Value of ADC measurements for nodal staging after chemoradiation in locally advanced rectal cancer-a per lesion validation study

Objectives To evaluate the performance of diffusion-weighted MRI (DWI) in addition to T2-weighted (T2W) MRI for nodal restaging after chemoradiation in rectal cancer. Methods Thirty patients underwent chemoradiation followed by MRI (1.5 T) and surgery. Imaging consisted of T2W-MRI and DWI (b0, 500, 1000). On T2W-MRI, nodes were scored as benign/malignant by two independent readers (R1, R2). Mean apparent diffusion coefficient (ADC) was measured for each node. Diagnostic performance was compared for T2WMRI, ADC and T2W+ADC, using a per lesion histological validation. Results ADC was higher for the malignant nodes (1.43 +/- 0.38 vs 1.19 +/- 0.27 *10(-3) mm(2)/s, p<0.001). Area under the ROC curve/sensitivity/specificity were 0.88/65%/93% (R1) and 0.95/71%/91% (R2) using T2W-MRI; 0.66/53%/82% using ADC (mean of two readers); and 0.91/56%/98% (R1) and 0.96/56%/99% (R2) using T2W+ADC. There was no significant difference between T2W-MRI and T2W+ADC. Interobserver reproducibility was good for T2W-MRI (kappa 0.73) and ADC (intraclass correlation coefficient 0.77). Conclusions After chemoradiation, ADC measurements may have potential for nodal characterisation, but DWI on its own is not reliable. Addition of DWI to T2WMRI does not improve accuracy and T2W-MRI is already sufficiently accurate.