival differences were tested by the log-rank test. RESULTS: A total of 25,645 (20.9%) were reported to have BM, and the prevalence in adenocarcinoma, squamous cell carcinoma, small cell lung cancer (SCLC), large cell lung cancer (LCLC), and non-small cell lung cancer/not otherwise specified lung cancer (NSCLC/NOS) were 24.4, 12.5, 24.7, 19.5 and 19.4%, respectively, with significant difference (P < 0.001). Male gender, more metastatic sites and lymphatic metastasis were positively associated with BM in all lung cancer subtypes. Larger tumor size was positively associated with BM in all the lung cancer subtypes except for NSCLC/NOS. Poorly differentiated histology was positively associated with adenocarcinoma, squamous cell carcinoma and NSCLC/NOS. The calibration curve and ROC curve exhibited good performance for predicting BM. The median survival of the bone metastatic lung cancer patients was 4.00 (95%CI 3.89-4.11) months. With the increased number of the other metastatic sites (brain, lung and liver metastasis), the survival significantly decreased (p < 0.001). CONCLUSION: Different lung cancer histological subtypes exhibited distinct prevalence and homogeneity and heterogeneity associated factors for BM. The nomogram has good calibration and discrimination for predicting BM of lung cancer.