| Background and objectiveLung cancer is the malignancy with the highest morbidity and mortality rate among all types of tumors.With the increase in low-dose spiral CT examinations,more and more peripheral lung nodules are being detected unexpectedly.Since 30%of early stage lung cancers are located in the peripheral third of the lung,an area that is difficult to reach with conventional bronchoscopy,the diagnostic efficacy of ultrafine bronchoscopy remains low because of the complex subsegmental bronchial structure and limited sampling specimens,even though smaller outside diameter ultrafine bronchoscopes have been developed.The emergence of electromagnetic navigation bronchoscopy(ENB),which is based on electromagnetic positioning technology,combined with virtual bronchoscopy,three-dimensional computed tomography imaging,and three-dimensional computed tomography imaging,has provided us with a new help.Electromagnetic navigation bronchoscopy is a new type of bronchoscopy based on electromagnetic positioning technology combined with virtual bronchoscopy,3D computed tomography imaging and respiratory gating technology.In this paper,we analyzed and compared the diagnostic efficacy and safety of electromagnetic navigation combined with radial ultrasound biopsy for the diagnosis of small peripheral pulmonary nodules with the diagnosis of ENB combined with X-ray,radial ultrasound and AI.This paper aims to provide guidance for the clinical application of electromagnetic navigation tracheoscopy by analyzing and comparing the diagnostic efficacy of electromagnetic navigation combined with X-ray,radial ultrasound and AI in the diagnosis of peripheral pulmonary nodules and its influencing factors.MethodsFifty-six patients with peripheral pulmonary nodules of less than 3 cm in diameter who underwent bronchoscopy at the First Affiliated Hospital of Soochow University and Dushu Lake Hospital of Soochow University from February 2019 to January 2022 were consecutively selected as the study subjects,including 24 patients who underwent ENB combined with X-ray and 32 patients who underwent ENB combined with rEBUS.Both groups used ENB as the guiding method,and the X-ray group combined with C-arm-X-ray performed biopsy and brush examination of pulmonary nodules under fluoroscopic guidance,and the rEBUS group combined with rEBUS performed biopsy and brush examination of pulmonary nodules after determining whether the lesion was reached.Using the results of surgery,puncture pathology,or clinical follow-up one year as the gold standard,the diagnostic sensitivity,specificity,negative predictive value(NPV),diagnostic coincidence rate,negative likelihood ratio(LR-),Youden index,missed diagnosis rate,success rate,and kappa value were compared between the two groups,and the occurrence of postoperative complications was also compared between the two groups.A multifactorial analysis of factors that may affect diagnostic efficacy such as lesion diameter and the presence of bronchial signs was performed.ResultsThe mean age of the 56 patients was 58.75 ± 12.77 years,32 were male and 24 were female,36 were never smokers and 20 were former or current smokers,and the final diagnosis of the patients was adenocarcinoma in 19 cases,squamous carcinoma in 6 cases,non-small cell carcinoma in 3 cases,metastatic carcinoma in 2 cases,lymphoma in 1 case,inflammation in 23 cases,fungal infection of the lung in 1 case,and non-tuberculous branching bacillus infection in 1 case.There was no significant difference in gender,age,height,weight and smoking history between the two groups(P>0.05),and the mean lesion size was 20.77 ± 9.05 mm,≤20 mm in 27 cases and>20 mm in 29 cases.rEBUS group had significantly smaller lesion diameters(P<0.001).The sensitivity,specificity,negative predictive value,diagnostic compliance,negative likelihood ratio,Youden index,miss rate,success rate and Kappa value in the X-ray group were 47.4%(9/19),100%(5/5),33.33%(5/15),58.33%(14/24),0.526,0.47,52.6%(10/19),respectively.95.8%(23/24),and 0.273.sensitivity,specificity,negative predictive value,diagnostic compliance,negative likelihood ratio,Youden index,miss rate,success rate,and kappa value in the rEBUS group were 50%(6/12),100%(20/20),76.92%(20/26),81.25%(26/32),respectively.0.5,0.5,50%(6/12),87.5%(28/32)and 0.556.There was no statistical difference in sensitivity(P>0.999),diagnostic compliance(P=0.060)and success rate(P = 0.279)between the two methods for peripheral type of nodules.negative predictive value was significantly better in the rEBUS group than in the X-ray group(P=0.006).The regression dominance ratio of lesion diameter on diagnostic efficacy in both groups was 4.059.ConclusionENB combined with rEBUS and X-ray had comparable diagnostic compliance,and ENB combined with rEBUS obtained a higher negative predictive value than X-ray.Lesion size may be a key factor affecting the diagnostic rate of ENB.Preliminary investigation of the diagnostic value of AI combined with ENB for peripheral pulmonary nodulesBackground and objectiveWith the development of artificial intelligence technology,AI-aided diagnostic software is being more and more widely used in medical imaging.Through artificial intelligence algorithms,medical images can be denoised,enhanced,segmented and aligned to improve doctors’ diagnostic ability of images.Using deep learning algorithms,tumor detection and analysis can be performed on medical images to help doctors improve diagnostic accuracy and efficiency.There are more studies on the application of AI software in the diagnosis of pulmonary nodules,and not many studies have combined it with electromagnetic navigation bronchoscopy,so it will be added as an auxiliary ENB tool in the study to explore the diagnostic value of both for peripheral pulmonary nodules.MethodsAI software analysis of pulmonary nodules was implemented byσ-Discover/SigmaCloud image processing software of pulmonary nodule assisted diagnosis cloud platform provided by Zhuhai Shengmei Bio-diagnostic Technology Co.The CT images(0.5-1.5 mm layer thickness)in dicom format of the first part of the patients were imported into the AI software,and all nodules were classified into benign and malignant groups according to the gold standard,and the subject operating characteristic(ROC)curves of AI software,electromagnetic navigation bronchoscopy combined with X-ray or radial ultrasound biopsy were plotted,and the area under the curve(AUC)was calculated.The combined diagnosis of AI software and electromagnetic navigation bronchoscopy combined with X-ray or radial ultrasound biopsy was achieved by building a binary logistic regression model and creating ROC curves for the model predicted values.ResultsThe sensitivity and specificity of AI software for the correct diagnosis of malignant lesions(primary lung cancer,metastatic lung cancer)were 51.6%,80%,respectively.The positive and negative predictive values for malignancy diagnosis were:76.2%,57.1%,respectively.the AUC value of AI software combined with electromagnetic navigation bronchoscopy biopsy for the diagnosis of peripheral lung nodules was 0.841,with a sensitivity of 58.1%and specificity of 100%for the diagnosis at the best cut-off value,both of which were statistically significant compared to AUC diagnosed by both alone(P=0.001;P=0.037).ConclusionAI software combined with ENB helps to improve the positive diagnosis rate of peripheral lung nodules. |
PDF Full Text Request