Histogram analysis based on unenhanced CT for identifying thymoma and lymphoma among prevascular mediastinal

Ming Liu1, Yang Zhang2,3, Li-Heng Liu4,5

  • 1Department of Radiation Oncology, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai, China.

Insights

Histogram analysis of unenhanced CT scans can aid in differentiating thymic tumors like thymoma and lymphoma from non-tumorous conditions such as hyperplasia and cysts. This technique shows promise for diagnosing incidental prevascular mediastinal lesions.

Area of Science:

  • Radiology
  • Medical Imaging
  • Oncology

Background:

  • Incidental prevascular mediastinal lesions require accurate differential diagnosis.
  • Distinguishing thymoma and lymphoma from thymic hyperplasia and cysts can be challenging.
  • Unenhanced CT (computed tomography) is a common imaging modality.

Purpose of the Study:

  • To evaluate the utility of histogram analysis on unenhanced CT for differentiating thymoma and lymphoma from thymic hyperplasia and cysts.
  • To assess the diagnostic performance of histogram parameters in classifying these lesions.

Main Methods:

  • Retrospective analysis of 192 asymptomatic patients with incidentally detected prevascular mediastinal lesions.
  • Calculation of thirteen histogram parameters from unenhanced CT images.
  • Comparison of histogram parameters between tumor (thymoma/lymphoma) and non-tumor (hyperplasia/cyst) groups.
  • Receiver operating characteristic (ROC) analysis to determine diagnostic accuracy.

Main Results:

  • Nine of thirteen histogram parameters showed significant differences between tumor and non-tumor groups.
  • No significant difference in mean CT attenuation was observed between groups.
  • Higher median histogram value was an independent predictor for distinguishing tumors, achieving an Area Under the Curve (AUC) of 0.785.

Conclusions:

  • Histogram analysis of unenhanced CT provides valuable information for the differential diagnosis of incidental prevascular mediastinal lesions.
  • This method can assist in differentiating thymic tumors from non-tumorous conditions.
  • Further research may refine the role of CT-based radiomics in thymic lesion characterization.
Abstract