Histogram analysis parameters identify multiple associations between DWI and DCE MRI in head and neck squamous cell...

Nowadays, multiparametric investigations of head and neck squamous cell carcinoma (HNSCC) are established. These approaches can better characterize tumor biology and behavior. Diffusion weighted imaging (DWI) can by means of apparent diffusion coefficient (ADC) quantitatively characterize different tissue compartments. Dynamic contrast-enhanced magnetic resonance imaging (DCE MRI) reflects perfusion and vascularization of tissues. Recently, a novel approach of data acquisition, namely histogram analysis of different images is a novel diagnostic approach, which can provide more information of tissue heterogeneity. The purpose of this study was to analyze possible associations between DWI, and DCE parameters derived from histogram analysis in patients with HNSCC.

Overall, 34 patients, 9 women and 25 men, mean age, 56.7±10.2years, with different HNSCC were involved in the study. DWI was obtained by using of an axial echo planar imaging sequence with b-values of 0 and 800s/mm². Dynamic T1w DCE sequence after intravenous application of contrast medium was performed for estimation of the following perfusion parameters: volume transfer constant (K_trans), volume of the extravascular extracellular leakage space (Ve), and diffusion of contrast medium from the extravascular extracellular leakage space back to the plasma (Kep). Both ADC and perfusion parameters maps were processed offline in DICOM format with custom-made Matlab-based application. Thereafter, polygonal ROIs were manually drawn on the transferred maps on each slice. For every parameter, mean, maximal, minimal, and median values, as well percentiles 10th, 25th, 75th, 90th, kurtosis, skewness, and entropy were estimated.

Сorrelation analysis identified multiple statistically significant correlations between the investigated parameters. Ve related parameters correlated well with different ADC values. Especially, percentiles 10 and 75, mode, and median values showed stronger correlations in comparison to other parameters. Thereby, the calculated correlation coefficients ranged from 0.62 to 0.69. Furthermore, K_trans related parameters showed multiple slightly to moderate significant correlations with different ADC values. Strongest correlations were identified between ADC P75 and K_trans min (p=0.58, P=0.0007), and ADC P75 and K_trans P10 (p=0.56, P=0.001). Only four K_ep related parameters correlated statistically significant with ADC fractions. Strongest correlation was found between K_ep max and ADC mode (p=-0.47, P=0.008).

Multiple statistically significant correlations between, DWI and DCE MRI parameters derived from histogram analysis were identified in HNSCC.