MRI MRS Biomarkers Robert E Lenkinski Ph D http

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

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

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

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

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

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

6

7

8

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

https: //nihpd. crbs. ucsd. edu/nihpd/info/ data_access. html 10

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

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

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

Dynamic Contrast Enhanced MRI (DCEMRI) Components – “High-field” MRI machine (1. 0 tesla or greater) – Phased array torso coil – Gadolinium contrast agent (Gd. DTPA) – Images taken at several time points (spatial vs temporal resolution – Software algorithm processes data for either parametric maps or semiquantitative plots

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 Contrastenhanced MR Imaging—Perfusion Differences in Malignant and Benign Lesions Radiology August 2004 232: 544 -553; 17

14 M-0 B 1 M-13 B 12 M-6 B 0 M-5 B 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.

56 yo w/ ↑ PSA (now 27. 7) repeat –bx x 3; 63 cores prior to MRI T 2 -W p. T 2 c Gleason 4+3 DCEMRI color map 19

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

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

RSNA QIBA DCE-MRI Technical Committee • 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

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

24

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

26

Anatomic T 2 W and T 1 W early subtraction images and parametric maps for r. BV 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

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

ASL Subtraction Experiment Control Image Labeled Imaged Slice Inversion Labeling

Background Suppression is Enabling for Abdominal ASL Control Label Difference

DAY 0 DAY 8 sorafenib + bevacizumab

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

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

Acknowledgments Beth Israel Deaconess Med. Ctr. Weiying Dai, Ph. D Philip Robson, Ph. D Cedric de Bazelaire, MD Guillaume Duhamel, Ph. D Dairon Garcia MS Barbara Appignani, MD Michael Atkins, MD Tamara Fong, MD Ph. D Daniel George MD David Hackney, MD Igor Koralnik, MD Barbara Klemm, RN Ivan Pedrosa, MD Daniel Press, MD Neil Rofsky, MD 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, Ph. D Ajit Shankaranarayanan, Ph. D Stanford University Greg Zaharchuk, MD Ph. D

35

Gradient time 36

37

38

39

functional Diffusion Maps (f. DM) • 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: Nonrigid registration will introduce the registration error into the f. DM computation Image from: Hamstra et al, Proc Natl Acad Sci U S A, 2005

Our approach • Nd. H/d. T: 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

Nd. H/d. T: Representative examples

RESONANCE ASSIGNMENTS SUPPRESSED H 2 0 -(CH 2)n- X 30 X 1500 Cho -C=C- Cho -CH 2 - -CH 3 - SUPPRESSED UNSUPPRESSED Cho H 2 0 X 1 X 525 Fat H 2 0 Fat

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

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

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

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