人工智能阅片对早期肺癌的诊断效果

The Diagnostic Effect of Artificial Intelligence Film Reading on Early Lung Cancer

摘要:

〔摘 要〕 目的：比较人工阅片与人工智能阅片诊断早期肺癌的效果。 方法：选取 2020 年 6 月至 2021 年 6 月在 赣州市人民医院行肺部计算机断层扫描（CT）检查的疑有肺癌的 65 例患者作为研究对象，患者的 CT 检查结果均 经过放射诊断医师和人工智能系统两种方式阅片，比较两种阅片方式的阅片用时、诊断效能及与金标准（病理检查 或手术结果）的一致性。 结果：人工智能阅片对每位患者的平均阅片用时为（0.62 ± 0.13）min，短于人工阅片的 （4.83 ± 0.34）min，差异具有统计学意义（P ＜ 0.05）；人工智能阅片的灵敏度、阴性预测值、准确度以及与金标准 的一致性均低于人工阅片，差异具有统计学意义（P ＜ 0.05）。 结论：相比于人工阅片，人工智能阅片的阅片用时显 著更快，但诊断效能更低，临床上将两种阅片方式相结合，可以有效提高早期肺癌的诊断效率。

Abstract:

〔Abstract〕 Objective To compare the effect of manual and artificial intelligence film reading in the diagnosis of early lung cancer. Methods Sixty-five patients with suspected lung cancer who underwent lung computed tomography (CT) examination in Ganzhou People's Hospital from June 2020 to June 2021 were selected as the research objects. CT examination results of patients were reviewed by radiologists and artificial intelligence system.The time, diagnostic efficacy and consistency with the gold standard (pathological examination or surgical results) of the two methods were compared. Results The average tablet reading time of artificial intelligence for each patient was (0.62 ± 0.13) min, which was shorter than that of manual tablet reading (4.83 ± 0.34) min, the difference was statistically significant (P < 0.05). The sensitivity, negative predictive value, accuracy and consistency with gold standard of artificial intelligence film reading were all lower than manual film reading, and the differences were statistically significant (P < 0.05). Conclusion Compared with manual film reading, the time of artificial intelligence film reading is significantly faster, but the diagnostic efficiency is lower. Clinical combination of the two methods of film reading can effectively improve the diagnostic efficiency of early lung cancer.