Independent predictors and prediction model of malignant micro-sized solitary pulmonary nodules
10.3969/j.issn.1000-8179.2018.10.162
- VernacularTitle:恶性孤立性肺微小结节的独立预测因子及预测模型
- Author:
Ying ZHU
1
,
2
;
Panfeng XU
;
Yake YAO
;
Jianfang PAN
;
Jianying ZHOU
Author Information
1. 浙江大学附属第一医院呼吸内科 杭州市310003
2. 浙江金华广福医院呼吸内科
- Keywords:
solitary pulmonary nodules;
Logistic regression analysis;
prediction model
- From:
Chinese Journal of Clinical Oncology
2018;45(10):497-502
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To evaluate the clinical factors affecting the probability of malignant micro-sized (≤10 mm) solitary pulmonary nodules (≤10 mm, micro-sized SPN), and established a clinical prediction model. Methods:Medical records from 102 patients with a pathological diagnosis of micro-sized SPN (Group A), established between June 2012 and March 2014, were reviewed. Clinical data were collected to evaluate the independent predictors of malignant micro-sized SPN using single factor analysis and Logistic regression analysis. A clinical prediction model was subsequently created. Receiver-operating characteristic (ROC) curves were constructed using the prediction model. Between January 2015 and August 2017, data from an additional 10 patients enrolled from January 2015 to August 2017 from Jinhua Guangfu Hospital (Group B) with a pathological y diagnosed micro-sized SPN were used to validate this clinical prediction model. The model was also compared with the Mayo Clinic Model. Results:Median age of 102 patients (Group A) was 55.31±10.77 years old. There were 75.5％malignant nodules and 24.5％benign ones. Logistic regression analysis identified six clinical characteristics (no symptoms, upper lobe, diameter>5 mm, no clear border, not irregular round, no calcification) as independent predictors of malignancy in patients with micro-sized SPN. The area under the ROC curve for our model was 0.922 (95％CI:0.857-0.986). In our model, the diagnosis sensitivity and specificity were 88.3％and 84.0％, respectively. The test power of the model was better compared with the Mayo Clinic Model. Conclusions:In this study, we had found the independent predictors of malignant micro-sized SPN, and developed a prediction model that could accurately identify malignant micro-sized SPN in patients.
