Development of a novel artificial intelligence algorithm to detect pulmonary nodules on chest radiography.

Autor:	Higuchi M; Department of Thoracic Surgery, Aizu Medical Center, Fukushima Medical University., Nagata T; University of Tsukuba School of Integrative and Global Majors.; Mizuho Research and Technologies, Ltd., Iwabuchi K; Mizuho Research and Technologies, Ltd., Sano A; Mizuho Research and Technologies, Ltd., Maekawa H; Mizuho Research and Technologies, Ltd., Idaka T; Mizuho Research and Technologies, Ltd., Yamasaki M; Mizuho Research and Technologies, Ltd., Seko C; Mizuho Research and Technologies, Ltd., Sato A; Fukushima Preservative Service Association of Health., Suzuki J; Fukushima Preservative Service Association of Health., Anzai Y; Aizuwakamatsu Medical Association., Yabuki T; Aizuwakamatsu Medical Association., Saito T; Department of Surgery, Aizu Medical Center, Fukushima Medical University., Suzuki H; Department of Chest Surgery, Fukushima Medical University School of Medicine.
Jazyk:	angličtina
Zdroj:	Fukushima journal of medical science [Fukushima J Med Sci] 2023 Nov 15; Vol. 69 (3), pp. 177-183. Date of Electronic Publication: 2023 Oct 17.
DOI:	10.5387/fms.2023-14
Abstrakt:	Background: In this study, we aimed to develop a novel artificial intelligence (AI) algorithm to support pulmonary nodule detection, which will enable physicians to efficiently interpret chest radiographs for lung cancer diagnosis. Methods: We analyzed chest X-ray images obtained from a health examination center in Fukushima and the National Institutes of Health (NIH) Chest X-ray 14 dataset. We categorized these data into two types: type A included both Fukushima and NIH datasets, and type B included only the Fukushima dataset. We also demonstrated pulmonary nodules in the form of a heatmap display on each chest radiograph and calculated the positive probability score as an index value. Results: Our novel AI algorithms had a receiver operating characteristic (ROC) area under the curve (AUC) of 0.74, a sensitivity of 0.75, and a specificity of 0.60 for the type A dataset. For the type B dataset, the respective values were 0.79, 0.72, and 0.74. The algorithms in both the type A and B datasets were superior to the accuracy of radiologists and similar to previous studies. Conclusions: The proprietary AI algorithms had a similar accuracy for interpreting chest radiographs when compared with previous studies and radiologists. Especially, we could train a high quality AI algorithm, even with our small type B data set. However, further studies are needed to improve and further validate the accuracy of our AI algorithm.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu