Quantitative computed tomography texture analysis: can it improve diagnostic accuracy to differentiate malignant...

Area of Science:

• Radiology
• Oncology
• Medical Imaging Analysis

Background:

• Mediastinal lymph node (LN) staging is crucial for non-small-cell lung cancer (NSCLC) treatment planning.
• Accurate differentiation between benign and malignant LNs is essential for appropriate patient management.

Purpose of the Study:

• To evaluate the diagnostic accuracy of computed tomography texture analysis (CTTA) in distinguishing benign from malignant mediastinal LNs in NSCLC patients.
• To compare the efficacy of CTTA with other imaging modalities.

Main Methods:

• Retrospective review of 132 pathologically confirmed mediastinal LNs (61 malignant, 71 benign) from 80 NSCLC patients.
• Analysis of chest CT, 18-fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT data, and endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) findings.
• CT texture analysis (CTTA) performed using AVIEW software.

Main Results:

• While certain EBUS features (size, shape, echogenicity, necrosis) and CTTA features (compactness, normalized standard deviation) showed differences between groups, only LN length was statistically significant in EBUS.
• Normalized standard deviation on CTTA showed a comparable area under the receiver operating characteristic curve (0.742) to high SUVmax on FDG PET/CT (≥5, 0.739) for distinguishing malignant LNs.
• Normalized SD on CTTA exhibited low sensitivity despite high specificity.

Conclusions:

• CT texture analysis (CTTA) has potential utility in differentiating benign and malignant mediastinal LNs.
• The standalone diagnostic value of CTTA is currently limited.
• Integrated evaluation with other imaging modalities, such as PET/CT, is necessary for improved diagnostic performance.