Scholarlink

Purpose <p>This study aimed to investigate the performance of an artificial intelligence (AI)-based lung nodule detection program in ultra-low-dose CT (ULDCT) imaging, with a focus on the influence of various image reconstruction methods on detection accuracy.</p> Methods <p>A chest phantom embedded with artificial lung nodules (solid and ground-glass nodules [GGNs]; diameters: 12&#xa0;mm, 8&#xa0;mm, 5&#xa0;mm, and 3&#xa0;mm) was scanned using six combinations of tube currents (160&#xa0;mA, 80&#xa0;mA, and 10&#xa0;mA) and voltages (120&#xa0;kV and 80&#xa0;kV) on a Canon Aquilion One CT scanner. Images were reconstructed using filtered back projection (FBP), hybrid iterative reconstruction (HIR), model-based iterative reconstruction (MBIR), and deep learning reconstruction (DLR). Nodule detection was performed using an AI-based lung nodule detection program, and performance metrics were analyzed across different reconstruction methods and radiation dose protocols.</p> Results <p>At the lowest dose protocol (80&#xa0;kV, 10&#xa0;mA), FBP showed a 0% detection rate for all nodule sizes. HIR and DLR consistently achieved 100% detection rates for solid nodules ≥ 5&#xa0;mm and GGNs ≥ 8&#xa0;mm. No method detected 3&#xa0;mm GGNs under any protocol. DLR demonstrated the highest detection rates, even under ultra-low-dose settings, while maintaining high image quality.</p> Conclusion <p>AI-based lung nodule detection in ULDCT is strongly dependent on the choice of image reconstruction method.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Application of a pulmonary nodule detection program using AI technology to ultra-low-dose CT: differences in detection ability among various image reconstruction methods

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年5月9日

查看全文
  • Nanae Tsuchiya,
  • Shifumi Kobayashi,
  • Ryo Nakachi,
  • Yukari Tomori,
  • Akira Yogi,
  • Gyo Iida,
  • Junji Ito,
  • Akihiro Nishie

摘要

Purpose

This study aimed to investigate the performance of an artificial intelligence (AI)-based lung nodule detection program in ultra-low-dose CT (ULDCT) imaging, with a focus on the influence of various image reconstruction methods on detection accuracy.

Methods

A chest phantom embedded with artificial lung nodules (solid and ground-glass nodules [GGNs]; diameters: 12 mm, 8 mm, 5 mm, and 3 mm) was scanned using six combinations of tube currents (160 mA, 80 mA, and 10 mA) and voltages (120 kV and 80 kV) on a Canon Aquilion One CT scanner. Images were reconstructed using filtered back projection (FBP), hybrid iterative reconstruction (HIR), model-based iterative reconstruction (MBIR), and deep learning reconstruction (DLR). Nodule detection was performed using an AI-based lung nodule detection program, and performance metrics were analyzed across different reconstruction methods and radiation dose protocols.

Results

At the lowest dose protocol (80 kV, 10 mA), FBP showed a 0% detection rate for all nodule sizes. HIR and DLR consistently achieved 100% detection rates for solid nodules ≥ 5 mm and GGNs ≥ 8 mm. No method detected 3 mm GGNs under any protocol. DLR demonstrated the highest detection rates, even under ultra-low-dose settings, while maintaining high image quality.

Conclusion

AI-based lung nodule detection in ULDCT is strongly dependent on the choice of image reconstruction method.