RSNA 2011 

Abstract Archives of the RSNA, 2011


SSA12-07

Nonlinear Relationship between Bone Marrow Signal Intensity and Apparent Diffusion Coefficient on Diffusion-weighted MRI

Scientific Formal (Paper) Presentations

Presented on November 27, 2011
Presented as part of SSA12: Molecular Imaging (Oncology)

Participants

Katherine van Ree MBBS, FRCR, Abstract Co-Author: Nothing to Disclose
Anwar Roshanali Padhani MD, Presenter: Consultant, F. Hoffmann-La Roche Ltd Consultant, IXICO Limited Consultant, Guerbet SA Advisory Board, Acuitus Medical Ltd Advisory Board, Bayer AG Advisory board, Siemens AG
Minal C. Jagtiani MBBS, MD, Abstract Co-Author: Nothing to Disclose
Shirley D'sa, Abstract Co-Author: Nothing to Disclose
Andreas Makris, Abstract Co-Author: Nothing to Disclose

PURPOSE

Diffusion-weighted MRI (DW-MRI) is helpful when evaluating the bone marrow for metastasis detection and therapy response. Metastases are easily detected using signal intensity (SI) on high b-value images. Disease progression causes increasing SI but ADC values can increase or decrease. Successful therapy also increases ADC values. These findings suggest a non-linear relationship between bone marrow SI and ADC. The purpose of this study was to evaluate the nature of the relationship between high b-value SI and ADC for normal and metastatic marrow before therapy, in-order to identify an ADC cut-off value beyond which therapy response could be inferred.

METHOD AND MATERIALS

A retrospective evaluation of 33 untreated patients undergoing abdominal & pelvic DW-MRI at 1.5T (EPI with STIR FS, bipolar gradients, 5mm slices, b-50 & b800-900 s/mm2). ADC values were calculated using mono-exponential curve fitting. Regions of interest (ROI) were placed in the capital femoral epiphysis (yellow bone marrow – YBM; n=13) and red bone marrow (RBM; n=16) of the pelvic bones in patients with no bony metastatic disease. 1-5 ROIs per patient were placed within metastatic breast (11 patients; 37 lesions) and multiple myeloma lesions (6 patients; 18 lesions). Area/lesion SI were normalized to psoas muscle. Non-parametric, ROC and non-linear regression analyses were performed.

RESULTS

YBM and RBM ADC values were lower than tumors (p <0.001; ROC 0.92; cut-off 686 µm2/s). SI of normal BM was also lower (p<0.001; ROC 0.94; cut-off 1.9) than tumors. Complete overlap of SI and ADC between myeloma and metastatic breast lesions was seen. 2nd order polynomial curve fitting provided the best fit to the observed data (R2=0.5). The maximum ADC reached by tumors was 1500 µm2/s.

CONCLUSION

Differences in muscle normalised SI and ADC allow the distinction between normal bone marrow and tumors. The results support the presence of a non-linear relationship between bone marrow SI and ADC. Post therapy tumor ADC values >1500 µm2/s are likely to indicate responding or necrotic lesions.

CLINICAL RELEVANCE/APPLICATION

By identifying the upper limit of ADC values in untreated bony tumors the effectiveness of successful tumor cell killing can be judged by ADC increases above the threshold value of 1500µm2/s

Cite This Abstract

van Ree, K, Padhani, A, Jagtiani, M, D'sa, S, Makris, A, Nonlinear Relationship between Bone Marrow Signal Intensity and Apparent Diffusion Coefficient on Diffusion-weighted MRI.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11004088.html