Assessment of image quality and lesion detection in liver spectral computed tomography based on image reconstruction algorithms and virtual tube voltage رسالة ماجستير

Abstract

Introduction The diagnostic capability of the radiological examination device is crucial for improving the diagnostic accuracy of liver disease. Spectral Detector Computed Tomography (SDCT) Innovation has shown high diagnostic performance in the assessment of liver diseases with better image quality compared to conventional computed tomography devices. Choosing the proper tube voltage and reconstruction algorithm type is crucial for maintaining image quality; improper settings can lead to increased noise and a reduced signal-to-noise ratio.

Purpose Firstly this study aimed to evaluate the diagnostic accuracy of spectral CT. Secondly assessing image quality enhancement based on different virtual tube voltages and reconstruction algorithms for diagnosing common liver diseases.

Methods This study is retrospective. Sixteen patients had previously been scanned by spectral CT and diagnosed with one of the following liver-related conditions: fatty liver, hemangiomas, or metastatic lesions. The virtual tube voltage and reconstruction algorithm were evaluated to assess the effect of tube voltage and algorithm type on the diagnostic efficiency of liver diseases in Spectral CT images. Images were reconstructed with different reconstruction algorithms such as iDose and IMR at various degrees with different virtual tube voltages. Three experienced radiologists analyzed eight combinations of reconstructed CT images, reviewing 128 scans from 16 patients.

Result Among liver diseases, the signal-to-noise ratio was the highest among spectral CT images scanned with the IMR3 algorithm in metastatic, hemangioma, and fatty liver images. Nevertheless, when imaging abnormal liver lesions, SNR for all reconstruction techniques increases as tube voltage decreases.

Conclusion Significant improvements were observed in several image quality metrics in SDCT when using low radiation doses. A specified strength of Iterative Reconstruction (IR) substantially enhanced signal-to-noise ratio (SNR) values for both normal and pathological liver tissues.