Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans
A reliable diagnosis and accurate monitoring are pivotal steps for treatment and prevention of COVID-19. Chest computed tomography (CT) has been considered a crucial diagnostic imaging technique for the injury assessment of the viral pneumonia. Furthermore, the automatization of the segmentation met...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SAGE Publishing
2022-04-01
|
| Series: | Dose-Response |
| Online Access: | https://doi.org/10.1177/15593258221082896 |
| _version_ | 1818475040831700992 |
|---|---|
| author | Marco Aiello Dario Baldi Giuseppina Esposito Marika Valentino Marco Randon Marco Salvatore Carlo Cavaliere |
| author_facet | Marco Aiello Dario Baldi Giuseppina Esposito Marika Valentino Marco Randon Marco Salvatore Carlo Cavaliere |
| author_sort | Marco Aiello |
| collection | DOAJ |
| description | A reliable diagnosis and accurate monitoring are pivotal steps for treatment and prevention of COVID-19. Chest computed tomography (CT) has been considered a crucial diagnostic imaging technique for the injury assessment of the viral pneumonia. Furthermore, the automatization of the segmentation methods for lung alterations helps to speed up the diagnosis and lighten radiologists’ workload. Considering the assiduous pathology monitoring, ultra-low dose (ULD) chest CT protocols have been implemented to drastically reduce the radiation burden. Unfortunately, the available AI technologies have not been trained on ULD-CT data and validated and their applicability deserves careful evaluation. Therefore, this work aims to compare the results of available AI tools (BCUnet, CORADS AI, NVIDIA CLARA Train SDK and CT Pneumonia Analysis) on a dataset of 73 CT examinations acquired both with conventional dose (CD) and ULD protocols. COVID-19 volume percentage, resulting from each tool, was statistically compared. This study demonstrated high comparability of the results on CD-CT and ULD-CT data among the four AI tools, with high correlation between the results obtained on both protocols (R > .68, P < .001, for all AI tools). |
| first_indexed | 2024-04-14T04:45:18Z |
| format | Article |
| id | doaj.art-3056a8461a59484d8f7c499635e0e789 |
| institution | Directory Open Access Journal |
| issn | 1559-3258 |
| language | English |
| last_indexed | 2024-04-14T04:45:18Z |
| publishDate | 2022-04-01 |
| publisher | SAGE Publishing |
| record_format | Article |
| series | Dose-Response |
| spelling | doaj.art-3056a8461a59484d8f7c499635e0e7892022-12-22T02:11:29ZengSAGE PublishingDose-Response1559-32582022-04-012010.1177/15593258221082896Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT ScansMarco AielloDario BaldiGiuseppina EspositoMarika ValentinoMarco RandonMarco SalvatoreCarlo CavaliereA reliable diagnosis and accurate monitoring are pivotal steps for treatment and prevention of COVID-19. Chest computed tomography (CT) has been considered a crucial diagnostic imaging technique for the injury assessment of the viral pneumonia. Furthermore, the automatization of the segmentation methods for lung alterations helps to speed up the diagnosis and lighten radiologists’ workload. Considering the assiduous pathology monitoring, ultra-low dose (ULD) chest CT protocols have been implemented to drastically reduce the radiation burden. Unfortunately, the available AI technologies have not been trained on ULD-CT data and validated and their applicability deserves careful evaluation. Therefore, this work aims to compare the results of available AI tools (BCUnet, CORADS AI, NVIDIA CLARA Train SDK and CT Pneumonia Analysis) on a dataset of 73 CT examinations acquired both with conventional dose (CD) and ULD protocols. COVID-19 volume percentage, resulting from each tool, was statistically compared. This study demonstrated high comparability of the results on CD-CT and ULD-CT data among the four AI tools, with high correlation between the results obtained on both protocols (R > .68, P < .001, for all AI tools).https://doi.org/10.1177/15593258221082896 |
| spellingShingle | Marco Aiello Dario Baldi Giuseppina Esposito Marika Valentino Marco Randon Marco Salvatore Carlo Cavaliere Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans Dose-Response |
| title | Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans |
| title_full | Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans |
| title_fullStr | Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans |
| title_full_unstemmed | Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans |
| title_short | Evaluation of AI-Based Segmentation Tools for COVID-19 Lung Lesions on Conventional and Ultra-low Dose CT Scans |
| title_sort | evaluation of ai based segmentation tools for covid 19 lung lesions on conventional and ultra low dose ct scans |
| url | https://doi.org/10.1177/15593258221082896 |
| work_keys_str_mv | AT marcoaiello evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans AT dariobaldi evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans AT giuseppinaesposito evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans AT marikavalentino evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans AT marcorandon evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans AT marcosalvatore evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans AT carlocavaliere evaluationofaibasedsegmentationtoolsforcovid19lunglesionsonconventionalandultralowdosectscans |