AI定量分析肺磨玻璃结节与病理相关性研究

Pathological relevance of AI-based quantitative analysis of pulmonary ground-glass nodules

目的:主要讨论人工智能(artificial intelligence,AI)对肺磨玻璃结节(ground-glass nodule,GGN)定量分析结果与病理相关性。方法:选取2017年5月—2022年3月于龙口市人民医院和青岛大学附属医院诊断和手术治疗的152例(其中女性104例,男性48例)GGN患者,共187个GGN。利用AI设备筛查并计算肺结节的定量参数,与手术后病理结果进行对照。结果:AI测量的最大直径、体积、最大密度、平均密度、恶性概率在不同病理类型间存在显著性差异,而AI测量的最小密度在不同病理类型间无显著性差异。结论:AI定量分析肺磨玻璃结节可以帮助区分结节的病理类型。

Objective To explore the pathological relevance of quantitative analysis of pulmonary ground-glass nodules(GGN)based on artificial intelligence(AI).Methods 152 patients(104 female and 48 male)diagnosed and operated in Longkou People's Hospital from May 2017 to March 2022 were selected,including 187 lesions.By using AI,the CT data were evaluated to measure quantitative parameters of CGN and compared with postoperative pathological results.Results By using AI,there were significant differences of major axis,volume,maximum density,mean density and probability of malignancy between different pathological types.However,there was no significant difference with minimum density.Conclusion The quantitative analysis based on AI may be helpful to identify the pathological type of pulmonary GGN.