(Monday, 7/30/2018) 9:30 AM - 10:00 AM
Room: Exhibit Hall | Forum 8
Purpose: To investigate if Diffusion Kurtosis Imaging (DKI) offers additional value beyond conventional DWI for predicting the response to neoadjuvant chemoradiation in esophageal cancer.
Methods: Twenty patients with esophageal cancer were included in the study. At baseline and interim of neoadjuvant chemoradiation, patients underwent separate MR scans that included conventional DWI and DKI studies. After therapy, patients were divided into pathologic complete response (pCR) and non-pCR based on either histologic (surgical) or FDG-PET findings (non-surgical). The monoexponential ADC protocol (MONO) had the following b-values: 0,200,800 s/mm². The DKI protocol (DKI) had the following b-values: 200,600,1000,1200,1600 s/mm². Images from the MONO scan were fit with a monoexponential model to generate a quantitative Apparent Diffusion Coefficient (ADC) map (ADC_MONO). Images from the DKI scan were fit to a kurtosis model to generate a quantitative map of apparent kurtosis (k_app). Tumor volumes were contoured and average diffusion parameter values were extracted. Diffusion parameters at baseline and as relative changes (Δ) at interim were compared between pCR and non-pCR. Average diffusion parameters were compared for correlation between MONO and DKI scans.
Results: Consistent with a previous report, patients with pCR had higher ΔADC_MONO as compared to non-pCR patients (AUC=0.85). Patients with pCR also had lower Δk_app as compared to non-pCR (AUC=0.81). Neither of the diffusion metrics achieved statistical significance. There was not a strong correlation between Δk_app and ΔADC_MONO (Spearman's Ï?=-0.098).
Conclusion: Our study does not show that DKI is more useful than ADC in predicting the response to neoadjuvant chemoradiation of esophageal cancer. However, the changes in ADC and kurtosis were not well correlated and both had similar ROC performance in separating pCR and non-pCR patients. We therefore expect that DKI may offer additional value when combined with ADC. Further validation is needed with a larger number of patients.
Keywords
Diffusion, Quantitative Imaging, MRI
Taxonomy
Contact Email
