PURPOSE:

We aimed to examine the utility of non-Gaussian diffusion kurtosis imaging (DKI) for assessment of chemotherapy response in patients with cervical (neck) non-Hodgkin lymphoma (NHL).

METHODS:

Patients with cervical NHL underwent 3.0 T magnetic resonance imaging with maximal b value of 2000 s/mm2 at baseline and seven days after chemotherapy onset. Apparent diffusion coefficient (ADC) value and diffusion kurtosis imaging maps for diffusion coefficient (D) and kurtosis (K) were calculated. Based on clinical examination, laboratory screening, and PET/CTs, patients were classified as responders or nonresponders.

RESULTS:

Twenty-six patients were enrolled. Among them, 24 patients were classified as responders and two as nonresponders. For responders, mean follow-up ADC and D increased significantly compared with baseline (ADC: 0.92±0.11 ×10-3 mm2/s vs. 0.68±0.11 ×10-3 mm2/s; D: 1.47±0.32 ×10-3 mm2/s vs. 0.98±0.21 ×10-3 mm2/s, P < 0.001 for both). Mean follow-up K decreased significantly compared with baseline (1.14±0.10 vs. 1.47±0.19, P < 0.001) for responders. Dratio showed significant positive correlation and high agreement with ADCratio (r = 0.776, P < 0.001). Likewise, Kratio showed significant negative correlation and high agreement with ADCratio (r = -0.658, P < 0.001).

CONCLUSION:

The new DKI model may serve as a new biomarker for the evaluation of early chemotherapy response in NHL.