Objective To explore the influence of artificial intelligence-assisted diagnosis on the detection of pulmonary nodules and the differentiation between benign and malignant.Methods A total of 128 patients with pul-monary nodules admitted to our hospital from July 2022 to May 2023 were examined by CT.The results were judged by artificial intelligence and imaging doctor's reading,Taking the pathological results of operation or puncture biopsy as the gold standard,the detection results,diagnostic efficiency,and diagnostic value of two ways of reading films for pulmonary nodules and nodules with different diameters were compared.Results A total of 216 nodules were detect-ed in 128 patients,of which 145 were malignant and 71 were benign by pathological examination.A total of 120 real malignant nodules were detected by artificial intelligence,and 75 real malignant nodules were detected by imaging doctors.The diagnostic sensitivity,accuracy,and negative predictive value of artificial intelligence film reading were higher than that of imaging doctors.The diagnostic specificity of artificial intelligence reading film was lower than that of imaging doctor reading film.141 positive nodules were detected by artificial intelligence reading,and 85 positive nodules were detected by imaging doctors reading.The two methods were compared,the detection rate of artificial in-telligence reading was higher for nodules with diameters less than 5 mm and 5 mm≤ 10 mm.The detection rate of ar-tificial intelligence reading was higher(P＜0.05).Conclusion Artificial intelligence-assisted diagnosis has higher de-tection rate and higher sensitivity,accuracy,and negative predictive value,but its specificity is low.It is suggested that artificial intelligence and imaging doctor's diagnoses should be combined in clinic to improve the diagnostic ac-curacy of pulmonary nodules.
关键词
人工智能/肺结节/良恶性
Key words
Artificial intelligence/Pulmonary nodules/Benign and malignant
维普
万方数据