I. Purpose

The purpose of this guideline is to assist nuclearmedicine practitioners in recommending, perform-ing, interpreting, and reporting the results of brainperfusion imaging to assist in confirming the diag-nosis of brain death.

II. Background Information and Definitions

The diagnosis of brain death is a clinical diagnosisthat is sometimes made with the help of cerebral per-fusion scintigraphy. It is important that all physi-cians be knowledgeable in the clinical requirementsfor the diagnosis of brain death, especially the needto establish irreversible cessation of all function ofthe cerebrum and brain stem. Institutions perform-ing scintigraphy for the evaluation of possible braindeath should develop clinical guidelines and proce-dures for the clinical diagnosis that incorporate bothclinical evaluations and the integration of ancillarytests such as perfusion scintigraphy.

III. Common Indications

A. Assess brain blood flow in patients suspected ofbrain death.

IV. Procedure

A. Patient Preparation1. No special preparation is necessary. 2. The patient should have a stable blood pres-

sure, and all correctable major systemic bio-chemical abnormalities should be addressed.

3. In some institutions, a tourniquet is placedaround the scalp, encircling the head justabove the eyebrows, ears, and around theposterior prominence of the skull. The tourni-quet can help diminish scalp blood flow, pre-venting it from being confused with brain

blood flow. However, a tourniquet should notbe used in patients with a history of headtrauma when there is a concern that thetourniquet will exacerbate the injury. Atourniquet may also raise intracranial pres-sure and, therefore, should not be used unlessthere is adequate monitoring of intracranialpressure or there is little reason to expect anelevation of intracranial pressure.

4. Patients should be normally ventilated to pre-vent changes in cerebral blood flow that maybe caused by hyperventilation.

B. Information Pertinent to Performing the Proce-dure1. History of head trauma or central nervous

system (CNS) injury should be obtained.Trauma or focal CNS ischemia or infectionmay cause abnormalities in blood flow thatmay complicate image interpretation.

2. It should be determined whether the patientcan be positioned as needed for brain perfu-sion imaging. Anterior or posterior imagesshould be properly aligned so that symmetryof blood flow to both sides of the head and su-perior sagittal sinus activity can be assessed.

3. Care should be taken to note whether the pa-tient has recently received barbiturates. Athigh levels, these agents may decrease cere-bral blood flow.

C. PrecautionsNone

D. RadiopharmaceuticalSeveral 9 9 mTc-labeled agents may be used, in-cluding:1. 99mTc-ethyl cysteinate dimer (99mTc-ECD)2. 9 9 mTc-hexamethylpropylene amine oxime

(99mTc-HMPAO)3. 9 9 mTc-diethylenetriaminepentaacetic acid

(99mTc-DTPA)Although brain-specific tracers such as 9 9 mT c -

Society of Nuclear Medicine Procedure Guideline forBrain Death Scintigraphyversion 1.0, approved February 25, 2003

A u t h o r s : Kevin J. Donohoe, MD (Beth Israel Deaconess Medical Center, Boston, MA); Kirk A. Frey, MD, PhD (Universityof Michigan Medical Center, Ann Arbor, MI); Victor H. Gerbaudo, PhD (Brigham and Women’s Hospital, Boston, MA);Giuliano Mariani, MD (University of Pisa Medical School, Pisa, Italy); James S. Nagel, MD (Veterans Affairs Boston Health-care System, West Roxbury, MA); and Barry Shulkin, MD (University of Michigan Medical Center, Ann Arbor, MI).

HMPAO and 99mTc-ECD are increasing in popu-larity, there is no clear evidence they are more ac-curate than nonspecific agents. Brain-specificagents are preferred by some institutions, be-cause their interpretation is far less dependenton the quality of the bolus and because delayedimages are usually definitive for the presence orabsence of cerebral blood flow.

The Brain Imaging Council of the Society ofNuclear Medicine believes that although individ-ual laboratories may have used and may continueto use agents such as DTPA, glucoheptonate, andpertechnetate (with perchlorate blockade), theseare much less favorable than HMPAO and ECDfor assessment of cerebral perfusion.

E. Image AcquisitionFlow images should be acquired. They are essen-tial for interpretation of studies using non-brain-binding agents such as 99mTc-DTPA. In studiesusing brain-specific agents, such as 9 9 mT c - H M-PAO and 99mTc-ECD, lack of visualization of thebrain on delayed images could conceivably becaused by improper preparation or instability ofthe radiopharmaceutical. Flow images will helpto confirm lack of brain blood flow when thebrain is not visualized on delayed images using99mTc-HMPAO and 99mTc-ECD.1. Instrumentation

a. Gamma camera with field of view largeenough to image entire head and neck.

b. Low-energy high-resolution (LEHR) or ul-

tra-high resolution (UHR) collimator. c. 15%–20% energy window centered around

140 keV.2. Flow images are acquired at the time of tracer

injection. a. 1–3 s per frame for at least 60 s. Flow im-

ages should start before the arrival of thebolus in the neck and end well after the ve-nous phase.

b. Use of high-resolution or UHR collimationis recommended. As a general rule, use thehighest resolution collimation available.

3. Static imagesa . If a non-brain-binding agent, such as 9 9 mT c -

DTPA, is used, static images are acquiredimmediately for 5 min in anterior, right lat-eral, left lateral, and, if possible, posteriorprojections for approximately 5 min perview. Zooming or magnification may behelpful, particularly in pediatric cases.

b. For brain-specific agents, planar andSPECT images should be obtained after atleast 20 min. Images should be obtained inanterior, right lateral, left lateral, and, ifpossible, posterior projections.

4. When using brain-specific agents, such as9 9 mTc-HMPAO and 9 9 mTc-ECD, SPECT im-ages may be obtained in addition to flow andplanar images as described above. SPECT al-lows better visualization of perfusion to theposterior fossa and brain stem structures.

114 • BRAIN DEATH SCINTIGRAPHY

Radiation Dosimetry in Adults

99mTc-DTPA1 555–740 0.065 0.0063intravenously Bladder wall

(15–20) (0.24) (0.023)99mTc-HMPAO2 370–1,110 0.034 0.0093

intravenously Kidneys(10–30) (0.0126) (0.034)

99mTc-ECD3 370–1,110 0.073 0.011intravenously Bladder wall

(10-30) (0.27) (0.042)

Radiopharmaceuticals Effective Dose

MSv/MBq(rem/mCi)

Organ Receiving theL a r g e s t Radiation Dose

MGy/MBq ( r a d / m C i )

Administered Activity

M B q( m C i )

1 International Commission on Radiological Protection and Measurements. Radiation Dose to Patients from Radiophar-maceuticals. ICRP report no. 53. London, UK: ICRP; 1988:188.

2 International Commission on Radiological Protection and Measurements. Radiological Protection in Biomedical Re-search. ICRP report no. 62. London, UK: ICRP; 1993:13.

3 Oak Ridge Institute for Science and Education. Radiation Dose Estimates for Radiopharmaceuticals. Oak Ridge, TN: Radi-ation Internal Dose Information Center. Available at: www.orau.gov/ehsd/DOSETABLES.doc. Accessed Feb.13, 2003.

However, SPECT is rarely, if ever, used inthese patients who are often unstable and onlife support equipment, which may be incom-patible with SPECT acquisition.a. Multiple-detector or other dedicated

SPECT cameras generally produce resultssuperior to single-detector, general-pur-pose units. However, with meticulous at-tention to procedure, high-quality imagescan be produced on single-detector unitswith appropriately longer scan times (5 x

106 total counts or more are desirable). b. LEHR or fanbeam collimators are pre-

ferred when SPECT images will be ac-quired. As a general rule, use the highestresolution collimation available.

c . Use the smallest possible radius of rotation.d. A 128 x 128 or greater acquisition matrix is

preferred.e. Angular sampling of 3° or better is pre-

ferred. Acquisition pixel size should be1/3–1/2 the expected reconstructed reso-lution. It may be necessary to use a hard-ware zoom to achieve an appropriate pixelsize. Different zoom factors may be usedwith in-plane and axial dimensions of afanbeam collimator.

f. The time per stop and number of counts ac-quired for the study will depend on the

amount of tracer activity in the brain andthe specific camera being used. It is sug-gested that the number of seconds per stopbe similar to that used on your equipmentfor acquiring other brain SPECT studies.

g. It is frequently useful to use detector panand zoom capabilities to ensure that the en-tire brain is included in the field of viewwhile allowing the detector to clear the pa-tient’s shoulders.

F. InterventionsNone

G. Processing: SPECT1. Filter studies in 3 dimensions. This can be

achieved either by 2-dimensionally prefilter-ing the projection data or by applying a 3-di-mensional postfilter to the reconstructed data.

2. Low-pass (e.g., Butterworth) filters should beused. Resolution recovery or spatially varyingthe filters should be used with caution, how-ever; they may produce artifacts.

3. Always reconstruct the entire brain. Use carenot to exclude the cerebellum or vertex.

4. Reconstruct data at highest pixel resolution(i.e., 1 pixel thick). If slices are to be summed,this should be done only after reconstructionand oblique reorientation (if performed).

H. Interpretation Criteria 1. For studies using brain-specific agents:

SOCIETY OF NUCLEAR MEDICINE PROCEDURE GUIDELINES MANUAL MARCH 2003 • 115

Radiation Dosimetry in Children(5 Years Old; Normal Renal Function)

99mTc-DTPA1 7.4 370 740 0.17 0.017intravenously (10) (20) Bladder wall

(0.2) (0.63) (0.063)99mTc-HMPAO2 11.1 185 740 0.14 0.026

intravenously (5) (20) Thyroid(0.3) (0.52) (0.096)

99mTc-ECD3 11.1 185 740 0.083 0.023intravenously (5) (20) Bladder wall

(0.3) (0.31) (0.085)

Radiopharmaceuticals Effective Dose

MSv/MBq (rem/mCi)

Organ Receiving theLargest Radiation Dose

MGy/MBq (rad/mCi)

Administered Activity

M B q / k g M i n i m u m M a x i m u mD o s e D o s e D o s e

( m C i / k g ) M B q M B q( m C i ) ( m C i )

1 International Commission on Radiological Protection and Measurements. Radiation Dose to Patients from Radiophar-maceuticals. ICRP report no. 53. London, UK: ICRP; 1988:188.

2 International Commission on Radiological Protection and Measurements. Radiological Protection in Biomedical Re-search. ICRP report no. 62. London, UK: ICRP; 1993:13.

3 Oak Ridge Institute for Science and Education. Radiation Dose Estimates for Radiopharmaceuticals. Oak Ridge, TN: Ra-diation Internal Dose Information Center. Available at: www.orau.gov/ehsd/pedose.doc. Accessed Feb. 13, 2003.

a. Flow images in brain death are character-ized by a lack of flow to the middle cerebralartery, the anterior cerebral artery, and theposterior cerebral artery. This often resultsin a lack of a “blush” of activity in the mid-dle of the head during flow images. Keepin mind that the external carotid artery willlikely remain patent and there will be someflow to the scalp, which can be mistakenfor brain flow in some instances. Anotherimportant sign in brain death is lack oftracer activity in the superior sagittal sinusduring the venous phase of the flow study.

b. Flow images are assessed for blood flow tothe brain.(1) Anterior views are preferred for imag-

ing blood flow. The head should beviewed straight on to allow for com-parison of right and left carotid flow.

(2) Tracer flow should be observed fromthe level of the carotids to the skull ver-tex. In the anterior position, the rightand left middle cerebral arteries appearalong the lateral aspects of the skull.The anterior cerebral arteries appearmidline and appear as 1 vessel.

(3) In brain death, blood flow superior tothe circle of Willis circulation is com-pletely absent. There may be an accom-panying blush of activity in the regionof the nose (“hot nose sign”). Caremust be taken to distinguish externalcarotid circulation to the scalp from in-ternal carotid circulation to the brain.

(4) The superior sagittal sinus is oftennoted during the venous phase ofblood flow in patients with intact bloodflow to the brain. However, low-levelsagittal sinus activity can come fromthe scalp. If no internal carotid flow orCNS perfusion is seen on the flowstudy and minimal sagittal sinus activ-ity is noted, these findings should benoted and a note of caution regardingthe accuracy of the interpretation in-cluded in the report.

(5) In cases of head trauma, hyperemicblood flow to injured scalp structuresmay mimic brain blood flow or supe-rior sagittal sinus activity.

(6) CSF shunts and intracranial pressuretransducers can cause hyperemia re-sulting in increased scalp flow, possi-bly causing a false-negative flow study.Disruptions in the skull and scalp, as

well as pressure on the portion of thescalp resting on a hard surface, canproduce a relatively photopenic areaon the flow study, falsely suggestingdiminished flow.

c. Delayed planar or SPECT images shoulddemonstrate no tracer uptake in the brainfor the diagnosis of brain death to be made.For SPECT studies, unprocessed projectionimages should be reviewed in cinematicdisplay before viewing of tomographic sec-tions. Projection data should be assessedfor target-to-background ratio and otherpotential artifacts. Inspection of the projec-tion data in sinogram form may also beuseful. The role and use of SPECT imagingis unclear. Both cerebral hemispheres andthe posterior fossa (cerebellum) should beevaluated for a complete study. Therefore,if performing planar scintigraphy an ante-rior-posterior or posterior-anterior view,separating left and right hemispheres andat least 1 lateral view to distinguish thecerebral flow from that of the cerebellumare commonly needed.

d. Images viewed on a computer screenrather than from film or paper copy permitinteractive adjustment of contrast, back-ground subtraction, and color table.

e. Grayscale is preferred to color tables. Atvery low levels of activity, color tables usu-ally designed for viewing near-normal ac-tivity may under-represent low activity,causing a false-positive study.

2. For studies using non-brain-binding agentsa. Delayed images using agents that are not

brain specific should not demonstrate su-perior sagittal sinus activity in patientswith brain death. Another finding thatmay be present in patients with braindeath is the “halo” sign. This is a pho-topenic defect caused by compression ofscalp blood flow that may be seen whenthe patient’s head is resting on a firm ob-ject, such as an imaging table.

I. Reporting1 . Reports should include the tracer used and ba-

sic imaging information, such as the acquisitionof SPECT and/or planar images. Flow imagesshould be reported in a separate paragraph.

2. Reports should describe the extent and sever-ity of brain perfusion deficits. If brain-specificagents are used, specific mention of perfusionto the posterior fossa and brain stem may bereported. Because this study is used in combi-

116 • BRAIN DEATH SCINTIGRAPHY

nation with other tests and physical examfindings, the final impression of a positivestudy should state that the study is “consis-tent with brain death” rather than “demon-strates brain death.”

3. Severely decreased brain perfusion is oftenprogressive. If there is a small amount of re-maining perfusion, consider recommending arepeat study in 24 h.

J. Quality Control1. If using brain-specific agents, quality control

of labeling and stability of the compound isessential to prevent false-positive results.Poor radiopharmaceutical labeling or stabilitywould result in minimal concentration oftracer in the brain. This could be falsely inter-preted as lack of cerebral perfusion.

2. See Society of Nuclear Medicine Procedure Guide-line for General Imaging.

K. Sources of Error1. Improper labeling of brain-specific radio-

pharmaceuticals or injection of the wrong ra-diopharmaceutical can result in false-positivestudies as described in section J.1.

2. Drainage of blood from the scalp into the su-perior sagittal sinus may cause a false-nega-tive flow study.

3. Hyperemic scalp structures may result infalse-negative flow studies if non-specificbrain agents are used.

4. Infiltration of tracer at injection site may causea false-positive study if the entire dose is infil-trated and not available to the vascular space.Absence of activity in the carotid vessels onflow images suggests complete infiltration ofthe dose.

V. Issues Requiring Further Clarification

A. The relative accuracies of brain-specific and non-specific agents.

B. The importance of SPECT imaging. C. The value of brain-specific agents for the detec-

tion of small areas of brain perfusion, such as inthe posterior fossa. Will this increased sensitivityfor small areas of perfusion change the ultimateprognosis?

D. The influence of open fontanels in small childrenupon the accuracy of flow studies.

VI. Concise Bibliography

Bonetti MG, Ciritella P, Valle G, Perrone E. 99m-TcHM-PAO brain perfusion SPECT in brain death.

Neuroradiology. 1995;375:365–369.Facco E, Zucchetta P, Munari M, et al. 99mTc-HM-

PAO SPECT in the diagnosis of brain death. In-tensive Care Med. 1998;249:911–917.

Flowers WM, Jr., Patel BR. Accuracy of clinical eval-uation in the determination of brain death. SouthMed J. 2000;93:203–206.

Haupt WF, Rudolf J. European brain death codes: acomparison of national guidelines. J Neurol.1999;246:432–437.

Hoch DB. Brain death: a diagnostic dilemma. J NuclMed. 1992;33:2211–2213.

Larar GN, Nagel JS. Tc-99m HMPAO cerebral perfu-sion scintigraphy: considerations for timelybrain death declaration. J Nucl Med.1992;33:2209–2211.

Lee VW, Hauck RM, Morrison MC, Peng TT, FischerE, Carter A. Scintigraphic evaluation of braindeath: significance of sagittal sinus visualization.J Nucl Med. 1987;28:1279–1283.

Lopez-Navidad A, Caballero F, Domingo P, et al.Early diagnosis of brain death in patients treatedwith central nervous system depressant drugs.Transplantation. 2000;70:131–135.

Mrhac L, Zakko S, Parikh Y. Brain death: the evalu-ation of semi-quantitative parameters and othersigns in HMPAO scintigraphy. Nucl Med Com-mun. 1995;16:1016–1020.

Spieth M, Abella E, Sutter C, Vasinrapee P, Wall L,Ortiz M. Importance of the lateral view in theevaluation of suspected brain death. Clin NuclMed. 1995;20:965–968.

VII. Disclaimer

The Society of Nuclear Medicine has written andapproved guidelines to promote the cost-effec-tive use of high-quality nuclear medicine proce-dures. These generic recommendations cannotbe applied to all patients in all practice settings.The guidelines should not be deemed inclusiveof all proper procedures or exclusive of otherprocedures reasonably directed to obtaining thesame results. The spectrum of patients seen in aspecialized practice setting may be quite differ-ent from the spectrum of patients seen in a moregeneral practice setting. The appropriateness ofa procedure will depend in part on the preva-lence of disease in the patient population. In ad-dition, the resources available to care for pa-tients may vary greatly from one medical facilityto another. For these reasons, guidelines cannotbe rigidly applied.

SOCIETY OF NUCLEAR MEDICINE PROCEDURE GUIDELINES MANUAL MARCH 2003 • 117