Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population

Abstract Background Accurate identification and characterization of lymph nodes (LNs) are essential in cancer staging and treatment planning. While artificial intelligence (AI) has shown potential in detecting lymphadenopathy on contrast-enhanced CT scans, its performance on non-contrast (plain) CT...

Full description

Saved in:
Bibliographic Details
Main Authors: Taku Takaishi, Tatsuya Kawai, Megumi Kita, Shunsuke Shibata, Akio Hiwatashi
Format: Article
Language:English
Published: SpringerOpen 2024-11-01
Series:The Egyptian Journal of Radiology and Nuclear Medicine
Subjects:
Online Access:https://doi.org/10.1186/s43055-024-01397-7
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850056445571104768
author Taku Takaishi
Tatsuya Kawai
Megumi Kita
Shunsuke Shibata
Akio Hiwatashi
author_facet Taku Takaishi
Tatsuya Kawai
Megumi Kita
Shunsuke Shibata
Akio Hiwatashi
author_sort Taku Takaishi
collection DOAJ
description Abstract Background Accurate identification and characterization of lymph nodes (LNs) are essential in cancer staging and treatment planning. While artificial intelligence (AI) has shown potential in detecting lymphadenopathy on contrast-enhanced CT scans, its performance on non-contrast (plain) CT scans remains less explored. This retrospective study aimed to assess the efficacy of AI for segmenting LNs on plain CT in patients with lung or gastrointestinal cancer. Eligible cases were selected based on plain CT scans with 5 mm slice thickness, showing mediastinal or hilar LNs with a short-axis diameter of ≥ 10 mm. A total of 144 patients (105 men, 39 women; mean age, 76 years) were included, comprising 342 LNs. A commercially available deep learning AI software was used for segmentation, with the ground-truth defined by two radiologists’ consensus on LNs measuring ≥ 10 mm. Results The AI software showed a sensitivity of 83.9%, with no statistically significant difference between mediastinal and hilar LNs (p = 0.84). The false-positive rate was 0.465 per patient. The Dice coefficient was 0.908, and there was no statistically significant difference between the mediastinum and hilum (p = 0.19). The sensitivity decreased in larger LNs, showing an inverse correlation (r = −0.122, p < 0.01). Conclusions This study demonstrated the robustness of the AI software in the segmentation of LNs in the mediastinum and hilum on plain CT scans.
format Article
id doaj-art-20f5655a36344d11befd4f16ea2ffad5
institution DOAJ
issn 2090-4762
language English
publishDate 2024-11-01
publisher SpringerOpen
record_format Article
series The Egyptian Journal of Radiology and Nuclear Medicine
spelling doaj-art-20f5655a36344d11befd4f16ea2ffad52025-08-20T02:51:42ZengSpringerOpenThe Egyptian Journal of Radiology and Nuclear Medicine2090-47622024-11-015511810.1186/s43055-024-01397-7Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer populationTaku Takaishi0Tatsuya Kawai1Megumi Kita2Shunsuke Shibata3Akio Hiwatashi4Department of Radiology, Nagoya City University Graduate School of Medical SciencesDepartment of Radiology, Nagoya City University Graduate School of Medical SciencesKasugai Municipal HospitalDepartment of Radiology, Nagoya City University Graduate School of Medical SciencesDepartment of Radiology, Nagoya City University Graduate School of Medical SciencesAbstract Background Accurate identification and characterization of lymph nodes (LNs) are essential in cancer staging and treatment planning. While artificial intelligence (AI) has shown potential in detecting lymphadenopathy on contrast-enhanced CT scans, its performance on non-contrast (plain) CT scans remains less explored. This retrospective study aimed to assess the efficacy of AI for segmenting LNs on plain CT in patients with lung or gastrointestinal cancer. Eligible cases were selected based on plain CT scans with 5 mm slice thickness, showing mediastinal or hilar LNs with a short-axis diameter of ≥ 10 mm. A total of 144 patients (105 men, 39 women; mean age, 76 years) were included, comprising 342 LNs. A commercially available deep learning AI software was used for segmentation, with the ground-truth defined by two radiologists’ consensus on LNs measuring ≥ 10 mm. Results The AI software showed a sensitivity of 83.9%, with no statistically significant difference between mediastinal and hilar LNs (p = 0.84). The false-positive rate was 0.465 per patient. The Dice coefficient was 0.908, and there was no statistically significant difference between the mediastinum and hilum (p = 0.19). The sensitivity decreased in larger LNs, showing an inverse correlation (r = −0.122, p < 0.01). Conclusions This study demonstrated the robustness of the AI software in the segmentation of LNs in the mediastinum and hilum on plain CT scans.https://doi.org/10.1186/s43055-024-01397-7CTArtificial intelligenceDeep learningSegmentation
spellingShingle Taku Takaishi
Tatsuya Kawai
Megumi Kita
Shunsuke Shibata
Akio Hiwatashi
Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population
The Egyptian Journal of Radiology and Nuclear Medicine
CT
Artificial intelligence
Deep learning
Segmentation
title Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population
title_full Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population
title_fullStr Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population
title_full_unstemmed Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population
title_short Artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain CT: a retrospective study in a cancer population
title_sort artificial intelligence deep learning software for segmentation of mediastinal and hilar lymph nodes on plain ct a retrospective study in a cancer population
topic CT
Artificial intelligence
Deep learning
Segmentation
url https://doi.org/10.1186/s43055-024-01397-7
work_keys_str_mv AT takutakaishi artificialintelligencedeeplearningsoftwareforsegmentationofmediastinalandhilarlymphnodesonplainctaretrospectivestudyinacancerpopulation
AT tatsuyakawai artificialintelligencedeeplearningsoftwareforsegmentationofmediastinalandhilarlymphnodesonplainctaretrospectivestudyinacancerpopulation
AT megumikita artificialintelligencedeeplearningsoftwareforsegmentationofmediastinalandhilarlymphnodesonplainctaretrospectivestudyinacancerpopulation
AT shunsukeshibata artificialintelligencedeeplearningsoftwareforsegmentationofmediastinalandhilarlymphnodesonplainctaretrospectivestudyinacancerpopulation
AT akiohiwatashi artificialintelligencedeeplearningsoftwareforsegmentationofmediastinalandhilarlymphnodesonplainctaretrospectivestudyinacancerpopulation