Procedure Manual for MRI of the Brain

Baxter Protocol 161003 SYN RC W H E R E S C I E N C E M E E T S S E R V I C E

Baxter Protocol 161003 A Phase 3 Randomized, Double-Blind, Placebo-Controlled Study of the Safety and Effectiveness of Immune Globulin Intravenous (Human), 10% Solution (IGIV, 10%) for the Treatment of Mild to Moderate Alzheimer s Disease (AD) Synarc Code: BAXT 2207 Version 1.0 10 November 2011 2011 SYNARC Inc. ALL RIGHTS RESERVED. No part of this work covered by the copyright hereon may be reproduced or used in any form or by any means - graphic, electronic, or mechanical, including photocopying, recording, taping, or information storage and retrieval systems - without permission of SYNARC Inc.

Procedure Manual Sign-Off Dear Study Coordinator and MRI Technologist(s): Synarc requires that the Study Coordinator and all MRI Technologists involved in the Baxter protocol 161003 study read and fully understand the Procedure Manual for MRI Imaging. This requirement should be completed before imaging subjects in the 161003 study. All new personnel, who join the study after site initiation, are also required to read and understand this manual. Please have all applicable study personnel sign and date this form to confirm completion of this requirement. Please fax or mail a copy of this form to Synarc and keep a copy for your records. Synarc Inc. Baxter 161003 Study Team (BAXT2207) 7707 Gateway Boulevard, 3 rd Floor Newark, CA 94560 Fax: 415-817-8999 Principal Investigator: Site Number: Please indicate who is responsible for submitting packages to Synarc: Study Coordinator Technologist Study Coordinator (Study Coordinators are responsible for reviewing and understanding the Logistics portion of this manual.) Printed Name Signature Initials MRI Technologist (MRI Technologists are responsible for reviewing and understanding the entire manual.) Printed Name Signature Initials Date (DD- MMM-YYYY) Date (DD- MMM-YYYY) - - Do not remove this page send a photocopy only- - i Version 1.0 10 November 2011

Table of Contents PROCEDURE MANUAL SIGN-OFF... I GLOSSARY AND ABBREVIATIONS... IV 1.0 INTRODUCTION... 1 2.0 STUDY INTRODUCTION... 2 2.1 STUDY OVERVIEW... 2 2.2 ROLE OF SYNARC... 2 2.3 RESPONSIBILITIES OF CLINICAL SITES... 3 CONTINUED RESPONSIBILITIES OF CLINICAL SITES:... 3 2.4 RESPONSIBILITIES OF MRI FACILITIES... 4 CONTINUED RESPONSIBILITIES OF IMAGING FACILITY:... 4 2.5 CONTRAINDICATIONS FOR THE MRI STUDY... 4 3.0 MRI STUDY SITE QUALIFICATION PROCESS... 5 3.1 PRE-TRIAL QUESTIONNAIRE... 6 3.2 TELEPHONE TRAINING FOR MRI SITES... 6 3.3 MRI INSTRUMENT ASSESSMENT... 6 3.3.1 Pre-study Qualification Phantom Scan... 7 3.3.2 First Patient Scan... 8 4.0 PROCEDURES FOR ACR PHANTOM SCANS... 9 4.1 PHANTOM SCHEDULING... 9 4.1.1 Pre-study Qualification Phantom Scan... 9 4.1.2 Longitudinal Phantom Scans... 9 4.2 EVALUATION OF PHANTOM SCANS... 9 4.2.1 Requests for Resubmission of Phantom Data... 9 4.2.2 Requests for Repeat Phantom Scans...10 4.3 LABELING PHANTOM IMAGES IN THE DIGITAL HEADER...10 4.4 PHANTOM POSITIONING...10 4.5 PHANTOM SCAN ACQUISITION TECHNIQUE...11 4.5.1 ACR Sagittal Locator...11 ACR SAGITTAL LOCATOR...11 4.5.2 ACR 11-slice Axial T1-weighted Spin Echo...12 4.5.3 Slices 1 to 11 of the ACR Axial T1 sequence...13 4.5.4 3D T1-weighted sequence...13 4.5.5 Criteria for Assessing Quality of Phantom MRI Scans...14 4.5.6 Clinical sequence parameter check during pre-qualification Phantom MRI Scan...14 4.6 COMMON PROBLEMS...14 4.6.1 Incorrectly (Mis-) prescribed Slices...14 4.6.2 Phantom not centered in the Field of View...15 4.6.3 Geometric Accuracy...15 4.6.4 High Contrast Spatial Resolution...16 4.6.5 Image Intensity Uniformity...16 4.6.6 Low Contrast Object Detection...17 5.0 PROCEDURES FOR SUBJECT MRI SCANS...18 5.1 PATIENT SCHEDULING...18 5.1.1 First Patient Scan...18 ii Version 1.0 10 November 2011

5.1.2 Patient scans...18 5.2 EVALUATION OF SUBJECT SCANS...18 5.2.1 Requests for Resubmission of Subject Data...19 5.2.2 Requests for Repeat Subject Scans...19 6.0 SUBJECT MRI EXAM PREPARATION...19 6.1 SUBJECT SAFETY AND MONITORING...19 6.2 SUBJECT POSITIONING...19 7.0 SUBJECT MRI ACQUISITION TECHNIQUE...22 7.1 LABELING DIGITAL HEADER FIELDS FOR SUBJECT MRI SCANS...22 7.2 PRE-SCAN ADJUSTMENTS...22 7.3 PROTOCOL FOR MRI OF THE BRAIN...22 7.3.1 Localizer: 3-Plane Gradient Sequence (Scouts) and Sagittal Spin Echo...23 7.3.2 3D T1 / MP-RAGE / IR-prepped fast SPRR / TFE...24 FIGURE 4. 3D T1 SAGITTAL ORIENTATION AND POSITIONING...25 7.3.3 Acquire a Second 3D T1 Sequence...26 7.3.4 Axial 2D PD/T2 Turbo Spin Echo (TSE) / Fast Spin Echo (FSE)...27 7.3.5 Axial 2D FLAIR / T2* Gradient Echo / Diffusion Weighted Imaging (DWI)...28 8.0 COMMON PROBLEMS SEEN WITH SUBJECT MRI SCANS...28 8.1 INCORRECT ORIENTATION / IMAGE PLANE MISALIGNMENT...28 8.2 SWAP OF READ AND PHASE ENCODING DIRECTIONS...29 8.3 GHOSTING ARTIFACTS...29 8.4 FLOW COMPENSATION...30 8.5 ALIASING (FOLDING) IN 3D T1...30 8.6 POOR SIGNAL-TO-NOISE RATIO (SNR) AND CONTRAST-TO-NOISE RATIO (CNR)...31 8.7 SIGNAL LOSS...31 8.7.1 Inferior Slices...31 8.7.2 Inhomogeneity or Shading Artifact...32 8.8 METAL ARTIFACT...32 8.9 INADEQUATE HEAD COVERAGE...33 9.0 DATA ARCHIVE ONTO DIGITAL MEDIA...34 10.0 DATA SUBMISSION PROCEDURES...34 10.1 SUBMITTING MRI DATA ARCHIVED ON DIGITAL MEDIA...34 10.2 DATA PREPARATION...35 10.3 DATA PREPARATION...35 10.4 SUBMITTING DATA VIA SYNARC S FTP WEBSITE (SYNARCCONNECT.COM)...35 11.0 SUPPLIES PROVIDED BY SYNARC...36 APPENDIX I: TECHNICAL PARAMETERS FOR MRI OF THE BRAIN...37 APPENDIX II: TECHNICAL PARAMETERS FOR ACR PHANTOM SCANS...44 APPENDIX III: INSTRUCTIONS FOR COMPLETING THE TRANSMITTAL FORM FOR MRI OF THE BRAIN...51 APPENDIX IV: TRANSMITTAL FORM FOR MRI OF THE BRAIN...52 APPENDIX V: TRANSMITTAL FORM FOR ACR PHANTOM MRI...52 APPENDIX VI: STUDY SUPPLY ORDER FORM...53 iii Version 1.0 10 November 2011

Glossary and Abbreviations Abbreviation ACR AWB CD CT DCF DICOM DWI FLAIR FTP GCP IGIV IPN MIRRN MRI NCR OS PI QC QRG SC SE TE TF TI TR YOB Definition American College of Rheumatology Airway Bill Compact Disk Computed Tomography Data Clarification Form Digital Imaging and Communications in Medicine Diffusion Weighted Images Fluid attenuation inversion recovery File Transfer Protocol Good Clinical Practice Immune Globulin Intravenous (Human) Incomplete Package Notification Medical Image Repeat Request Notification Magnetic Resonance Imaging Non Carbon Required Operating System Principal Investigator Quality Control Quick Reference Guide Study Coordinator Spin Echo Echo Time Transmittal Form Inversion Time Repetition Time Year of Birth iv Version 1.0 10 November 2011

1.0 Introduction The purpose of this manual is to standardize MRI image acquisition procedures between the sites participating in the Baxter 161003 study. All radiologists and technologists contributing to this study are expected to have had appropriate theoretical and practical training in MRI. Study personnel should also satisfy all local requirements for radiology licensing and registration. For the safety of subjects and technologists alike, an understanding of radiation risks and radiation safety procedures is also required. Utilizing qualified radiology personnel is the first step toward the successful use of medical imaging in this study. The procedure manual is designed for the study coordinator and the MRI technologists involved in this study. All new personnel, who join the study after site initiation, are also required to read and understand the manual, and return the manual sign-off on Page i. This manual, taken alone, should not be considered as sufficient training in the proper technique for acquiring MRI images. The goal of the manual is to define a standard procedural approach for acquiring MRIs of sufficient quality for achieving the study goals. Questions regarding this manual or MRI techniques should be directed to: BAXTER 161003 Study Team (BAXTER 2207) Synarc Inc. Baxter 161003 7707 Gateway Boulevard, 3 rd Floor Newark, CA 94560 Tel: 415-817-8900 Fax: 415-817-8999 Email : BAXT2207@synarc.com HOURS: 9.00 17.00 Pacific Time 1 of 54 Version 1.0 10 November 2011

2.0 Study Introduction 2.1 Study Overview Structural imaging will be used in the 161003 protocol to evaluate the effect of IGIV, 10% on the rate of Alzheimer s disease progression in the brain using various volumetric measurements. It is anticipated that approximately 402 subjects at 65 sites worldwide will participate in the study. Each subject will have MRI scans acquired at Baseline, 9-Month, and 18-Month, or upon Early Termination. Unscheduled MRI scans might also be acquired during the course of the study and sent to Synarc for evaluation. All Unscheduled visits should be pre-approved by Baxter. For all unscheduled MRI visits, the site should contact their Field CRA s in order to obtain approval from Sponsor. MRI is the most sensitive method for determining the incidence of brain atrophy and for monitoring the progression of neurodegeneration both across the whole brain and in specific structures. The objective of the MRI study is to provide supporting evidence that IGIV, 10% attenuates the underlying pathologic process of AD by assessing rate of decline in Whole Brain Atrophy, Ventricular Volume, Hippocampal Volume, and Entorhinal Cortical Volume. 2.2 Role of Synarc Synarc s Primary Responsibilities to Site Qualify MRI facilities into the study. Synarc will conduct an initial MRI Instrument Assessment through the evaluation of Pre-Trial Questionnaires, American College of Radiology (ACR) phantom scans, and First Patient Qualification scans. Conduct trainings for MRI Technologists on the MRI acquisition procedures specific to the Baxter MRI study. Provide Study Manual and Quick Reference Guides (QRG) for subject and phantom scanning. Evaluate subject scans and submit timely Quality Control Reports (as defined in Section 5.0, Evaluation of Subject MRI Scans Continued responsibilities are to: Collect and archive exams. Verify that the sequence parameters used to acquire phantom and subject MRI exams are in agreement with the study procedure manual. Review the quality of the subject MRI images for adequate anatomical coverage, signal-tonoise ratio, and the presence of artifacts. Submit quality control (QC) reports to the study coordinator and MRI facility detailing any issues regarding image quality found during phantom and subject scan QC along with suggestions for improvement. Provide ongoing support and feedback to clinical sites and MRI facilities. 2 of 54 Version 1.0 10 November 2011

Synarc personnel will review the quality of all MRI data submitted. It is expected that the majority of examinations received will be of acceptable quality. If any problems are detected related to image quality, Synarc will notify the responsible clinical site and the MRI facility via fax, email or phone, suggest possible causes of the problem, and offer potential solutions. The MRI facility should try to correct these errors and avoid them in future exams. Despite this review, the acquisition of MRI scans of acceptable quality remains the responsibility of the MRI facility. 2.3 Responsibilities of Clinical Sites Primary Responsibilities of Clinical Sites Ensure that patients enrolling in this study do not have any MR contraindications and that patients are good candidates for tolerating an MRI scan and possible repeat exams. Schedule subject exams for all visits in conjunction with the MRI technologist. During site initiation, patients should only be scheduled after the site has received training from Synarc and submitted a passing phantom scan and a passing first patient scan (see Sections 3.3.1 and 3.3.2 of this manual). For the duration of the study, patients should avoid being scheduled and scanned when there is a pending repeat request for a phantom scan (see Section 4.2.2). Provide all subject demographic and exam information to the MRI technologist so that this information is entered completely and correctly on the MRI Transmittal Form that is submitted with the data. Ensure that MRI data and corresponding Transmittal Form are submitted to SYNARC within 24-hours of acquisition. Continued responsibilities of Clinical Sites: Confirm receipt of Synarc supplies and distribute these supplies to the appropriate study personnel, based on defined roles in the data acquisition and submission process. Notify Synarc when all subjects enrolled in the study have completed an 18-Month or Early Termination MRI scan. Ensure that the MRI personnel have a copy of the Quick Reference Guide for Acquiring MRI of the Brain. Do not proceed with scanning or scheduling additional subjects until MRI facility has been approved by Synarc. Ensure that MRI facility is performing and submitting longitudinal phantom scans. Notify Synarc about planned upgrades at the MRI facility or other issues that might compromise the consistency MRI scanning over time. 3 of 54 Version 1.0 10 November 2011

2.4 Responsibilities of MRI Facilities Primary Responsibilities of MRI Facility Acquire MR scans for phantoms and subjects in compliance with Synarc s procedures and imaging protocol detailed in the Procedure Manual for MRI Imaging of the Brain. Verify that all subject demographic and exam information are entered completely and correctly on the MRI Transmittal Form that is submitted with the MRI data. Verify that the electronic MRI header is entered completely and correctly and in compliance with privacy laws to ensure and protect the confidentiality of the patient (no patient names or identifiers). Send MRI data and corresponding Transmittal Form to SYNARC within 24-hours of acquisition. Submit Longitudinal ACR Phantom scan every three months until the end of the study (as defined in Section 4.1.2, Longitudinal Phantom Scans ). Continued responsibilities of imaging facility: Confirm that all MRI technologists who will be performing scans for the Baxter 161003 protocol are properly trained on the study-specific acquisition parameters and data submission procedures. Scan ACR phantom and perform system calibrations at the beginning of the study. Do not scan patients until an approved pre-study phantom QC report has been received from Synarc. Do not scan additional patients until a passing First Patient Qualification Scan QC report has been received from Synarc. Review exam quality and obtain repeats as necessary. Maintain an archive of subject and phantom MRI exams. Notify Synarc immediately as soon as they are aware that a hardware or software upgrade is scheduled. Synarc will evaluate the planned upgrade and provide feedback to the facility as to whether this will compromise volumetric analysis. * In some cases, the responsibilities of the Clinical Site and the MRI facility overlap. We ask that the Study Coordinator work with the lead technologist to decide how overlapping responsibilities will be assigned (for example, who will fill out Transmittal Forms and mail data to Synarc). At site training, Synarc will document which party is responsible for these activities for the duration of the study. 2.5 Contraindications for the MRI Study Subjects may not undergo MR scans with certain contraindications. It is the responsibility of the MR imaging technologist or Radiologist to check for all contraindications. If a contraindication is found for a subject, that subject shall not be scanned and the MR facility shall report this finding to the Clinical site. 4 of 54 Version 1.0 10 November 2011

3.0 MRI Study Site Qualification Process Synarc will assist in identifying and qualifying facilities for participation in the study. considered for participation, MRI facilities must meet the requirements listed in the box below: To be Preliminary Requirements for Imaging Facilities MRI scanner must have a magnetic field strength of 1.5 Tesla ONLY. MRI scanner must meet certain requirements with regards to hardware and/or software. No major hardware or software upgrades that would compromise volumetric analysis of existing subject scans. MRI scanner must be manufactured by Siemens (Symphony, Espree, Avanto or Aera), General Electric (Excite or MR450), or Philips (Intera, Achieva or Ingenia). Certain older vendor release levels will not be allowed. Facilities must have a volumetric head coil. Phased-array head (multiple channels) coils may be allowed under conditions as defined by Synarc. The site may NOT change RF coil once approved. Surface coils cannot be used to acquire any scans for this study. MRI software must be capable of acquiring the 3-D sequence as specified for a given scanner. Facilities must be able to digitally archive data on CD or FTP in uncompressed DICOM format. Facilities must be able to perform the required MRI scans using the Technical Parameters specified in Appendices I and II of the MRI Procedure manual. MRI Scanner to be used for the study should not be a mobile MRI scanner or must not be moved during the study. Only one MR scanner may be used at a facility and the scanner may not be switched during the length of the study. MRI facilities that meet the requirements listed above will then advance through subsequent qualification procedures. An overview of the complete Site Qualification procedure is detailed below: 1. Pre-Trial Questionnaire: MRI facility must complete and submit a Synarc Pre-Trial Questionnaire. 2. Telephone Training: Primary MRI technologists must complete a telephone-training with Synarc to ensure comprehension of and adherence to MRI protocol and procedures. 3. MRI Instrument Assessment: a. Pre-Study ACR Phantom Scan Upon completion of telephone training, site will be asked to perform an ACR Phantom Scan. b. Patient Scan Upon the approval of an ACR phantom scan, the site will be asked to perform a First Patient Scan. The clinical site and MRI facility will be notified of its complete eligibility to participate in the MRI study once Synarc has received and reviewed all of the MRI Instrument assessment data. 5 of 54 Version 1.0 10 November 2011

3.1 Pre-Trial Questionnaire Site Questionnaires will be collected from each MRI facility considered for participation in the Baxter 161003 study. The questionnaire will ask for information about the MRI technologists as well as the MRI equipment. Although several technologists may work on this study from one site, the MRI facility must designate one MRI technologist to be the Primary MRI Technologist to work on this study. The responsibility of the Primary Technologist is to ensure that the imaging protocol is followed when acquiring any subject or phantom scans for this study, as well as to ensure that other MRI technologists acquiring images for this study read and understand the procedures detailed in this manual. IMPORTANT If the Primary MRI Technologist leaves the MRI facility, it is the responsibility of the MRI facility to ensure that he/she trains the replacement technologist on the protocol provided by this manual. A complete manual sign-off is required from the new personnel. Should the new personnel have any questions, they may contact Synarc. 3.2 Telephone Training for MRI Sites Once Synarc has received the completed Pre-Trial Questionnaire and determined that the site meets the Preliminary Requirements for MRI Facilities (as defined in Section 3.0, MRI Study Site Qualification Process ), a 45-60 minute telephone training will be scheduled with the technologists at each site. The goal of the telephone training is to: Instruct sites on study specific protocols for acquiring acceptable MRI images of the brain and phantom. Allow for troubleshooting of potential common problems. Explain the submission process of MRI data to Synarc. IMPORTANT Sites are not allowed to acquire subject MRI exams until they have received training from Synarc and have submitted an acceptable ACR Phantom scan. 3.3 MRI Instrument Assessment The overall purpose of the MRI Instrument Assessment is to detect and reduce measurement noise on one scanner per site, and to minimize scanner differences between sites, ensuring the most standard and reliable data for volumetric analysis. The Phantom and First Patient assessments are critical, because each provides unique information about scanner performance that allows the evaluation, optimization and correction of scan parameters. 6 of 54 Version 1.0 10 November 2011

3.3.1 Pre-study Qualification Phantom Scan Each site is required to use an American College of Radiology (ACR) phantom to perform all phantom test-run scans. If a site s MRI facility does not have an ACR phantom, Synarc will provide one for the duration of the study. An ACR phantom test scan is collected for the following purposes: To assess important instrument parameters, including gradient fidelity and signal-to-noise ratio (SNR), and to determine whether scanner performance is acceptable for the study. The purpose of phantom test scans is to reduce the measurement noise due to instrument variability. To calibrate important instrument parameters, including image resolution and gradient linearity over field of view, and to minimize differences between scanners. The purpose of system calibration is to minimize these measurement errors. Check the clinical scanning parameters on the ACR phantom during pre-qualification to insure that: 1) the scanner can perform the desired clinical sequences and 2) the correct clinical sequences are stored properly in the scanner s preset protocol section. The following sequences will be run for the Pre-qualification scan: Localizer 3-Plane Gradient (optional, used to assess phantom positioning) ACR single slice Sagittal Spin Echo ACR 11 slice Axial T1-weighted Spin Echo 3D T1-weighted sequence as specified in this manual and/or Quick Reference Guide (QRG): MP-RAGE (Siemens), IR-prepped fast SPGR (GE), 3D TSE (Philips). Note: All GE scanners use the coronal plane, while all Philips and Siemens scanners use the sagittal plane. 2D Axial TSE/FSE dual echo 2D Axial FLAIR 2D Axial T2* gradient echo The imaging protocol required for the ACR phantom scan is different from the protocol used for subject scans. Refer to Appendix II, Scanning Protocol for ACR Phantom Scans for the exact scanning parameters of each sequence. * Please note that once a facility qualifies for participation in the MRI study, it will need to repeat the phantom test scans every three months (as described in Section 4.1.2, Longitudinal Phantom Scans ) until the last patient has received an 18-Month / Early Termination scan. The longitudinal phantom scans only require the first 4 sequences above, eliminating the last 3 which are only used to evaluate the safety sequence parameters before the first patient is scanned. The pre-study ACR phantom test scan sent to Synarc will be evaluated by a radiologist or speciallytrained technologist who will assess the phantom images. The results of the pre-study ACR phantom test scan will be sent to the site within seven (7) business days after receipt of the phantom test scan. Synarc may request up to two (2) repeats of the pre-study ACR phantom test scan. Repeat requests will be made by contacting the site s Study Coordinator via telephone and/or email, and followed up by a 7 of 54 Version 1.0 10 November 2011

Phantom QC Report sent to both the study coordinator and MRI facility. The Phantom QC Report will state a request for the ACR phantom repeat, the reason for the repeat request and detailed information regarding the image quality and suggestions for improvement. Repeat exams requested by Synarc should be performed as quickly as possible. A site is permitted only two repeats for the pre-study ACR phantom scan. Failure of a second repeated ACR Phantom scan may result in exclusion from the volumetric MRI study. Only upon receipt of a passing Phantom QC report may a site continue with the next qualification procedure, the First Patient Scan. 3.3.2 First Patient Scan After an acceptable phantom test scan has been received, Synarc will request one patient scan. This scan will undergo a specific set of analyses to determine measurement reliability and reproducibility, thus ensuring the most accurate data possible across multi-site measurements. These scans differ from the ACR Phantom scans in their ability to account for the interactions between patient and scanner (i.e., local intensity variations due to dielectric tissue effects or local geometrical image distortions due to tissue susceptibility), which can induce substantial noise. The First Patient Scan will be done on a study patient and uses all the specified imaging sequences for scanning study patients. Refer to Appendix I, Technical Parameters for MRI of the Brain, for the exact scanning parameters for each sequence. 1. Localizer: 3-plane gradient echo (GRE) 2. 2D Sagittal Spin Echo (SE) 3. 3D T1/ MP-RAGE (Siemens)/ (GE) IR-prepped fast SPGR (GE) / TFE (Philips) 4. Repeat 3D T1 sequence with identical parameters as in sequence #3. 5. 2D Axial Dual echo PD/T2 TSE (Siemens and Philips)/FSE (GE) 6. 2D Axial Fluid-Attenuated Inversion Recovery sequence (FLAIR) 7. 2D Axial T2* Gradient Echo 8. 2D Axial Diffusion Weighted Imaging (DWI) The results of the First Patient Scan will be faxed to the site within seven (7) business days after Synarc receives the data. This report, if appropriate, will note any technical deviations from protocol and offer potential suggestions for the improvement of image data. Synarc may request up to two repeats of the First Patient Scan. Repeat requests will be made by contacting the site s study coordinator by telephone and/or email, and followed up by a Patient QC Report sent to both the Study Coordinator and the MRI facility. Detailed information regarding the image quality, as well as suggestions for improvement, will be noted on the Patient QC Report. In the event that a patient is not recommended for a repeat, a scan from a different subject may be submitted to Synarc for purposes of qualification. IMPORTANT A site may not proceed in scanning any additional patient until an acceptable Patient QC Report is received from Synarc. Once the First Patient Scan requirement is satisfied, the site can start scheduling additional patients. 8 of 54 Version 1.0 10 November 2011

4.0 Procedures for ACR Phantom Scans The phantom MRI scan will be acquired using a uniform scanning procedure in order to account for and minimize differences in MRI systems across study sites. The images should be acquired in strict accordance with Synarc s Technical Parameters, as described in Appendix II, Technical Parameters for ACR Phantom Scans. This phantom has been designed to be used for accreditation of sites in the US and is used to measure a standard set of parameters that relate to the performance of the scanner and further compare with other similar scanners. This is critical to evaluating data from multi-center trials. Each site is REQUIRED to use an ACR phantom to perform all phantom test scans. If the MRI facility does not have an ACR phantom, Synarc will provide one to use for the duration of the study. IMPORTANT THE SAME ACR PHANTOM MUST BE USED WHEN PERFORMING THE PHANTOM SCANS THROUGHOUT THE COURSE OF THIS STUDY. 4.1 Phantom Scheduling 4.1.1 Pre-study Qualification Phantom Scan The pre-study Qualification Phantom scan will be preformed and submitted upon Synarc s acknowledgement of an acceptable Pre-Trial Questionnaire and successful completion of the primary MRI Technologist s telephone training. Synarc will then evaluate the qualification scan and provide a QC report to the Site/MRI facility. Upon receiving a passing Pre-study Phantom QC report, the site will need to acquire the same scan at three month intervals for the duration of the study. 4.1.2 Longitudinal Phantom Scans Longitudinal phantom scans will be performed every three (3) months for the duration of the MRI portion of the study as an assessment of the reproducibility of protocol measurements over time. These repeated calibration checks will provide evidence of systematic measurement errors due to system drift degradation or change. The planned acquisition dates of Longitudinal Phantom scans can be calculated at three month intervals from the scan date of the first accepted Pre-Study Qualification Phantom Scan. A failed Phantom QC will most likely result in a repeat request to verify if there is a problem with the scanner. If the problem turns out to be legitimate, patient data submitted during the period from the last successful phantom QC may not be useable for volumetric analysis. 4.2 Evaluation of Phantom Scans If the phantom data can be loaded and viewed, a QC report listing results from the image quality review will be communicated to the MRI facility within seven (7) business days of receipt for both Pre-Study Qualification and Longitudinal phantom data. 4.2.1 Requests for Resubmission of Phantom Data If the submitted phantom data cannot be loaded and viewed on the image viewing stations at Synarc, a Resubmission Request will be made to the site s study coordinator and MRI facility identifying the 9 of 54 Version 1.0 10 November 2011

specific scan and the need for resubmission. The MRI data must be resubmitted to Synarc as quickly as possible. 4.2.2 Requests for Repeat Phantom Scans If the data is able to be loaded and viewed, and is still deemed unacceptable for use in this study, Synarc will make a Repeat Request. The Phantom QC Report will be sent to both the Study Coordinator and MRI facility. The Phantom QC Report will state a request for the phantom MRI repeat, the reason for the request, detailed information regarding the image quality, and suggestions for improvement. Software Upgrades If a software upgrade is planned for a scanner, Synarc may request that additional phantom scans be acquired before and after the upgrade to ensure consistency of scanner performance. 4.3 Labeling Phantom Images in the Digital Header On the device, subject and visit information should be entered as follows: FIELD: SUBJECT NAME: Enter Site Number - [PHANTOM] Example: 02 PHANTOM DATE OF BIRTH: WEIGHT: STUDY DESCRIPTION: Enter scan date Enter 100 lbs or 50 kg. Enter [ACR PHANTOM SCAN] 4.4 Phantom Positioning In order to ensure the reproducibility of the measurements that are taken throughout the study trial period, the ACR phantom must be placed inside the MRI magnet in the exact same position every time it is scanned. This should be accomplished by: Aligning the phantom as a head would be aligned. On the phantom, the words NOSE and CHIN should be positioned where the nose and chin would be for a head study. Using the crossed black lines on the phantoms anterior side as a landmark to line up the phantom in the center of the head coil. Leveling to ensure that the phantom is horizontal and that the plastic bar on the CHIN side is horizontal. Use a plastic bubble level, if available. Ensuring that the crosshairs engraved on the phantom are placed at the magnet s isocenter. If using a phased array coil which is not open at the upper end, push the phantom in as far as possible while still using the patient headrest holder. The inferior end of the phantom should be at least flush with the end of the coil if not inside the coil. If necessary, remove the head cradle and pad from the coil. 10 of 54 Version 1.0 10 November 2011

Note: for phased array coils with non-removable head holder sections, it may not be possible to use this coil for the phantom scans. In this case, use your quadrature type coil for the phantom but you can use the phased array coil for the patients. 4.5 Phantom Scan Acquisition Technique Ensure that the phantom is at room temperature before scanning (T1 and T2 are temperature dependent). If this is not done, ghosting artifacts may be present in the images. Position slices in the center of the phantom. Note: the last 4 scans in the list below are ONLY DONE FOR PRE- QUALIFICATION and need not be included for longitudinal submissions. The phantom MRI scans will consist of the following imaging sequences: 1. Localizer 3-Plane Gradient 2. ACR single slice Sagittal Spin Echo 3. ACR 11 slice Axial T1-weighted Spin Echo 4. 3D T1 MP-RAGE (Siemens), IR-prepped fast SPGR (GE), 3D TSE (Philips). NOTE: All GE scanners use the coronal plane, while all Philips scanners, Siemens scanners use the sagittal plane. 5. 2D Axial TSE/FSE dual echo (Pre-qualification only) 6. 2D Axial FLAIR (Pre-qualification only) 7. 2D Axial T2* gradient echo (Pre-qualification only) 8. 2D Axial DWI (Pre-qualification only) Note: The imaging protocol required for the ACR phantom scan is different from the protocol used for subject scans. Refer to Appendix II, Technical Parameters for ACR Phantom Scans for the exact scanning parameters of each sequence. 4.5.1 ACR Sagittal Locator The ACR Sagittal Locator acquisition is a 20 mm thick single-slice spin-echo acquisition through the center of the phantom and should appear as pictured in the image below. ACR Sagittal Locator Notice that the 45 crossed wedges on the left of this image appear with equal intensity. If the image you acquire does not appear like this, the phantom must be re-centered and reimaged. 11 of 54 Version 1.0 10 November 2011

4.5.2 ACR 11-slice Axial T1-weighted Spin Echo The 11-slice ACR Axial T1 series must be prescribed on the ACR Sagittal Locator as pictured in the image below. This image illustrates the correct alignment of the slices of the ACR Axial T1 sequences on the ACR Sagittal Spin Echo. The center of the first slice must be aligned with the vertex of the crossed wedges and through the center of the dark chemical shift and resolution insert. The centers of slices 8 through 11 must align with the four low-contrast disc inserts. The center of slice 11 must be aligned with the vertex of the crossed wedges. If the slices of the ACR Axial T1 Sequence are not prescribed as pictured above, the slices must be re-prescribed and the acquisition must be repeated. 12 of 54 Version 1.0 10 November 2011

4.5.3 Slices 1 to 11 of the ACR Axial T1 sequence The collection of images above shows the correctly positioned 11 slices of the ACR Axial T1 sequence. Note that it is possible that the phantom may look different in slices 1-4 if the spacers are different. This makes no difference to the analysis. Examine all 11 images and verify that all are present, properly positioned and free of image artifacts. Quality Evaluation Slice 1: High Contrast Spatial Resolution, Slice Thickness, Geometric Accuracy Slice 5: Geometric Accuracy Slice 7: Image Intensity Uniformity, Percent Signal Ghosting Slice 8-11: Low Contrast Object Detectability 4.5.4 3D T1-weighted sequence Acquire the 3D sequence (3D T1 MP-RAGE (Siemens), IR prepped fast SPGR (GE) or 3D TSE (Philips) Sequence) listed in Appendix II or the QRG for your scanner. All GE scanners use a coronal acquisition. All Philips and Siemens scanners use the sagittal plane. Be sure to use the exact sequence listed in Appendix II or the QRG without change. 13 of 54 Version 1.0 10 November 2011

Use the 3 Plane localizer, sagittal or axial images to plan the 3D sequence. Be sure to cover the entire volume of the phantom. 4.5.5 Criteria for Assessing Quality of Phantom MRI Scans Specially trained SYNARC technologists will evaluate if the MRI scanner is performing according to the Synarc s specifications. The following four (4) standard ACR measurements will be performed on the data from each ACR phantom scan: 2D- and 3D-Geometric Accuracy: Measures the degree of geometrical distortions in an image, generated by the MRI system. High Contrast Spatial Resolution: Assesses the scanner s ability to resolve small structures. Image Intensity Uniformity: Assesses the scanners ability to yield constant uniform image intensity in uniform regions of the brain. Low Contrast Object Detection: Assesses the extent to which objects of low contrast are discernable in the images. Each measurement must fall within a range defined by the American College of Radiology in order for the phantom test run to be of acceptable quality. Results of the phantom scan QC will be faxed and/or emailed to the study coordinator and MRI technologist, within seven (7) business days after receipt of the phantom image data at Synarc. 4.5.6 Clinical sequence parameter check during pre-qualification Phantom MRI Scan In addition to the ACR and 3D T1 sequences to be used for performance checks, the clinical sequences for dual echo TSE/FSE, FLAIR, and T2* must also be run on the phantom to check for correct parameters. The accepted parameters for the clinical scans plus the 3D T1 can then be stored in the protocol section for use with clinical subjects. 4.6 Common Problems The purpose of this section is to enable sites who have received a failed Phantom QC Report from Synarc to understand the significance of the failure and understand the steps necessary to correct problems detected. Errors can occur in any of the measurements listed in 4.5.5, Criteria for Assessing Quality of Phantom MRI Scans, as well as in the positioning of the phantom and of the sequences. 4.6.1 Incorrectly (Mis-) prescribed Slices Each of the measurements performed on the ACR phantom images requires a specific imaging slice to make the measurement on. Therefore, mis-prescribing the slices on the phantom may prevent proper measurement of the phantom since a particular measurement may not be possible if a sequence slice was not obtained correctly. Possible Remedies: Please refer to Section 4.5.1 ACR Sagittal Locator and Section 4.5.2, ACR Axial T1-Weighted Sequence, which details the correct method for aligning the slices in the phantom. 14 of 54 Version 1.0 10 November 2011

4.6.2 Phantom not centered in the Field of View This image demonstrates an example of a phantom that is incorrectly positioned in the magnet. It is off center, and is too low in the field of view. If the ACR phantom is not correctly positioned within the magnet, the correct slices needed to make measurements can easily be mis-prescribed. Possible Remedies: Please refer to Section 4.4, Phantom Positioning, which details the proper method for positioning the ACR phantom within the scanner. 4.6.3 Geometric Accuracy The Geometric Accuracy measurement is performed on images from the ACR Sagittal Spin Echo and the ACR Axial T1-weighted sequence. These measurements detect any distortions in the MR images obtained which result in the image dimensions being either smaller or larger than the object truly is. This distortion compromises the sensitive measurements that are being assessed for this MRI study. 190 mm +/- 2 Possible Remedies: A geometric accuracy failure is most commonly due to mis-calibrated gradients. Gradient calibration can drift over time. Therefore, it is important that the scanner s service engineer recalibrate the gradients after a failed phantom scan result. This recalibration must match exactly 15 of 54 Version 1.0 10 November 2011

the pre-screening scan measurements to ensure accurate co-registration of the patient images. 4.6.4 High Contrast Spatial Resolution The High Contrast Spatial Resolution measurement assesses the magnet s ability to resolve small objects. This measurement is highly important for this MRI study as the volumetric measurements made on patient images will be done on small structures of the brain such as the hippocampus and entorhinal cortex. The images below show an example of phantom images that (a) Passed the High Contrast Spatial Resolution test and one that (b) Failed the test. (a) Passed (b) Failed Possible Remedies: Make sure the phantom is stable in the head coil and not free to move or vibrate. Image ghosting due to maladjustments and instabilities can cause the resolution of small images to be compromised. As mentioned in Section 8.3, Ghosting Artifacts it may be necessary to request the service engineer to determine and correct the cause of the ghosting. Check that any user selected image filters are turned off. Excessive image filters may make the images appear less noisy, but will smooth the images and compromise the resolution. Therefore, turn off the filters to correct errors in high contrast spatial resolution. Check that the phantom is not tilted; this can cause blurring of these objects. Check that the correct number of phase encoding steps is used (see Appendix II or QRG). 4.6.5 Image Intensity Uniformity The Image Intensity Uniformity measurement assesses image uniformity in a large uniform circular region of the phantom. Failure of this test may indicate a defective head coil or problem in the radiofrequency subsystems. Possible Remedies: Make sure the phantom is centered in the head coil. If the phantom is closer to one side of the head coil than the other, uneven image intensities can result. Make sure the phantom is stable in the head coil and not free to move or vibrate. 16 of 54 Version 1.0 10 November 2011

Image ghosting due to maladjustments and instabilities can cause image intensity variations. If ghosting is suspected, call your service engineer (as defined in Section 8.3, Ghosting Artifact and Section 4.6.4, High Contrast Spatial Resolution ). If the images also appear grainy (i.e. a low signal-to-noise ratio), then the problem could be with the components of the head coil. Call the service engineer to diagnose and correct the problem. Bright intensity areas at the edge of images may occur due to the use of a phased array coil without using proper intensity corrections. Note also, if the phased array coil is used, a calibration scan may be required for the filter to work properly. See appropriate Protocol section to see if phase array coil can be used for your scanner. A swirling artifact may indicate that the temperature in the phantom was not equilibrated. Be sure the phantom is in the scanner room for at least an hour before running the scans. 4.6.6 Low Contrast Object Detection The Low Contrast Object Detection measurement assesses the magnets ability to discern low contrast objects. The ability to detect low contrast objects is reflective of the signal-to-noise performance of the scanner. A failure in this measurement signifies that the scanner produces images with fewer low contrast objects. The images below show images of a (a) good phantom image showing low contrast object detection test and a phantom image that (b) failed the low contrast object detection test. (a) Good Low Contrast Object Detection (b) Poor Low Contrast Object Detection Possible Remedies: Make sure that the image slices are correctly positioned. Refer to Section 4.5.1, ACR Sagittal Locator and Section 4.5.2, ACR Axial T1-weighted Sequence to correctly position the slices. Make sure that the phantom is not tilted in the scanner. A tilted phantom can result in parts of the slices being out of their proper location. If the phantom doesn t look square with the edges of the field of view on the localizer, reposition the phantom in the head coil. Make sure the image is stable in the head coil and cannot move or vibrate. If ghosting artifact due to maladjustments and instabilities is apparent on the images, call your service engineer (as defined in Section 8.3. Ghosting Artifacts and Section 4.6.4, High Contrast Spatial Resolution ). 17 of 54 Version 1.0 10 November 2011

5.0 Procedures for Subject MRI Scans 5.1 Patient Scheduling 5.1.1 First Patient Scan As stated in Section 3.3.2, First Patient Scan, Synarc will only allow patients to be scanned after a passing Pre-study Qualification Phantom data set has been received and reported. Once an acceptable phantom report has been received (and the clinical site has obtained complete IRB (or Ethics committee) approval to scan these clinical research patients), the site may schedule and submit a First Patient Scan. Please allow at least ten (10) business days in between the submission of a First Patient Scan to Synarc and scanning additional patients. Scanning of additional patients for the MRI study may need to be rescheduled or repeated based on the amount of time required to obtain a passing first patient scan. After the first patient scan is accepted, additional patients may be scheduled and scanned as needed. IMPORTANT The study coordinator must receive written confirmation from Synarc stating acceptance of the first patient s baseline MRI prior to scanning another patient for the volumetric MRI study. 5.1.2 Patient scans Upon the acceptance of the First Patient Scan, patients will be scanned at the following visit intervals: Baseline 9-Month 18-Month Unscheduled Early Termination IMPORTANT All Unscheduled visits should be pre-approved by Baxter. For all unscheduled MRI visits, the site should contact their Field CRA s in order to obtain approval from Sponsor. If a patient discontinues from the study, an Early Termination scan should be scheduled. 5.2 Evaluation of Subject Scans If the subject exam data can be loaded and viewed on image viewing systems at Synarc the turnaround times are as follows: The first subject from each site that has completed the telephone training will be used to perform an in vivo assessment. The turnaround time for reporting on the quality of these sequences will be seven (7) business days after receipt of complete package at Synarc. Quality report of all subsequent subject scans will be sent to the sites within five (5) business days after receipt of complete package at Synarc. The Subject QC Report will be sent to both the Study Coordinator and the MRI facility. 18 of 54 Version 1.0 10 November 2011

5.2.1 Requests for Resubmission of Subject Data If subject data cannot be loaded and viewed on our image viewing stations, a Resubmission Request will be made via fax or email to the site s study coordinator and MRI facility identifying need for resubmission. The data must be resubmitted to Synarc as quickly as possible. 5.2.2 Requests for Repeat Subject Scans If subject data are unacceptable for this study due to poor quality for example, Synarc will require a Repeat of the exam. The request for the subject MRI repeat exam will be noted in the Subject QC Report, the reason for the request detailing information regarding the image quality issues and suggestions for improvement. The Study Coordinator should reply to Synarc within one (1) business day of receiving a Repeat Request to acknowledge receipt and implications. It is important to note that if Synarc requests a repeat, the repeat exams should be done as quickly as possible in accordance with the study protocol visit window. Good communication between the study coordinator, MRI facility and Synarc is critical throughout the study to ensure subjects do not fall out of the protocol visit window. 6.0 Subject MRI Exam Preparation Preparation for the MRI exam prior to the subject s arrival is critical in order to ensure that all scans can be acquired within the allotted time frame and for accommodation of any unforeseen delays. The total scan time, not including subject positioning, is approximately 45 minutes. Make certain that positioning aids are present in the procedure room, have the transmittal form on-hand for completion, and have supplies ready to label the digital media immediately after the MRI exam is complete. 6.1 Subject Safety and Monitoring Remember to follow all standard subject consent protocols approved by the Institutional Review Board (IRB) or Ethics Committee. Ensure that the subject does not have any of the MRI contraindications and comply with the local requirements outlined at your imaging facility. The imaging site is responsible for the MRI safety of subjects who are scanned at the site, all procedures and guidelines for safety consideration should be followed. Be sure to explain the examination procedure to the subject and caregiver, if applicable. The caregiver should be present for subject consent, MRI exam preparation, and scan acquisition. In some cases, it may be helpful to ask the caregiver to hold the subject s hand and offer reassurance during the MRI exam. If the caregiver will be going into the MRI scanner room during the examination, the caregiver must also be checked to ensure that they do not have any MRI contraindications. If sedation is administered to the patient, it is important to note the type of sedation administered on the Transmittal Form for MRI of the Brain (as defined in Appendix IV, Transmittal Form for MRI of the Brain ). It is also advisable to monitor the patient s pulse, respiration and O 2 levels through use of a standard on-site monitoring device during the scan, even if a caregiver is present. Note: MRI scan sedation requires medical monitor consultation. 6.2 Subject Positioning Proper subject positioning is critical for obtaining high quality images. Correct, consistent and comfortable positioning of the subject within the MRI scanner will limit artifacts and maximize the acquisition of good quality images. 19 of 54 Version 1.0 10 November 2011

As part of the normal subject pre-screening routine, make sure all removable dental bridgework or other metallic objects (removable dental plates, belts, zippers, etc.) are removed prior to entering the scanner room. The metal (even though not ferromagnetic) may cause artifacts that can affect volumetric analysis. Consistency of the environment (e.g., no metal objects) around the subject is also very important for consistently achieving high scan quality. The subject s head must be placed in a volumetric radio frequency (RF) head coil (no surface coils are allowed). Maximize the subject s comfort in the RF head coil. Once the subject has been placed into the RF coil and comfort is maximized, ensure that the center of the RF coil to be used for landmarking, is approximately 1 finger width above the eyebrows. The images (from first set of scout images which identify how landmarkings were performed) below show a correct landmarking (left) and an incorrect landmarking (right). However, if the subject was landmarked as shown on the right for the baseline, landmarking for all follow up visits must be done identical to the baseline to insure that the 3D T1 images can be properly coregistered. Correct Landmarking Incorrect Landmarking 20 of 54 Version 1.0 10 November 2011

Ensure consistency of position of the volume of interest and field of view of the scan with respect to the magnet s isocenter. 15 mm from head to top of FOV 45 mm from head to top of FOV The issue of consistent positioning of the 3D T1 scans is most important for GE scanners but everyone should be careful to be consistent during the graphic positioning phase to insure maximum reproducibility. It matters little whether the subjects are position like the image on the left or like the right but the site must be consistent throughout the study for each patient. Having a general plan for how this is done and doing it for all subjects, makes QC simpler when sites use a variety of technologists during the study. Consistency between the MRI exam taken at the Baseline and at all follow-up visits is extremely important. If there are any deviations from these instructions (i.e., positioning, or parameters) to accommodate a subject during the Baseline visit, the MRI technologist must note these on the Transmittal Form (as defined in Appendix III). General guidelines for positioning of the head in the FOV is + 5 mm difference from baseline to all follow ups. For follow up scans, always check the position of the baseline not the last scan. Maximizing comfort through proper head support will not only help to restrict head movement, but will also provide greater compliance in completing the entire exam within the allotted exam time. Proper head support can be achieved through the use of a vacuum-molded head holder, foam wedges or padding at the sides of the head, or a neck brace. Placing a Velcro strap or tape over the forehead can also provide stability and feedback to the subject and decrease movement. It is imperative that the subject s head remains stable during acquisition. Imaging data degradation due to motion artifacts will almost always result in data rejection. 21 of 54 Version 1.0 10 November 2011

7.0 Subject MRI Acquisition Technique Image quality criteria for MRI of the brain in a clinical trial are stricter than in standard clinical practice. The measurements that will be performed on this MRI data depend highly upon the quality of images. Small changes in volume are expected between the screening and follow up visits. Thus, it is imperative that the acquisition of the MRI data be extremely precise and consistent in order to detect these changes. In order to achieve the most reliable evaluations of MRI of the brain, strict adherence to a uniform acquisition protocol and quality standards is required. 7.1 Labeling Digital Header Fields for Subject MRI Scans In compliance with privacy laws to ensure and protect the confidentiality of the subject, no subject names or identifiers should be entered into the electronic MRI header. The information in the box below should be entered into the electronic MRI header in lieu of subject identifiers. FIELD: SUBJECT NAME: DATE OF BIRTH: STUDY DESCRIPTION: Enter Site Number and Subject Number [2 digits Site ID Number] [4 digits Subject ID Number] Example: 02-1004 Enter subject s date of birth in alphanumeric format (01-Jan-yyyy) Example:01-Jan-1939 Select the Visit name (i.e., Baseline, 9-Month, 18-Month, Early Termination, Unscheduled) Indicate if this is a repeat exam requested by Synarc. At the end of the exam, export MRI images to digital media (CD) or upload to Synarc FTP website in uncompressed DICOM format. MRI images should be submitted to Synarc within 24 hours of acquisition. In addition to archiving the data to digital media, the imaging facility will need to locally archive this data. 7.2 Pre-scan Adjustments Most modern MRI scanners provide automated adjustment procedures for RF coil tuning and frequency adjustments after the subject is positioned in the magnet. Follow the adjustment procedures provided by the manufacturer. Image quality is usually unacceptable without proper adjustment of the RF coil and the transmit/receive equipment. Furthermore, without frequency adjustment, problems can occur with signal acquisition and proper localization of image FOV and slices. 7.3 Protocol for MRI of the Brain The following is the list of the MRI sequences recommended for this protocol along with the approximate scan time for each sequence. This protocol is to be complied for all Baseline and follow up visits. All scheduled scans for each subject should be performed on the same scanner used for Baseline visit. 22 of 54 Version 1.0 10 November 2011

This list is organized in the chronological order of acquisition: 1. Localizer: 3-Plane Gradient Echo (GRE) (20 seconds) 2. 2D Sagittal Spin Echo (SE) (2 minutes) 3. 3D T1/(MP-RAGE (Siemens)/IR-prepped fast SPGR (GE)/or TFE (Philips) (7-8 minutes) 4. Repeat 3D T1 sequence with exact parameters as in #3. Use identical parameters. 5. 2D Axial PD/T2 TSE (Siemens and Philips)/FSE (GE) (2-4 minutes) 6. 2D Axial Fluid-Attenuated Inversion Recovery (FLAIR) (2.5 4.5 minutes) 7. 2D Axial T2* Gradient Echo (3-4 minutes) 8. 2D Axial Diffusion Weighted Imaging (DWI) (<1 minutes) The purpose of these sequences is as follows: 1. Localizer Used to position the imaging planes. 2. Sagittal Spin Echo -- Used to position the volume of the 3D-T1 sequence (either coronal or sagittal) and all axial sequences in conjunction with the localizer. 3. 3D T1 Used to perform volumetric measurements of specific structures of the brain. 4. 2 nd 3D T1 is used to better define whole brain volume for each patient. Note that only one 3D sequence needs to pass QC for the study to pass but 2 passing QC scans are preferable. 5. Axial 2D TSE/FSE Used to detect white matter lesions, lacunes, and infarcts. 6. Axial FLAIR Used to enhance visualization of white matter lesions, tumors, lacunes, and infarcts. Vasogenic edema is best seen on this sequence. 7. T2* Gradient Echo Used to detect microhemorrhages. 8. Diffusion Weighted Imaging Used to detect signs of stroke. IMPORTANT The order of the sequences is critical! Volumetric analysis will be performed on the 3D T1 sequence. This sequence is most sensitive to motion artifact, therefore it is crucial to obtain this sequence at the beginning of the scan while the subject is most comfortable. 7.3.1 Localizer: 3-Plane Gradient Sequence (Scouts) and Sagittal Spin Echo a. Localizer: 3-Plane Gradient Sequence: This shows a quick acquisition in 3 orthogonal planes for anatomical orientation. Acquire one slice in the middle of each plane (sagittal, coronal, and axial) plus additional slices as needed. Ensure that the subject is positioned properly. [fig. 1] FIGURE 1. 3-Plane Gradient Sequence a) Mid-plane axial slice b) Mid-plane sagittal slice c) Mid-plane coronal slice 23 of 54 Version 1.0 10 November 2011

b. Sagittal Spin Echo: This scan should consist of approximately 22 slices to cover the midsection of the brain as well as the Left and Right medial temporal lobes. Use the axial slice from the triplanar scout to position stack in the middle of head. See Figure 2. FIGURE 2. Placement of slices for Sagittal Spin Echo 7.3.2 3D T1 / MP-RAGE / IR-prepped fast SPRR / TFE a. Orientation: For all GE scanners, use the orthogonal Coronal plane. FIGURE 3. 3D T1 Coronal orientation and positioning b. Positioning: The positioning is best done on the mid-sagittal slice (see above). The whole head (including the top part of the cranium) should be included in the volume box. The entire brain plus the skull must be included in the slice direction. Check that if the nose is outside the volume that it will not fold into the region of the brain. DO NOT use foldover suppression techniques in the slice direction. 24 of 54 Version 1.0 10 November 2011

c. Orientation: For Philips Intera and Achieva scanners, Siemens Symphony, Espree and Avanto: use the orthogonal Sagittal plane. FIGURE 4. 3D T1 sagittal orientation and positioning d. Positioning: Use an axial slice that shows the inferior part of the brain and nose as well as midsagittal slice to fully define the volume of the 3D acquisition. Leave approximately 10-15 mm from top of the head to the top of the FOV. For an external landmark, use the eyes and for an internal landmark, use the thalamus to position the middle of the volume. For follow up scans, refer to the screening for positioning. Consistent positioning is critical to the volumetric analysis of this longitudinal study. Scans that do not contain the whole brain cannot be processed. For coronal images, if the acquisition box does not extend past the ears, increase Phase FOV to 100%. If the acquisition box does not cover the head from front to back, add slices if possible. If not, make sure the nose will not wrap (alias) into the brain. For sagittal images, if the acquisition box does not completely encompass the head from front to back, make sure the posterior brain is included. If the nose extents outside the FOV, it will wrap into the back of the image. Make sure that this does not impinge on the brain. If the subject s head is larger than the field of view (FOV), a small amount of oversampling (20%) can be added to eliminate any aliasing. If partial oversampling is not available, the FOV may be increased to a maximum of 250 mm (this change should be noted on the Transmittal Form and must be utilized for all follow up scans for that subject). ONLY use this option if Phase FOV is already at 100% and as a last resort. IMPORTANT Any deviations from the 3D T1 protocol should be: 1. Documented locally; 2. Reported to Synarc; 3. Implemented identically for every subsequent scan for the subject. 25 of 54 Version 1.0 10 November 2011

e. Read and Phase Directions: The read direction (in blue) must be along the superior-inferior plane to avoid aliasing of neck and shoulder into the FOV. The phase direction for sagittal (in solid red) should be anterior-posterior; the phase encoding direction for the coronal in dashed red) should be left-right. Superior (Read Direction) Left (Phase encode Coronal) Posterior Anterior (Phase encode Sagittal) Right Inferior f. 3D Partition: For GE Excite and MR450 scanners, use approximately 184 slices with 1.2 mm slice thickness. For a larger head, you may add slices to get complete coverage; remember that 4 slices are not reconstructed and also that there is some aliasing of the volume so be sure to cover the head completely. Remember, however, adding slices adds scan time. For all Siemens scanners use 176 slices with a slice thickness of 1.2 mm. For Philips scanners use 170 slices with a slice thickness 1.2 mm to fully cover the head from left to right. Do not worry that the area covered by the slices extends outside the head. Using a Phase FOV less than 100% will most likely not be an option because of aliasing in the AP direction. If the head is very large, you may add 20% oversampling to avoid serous aliasing of the face/nose into the brain. Note this modification on the Transmittal Form. Do NOT change the number of slices as this will affect contrast. IMPORTANT If any motion artifact is detected, the 3D T1 should be repeated before proceeding with additional sequences. It is highly recommended to reformat the coronal images into the axial and sagittal planes or sagittal into axial and coronal planes to confirm there is no: Aliasing (see section 8.5) Image shading (see section 8.7.2) 7.3.3 Acquire a Second 3D T1 Sequence In order to improve the accuracy of the volumetric analysis and to reduce the numbers of repeat requests for motion, each site is to run a second 3D T1 sequence, identical to the one described above (see also Appendix I or the QRG). To ensure that the patient is properly positioned for this second scan, use the vendor specific method to copy the graphics of the previous 3D scan. 26 of 54 Version 1.0 10 November 2011

It is best to have two acceptable 3D sequences, and you may repeat the second scan if there are noticeable motion artifacts, if time is available and the patient is cooperative but only one 3D T1 scan passing QC is needed to have a passing exam for each visit. 7.3.4 Axial 2D PD/T2 Turbo Spin Echo (TSE) / Fast Spin Echo (FSE) a. Orientation: Angulate Axial Dual Echo TSE/FSE parallel to anterior commissure and posterior commissure (AC-PC) line. [fig. 5] FIGURE 5. Orientation of Axial Turbo Spin Echo/Fast Spin Echo FIGURE 6. Axial 2D TSE/FSE Stack Placement b. Positioning: Position on mid-sagittal slice. Be sure to obtain coverage of the entire brain by including one slice of air above the skull. The acquisition stack should be placed at the most superior point of the brain, and fully cover the cerebellum as well as all the brain in the lateral and the anterior-posterior planes. If extra axial slices are required to achieve this coverage, please acquire those slices. [fig. 6] NOTE: Acquisition stack for PD/T2 sequence must fully cover the subject s brain. The flow saturation band should be placed approximately 15 mm below acquisition stack. c. Slices: 48 to 50, 3-mm, slices covering the entire brain, with no gap between slices, interleaved acquisition. If necessary, add 2-4 additional slices in order to obtain full brain coverage. This may require an increase of the repetition time (TR), which will result in an increase of total acquisition time. Please be sure to avoid altering additional parameters. You may optimize the TR time (maintaining 2 groups of concatenations) to minimize scan time. d. Flow Saturation Band: 50-mm flow saturation band inferior to the MRI slices and positioned about 15-mm from the bottom of the MRI stack to reduce inflow effects. e. Fat Saturation to be turned on (Vendor specific terms include Fat Sat and SPIR) 27 of 54 Version 1.0 10 November 2011

f. Parallel imaging is allowed for scanners with phased array coils: for GE this means ASSET, for Philips this means SENSE and for Siemens, ipat may be turned on. Factors from 1.5 to 2 are allowed. 7.3.5 Axial 2D FLAIR / T2* Gradient Echo / Diffusion Weighted Imaging (DWI) Please refer to the protocols found in Appendix I Technical Parameters for MRI of the Brain for direction on how to acquire these sequences. Synarc should be notified if your facility will not be able to follow the parameters listed but other than the 3D T1 sequence, many of the values are only approximate and may vary for your specific scanner. Please note that the FLAIR must have dark CSF to allow white matter lesions to be highlighted. Note that the FLAIR, T2* and DWI orientation can be copied from the Dual echo sequence. For scanners with phased array coils, it is allowable to use factor of 1.5 to 2 parallel imaging for the the FLAIR, T2* and the DWI sequence. 8.0 COMMON PROBLEMS SEEN WITH SUBJECT MRI SCANS Due to the nature of the disease, it may be challenging for the subject to properly complete the entire MRI exam. In order to get a good measure of brain volume, the imaging protocol is designed to give as clear a boundary between brain and CSF as possible. Anything that alters this (movement, inhomogeneity, chemical shift) potentially compromises such measures. It is highly recommended that the raw and reformatted images are viewed after each acquisition to determine whether quality is sufficient for submission. If problems are detected, the scan should be repeated while the subject is at the MRI facility. The following sections illustrate examples of common problems that can be encountered while acquiring MRI images and possible solution. 8.1 Incorrect Orientation / Image Plane Misalignment In this example, the 3D coronal T1 sequence was acquired with angulation. Note that the 3D T1 sequences are to use orthogonal orientation (i.e. no angulation). Possible Remedies: Check that orthogonal plane was used. If subject moved significantly, reposition subject and reacquire ALL scans. 28 of 54 Version 1.0 10 November 2011

8.2 Swap of Read and Phase Encoding Directions In this example, the read and phase encodings were swapped. The read direction was in the left-right plane instead of the superiorinferior plane, which resulted in the aliasing (folding) of the neck into image field of view. Note that GE scanners list read direction while Siemens, Toshiba and Philips scanners list phase encoding direction. Possible Remedy: Check the read direction (or phase encoding direction) is correctly specified. 8.3 Ghosting Artifacts In these examples the images exhibit ghosting. Ghosting can be created by motion (subject s movement, table bed vibrations, CSF or vessel flow, etc) or they can be related to maladjustments and instabilities in acquisition. Possible Remedies: Ensure that the subject is comfortable. If motion artifacts are the result of subject head motion, reacquire the sequence after tightly securing the subject s head with additional restraints (as defined in Section 6.2, Subject Positioning ) and emphasize the importance of remaining still. If a break is needed and time allows, permit the subject to exit the scanner and then reacquire all images once the subject feels ready to re-enter the magnet. If motion artifacts are not due to mechanical problems (i.e. table bed vibrations) or subject movement, the problem is most likely related to maladjustments and instabilities in acquisition. Maladjustments and instabilities should be suspected if all measures to eliminate motion artifacts have been exhausted. A large number of service or adjustment problems can be encountered with the MRI scanner. Among them are: Receiver gain too high Errors in phase encoding gradients Unbalance of receivers RF transmitter instability Remember to rerun the Pre-scan routine if you encounter artifacts that you think might be machine related. If all else fails, contact your service engineer. 29 of 54 Version 1.0 10 November 2011

8.4 Flow Compensation In the T2-weighted image shown to the right, the flow compensation was not applied during acquisition resulting in artifacts from cerebrospinal fluid (CSF) motion running through the anatomy of the brain. The artifacts running through the eyes come from the eyes motion and cannot be avoided, but they do not have any influence on brain anatomy depiction. Possible Remedies: Use flow compensation 8.5 Aliasing (Folding) in 3D T1 These images show aliasing in the coronal images. Foldover generally occurs when the subject s head size is larger than the acquisition box. Possible Remedies: If the acquisition box does not fully cover the subject s head, increase the Phase FOV up to 100%. For Siemens and Philips scanner, you may use 20% oversampling to eliminate this foldover. If you cannot use 20% oversampling you may increase your FOV to 250 mm (notify Synarc) but this cannot be done on follow ups only on a screening scan. If the foldover is in the slice direction, adding more slices (if possible) or repositioning the head within the volume are the only options. 30 of 54 Version 1.0 10 November 2011

8.6 Poor Signal-to-Noise Ratio (SNR) and Contrast-to-Noise Ratio (CNR) Below are two examples showing good SNR (a) and poor SNR (b). (a) Good SNR (b) Poor SNR (c) Poor CNR Another example of poor contrast-to-noise (CNR) is shown in example (c) where the contrast between the gray and white matter is indistinguishable. In addition, the image is grainy and has an overall low signal. Possible Remedies for poor SNR and CNR: Check to ensure that the appropriate head coil was used (do NOT change coils from subject s baseline to follow up) Ensure that the head coil is properly connected. Check that the flip angle value corresponds to the protocol (see Appendix I, Technical Parameters for MRI of the Brain ). Check time parameters: repetition time (TR), inversion time (TI), and echo time (TE). Check the receiver gain. The gain may be too low (insufficient amplification). Rerun the Pre-scan routine in manual or automatic mode. 8.7 Signal Loss 8.7.1 Inferior Slices In this example, there was a signal drop at the lower part of the brain. This is usually due to improper positioning of the subject s head, an incorrectly tuned RF coil or an improper or bad connection. 31 of 54 Version 1.0 10 November 2011

Possible Remedies: Ensure that the subject is positioned correctly in the head coil. Rerun the localizer, 3-plane scout and verify that the prescription of the mid-plane sagittal slice is centered at the thalamus. Reposition subject s head if necessary and rescan. Check that the RF coil is properly connected. (This may require that you enter the magnet room and check a cable). 8.7.2 Inhomogeneity or Shading Artifact Here is an image from an axial reformat of the off-coronal 3D T1 acquisition: The images are from the same slice with two different window/level settings. It is apparent that the signal is stronger on the posterior portion of the head. Possible Remedies: Ensure volumetric radio frequency (RF) head coil is used. Surface coils are not permitted. If a phased array coil was used, the proper vendor specific homogeneity filter must be used (PURE for GE, CLEAR for Philips and Pre-scan Normalization for Siemens). DO NOT use phased array coils without these filters. Ensure that the subject is positioned correctly in the head coil. Rerun the localizer, 3-plane scout and verify that the prescription of the mid-plane sagittal slice is centered at the thalamus. Reposition subject s head if necessary and rescan. Rerun automatic or manual pre-scan. Investigate with your field engineer whether the coil or receiver gain needs service. 8.8 Metal Artifact Magnetic field distortions: In this example there is blacking out due to the presence of metal near the participant s head. 32 of 54 Version 1.0 10 November 2011

Possible Remedy: Make sure the participant is not wearing any metal. Check for hair clips, metallic makeup (i.e. permanent eyeliner), necklace, safety pins, removable dentures and bridges, and facial jewelry. Remove metal and rescan. Certain mascara and tattoos will also cause artifacts. If the metal cannot be removed and the artifacts hinder anatomical depiction of the brain, the subject may have to be excluded. NOTE: After repositioning subject, rerun the Pre-scan routine before re-scanning. 8.9 Inadequate Head Coverage (a) Image is cut too close superiorly (b) Image is inadequate inferiorly (c) Adequate head coverage (d) Adequate head coverage In the figure above, images (a) and (b) illustrate examples of inadequate head coverage. Images (c) and (d) exemplify adequate head coverage. Please make certain that the field of view is large enough to include the whole head in the image. Ideally there should be air visible posteriorly, laterally, and superiorly beyond the full extent of the scalp. In the anterior portion of the brain, please ensure that the full extent of the brain and CSF are visible. Centering the subject in both the center of the RF coil and FOV is critical. 33 of 54 Version 1.0 10 November 2011