The Application Value Of The Scoring System Based On Endobronchial Ultrasound Image In The Diagnosis Of Pulmonary Nodules

Background The morbidity and mortality of lung cancer have long been the first cancer in China,and its prognosis is poor.Early stage lung cancer usually appears as pulmonary nodules.At present,there are few studies on the correlation between the ultrasonic image features of pulmonary nodules and the nature of pulmonary nodules,and no ultrasonic image feature of pulmonary nodules has been included in the prediction model.Objective This paper aims to retrospectively analyze the ultrasonographic characteristics of pulmonary nodules,search for the risk factors of benign and malignant pulmonary nodules,establish the malignant and predictive model of pulmonary nodules,and establish a scoring system based on endobronchial ultrasonography(EBUS)images to distinguish benign and malignant pulmonary nodules.Methods data of patients who undergone bronchoscopy in the Bronchoscopy Room of XX Hospital during May 1,2018 and March 1,2020 were retrospectively included.all of the patients’ chest CT scan showed pulmonary nodules<3cm,while routine bronchoscopy showed no obvious abnormalities.The morphological features of the images were recorded at the lesions under the guidance of the ultrasound probe,and biopsy and brush examination were performed to assist the diagnosis.Patients with no definite diagnosis were further subjected to invasive surgeries or observations for at least 3 months to obtain the final diagnosis.Eight different EBUS image characteristics including lesion shape,margin,border,internal echo,homogeneous or heterogeneous internal echoe,air bronchogram,irregular anechoic area,concentric circles were observed,and the relationship between these image characteristics and the nature of pulmonary nodules were analyzed,the prediction model and a simple scoring system were established.datas were analyzed using SPSS software.Results Among 114 patients with pulmonary nodules,65 were benign and 49 were malignant.Three image features in EBUS images,including round or round shape of lesions,discontinuous margin,and heterogeneous echogenicity,with statistically significant(p<0.05).The edge of the lesion,homogeneous or heterogeneous echo of the lesion and serum CEA were independent risk factors for predicting benign and malignant pulmonary nodules.Logistic prediction model based on this model has high sensitivity(81.63%)and specificity(80.00%).According to the simple scoring system established,ROC curve was drawn.When the score was set at 7,the sensitivity(65.3%)and specificity(78.5%)were the highest,and this was the best point for diagnosis.When the score is>7,the diagnosis of pulmonary malignant lesions was more accurate.Kappa consistency was calculated with a coefficient of 0.441(95%CI 0.274-0.607,p<0.01),showing moderate strength consistency.Conclusion EBUS can provide image characteristic information to differentiate the nature of pulmonary nodules,and the scoring system based on the images has a good application value in differentiating benign from malignant pulmonary nodules.