| Objective: To eastablish a model to estimate the probability of malignancy in solitary pulmonary nodules (SPNs) and evaluate its test effectiveness in the differential diagnosis in solitary pulmonary nodules (SPNs).Patients and Methods: A retrospective study at a multispecialty group practice included 339 patients with newly discovered (between October,2006 and October,2009) radiologically solitary pulmonary nodules on chest radiography. Clinical data,includeing sex,age,cigarette-smoking status and history of extrathoracic malignant neoplasm; chest radiological data, including the size,location,edge characteristics,vacuole sign, enhanced CT value,spiculation,lobulation,pleural indentation and calcification,and the final pathological diagnosis of the SPNs were recorded and assigned with value.We define the final pathological diagnosis as dependent variable,and the clinical characteristics as well as pathological characteristics as independent variable. Predictors were identified by multivariate logistic regression analysis and a concise model was established to calculate and estimate the probability of malignancy solitary pulmonary nodules. A ROC(receiver operating curve) was subsequently obtained to study the value of the model in differential diagnosis of SPNs.Results: 188 cases,55.5% of the nodules,were malignant,and 151 cases,44.5% of the nodules were benign.A clinical characteristics(age), and 3 radiological characteristics (enhanced CT value,spiculation and lobulation of the SPNs)were independent predictors of malignancy with their odds ratio (OR) and 95%CI are2.252 (95% CI: 1.605-3.160),6.534 (95%CI: 3.348-12.75),2.260 (95%CI:1.257-4.065) and 7.514 (95%CI:3.882-14.543) respectively.A concise model was established as following: P=1/(1+e-x), X=-4.487+0.812*age+1.877*enhanced CT value+0.816*spiculation+2.017*loubulation. where e is the base of the natural logarithm,age is age in years divided by 10 into 4 groups:...
|
PDF Full Text Request