Abstracts

RETURN TO ABSTRACT LISTING


E2455. Imaging and Radiomics in Differentiating Lipid Poor Angio-Myolipoma and Renal Cell Carcinoma
Authors
  1. Mohab Elmohr; Baylor College of Medicine
  2. Jeffrey Guccione; UT McGovern Medical School
  3. Moataz Soliman; UT MD Anderson Cancer Center
  4. Aline Khatchikian; McGill University
  5. Ahmed Moawad; UT MD Anderson Cancer Center
  6. Serageldin Kamel; UT MD Anderson Cancer Center
  7. Khaled Elsayes; UT MD Anderson Cancer Center
Background
It’s estimated that solitary renal angiomyolipoma (AML) can be found in up to 2% of normal population. Moreover, they are being more frequently detected due to rapidly increasing imaging use of imaging in healthcare. In addition, 5% of these are fat poor, leading to confusion with renal cell carcinoma. This confusion can lead to increased healthcare costs & undue emotional stress for these patients. It’s important for abdominal radiologists to know the differentiation between these lesions very well. The rising role of AI in imaging will likely prove to be useful as an assisting tool for radiologists in this regard.

Educational Goals / Teaching Points
First, we will explain the pathological background of AML and its variants i.e. fat poor AML. Second, we will also briefly explain the pathological background of RCC and its pathological subtypes. Then, we will explore the common imaging features of both fat poor AML and RCC. We will teach the audience about the importance of differentiating the two lesions. After that, we will demonstrate the different radiological features using cross-sectional imaging. We will discuss the application of AI to differentiate fat poor AML and RCC. We will finally illustrate the impact of differentiation on patient outcome.

Key Anatomic/Physiologic Issues and Imaging Findings/Techniques
First, it’s paramount to explain the pathologic background of all the discussed lesions, as this explains many imaging features of these tumors. Second, we will explain & illustrate the common imaging features shared by all these lesions. This includes CT & MRI features as well as imaging-based modality differentiation. Then we will move on to AML specific features and discuss the role of macroscopic fat detection by CT. We will also go over the role of US in AML diagnosis. In contrast, we will illustrate how the lack of macroscopic fat appears on CT. We will also discuss the MRI role of fat suppression sequences. Regarding RCC, we will explain what CT features suggest malignancy. We will also discuss the PET CT added value in staging. Finally, we will summarize the promising role of AI in the differentiation between these lesions. A brief explanation of radiomics features will follow. We will conclude by going over which radiomics features suggest the diagnosis of fat poor AML.

Conclusion
AI & radiomics features would likely be very useful in the differentiation between fat poor AML and RCC. This would significantly improve the approach to the incidental finding of these lesions on imaging studies. Abdominal radiologists should be aware of the potential clinical application of this in the upcoming years.