2022 World Conference on Lung Cancer (ePosters)-EP13.01-005. Role of Artificial Intelligence on Chest Radiographs for Detecting Resectable Early Lung Cancer

EP13.01-005. Role of Artificial Intelligence on Chest Radiographs for Detecting Resectable Early Lung Cancer

Pdf Summary

This study aimed to assess the role of artificial intelligence (AI) in detecting resectable early lung cancer on chest radiographs. The researchers conducted a retrospective evaluation of AI-based lesion detection software on patients who underwent lung cancer surgery at three referral hospitals between March 2020 and February 2021.<br /><br />Of the 827 patients with resectable lung cancer, the majority were male (51.3%) and had adenocarcinoma (87.3%) or squamous cell carcinoma (8.6%). Most patients (77.4%) had stage I lung cancer, while stage II and III accounted for 11.4% and 9.7%, respectively. 44.9% of patients had EGFR mutation. <br /><br />The study found that 55.9% of patients were detected by the AI lung nodule software, with an average nodule abnormality score of 40.4. Male gender, smoking history, squamous cell carcinoma, higher stage, and larger tumor size were associated with lung nodule detection by the AI software. <br /><br />The mean size of resected lung cancers was 2.3 cm, and there was a statistically significant correlation between tumor size and nodule detection abnormality score. <br /><br />Overall, the results indicated that a significant number of resectable early lung cancers were detected by the AI lung nodule software. However, the researchers emphasized the need for further large-scale and well-designed studies to fully understand the role of AI chest radiographs in early lung cancer detection. <br /><br />Table 1 provides a comparison of baseline characteristics of resectable lung cancer patients according to AI nodule detection, and Figure 1 demonstrates the integration of AI results in the picture archiving and communicating system (PACS) and provides an example of lung nodule detection using AI software.

Asset Subtitle

Eunhye Lee

Meta Tag

Speaker Eunhye Lee

Topic Pulmonology, Radiology, and Staging

Keywords

artificial intelligence

detecting resectable early lung cancer

chest radiographs

retrospective evaluation

AI-based lesion detection software

lung cancer surgery

referral hospitals

adenocarcinoma

squamous cell carcinoma

stage I lung cancer