Detectability of Small Low-Attenuation Lesions With Deep Learning CT Image Reconstruction: A 24-Reader Phantom Study.
| Title: | Detectability of Small Low-Attenuation Lesions With Deep Learning CT Image Reconstruction: A 24-Reader Phantom Study. |
|---|---|
| Authors: | Toia GV; Department of Radiology, University of Wisconsin School of Medicine and Public Health, 600 Highland Ave, Mailbox 3252, Madison, WI 53792.; Zamora DA; Department of Radiology, University of Washington School of Medicine, Seattle, WA.; Singleton M; Institute of Translational Health Sciences, University of Washington, Seattle, WA.; Liu A; Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, WA.; Tan E; Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, WA.; Leng S; Department of Radiology, Mayo Clinic, Rochester, MN.; Shuman WP; Department of Radiology, University of Washington School of Medicine, Seattle, WA.; Kanal KM; Department of Radiology, University of Washington School of Medicine, Seattle, WA.; Mileto A; Department of Radiology, Mayo Clinic, Rochester, MN. |
| Source: | AJR. American journal of roentgenology [AJR Am J Roentgenol] 2023 Feb; Vol. 220 (2), pp. 283-295. Date of Electronic Publication: 2022 Sep 21. |
| Publication Type: | Journal Article |
| Language: | English |
| Journal Info: | Publisher: American Roentgen Ray Society Country of Publication: United States NLM ID: 7708173 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1546-3141 (Electronic) Linking ISSN: 0361803X NLM ISO Abbreviation: AJR Am J Roentgenol Subsets: MEDLINE |
| Imprint Name(s): | Publication: : Leesburg, VA : American Roentgen Ray Society; Original Publication: Springfield, Ill., Thomas. |
| MeSH Terms: | Deep Learning*; Radiographic Image Interpretation, Computer-Assisted/methods ; Tomography, X-Ray Computed/methods ; Humans ; Radiation Dosage ; Algorithms ; Phantoms, Imaging ; Image Processing, Computer-Assisted |
| Abstract: | BACKGROUND. Iterative reconstruction (IR) techniques are susceptible to contrast-dependent spatial resolution, limiting overall radiation dose reduction potential. Deep learning image reconstruction (DLIR) may mitigate this limitation. OBJECTIVE. The purpose of our study was to evaluate low-contrast detectability performance and radiation-saving potential of a DLIR algorithm in comparison with filtered back projection (FBP) and IR using a human multireader noninferiority study design and task-based observer modeling. METHODS. A dual-phantom construct, consisting of a low-contrast detectability module (21 low-contrast hypoattenuating objects in seven sizes [2.4-10.0 mm] and three contrast levels [-15, -10, -5 HU] embedded within liver-equivalent background) and a phantom, was imaged at five radiation exposures (CTDIvol range, 1.4-14.0 mGy; size-specific dose estimate, 2.5-25.0 mGy; 90%-, 70%-, 50%-, and 30%-reduced radiation levels and full radiation level) using an MDCT scanner. Images were reconstructed using FBP, hybrid IR (ASiR-V), and DLIR (TrueFidelity). Twenty-four readers of varying experience levels evaluated images using a two-alternative forced choice. A task-based observer model (detectability index [d']) was calculated. Reader performance was estimated by calculating the AUC using a noninferiority method. RESULTS. Compared with FBP and IR methods at routine radiation levels, DLIR medium and DLIR high settings showed noninferior performance through a 90% radiation reduction (except DLIR medium setting at 70% reduced level). The IR method was non-inferior to routine radiation FBP only for 30% and 50% radiation reductions. No significant difference in d' was observed between routine radiation FBP and DLIR high setting through a 70% radiation reduction. Reader experience was not correlated with diagnostic accuracy (R2 = 0.005). CONCLUSION. Compared with FBP or IR methods at routine radiation levels, certain DLIR algorithm weightings yielded noninferior low-contrast detectability with radiation reductions of up to 90% as measured by 24 human readers and up to 70% as assessed by a task-based observer model. CLINICAL IMPACT. DLIR has substantial potential to preserve contrast-dependent spatial resolution for the detection of hypoattenuating lesions at decreased radiation levels in a phantom model, addressing a major shortcoming of current IR techniques. |
| Comments: | Comment in: AJR Am J Roentgenol. 2023 Feb;220(2):296. doi: 10.2214/AJR.22.28554.. (PMID: 36169549); Erratum in: AJR Am J Roentgenol. 2023 Apr;220(4):614. doi: 10.2214/AJR.23.29111.. (PMID: 36947732) |
| Contributed Indexing: | Keywords: CT; image processing; radiation dosage |
| Entry Date(s): | Date Created: 20220921 Date Completed: 20230124 Latest Revision: 20241017 |
| Update Code: | 20260130 |
| DOI: | 10.2214/AJR.22.28407 |
| PMID: | 36129222 |
| Database: | MEDLINE |
Journal Article