Comparison of diagnostic accuracy of radiomics parameter maps and standard reconstruction for the detection of liver lesions in computed tomography

• DOI: 10.3389/fonc.2024.1444115
• Publication Date: 2024-10-07T04:10:23Z
• Authors: Alexander Hertel; Mustafa Kuru; Fabian Tollens; Hishan Tharmaseelan; Dominik Nörenberg; Nils Rathmann; Stefan O. Schoenberg; Matthias F. Froelich
• Keywords: Oncology; radiomics; Neoplasms. Tumors. Oncology. Including cancer and carcinogens; radiomics maps; colorectal cancer; liver metastases; lesion detectability; RC254-282

Background The liver is a frequent location of metastatic disease in various malignant tumor entities. Computed tomography (CT) the most frequently employed modality for initial diagnosis. However, metastases may only be delineated vaguely on CT. Calculating radiomics features feature maps can unravel textures not visible to human eye standard CT reconstruction (SCTR). This study aimed investigate comparative diagnostic accuracy and SCTR metastases. Materials methods Forty-seven patients with hepatic colorectal cancer were retrospectively enrolled. Whole-liver original generated. A representative was selected each class based visualization example lesions from five patients. These conventional image data viewed evaluated by four readers terms parenchyma, number lesions, visual contrast confidence. T-tests chi²-tests performed significance cut off p<0.05 compare SCRT, visualized as boxplots. Results Regarding detected, showed superior performance compared maps. map firstorder RootMeanSquared ranked very high 57.4% cases, 41.0% reconstructions (p < 0.001). All other significantly lower when SCTR. For confidence, reached ratings 47.9% 62.8% images results all categories investigated. Conclusion application help visualize faintly demarcated increasing contrast. reading remains necessary