| Background:Lung cancer is a malignant tumor with the highest morbidity and mortality,which seriously endangers people’s health.Therefore,it is of great significance to improve the early diagnosis of lung cancer to reduce the mortality and prolong the survival of patients.Studies have shown that screening high-risk populations with low-dose computed tomography(LDCT)can timely detect early stage lung cancer and reduce the mortality rate of lung cancer patients by about 20%.However,when the survival rate of lung cancer patients is improved,the detection rate of pulmonary nodules is also greatly increased.The subsequent problem is how to effectively distinguish the benign and malignant pulmonary nodules.At present,with the progress of science and technology,artificial intelligence(AI)image assisted diagnosis system can find lung nodules in LDCT more efficiently,and determine the nature of nodules by extracting and analyzing nodule image information.However,the imaging features of early lung cancer are usually not typical,and the AI-assisted diagnosis alone still has its limitations.Liquid biopsy also holds great potential in the early diagnosis of lung cancer.Anderson Cancer Center found that cells with circulating chromosomal abnormalities(CACs)were cells with tumor-specific chromosomal loci in the peripheral blood and were positively expressed in lung cancer.This paper discusses the diagnostic value of the combined detection of benign and malignant pulmonary nodules by circulating chromosomal abnormal cells(CACs)and AI image-assisted diagnostic system,and provides a basis for the selection of clinical detection methods.Methods:In this study,a total of 77 patients with isolated pulmonary nodules whose lung lesions were less than30 mm by chest CT were enrolled in our hospital from November 2020 to November 2021.Before surgical resection or tissue biopsy(tracheoscopy,lung puncture,etc.),10 m L peripheral blood was extracted from patients in the group,and the number of CACs in peripheral blood samples was detected and counted by fluorescence in situ hybridization(FISH).At the same time,chest CT scan images and serum tumor markers(CEA,NSE,SCC)levels were collected,and demographic characteristics were analyzed.Chest CT images were imported into AI image assisted diagnosis system to evaluate the malignant probability of nodules.To compare the sensitivity and specificity of single detection and combined detection of benign and malignant pulmonary nodules based on CACs and AI image assisted diagnosis system.The diagnostic effect was analyzed by area under ROC curve(AUC).Results:1.There was no correlation between the nature of pulmonary nodules and gender,smoking history,family history of tumors,or history of chronic lung disease,location of pulmonary nodules,pulmonary nodule type,and burr sign(P>0.05);Correlation with age and nodule diameter(P<0.05).2.There was no correlation between the nature of pulmonary nodules and the expression level of tumor markers,with no statistical significance(P>0.05).3.The expression level of circulating chromosome abnormal cells(CACs)was the median,which was1.51 in the benign pulmonary nodules group and 3.53 in the non-benign pulmonary nodules group,and there were significant differences between the two groups(P<0.05).4.ROC curve of circulating chromosomal abnormal cells(CACs)was drawn to distinguish benign and malignant pulmonary nodules.The results showed that the AUC value was 0.8138,the cut-off value was 3,and the sensitivity and specificity were 78.37% and 83.07%,respectively.5.The ROC curve of circulating chromosomal abnormal cells(CACs)was drawn to distinguish benign and malignant pulmonary nodules of different diameters.The results showed that when d≤10mm,the AUC value was 0.6810,the sensitivity and specificity were 75.82% and 71.01%,respectively.When10 mm < d≤20mm,the AUC value was 0.8720,the sensitivity and specificity were 79.08% and 90.34%,respectively.When 20 mm < d≤30mm,the AUC value was 0.8961,the sensitivity and specificity were80.59% and 92.65%,respectively.All the three groups had good diagnostic value,and the larger the nodule diameter,the higher the diagnostic efficiency.6.The ROC curve of AI image-assisted diagnosis system was drawn to distinguish benign and malignant pulmonary nodules.The results showed that AUC value was 0.7781,cut-off value was 74.80%,sensitivity and specificity were 81.51% and 76.42%,respectively.7.Consistency analysis of AI image-assisted diagnosis system and pathological test results showed that Kappa value =0.752,and Kappa test had good consistency(P < 0.05).8.ROC curves were drawn to distinguish benign and malignant pulmonary nodules by AI image-assisted diagnosis system.The results showed that for solid nodules,AUC value was 0.812,sensitivity and specificity were 82.02% and 80.51%,respectively.For partial solid nodules,the AUC value was 0.840,the sensitivity and specificity were 83.03% and 81.02%,respectively.When grinding glass nodules,the AUC value was 0.790,the sensitivity and specificity were 88.5% and 74.71%,respectively.All three groups have good diagnostic value.The diagnostic efficacy of partial solid nodules is particularly significant.9.The ROC curve of circulating chromosomal abnormal cells(CACs)combined with AI image-assisted diagnosis system was drawn to distinguish benign and malignant pulmonary nodules.The results showed that the AUC value was 0.8362,and the sensitivity and specificity were 81.61% and 84.42%,respectively.Combined detection is better than single detection.Conclusion:1.When circulating chromosomal abnormal cells(CACs)≥3,it is of high value in the diagnosis of malignant pulmonary nodules,and CACs may be a new biomarker to distinguish benign and malignant pulmonary nodules.2.Artificial intelligence(AI)imaging-assisted diagnostic system has good sensitivity and specificity for the differentiation of benign and malignant pulmonary nodules,especially in some solid nodules.3.For the judgment of benign and malignant pulmonary nodules,the combined detection of circulating chromosomal abnormal cells(CACs)and AI image-assisted diagnosis system is superior to the single detection,which is expected to become a new model for the diagnosis of pulmonary nodules. |
PDF Full Text Request