The purpose of this study was to perform evaluation of a deep learning image reconstruction (DLIR) algorithm in native CT of adrenal lesions and reveal the potential of radiation exposure reduction

Retrospective review (September–December 2020) of 32 cases of adults undergoing native CT examination of adrenal glands were conducted for evaluation of standard adaptive statistical iterative reconstruction V (50% ASIR-V) reconstruction compared with DLIR at high strength. All patients were divided into 2 groups. The first group included 19 patients, who were scanned using the following parameters (low dose protocol): gantry speed, 0.6 second; pitch, 0,984:1; beam collimation, 40 mm; detector configuration, 128 × 0.625 mm; 100 kVp; tube current modulation range, 120-700 mA; noise index,...

The imaging findings obtained from groups 1A (TrueFidelity) and 1B (Asir V50) are summarized in Table 1. The reconstructions in Table 2 had significantly different noise and CNR: high-strength DLIR provided 52% reduction in noise compared with 50% ASIR-V. Radiation dose: DLP in group 1: 574,4±332 (range, 184-1467), Ме 513,495 мГр·см (351,13 : 768,3), in group 2: 775 ±464 (range, 197-1562), Ме 872 мГр·см (317,63 : 1255).

Нistorically lower-noise images have been associated with higher radiation doses [1]. In our study, we reveal that the use of DLIR in patients, scanned with the low-dose protocol not only reduced the radiation dose while maintaining image quality, but also increased the quality of the obtained images. Our investigation of the vendor-specific DLIR tool TrueFidelity showed that it improved the image quality of noncontrast CT scans of adrenal lesions relative to that obtained by our standard 50% ASIR-V. DLIR was scored significantly better for overall...

1.Jensen CT, Wagner-Bartak NA, Vu LN, et al. Detection of colorectal hepatic metastases is superior at standard radiation dose CT versus reduced dose CT. Radiology 2019; 290:400–409 2.Jensen CT, Telesmanich ME, Wagner-Bartak NA, et al. Evaluation of abdominal computed tomography image quality using a new version of vendorspecific model-based iterative reconstruction. J Comput Assist Tomogr 2017; 41:67–74. 3.Telesmanich ME, Jensen CT, Enriquez JL, et al. Third version of vendor-specific model-based iterative reconstruction (Veo 3.0): evaluation of CT image quality in the abdomen using new noise...

S. A. Buryakina: Author: writing text, design creation N. Tarbaeva: Author: study design N. Volevodz: Author: design and text correction N. Mokrysheva: Author: design approval, text correction