Stefan Schmidt, Presenter: Nothing to Disclose

Ken Herrmann, Abstract Co-Author: Nothing to Disclose

Hinrich Wieder, Abstract Co-Author: Nothing to Disclose

Bernd Krause, Abstract Co-Author: Nothing to Disclose

Markus Schwaiger MD, Abstract Co-Author: Research grant, Siemens AG

Ernst Josef Rummeny MD, Abstract Co-Author: Nothing to Disclose

The aim of this study was to compare different analysis methods of F-18-FDG PET/CT and MRI data for prediction of histopathological response (HPR) to neoadjuvant radiochemotherapy (RCTx) in patients with rectal cancer.

28 patients with advanced rectal cancer (uT3/4) underwent serial FDG-PET/CT scans at baseline, 14 days after initiation and after completion of RCTx. In addition MRI was performed at baseline and after the end of therapy. FDG uptake was assessed by calculating the mean SUV within different geometries: standard 2D region of interest (ROI, diameter 1.5 cm) and 3D volumes of interest (VOIs) with a fixed diameter as well as threshold based isocontours and the total lesion glycolysis (TLG) as combined morphological and metabolic parameter using the tumor volume defined in MRI. Changes between the parameters at the defined time points were calculated and analysed for their potential to predict HPR to RCTx (ROC analysis). Additionally the interobserver variability of some methods was analysed.

Histopathology classified eight of 28 patients as non-responders and 20 patients as responders to RCTx. ROC analysis of the standard 2D ROI technique revealed areas under the curve (AUCs) of 0.64 and 0.71 for the early and late time point. Corresponding AUCs for 3D VOI techniques resulted in AUCs ranging from 0.70 - 0.75 (early time point) and 0.69 - 0.76 (later time point), respectively. AUCs of isocontours with an organ-defined cut-off showed values of 0.68 - 0.69 (early time point) and 0.61 - 0.63 (later time point). AUCs of the different TLGs were between 0.63 and 0.66 for the later time point, AUC for MRI lesion volume change as only pure anatomical parameter was 0.46.

Our study demonstrates that 3D based approaches for assessing SUV values result in better AUC values for prediction of HPR to neoadjuvant RCTx in patients with rectal cancer. Isocontour methods showed less influence of interobserver variability than methods with fixed size volumes of interest. MRI showed the worst result for therapy response assessment, hence the combined anatomic and metabolic parameters showed poor results too.          

Methods for early response prediction to neoadjuvant radiochemotherapy in locally advanced rectal cancer by comparing FDG-PET/CT and MRI as a routine diagnostic procedure.

Schmidt, S, Herrmann, K, Wieder, H, Krause, B, Schwaiger, M, Rummeny, E, Comparison of Different SUV-based Methods for Response Prediction to Neoadjuvant Radiochemotherapy in Locally Advanced Rectal Cancer by FDG-PET/CT and MRI  .  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8012674.html