Background
Primary pulmonary osteosarcoma (PPOS) is an uncommon malignancy originating from the lung with low incidence, and its clinical characteristics and prognosis have not been systematically reported. Therefore, we aimed to recognize the prognostic factors and constructed a survival prediction model for PPOS.
Methods
We collected the data from the Surveillance, Epidemiology, and End Results database and systematic review of previous studies. Demographical and clinical characteristics, radiographic manifestations, treatment modalities, and prognosis were analyzed. A prediction model via nomogram was constructed and then evaluated by the concordance index (C-index) and the receiver operating characteristic (ROC) curve.
Results
A total of 49 cases were included for analysis with a median age of 67 years old (range 33–94 years), of which 32 (65.3%) were male. The median size was 6 cm (range 1.8-25 cm), and the median overall survival (OS) was eight months (interquartile range 4.5–12 months) with a 1-year OS rate of 30.8%. Tumor size over 7 cm (hazard ratio [HR] = 2.98; P = 0.018) and those without microscopic findings of osteoid found in the tumors (HR = 2.11; P = 0.048) were referred to a poor OS, while surgery was associated with an improved OS (HR = 0.20; P < 0.001). The C-index of the nomogram prediction model was 0.771, and the area under curve, sensitivity and specificity of the ROC curve were 0.818, 0.848 and 0.800, respectively.
Conclusions
Patients with PPOS had a poor prognosis, and tumor size was mostly prognostic. Surgery seemed to be an effective treatment, and the prediction model with a nomogram in our study could effectively predict the prognosis of patients with PPOS.