Prediction Model Of The Value Of CT Value In The Diagnosis And Differential Diagnosis Of Pulmonary Cryptococcosis

Objective: The incidence of pulmonary cryptococcosis(PC)is gradually increasing,and because it has no specific imaging manifestations,it is easily confused with other types of pulmonary occupying lesions,and it is important to carry out diagnostic prediction of the occurrence of PC patients for early clinical diagnosis and treatment.The purpose of this study was to investigate the value of the application of CT values in the identification and diagnosis of PC risk factors.The validity of the model was evaluated while constructing a clinical prediction model of relevant risk factors.Methods: This study was conducted by collecting clinical data from 73 patients who presented with nodule/mass type occupancy on CT of lung and confirmed by histopathology from January 2019 to May 2022 at the First Affiliated Hospital of Anhui Medical University.Based on the pathological findings,the group was divided into a pulmonary cryptococcosis group(23 cases)and a non-pulmonary cryptococcosis group(50 cases),which were retrospectively analyzed,and the corresponding values of each region of tissue on the CT images of each patient’s lesion with comparable X-ray attenuation coefficients,that is,CT values,were determined.The two groups were compared for differences in age,sex,symptoms,lesion involvement in one/both lungs,lung lobe distribution,number of lesions,maximum lesion diameter,lesion margin condition,and CT value results.For the indicators with statistically significant differences,independent risk factors for pulmonary cryptococcosis were analyzed,clinical prediction models were constructed and plotted in columns,and the validity of the prediction models was assessed by calculating C(correction)indices,plotting calibration curves,subject work characteristic(ROC)curves and clinical decision curve analysis.Results: Univariate and multivariate logistic regression analyses showed no significant differences in gender,age,clinical presentation,distribution,size and morphology of the lesions in PC compared with pulmonary malignancies,pulmonary malignancies,tuberculosis and other lung infections(P>0.05).Univariate analysis showed statistically significant differences(P<0.05)in the central CT values(P=0.0049),peripheral CT(P=0.0003)and overall CT(P=0.0208)values of the lesions measured in both groups,suggesting that the central CT,peripheral CT and overall CT values of the lesions could be independent risk factors for the diagnosis and differential diagnosis of PC.The results of the Nomo prediction model created based on the above factors showed that the area under the ROC curve(AUC)for pulmonary cryptococcosis was 0.814(95%interval: 0.7011-0.9267),and the corrected C-index(Bootstrap=1000)was 0.781;the calibration curve suggested that the overlap between the predicted and actual curves was relatively good;the results of the decision curve analysis suggested The value of the column line plot model for clinical application is high,showing that the net benefit of using column line plots to predict the risk of diagnosis in patients with pulmonary cryptococcosis is higher when the threshold probability of the patient is 0-1.0.Conclusions: The CT value of the lesion can be used as an independent risk factors for pulmonary cryptococcosis.The clinical prediction model based on the above factors has certain predictive effect on the diagnosis and differential diagnosis of PC,which is helpful to distinguish PC from other lung space-occupying lesions.