Page 101 - TD-3-3

Tumor Discovery

ORIGINAL RESEARCH ARTICLE
Evaluating the effectiveness of artificial

intelligence imaging in the qualitative diagnosis
of pulmonary nodules

Chunlan Hu 1 , Dan Yang 1 , Xiangwen Luo 1 , Chao Lv 1 , Juan Li 1 ,
2
Yaya Zhang 1 , Xinrong Xiong 1 , and Long Zhou *
1 Cancer Early Diagnosis and Treatment Center, Clinical Research Center, Medical Pathology
Center, and Translational Medicine Research Center, Chongqing University Three Gorges Hospital,
Chongqing University, Chongqing, China
2 Department of Laboratory Medicine, Chongqing University Three Gorges Hospital, Chongqing
University, Chongqing, China
(This article belongs to the Special Issue: New Developments in Lung Cancer Research, Diagnosis,
Treatment, and Prognosis)

Abstract

Our study aimed to evaluate the effectiveness of artificial intelligence (AI) image
diagnostic systems in the qualitative diagnosis of pulmonary nodules. We analyzed
291 cases from June 2023 to January 2024 at Chongqing University Three Gorges
Hospital. All patients in the study underwent low-dose chest computed tomography
scans, which identified lung nodules, followed by thoracic surgery for pathological
confirmation. We compared the predictive accuracy of AI-based diagnosis with
that of physician-based diagnosis in distinguishing between benign and malignant
*Corresponding author:
Long Zhou lung nodules. Among the 291 lung nodules examined, 226 were cancerous, and 65
(goldfall2004@163.com) were benign. Receiver operating characteristic (ROC) curves, plotted based on the
malignancy probabilities predicted by both methods, revealed that the AI group
Citation: Hu C, Yang D, Luo X,
et al. Evaluating the effectiveness achieved an area under the ROC curve (AUC) of 0.727, with a sensitivity of 90.27% and
of artificial intelligence imaging a specificity of 58.46%. In comparison, the physician-reading group had an AUC of
in the qualitative diagnosis of 0.737, with a sensitivity of 83.19% and a specificity of 66.15%. Our findings demonstrate
pulmonary nodules. Tumor Discov.
2024;3(3):4718. that the AI diagnostic system effectively calculates malignancy probabilities for lung
doi: 10.36922/td.4178 nodules, highlighting its significant predictive potential. This system can serve as a
Received: July 9, 2024 valuable adjunct tool for clinicians and imaging physicians in the diagnostic process.
Accepted: September 2, 2024
Keywords: Artificial intelligence; Pulmonary nodule; Assisted diagnosis
Published Online: September 25, 2024
Copyright: © 2024 Author(s).
This is an Open-Access article
distributed under the terms of the
Creative Commons Attribution 1. Introduction
License, permitting distribution,
and reproduction in any medium, Lung cancer exhibits the highest morbidity and mortality rates among malignancies
1
provided the original work is worldwide, posing a significant threat to human health. In 2022, China reported
properly cited. approximately 1,060,600 new cases of lung cancer and 733,300 deaths related to the disease,
Publisher’s Note: AccScience underscoring its substantial burden on public health. Despite recent advances in lung cancer
Publishing remains neutral with treatment technology, about 75% of patients miss the optimal window for treatment due to
regard to jurisdictional claims in 2
published maps and institutional delayed diagnosis. Early diagnosis and appropriate treatment are essential for improving
affiliations. the survival rates of lung cancer patients, particularly through early detection.

Volume 3 Issue 3 (2024) 1 doi: 10.36922/td.4178