Role of Chest CT Radiomics in Differentiating Tumorlets and Granulomas: A Preliminary Study

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Distinguishing between granulomas and tumorlets (TL) is crucial for patient management.
• Tumorlets can potentially evolve into malignancy, necessitating early and accurate diagnosis.
• Radiomics offers a non-invasive method for characterizing small lung nodules.

Purpose of the Study:

• To identify radiomics features differentiating granulomas from tumorlets (TL).
• To assess the diagnostic performance of radiomics in distinguishing these two conditions.
• To explore the potential of radiomics in early detection and management of lung nodules.

Main Methods:

• Retrospective analysis of 55 patients with pathologically proven granulomas or TL.
• Manual segmentation of lesions on pre-operative chest CT scans.
• Automatic extraction of radiomic features using dedicated software and performance evaluation via ROC curves.

Main Results:

• Significant differences were found in 16 out of 107 radiomic features between granulomas and TL.
• Flatness and Long Run High Gray Level Emphasis demonstrated high discriminatory power (AUCs of 0.903 and 0.896, respectively).
• High sensitivity and specificity were achieved in differentiating tumorlets from granulomas.

Conclusions:

• Radiomics serves as a promising non-invasive tool for characterizing small lung nodules.
• It can effectively differentiate between granulomas and tumorlets, aiding in clinical decision-making.
• This approach may help prevent delayed diagnosis and the potential malignant progression of tumorlets.