1. MRI/MRS Biomarkers Robert E. Lenkinski, Ph.D. http://catalyst.harvard.edu

2. 1 The major focus will be on non-neuro applications Issues-quantitation, relationship to physiology, metabolism Examples-DCEMRI, ASL, DWI, MRS

3. Issues Quantitation-accuracy, precision, detection limits Standardization-acquisition, processing/analysis Validation ADNI, NIHPD, OAI, RSNA-QIBA, ACRIN 2

4. Ashton, JMRI 31:279-288 (2010) 3

5. 4 http://www.adni-info.org/ ✖ ✖ ✖ RF Pulse Signal

6. 5 http://www.adni-info.org/

7. 6

8. 7

9. 8

10. https://nihpd.crbs.ucsd.edu/nihpd/info/dat a_access.html 9

11. 10

12. https://nihpd.crbs.ucsd.edu/nihpd/info/ data_access.html 11

13. Contrast Enhanced MRI of Has Diagnostic Value in Oncology Breast, Lung, Prostate, etc.

15. Behavior of Contrast Agent in the Body B O D Y Injection of Contrast Agent Kidney  Depends on: Cellular density or “Extracellular Volume Fraction” Blood vessel permeability “Microvascular Permeability” Blood Stream

16. Dynamic Contrast Enhanced MRI (DCEMRI) Components –“High-field” MRI machine (1.0 tesla or greater) –Phased array torso coil –Gadolinium contrast agent (GdDTPA) –Images taken at several time points (spatial vs temporal resolution –Software algorithm processes data for either parametric maps or semi- quantitative plots

18. 17 Juergen F. Schaefer, Joachim Vollmar, Fritz Schick, Reinhard Vonthein, Marcus D. Seemann, Herrmann Aebert, Rainer Dierkesmann, Godehard Friedel, and Claus D. Claussen Solitary Pulmonary Nodules: Dynamic Contrast- enhanced MR Imaging—Perfusion Differences in Malignant and Benign Lesions Radiology August 2004 232:544-553;

19. Fourteen malignant and no benign nodules demonstrated curve type A, demonstrating a fast initial SI increase and a marked decrease after 20 or 30 seconds. No SI decrease after the first bolus transit is present in curve type B, which was demonstrated by 18 nodules (12 malignant and six benign nodules). A more continuous SI increase without any sharp early peak after the first bolus transit demonstrates that curve type C was found in 14 nodules (one malignant and 13 benign nodules). Note that ultimately, the curves demonstrate a plateau. No relevant enhancement was calculated for the five benign nodules with curve type D. 14M-0B12M-6B 1M-13B0M-5B

20. 56yo w/ ↑ PSA (now 27.7) repeat –bx x3; 63 cores prior to MRI 19 T2-W DCEMRI color map pT2c Gleason 4+3

21. Tofts Model Equation SI= [a 1 *(e (-ktrans*t/ve) -e (-m1*t))/(m1-ktrans/ve) + a 2 *(e (-ktrans*t/ve) -e (-m2*t))/(m2-ktrans/ve )+ a 1 *e (-m1*t) +a 2 *e (-m2*t) ] * d*k trans Two Compartment Model Negligble vascular space Idealized arterial input function (SI linear with Gd Agent Concentration) 20

22. Challenges/Issues How accurate is the model? How precise are the values of Ktrans and Ve? What is the influence of SNR and temporal resolution? 21

23. Modified ADNI/IRAT phantom for DCE-MRI Defined generic DCE-MRI acquisition protocols Conduct multi-center phantom reproducibility study Define procedure for routine phantom use Develop simulated data set for algorithm testing RSNA QIBA DCE-MRI Technical Committee

24. http://www.rsna.org/research/qiba.cfm 23

25. 24

26. http://www.rsna.org/research/qiba.cfm 25

27. 26

28. Anatomic T2W and T1W early subtraction images and parametric maps for rBV and Ktrans at baseline and following 2 cycles of neoadjuvant chemotherapy in a clinically and pathologically (A) responding patient and (B) nonresponding patient. Ah-See M W et al. Clin Cancer Res 2008;14:6580-6589 ©2008 by American Association for Cancer Research

29. Change in MRI-derived tumor size and DCE-MRI kinetic parameters according to (A) clinical tumor response and (B) pathologic tumor response. Ah-See M W et al. Clin Cancer Res 2008;14:6580-6589 ©2008 by American Association for Cancer Research

30. ASL Subtraction Experiment Control Image Labeled Image Imaged Slice Inversion Labeling

31. Background Suppression is Enabling for Abdominal ASL Control Label Difference

32. DAY 0 DAY 8 sorafenib + bevacizumab

33. Early changes at 1 mo in blood flow and tumor size compared with tumor size changes at 4 mo with a Spearman rank order correlation. de Bazelaire C et al. Clin Cancer Res 2008;14:5548-5554 ©2008 by American Association for Cancer Research

34. Early changes at 1 mo in blood flow and tumor size compared with delay of progression of the disease after initiation of the treatment. de Bazelaire C et al. Clin Cancer Res 2008;14:5548-5554 ©2008 by American Association for Cancer Research

35. Beth Israel Deaconess Med. Ctr. Weiying Dai, PhD Philip Robson, PhD Cedric de Bazelaire, MD Guillaume Duhamel, PhD Dairon Garcia MS Barbara Appignani, MD Michael Atkins, MD Tamara Fong, MD PhD Daniel George MD David Hackney, MD Igor Koralnik, MD Barbara Klemm, RN Ivan Pedrosa, MD Daniel Press, MD Neil Rofsky, MD Acknowledgments Gottfried Schlaug, MD Magdy Selim, MD Eric T. Wong, MD University of Pennsylvania John Detre, MD Jiongjiong Wang, MD GE Global Applied Sciences Lab Ananth Madhuranthakam, PhD Ajit Shankaranarayanan, PhD Stanford University Greg Zaharchuk, MD PhD

36. 35

37. 36 Gradient time

38. 37

39. 38

40. 39

41. functional Diffusion Maps (fDM) ADC maps from pre and post treatment are aligned Voxel-wise statistical analysis highlights the voxels with significant change over time  Results only for brain and animals models  Abdomen application: Non- rigid registration will introduce the registration error into the fDM computation Image from: Hamstra et al, Proc Natl Acad Sci U S A, 2005

43. Our approach NdH/dT: Normalized cumulative histogram difference over time –Difference between the Cumulative histograms of the tumor ADC values –Area Under the Curve (AUC) represent the overall change in tumor diffusivity –Normalization by the AUC of healthy tissue sample produce absolute global measure Single number Intuitive to interpret No non-rigid registration is required Capture tumor heterogeneity

44. NdH/dT: Representative examples

45. RESONANCE ASSIGNMENTS SUPPRESSED X 30 UNSUPPRESSED X 1 SUPPRESSED X 525 X 1500 Cho H20H20Fat Cho H20H20 Fat H20H20 -CH 2 - Cho -CH 3 - -(CH 2 ) n - -C=C-

46. Journal of the National Cancer Institute, Vol. 94, No. 16, 1197-1203, August 21, 2002 © 2002 Oxford University Press ------------------------------------------------------------------------ REVIEW Clinical Utility of Proton Magnetic Resonance Spectroscopy in Characterizing Breast Lesions Rachel Katz-Brull, Philip T. Lavin, Robert E. Lenkinski

50. Copyright ©Radiological Society of North America, 2004 Meisamy, S. et al. Radiology 2004;233:424-431 Figure 2

51. Copyright ©Radiological Society of North America, 2004 Meisamy, S. et al. Radiology 2004;233:424-431 Figure 5

52. MRI/MRS is Complicated Navigating through the maze to reach quantitation requires a systematic approach 51