Non-Gaussian diffusion imaging for malignant and benign pulmonary nodule differentiation: a preliminary study

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Standard diffusion-weighted imaging (DWI) apparent diffusion coefficient (ADC) shows inconsistent results for pulmonary nodule differentiation.
• Diffusion kurtosis imaging (DKI) quantifies non-Gaussian diffusion, potentially offering better micro-structural characterization than conventional DWI.

Purpose of the Study:

• To evaluate the feasibility of DKI in human lungs.
• To compare the diagnostic performance of DKI with standard DWI in distinguishing malignant from benign pulmonary nodules.

Main Methods:

• DKI and conventional DWI were performed on 35 pulmonary nodules in 32 patients.
• Acquired data included DKI with b-values of 0, 500, and 1000 s/mm² and DWI with b-values of 0 and 800 s/mm².
• Two independent observers assessed the diagnostic accuracy of mean kurtosis (MK) and ADC values, along with intra- and inter-observer repeatability (ICC).

Main Results:

• Diagnostic accuracy and AUC for differentiating malignancies were not significantly higher for MK compared to ADC.
• Observer 1a: MK (85.70%, AUC 0.87) vs. ADC (77.14%, AUC 0.81).
• Observer 1b: MK (80.00%, AUC 0.80) vs. ADC (80.00%, AUC 0.85).
• Observer 2: MK (82.80%, AUC 0.91) vs. ADC (77.14%, AUC 0.85).
• Substantial intra- and inter-observer agreement was observed for both MK and ADC values.

Conclusions:

• Diffusion kurtosis imaging (DKI) is feasible for use in human lungs.
• DKI-derived MK values did not demonstrate a significant advantage over ADC values in differentiating malignant from benign pulmonary nodules.