William G. Bradley Jr., MD, PhD

CT perfusion for stroke leaped from clinical discussion forums to the front pages in the last 13 months. First, the FDA revealed in October 2009 that Cedars-Sinai Medical Center may have exposed more than 200 stroke patients to eight times the normal dose of ionizing radiation for CT perfusion brain scans. Then, in July an article featured in The New York Times concluded that overdoses from CT perfusion brain scans are far more common than even the FDA recognized.

The New York Times article prompted a letter to the editor from Dr. George Lantos, vice chair of radiology and director of neuroradiology at North Bronx Healthcare Network, arguing that CT perfusion for stroke should be performed only in the setting of a clinical trial. He agreed to expand on his comments for Diagnostic Imaging. To get the other side of the story we turned to Dr. Carol Geer and Dr. Christopher Whitlow, assistant professors of radiology in the neuroradiology division at Wake Forest University.

Finally, we asked Dr. William G. Bradley Jr., chair of radiology at the University of California, San Diego what it will take to move MR into a more prominent position on stroke imaging.

How to treat stroke is a crucial issue in ER departments across the nation, and CT imaging remains a common gateway. It is our hope that this contrast of views will contribute to the discussion of this topic.

NO: Routine clinical use of CT perfusion, 				MR diffusion/perfusion is not ready for prime time

Until phase III results are in, perfusion scans should not be used except within the context of clinical trials

In our field of diagnostic imaging, no one wants to be involved in a controversy and land on the front page of The 

. Yet that is what happened in the case of CT perfusion this summer.

 reported that in the investigation of acute stroke, a number of patients undergoing this procedure, known to involve radiation levels higher than most other diagnostic imaging tests even when properly performed, actually suffered radiation burns to their skin. Some of the subjects received far more radiation than intended. Clearly, there is potential for permanent brain damage in these patients as well as radiation-induced neoplasms in the decades to come.

The article prompted me to write a letter to the editor.

 My intent was not to join others decrying the excessive radiation; it’s obvious that CT scanners must be programmed and radiologists and technologists must be trained to avoid overdoses of radiation. Rather, I wanted to point out that since there is no current FDA-approved therapy that utilizes the information from CT perfusion or MRI perfusion/diffusion studies in the setting of acute stroke therapy, these advanced imaging techniques should be performed only in the context of clinical investigation. I thank the editors of Diagnostic Imaging for the opportunity to expand on these views.

To understand the role of imaging in acute stroke management, it’s useful to review some pathophysiology. It is well known that cerebral ischemia initiates a number of damaging cellular events called the ischemic cascade.

 These cellular events include acidosis due to a switch from aerobic to anaerobic metabolism, release of the amino acid glutamate and consequent excitotoxicity, elevation of intracellular calcium, and production of toxic substances including nitric oxide and free radicals. Ultimately, cell death results from these events.

The brain tissue that suffers irreversible necrosis is referred to as the ischemic core. Reviewing their own and others’ physiologic studies of cerebral ischemia, the concept of a penumbra in acute stroke was first advanced by Astrup and his colleagues in 1981.

 Penumbra refers to the poorly perfused and therefore nonfunctioning brain substance surrounding the core; but though the tissue is nonfunctioning, it has not suffered permanent damage. The situation is similar to that in the case of myocardial infarction, where permanently damaged myocardium is surrounded by nonfunctioning (or hibernating), but nonetheless viable, tissue.

Strategies directed at limiting ischemic brain damage as much as possible can therefore be divided into two broad categories. One is neuroprotection. The various stages in the damaging ischemic cascade, for example, provide attractive targets for neuroprotective agents that could potentially be deployed against one or more of the toxic chemical species produced during a stroke. Various strategies have been tried to limit these cell-damaging events. However, despite much effort in both the basic science and clinical trial spheres, and for a variety of reasons including timing of therapy, to date no neuroprotective agent has been shown to have definite clinical efficacy.

The other broad category of strategies is the attempted restoration of cerebral blood flow and perfusion, the earlier the better, to minimize irreversibly damaged tissue. In 1995, the National Institute of Neurological Disease and Stroke published a landmark multi-institution study on the intravenous injection of recombinant tissue plasminogen activator (rtPA, or simply tPA) in acute stroke therapy.

 This was followed the next year by FDA approval of the drug. Fifteen years later, tPA is still the only FDA-approved drug for the treatment of acute stroke. A variety of clinical parameters must be met prior to administration of tPA. These include the passage of no more than three hours since the patient was last known to be neurologically normal, as well as absence of significant bleeding diathesis and severe hypertension and other clinical features.

It turns out that the three-hour window for administration of tPA is the most limiting factor, making most patients who present with ischemic stroke ineligible.

 Although recent evidence indicates that this window can be extended to 4.5 hours,

 many patients will likely still be disqualified because they present too late.

Imaging needed for the use of tPA is quite simple. A noncontrast CT scan is performed to exclude hemorrhage and nonstroke pathology. In addition, tPA is not administered if there is a well-formed infarct in the area of clinical concern not consistent with a duration less than three hours. Evidence from the European Cooperative Acute Stroke Study indicates that involvement of more that one-third of the middle cerebral artery territory results in a 3.5-fold risk of parenchymal hemorrhage following administration of tPA.

A routine noncontrast CT scan is a blunt instrument with which to assess the rapidly changing pathophysiologic events involved in an ischemic stroke. It cannot distinguish tissue destined to undergo irreversible damage from potentially salvageable surrounding penumbra. Although noncontrast CT can disclose the presence of thrombus in a major vessel-for example, the dense middle cerebral artery sign first reported by Gacs et al

-it cannot depict the level of detail of cerebral arterial supply that one would like to have in order to restore regional perfusion by opening a cerebral arterial branch via either thrombolysis or mechanical clot retrieval.

Both CT and MRI methods have been developed for the study of regional brain perfusion as well as depiction of the details of brain arterial supply. The merits of each modality have been debated in the recent literature.

 While these tests undoubtedly provide the radiologist and clinician with a great deal of information, such diagnostic methods are currently far ahead of approved therapy. There is a strong rationale for treating patients with a mismatch between an already infarcted core and surrounding poorly perfused but viable penumbra, particularly with drugs that would work outside the tPA window of up to 4.5 hours. Unfortunately, there has already been a false start along this road. Hacke and colleagues conducted promising trials with the novel thrombolytic agent desmoteplase, selecting patients for treatment on the basis of a mismatch between diffusion and perfusion on MRI.

 But later studies with larger numbers of patients failed to confirm the benefit of desmoteplase in the initial trials.

There is also a strong rationale for using the information from MR or CT angiography to treat patients with major intracranial arterial occlusions or stenoses with thrombolysis, stenting, or clot retrieval. Again, strong rationale does not equate to solid evidence from well-done clinical trials.

I believe that before we put a diagnostic modality or therapy into routine clinical use, we should require the best possible evidence for its efficacy. Guidelines for the evaluation of various acute stroke therapies have been published (see table). Lest we fall into the trap of “knowing” that a particular approach is correct, we must remember that the history of medicine is replete with things we thought we knew that turned out to be wrong.

One example in the stroke field is extracranial-intracranial bypass surgery for carotid occlusive disease, first described by Donaghy and Yasargil in the 1960s.

 I worked with a neurosurgeon in the 1980s who performed these operations and who “knew” they were effective in reducing the long-term risk of stroke. He was so certain that he refused to include his patients in the definitive prospective randomized clinical trial of this procedure,

 although two of our patients were randomized to medical therapy in that study. Unfortunately for stroke patients, the study showed that after the initial surgical mortality and morbidity is taken into account, the benefits of surgery never caught up with those accorded to the medically treated group.

This operation is still being performed for stroke in selected situations and various other conditions, including cases where the internal carotid or other major artery has to be sacrificed in aneurysm or skull-base surgery. There is presently no justification for performing it in otherwise unselected patients with occlusive disease. However, it remains possible that there is a subset of patients who could be identified by modern physiologic perfusion imaging and benefit from this procedure. A recent study published by the Cochrane group made the recommendation that this possibility should be investigated.

Getting back to the issue of routine use of perfusion imaging in the setting of acute stroke, I’d like to summarize by quoting the conclusion of a recent meta-analysis of the literature on mismatch-based stroke therapy conducted by several leading neurologists working in this field:

Delayed thrombolysis amongst patients selected according to mismatch imaging is associated with increased reperfusion/recanalization. Recanalization/reperfusion is associated with improved outcomes. However, delayed thrombolysis in mismatch patients was not confirmed to improve clinical outcome, although a useful clinical benefit remains possible. Thrombolysis carries a significant risk of symptomatic intracerebral hemorrhage and possibly increased mortality. Criteria to diagnose mismatch are still evolving. Validation of the mismatch selection paradigm is required with a phase III trial. Pending these results, delayed treatment, even according to mismatch selection, cannot be recommended as part of routine care.

At this writing, I’m not aware of any prospective, randomized, controlled phase III studies that have shown a superior outcome in acute stroke patients selected with either CT perfusion or MR perfusion for treatment with any modality. Of course, if treatment doesn’t depend on the results, then there’s really no reason to perform these tests in routine clinical practice outside the context of clinical trials.

1. Bogdanich W. The radiation boom-after stroke scans, patients face serious health risks. NYT, July 31, 2010; 

www.nytimes.com/2010/08/01/health/01radiation.html

. Accessed Oct. 4, 2010.
 2. Lantos G. Excess radiation from CT scans; letters to the editor. NYT, Aug. 8, 2010; 

www.nytimes.com/2010/08/09/opinion/l09radiation.html

. Accessed Oct. 4, 2010.
 3. Brouns R, De Deyn PP. The complexity of neurobiological processes in acute ischemic stroke. Clin Neurol Neurosurg 2009;111(6):483-495.
 4. Astrup J, Siesjo BK, Symon L. Thresholds in cerebral ischemia - the ischemic penumbra. Stroke 1981;12(6):723-725.
 5. Slezak J, Tribulova N, Okruhlicova L, et al. Hibernating myocardium: pathophysiology, diagnosis, and treatment. Can J Physiol Pharmacol 2009;87(4):252-265.
 6. Ginsberg MD. Neuroprotection for ischemic stroke: past, present and future. Neuropharmacology 2008;55(3):363-389.
 7. Tissue plasminogen activator for acute ischemic stroke. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. N Engl J Med 1995;333(24):1581-1587.
 8. Barber PA, Zhang J, Demchuk AM, et al. Why are stroke patients excluded from TPA therapy? An analysis of patient eligibility. Neurology 2001;56(8):1015-1020.
 9. Hacke W, Kaste M, Bluhmki E, et al. Thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Engl J Med 2008;359(13):1317-1329.
 10. Larrue V, von Kummer R, del Zoppo G, Bluhmki E. Hemorrhagic transformation in acute ischemic stroke. Potential contributing factors in the European Cooperative Acute Stroke Study. Stroke 1997;28(5):957-960.
 11. Gacs G, Fox AJ, Barnett HJ, Vinuela F. CT visualization of intracranial arterial thromboembolism. Stroke 1983;14(5):756-762.
 12. Köhrmann M, Schellinger PD. Acute stroke triage to intravenous thrombolysis and other therapies with advanced CT or MR imaging: pro MR imaging. Radiology 2009;251(3):627-633.
 13. Wintermark M, Rowley HA, Lev MH. Acute stroke triage to intravenous thrombolysis and other therapies with advanced CT or MR imaging: pro CT. Radiology 2009;251(3):619-626.
 14. Hacke W, Albers G, Al-Rawi Y, et al. The Desmoteplase in acute ischemic stroke trial (DIAS): a phase II MRI-based 9-hour window acute stroke thrombolysis trial with intravenous desmoteplase. Stroke 2005;36(1):66-73.
 15. Furlan AJ, Eyding D, Albers GW, et al. Dose escalation of desmoteplase for acute ischemic stroke (DEDAS): evidence of safety and efficacy 3 to 9 hours after stroke onset. Stroke 2006;37(5):1227-1231.
 16. Hacke W, Furlan AJ, Al-Rawi Y, et al. Intravenous desmoteplase in patients with acute ischaemic stroke selected by MRI perfusion-diffusion weighted imaging or perfusion CT (DIAS-2): a prospective, randomised, double-blind, placebo-controlled study. Lancet Neurol 2009;8(2):141-150.
 17. Donaghy RMP. Patch and bypass in microangional surgery. In: Donaghy RMP, Yasargil MG, eds. Microvascular surgery. St. Louis: CV Mosby, 1967:75-86.
 18. Yasargil MG. Anastomosis between the superficial temporal artery and a branch of the middle cerebral artery. In: Yasargil MG, ed. Microsurgery applied to neurosurgery. Stuttgart, Germany: Georg Thieme, 1969:105-115.
 19. Failure of extracranial-intracranial arterial bypass to reduce the risk of ischemic stroke. Results of an international randomized trial. The EC/IC Bypass Study Group. N Engl J Med 1985;313(19):1191-1200.
 20. Fluri F, Engelter S, Lyrer P. Extracranial-intracranial arterial bypass surgery for occlusive carotid artery disease. Cochrane Database Syst Rev 2010;2:CD005953.
 21. Mishra NK, Albers GW, Davis SM, et al. Mismatch-based delayed thrombolysis: a meta-analysis. Stroke 2010;41(1):e25-e33.
 22. Adams HP Jr, Adams RJ, Brott T, et al. Guidelines for the early management of patients with ischemic stroke: A scientific statement from the Stroke Council of the American Stroke Association. Stroke 2003;34(4):1056-1083.

YES: CT perfusion proves its value in planning treatment for acute ischemic stroke

Careful patient selection is crucial, but CT perfusion can be key in making a decision whether to proceed with intravascular tPA

BY CAROL P. GEER, M.D., AND CHRISTOPHER T. WHITLOW, M.D., Ph.D.

Ischemic stroke remains a common cause of substantial morbidity and mortality in the U.S., and great effort has been expended to better evaluate and treat affected patients. The FDA has approved the use of intravenous tissue plasminogen activator in eligible patients presenting for treatment of ischemic stroke within three hours of symptom onset and without evidence of intracerebral hemorrhage on CT of the brain, which has improved clinical outcomes in some patients.

However, there remain a significant number of patients who do not derive benefit from IV tPA, possibly due to substantial clot burden in larger vessels, such as the internal carotid artery and proximal middle cerebral artery (MCA). There are also a number of patients who present to the healthcare system after the three-hour-or now the 4.5-hour-window has passed.

 Although many patients present outside of the standard treatment window for IV tPA, there is evidence that delayed recanilization may have clinical benefit in some of these patients. Of these, patients who have brain tissue that is ischemic but not irreversibly damaged, or so-called ischemic penumbra, are those who stand to benefit most. Diagnostic imaging methods designed to accurately measure penumbra and therapies directed at preserving this tissue in patients who present after the IV tPA window are targets of active research across the globe.

CT perfusion has become a key imaging modality to qualitatively determine the amount of penumbra present relative to completed infarct in patients with occlusive cerebrovascular thrombus.

 CT perfusion defines penumbra as a mismatch between the mean transit time (MTT) and cerebral blood volume (CBV).

 Cerebral blood flow (CBF) is subsequently calculated from the measured MTT and CBV. MRI diffusion and perfusion imaging has also been used to evaluate for significant salvageable brain tissue.

 But in most emergency departments, CT perfusion can be obtained more quickly.

Patients presenting with stroke symptoms generally first undergo head CT without contrast to evaluate for the presence of intracranial hemorrhage and imaging signs of acute ischemia. In many centers, including ours, patients then undergo head and neck CT angiography to evaluate for occlusive cerebrovascular thrombus. Because the patient is already on the CT table, obtaining CT perfusion data to evaluate for penumbra, as opposed to transferring the patient for MRI examination, makes good clinical sense, given that time is the critical factor for preserving brain. The main disadvantages to obtaining CT perfusion rather than MRI diffusion-perfusion data are the radiation exposure and the IV contrast administration. The critical question faced by diagnostic and treating clinicians is whether the decrease in radiation and contrast exposure gained by performing MRI diffusion-perfusion rather than CT perfusion is worth the increase in time to reperfusion and potential infarct expansion.

Indeed, several studies have demonstrated that there is an increased risk of infarct expansion if there is a large penumbra.

 Furthermore, some recently published data suggest that patients presenting with large penumbra and relatively small completed infarcts longer than eight hours after symptom onset, as well as some patients with wake-up strokes, may benefit from intra-arterial thrombolysis.

 This being so, many centers, including our own, have started to be more aggressive in treating patients with substantial ischemic penumbra, using intra-arterial thrombolysis to recanalize obstructed arteries.

The critical information necessary for evaluating the viability of intra-arterial thrombolysis as a treatment option is based on perfusion imaging data and includes penumbra size and degree of completed tissue infarction. Endovascular recanalization of patients who already have a large completed infarct increases their risk of complications due to edema and reperfusion hemorrhage.

 CT perfusion and MRI data, therefore, are used to identify patients who may benefit from intra-arterial thrombolysis, and those in whom recanalization may be of questionable benefit or even be harmful. Although CT and MRI can both be useful for evaluation of penumbra and infarct completion, in our hospital and many others, it is much faster to obtain CT perfusion. MRI scanners at our institution are not located in, or even very close to, our emergency department, making transport time a serious issue.

Given the necessity for CT perfusion in the evaluation of ischemic stroke, at our institution we attempt to employ strict adherence to ALARA (as low as reasonably achievable) principles. In particular, we have optimized our CT scan technique to use a low-dose perfusion protocol that maintains a radiation dose at less than 200 mGy. Attention to technique, as well as oversight by experienced radiologists and a dedicated physicist, are critical elements we have employed to avoid the complications that other centers have reported with radiation exposure, including hair loss and scalp injury.

Furthermore, we work in close collaboration with the primary clinical referral services at our institution (emergency medicine, neurology, and neurosurgery) to carefully select patients in whom the use of CT perfusion is important for treatment planning. Indeed, performing CT perfusion on every stroke patient as a standard protocol is probably not good medicine and is not being responsible with radiation exposure. Using CT perfusion in a careful and judicious manner, when the results may change clinical management, is likely to be the safest and most efficacious way to use this technology.

To elucidate the above discussion, we present examples in which CT perfusion data were collected during the evaluation of two patients presenting to our hospital for treatment of acute ischemic stroke 4.5 hours after symptom onset.

: A 64-year-old woman presented with acute left hemiplegia and neglect four hours and 46 minutes after symptom onset. Head CT was negative for acute hemorrhage, and no large area of low density was identified to suggest infarct. Head and neck CTA demonstrated a large occlusive thrombus in the M1 segment of the right MCA. CT perfusion demonstrated elevated MTT in the entire right MCA, with only a small area of decreased blood volume in this vascular territory, compatible with a large penumbra (

). Studies have shown these patients to have an increased risk of infarct expansion. This patient was, therefore, recanalized using a combination of intra-arterial tPA and mechanical thrombolysis, including the Penumbra System (Penumbra, Alameda, CA) and balloon angioplasty. She made a complete recovery.

: An 84-year-old man presented with acute right hemiplegia, aphasia, and neglect four hours and 30 minutes after symptom onset. Head CT was negative for acute hemorrhage, and no large area of low density was identified to suggest infarct. Head and neck CTA demonstrated a large occlusive thrombus in the left M1 segment. CT perfusion was obtained, but due to a computer malfunction immediately after the study was acquired, the data could not be processed and viewed. The patient was nonetheless taken for intra-arterial recanalization, as the head CT suggested the absence of a large completed infarct, and we could not wait any longer for the perfusion data. Recanalization of the left MCA was performed using mechanical thrombolysis.

Following the procedure, the computer was fixed and the CT perfusion data were processed and subsequently reviewed. They demonstrated elevated MTT in the entire distribution of the left MCA with a matched decrease in CBV and CBF, suggesting completed infarct (

). Postoperatively, the patient developed severe edema in the left MCA distribution and expired.

Although outcomes from these two cases, and others in the literature, are certainly insufficient to direct policy and change standards of care, they do highlight the potential clinical utility of CT perfusion data in directing the delayed recanalization of occlusive cerebrovascular thrombus, and serve as a basis for further evaluation of these technologies. Case 2 is particularly important, as it demonstrates the importance of careful patient selection before application of endovascular recanalization techniques. Indeed, had the CT perfusion data driven subsequent management of this patient, intra-arterial thrombolysis would not have been performed, given the presence of a large completed infarct. The patient in case 2 was obviously not helped by recanalization, and it is possible that intra-arterial treatment resulted in worsening cerebral edema, among other negative consequences, such as raising false hope among the family, and adding unnecessary expense to the healthcare system.

Articles

CT perfusion for stroke leaped from clinical discussion forums to the front pages in the last 13 months.

CT perfusion for stroke: Should you use it?

When considering which is the most sensitive MRI technique for detecting disease in the brain, the choice is usually between contrast-enhanced T1-weighted imaging and T2-weighted fluid-attenuated inversion recovery (FLAIR) imaging. Twelve years ago, however, we found that contrast-enhanced FLAIR was actually more sensitive than either,1 particularly for subtle cortical lesions (Figure 1). This sensitivity is likely due to the combination of T2 prolongation, the usual mechanism for parenchymal hyperintensity on FLAIR images, and T1 shortening from the gadolinium chelate. While we normally think of T2 prolongation as the only mechanism for increasing signal intensity on T2-weighted FLAIR images, T1 shortening can also increase signal if it hasn't maxed out yet. Since T2 prolongation and T1 shortening are synergistic, contrast-enhanced FLAIR is more sensitive for subtle abnormalities than either FLAIR alone or postcontrast T1-weighted imaging alone.

The greatest advantage of contrast-enhanced FLAIR seems to be for detecting subtle cortical abnormalities such as leptomeningeal carcinomatosis (Figure 1), where there is no mass effect. However, it also has an advantage when compared with conventional FLAIR or T2-weighted imaging for deep lesions (Figure 2), where the contrast between the enhancing lesion and the surrounding vasogenic edema is greater than with conventional techniques, and produces very remarkable images.

A reason contrast-enhanced FLAIR is especially sensitive for subtle cortical abnormalities may be that a small amount of gadolinium leaks into the adjacent sulcus. While this may not produce enough T1 shortening to be seen on a T1-weighted image, it may shorten the T1 of cerebrospinal fluid so that it is no longer nulled by the initial 180o pulse in the inversion recovery FLAIR technique.

Contrast-enhanced FLAIR need not take longer than a conventional contrast-enhanced brain study. Since unenhanced FLAIR is usually a part of an MR brain exam, we merely acquire it after contrast is given, rather than before, as is usually done. In the beginning we would acquire T1-weighted images with and without contrast as well as FLAIRs with and without contrast. However, with time, we found we did not need the precontrast FLAIR. Comparing T1-weighted images pre- and postcontrast demonstrated enhancement, and any additional signal on the contrast-enhanced FLAIR we found to be due to T2 prolongation. Since most enhancing lesions generally also produce vasogenic edema, which prolongs T2, the two contrast mechanisms are synergistic.

When reading out brain studies performed with contrast enhancement, I usually read the contrast-enhanced FLAIRs first. If they are normal, it is unlikely that the enhanced T1-weighted images or the unenhanced FLAIRs will add anything. If they are abnormal, the pre- and post–T1-weighted images can sort out whether the brightness on the contrast-enhanced FLAIRs is due more to contrast enhancement or T2 prolongation.

Contrast-enhanced FLAIR is more sensitive than other MRI techniques due to T2 prolongation and T1 shortening.

Contrast FLAIR tops FLAIR and contrast T1-weighted images

When considering which is the most sensitive MRI technique for detecting disease in the brain, the choice is usually between contrast-enhanced T1-weighted imaging and T2-weighted fluid-attenuated inversion recovery (FLAIR) imaging. Twelve years ago, however, we found that contrast-enhanced FLAIR was actually more sensitive than either,

 particularly for subtle cortical lesions (Figure 1). This sensitivity is likely due to the combination of T2 prolongation, the usual mechanism for parenchymal hyperintensity on FLAIR images, and T1 shortening from the gadolinium chelate. While we normally think of T2 prolongation as the only mechanism for increasing signal intensity on T2-weighted FLAIR images, T1 shortening can also increase signal if it hasn’t maxed out yet. Since T2 prolongation and T1 shortening are synergistic, contrast-enhanced FLAIR is more sensitive for subtle abnormalities than either FLAIR alone or postcontrast T1-weighted imaging alone.

A reason contrast-enhanced FLAIR is especially sensitive for subtle cortical abnormalities may be that a small amount of gadolinium leaks into the adjacent sulcus. While this may not produce enough T1 shortening to be seen on a T1-weighted image, it may shorten the T1 of cerebrospinal fluid so that it is no longer nulled by the initial 180

 pulse in the inversion recovery FLAIR technique.

When considering which is the most sensitive MRI technique for detecting disease in the brain, the choice is usually between contrast-enhanced T1-weighted imaging and T2-weighted fluid-attenuated inversion recovery (FLAIR) imaging.

I have had the good fortune of being involved with MRI since the end of my first year of residency at the University of California, San Francisco in 1979. Over the last quarter century, whenever I thought MRI technology had plateaued, out came another application. Like an expanding bubble, the number of MR applications continues to rise exponentially.

Looking back over the last 27 years, I see several major MRI epochs: low- to midfield systems (late 1970s to mid-1980s), 1.5T with 10 mT/m gradients (mid-1980s to mid-1990s), and 1.5T with echo-planar gradients (mid-1990s to early 2000s). We entered a new epoch a few years ago: 3T with echo-planar gradients. Examining changes currently occurring, and understanding why they occur, can help us predict further changes to come over the next decade.

The first 3T magnets were introduced 10 years ago. Despite the better spatial resolution, early users such as Dr. Jeffrey S. Ross at the Cleveland Clinic were not thrilled. In fact, he wrote a fairly scathing editorial against 3T just three years ago in the American Journal of Neuroradiology.

 So why the stampede to 3T now? What changed? What changed was the appearance of two synergistic technologies: parallel imaging and phased-array coils.

Parallel imaging is also known as SENSE (Philips Medical Systems), ASSET (GE Healthcare), or iPAT (Siemens Medical Solutions). In order to do it, you need phased-array coils. Besides enabling parallel imaging, phased-array coils increase the signal-to-noise ratio. A typical eight-channel phased-array coil increases SNR 40%, independent of field strength. A prototype 32-channel head coil developed at Massachusetts General Hospital increased SNR in the cortex 3.5 times and that in the center of the brain almost 1.4 times.

 Parallel imaging allows users to trade off the extra SNR of 3T to go faster. Since 3T has roughly twice the SNR of 1.5T, it permits better spatial resolution and/or thinner slices. Users can go four times faster with the same spatial resolution and slice thickness, as long as they have at least four coil elements in the phased array.

If you cover every other line of k-space, the acquisition time is halved but so is your field-of-view, leading to aliasing or wraparound. By knowing the local sensitivity of each coil in the phased array, you can "unwrap" the image.

For the first time, unenhanced 1024 x 1024 (matrix) MR angiography of the brain becomes possible (Figure 1). Actually, by using a 1024 x 680 matrix over a 16 x 12-cm field-of-view, 160 x 200-micron pixels can be produced. This compares with 250-micron resolution for digital subtraction angiography. The 100-micron-diameter ophthalmic and lenticulostriate arteries are now commonly seen.

MRI of the breast can be performed with greater spatial resolution and thinner slices. Breast cancer that might appear smoothly marginated, and thus benign, on a 256 x 256 1.5T scan would have a greater chance of appearing spiculated on a 512 x 400 matrix at 3T (Figure 2).

Since T1 increases with field strength, the amount of T1 weighting (measured by the ratio TR/T1) will increase at constant TR, as will contrast enhancement. Greater T1 weighting will also lead to greater flow-related enhancement

 and allow better noncontrast perfusion imaging using arterial spin labeling.

Magnetic susceptibility effects such as those due to acute hemorrhage (deoxyhemoglobin) are more obvious at 3T. This sensitivity basically doubles the blood oxygen level-dependent effect, which is the basis for functional MRI. Since MRI is likely to be used for acute stroke triage in the future, it must be able to do what is currently done by CT: exclude hemorrhage. As described above, 3T MRI is exquisitely sensitive to acute hemorrhage. Since it can also perform diffusion and perfusion imaging on the entire brain, while 64-slice CT cannot do diffusion at all and can cover only 4 cm of brain for perfusion, MRI should replace CT for this purpose in the future. We will be placing 3T MR scanners near the emergency departments of both of our university hospitals over the next one to two years.

While 3T is also more sensitive to diamagnetic susceptibility artifacts, parallel imaging can be used to decrease these. With echo-planar diffusion and perfusion imaging, the large artifacts over the petrous bones result from the difference in diamagnetic susceptibility between air and water. These tend to be worse at 3T, but parallel imaging can be used to decrease the TE, reducing the diamagnetic susceptibility artifacts to 1.5T levels.

Chemical shift artifact between fat and water doubles in the move from 1.5T to 3T, all else (e.g., bandwidth) staying the same. Thus, use of STIR and fat saturation will probably be increased at 3T. The flip side of increased chemical shift artifact is increased spectral separation in MR spectroscopy.

In MRS, the SNR and spectral separation are twice at 3T what they are at 1.5T. The first property will allow smaller voxels and improved spatial resolution, down to ~5 mm for 3D techniques. The second may allow us to see species such as glutamate not currently seen. Glutamate, released by 90% of the excitatory neurons in the brain, is involved in memory, learning, and neurodegeneration. Being able to quantitate glutamate will be a huge step. Glutamate is broken down to glutamine, but, unfortunately, the two are tightly coupled spin systems currently combined as one low SNR peak labeled Glx. The spin coupling and loss of peak height increase with TE. In order to see glutamate separately from glutamine, the TE must be less than 5 msec, which is not currently possible on commercial scanners but has been achieved in a research setting.

While 3T brings many advantages, it also has potential problems. Radiofrequency heating is four times greater at 3T compared with 1.5T. For sequences like fast spin-echo, the increased RF heating could exceed the FDA's specific absorption rate limits. But tricks to decrease RF heating are available. The 180

 refocusing pulses, for example, can be reduced to 125

. The magnetization transfer pulses that are part of every MR angiogram of the brain increase RF heating but can be limited to the central 20% of k-space, reducing the SAR by 80%.

One problem that hasn't been entirely solved yet is the banding on body images due to standing waves from dielectric effects. This leads to local bright or dark stripes, which can be reduced with dielectric pads. While this remains a problem, 3T body images are still better than those at 1.5T, according to our chief of body MRI. Double-contrast livers with ultrasmall superparamagnetic iron oxide particles and gadolinium are definitely superior to anything at 1.5T for characterizing cirrhosis (Figure 3).

While some minor problems with 3T may be a challenge, I am reminded of that period 10 years ago when echo-planar units first came out. I purchased my first EPI upgrade knowing only that we would be able to scan in 100 msec. I had no idea that EPI diffusion, which we now do on every brain, and EPI perfusion, which we try to do on every contrast-enhanced brain, were in the wings or that the strong, fast gradients required for EPI would also enable the short TR and TE required for contrast-enhanced MRA.

Thus, whenever you are at a technology inflection point, as we are now with 3T, I would urge you to adopt the new technology even if you can't fully justify the added cost-yet. New applications will come along as they did for the systems with EPI gradients. Staying with current technology just means you will be married to the equivalent of a 0.5T system for the next 10 years.

Ultrashort TE (UTE) was first described almost 15 years ago by John Pauly, a professor of electrical engineering in the information systems laboratory at Stanford University. Prof. Graeme Bydder of STIR and FLAIR fame started working with the UTE technique almost five years ago when he was at Hammersmith Hospital in London, and he has shifted the work to 3T since moving to UCSD three years ago.

When most of us think short TE, we think 1 msec or so with a fractional echo; that is, a partial Fourier in read. Bydder is down to a TE of 8 microsec! At that short a TE, you can pick up signal from short T2 solids. Tissues with short T2s also have short T1s, so calcium and cortical bone appear bright on UTE images due to rapid T1 recovery.

Bydder has applied this technique to see the calcified deep layer of cartilage (Figure 4). Normally, this layer is dark and merges imperceptively with cortical bone. Since it has a short T1, it appears bright, so it can now be seen directly. This is significant because many arthridites start in the calcified deep layer of cartilage. This objective method to image early arthritis requires imaging the deep layer can be done only by UTE.

Motion has been a major cause of artifact in MRI. Several years ago, Jim Pipe, senior staff scientist in the MRI department at the Barrow Neurologic Institute, introduced a technique called Propeller, which samples k-space in a radial rather than the usual rectilinear fashion.

 If motion occurs, the high-intensity dots representing the center of k-space can be realigned, eliminating the motion artifact. Although 100-msec EPI diffusion techniques are relatively immune to motion, they are susceptible to diamagnetic susceptibility artifacts over the petrous bones. Combining diffusion imaging with a fast spin-echo readout minimizes these susceptibility effects, although it takes longer. Further combination with Propeller eliminates motion artifacts (Figure 5).

Several other techniques are enhanced at 3T:

 High-intensity focused ultrasound (HIFU) has been around for half a century, but there has never been a way to measure the temperature rise. In order to kill tissue, the temperature has to be at 60oC for at least a minute. MR can measure temperature changes based on the phase shift on a gradient-echo image. When HIFU is performed under MR guidance, it is known as MR-guided focused ultrasound (MRgFUS). This is a new technique that enables the thermal ablation of tumors throughout the body without breaking the skin.

A major application for MRgFUS is fibroid ablation (Figure 6). Women with symptomatic fibroids have a choice of hysterectomy, which is expensive and painful and requires a long recovery; uterine artery embolization, which is somewhat less expensive but still painful; and MRgFUS. Many women making the third choice are out riding a bicycle or a horse on the afternoon of their MRgFUS procedure.

In addition to fibroid ablation at 1.5T (FDA-approved) and 3T (FDA approval pending), MRgFUS can be used to ablate benign and malignant breast tumors and bone tumors. Recent work in Dr. Ferenc Jolesz's lab at Brigham and Women's Hospital in Boston has focused on MRgFUS of the brain. Since 90% of the ultrasound energy is reflected back by the calvarium, a waterbag heat sink has been added. Deep-seated tumors can be zapped without cracking the skull or breaking the skin. Glioblastoma multiforme can be palliated. Low-grade gliomas can potentially be cured. Since no radiation is involved, radiation necrosis that occurs with gamma knife and other radiosurgery techniques should be avoided.

When MRgFUS of the brain is run in low power mode, the blood-brain barrier can be temporarily opened, according to a recent finding.

 This could be important for patients with epilepsy (1% of the population). If magnetoencephalography is used to localize the initial seizure focus to within a centimeter, that region of the brain can be swept with low-power MRgFUS, temporarily opening up the blood-brain barrier. If an intravenous drip is going with Versed-laden liposomes, the MRgFUS would both open up the blood-brain barrier and lyse the liposomes, putting that point of the brain to sleep. If an EEG shows a drop in the seizure spike, the MRgFUS energy can be turned up to zap the seizure focus.

 MR has played a role in the operating room to monitor brain tumor resection since the initial 0.5T double-donut system was installed at the Brigham over a decade ago.

 The major finding at the Brigham and at Long Beach Memorial when I was there was that 80% of the time a neurosurgeon thinks a gross total resection has been achieved, MR-visible tumor remains to be removed. For low-grade gliomas, getting all the tumor out amounts to a complete cure. If even a small nodule is left behind, however, it will eventually degenerate into a glioblastoma multiforme, which currently has no cure.

While complete cure of a brain tumor is important to the individual patient, not that many new glioma patients emerge each year (maybe 20,000 low-grade tumors), and this is not a large market for vendors. They need a more common disease that can be cured only with intraoperative MRI. An exciting prospect being developed at UCSD by neuroscientist Dr. Mark Tuszynski is use of gene therapy (nerve growth factor) to treat Alzheimer's and Parkinson's diseases.10 To have a favorable effect in Alzheimer's and Parkinson's, however, the nerve growth factor must be deposited exactly in the nucleus basalis and the subthalamic nucleus, respectively. If it is deposited a few millimeters away from the target, the outcome can be negative rather than potentially positive.

Current techniques are based on frame-based stereotaxy, which has two problems. The first is the brain shift that occurs following the burr hole and leakage of cerebrospinal fluid. The second is that the performance of stereotaxy in Tallairach space is based on the brain of one alcoholic 40-year-old Frenchwoman. Performing the procedure under direct visualization using intraoperative MRI will ensure proper localization of the injection of the gene therapy today and the stem cell therapy tagged with superparamagnetic iron oxide particles (SPIO) in the future.

. Thirty years ago, clinical ultrasound units were large, cumbersome, expensive consoles with articulated arms. Today they are portable and cheap. The same evolution could happen with MRI. The most expensive part of the MR system is the magnet, which is wound and shimmed to produce a very uniform main magnetic field (Bo). MagneVu (Carlsbad, CA) has been working with cheap nonuniform Bo magnets. A prototype handheld MRI was produced several years ago. The technology has already been commercialized for in-office portable MR scanning of arthritis by rheumatologists. The 3D STIR images produced (Figure 7) are clearly superior to plain films for detecting early synovitis or changes with treatment.

Hyperpolarized carbon-13 MR spectroscopy.

 One of the most exciting prospects on the horizon is hyperpolarized C-13 MR spectroscopy. While hyperpolarized helium and neon have been used in the research environment for lung imaging, hyperpolarized C-13 gives us the opportunity to rapidly evaluate not only the structure but the metabolism of a tissue. Hyperpolarization refers to a process whereby the spins can acquire up to 100,000 times more signal than that achievable with conventional scanning.

Since there is no background signal, the SNR is extremely high, allowing rapid temporal spectroscopy at high resolution. There are two techniques for hyperpolarizing C-13: one using parahydrogen (the dominant form of hydrogen at low temperatures) and the other known as DNP (dynamic nuclear polarization), which involves doping hydrocarbon compounds with paramagnetic species and then separating them out before injection.

Each technique has its pros and cons. But using either technique will eventually make it possible to measure the kinetics as, for example, pyruvate is metabolized to lactate. Distinguishing normal from neoplastic tissue will potentially become possible.

Dr. Bradley is chair of radiology at the University of California, San Diego. He has received departmental support from GE Healthcare and is a consultant for GE and a participant in GE's speakers' bureau.

 Ross JS. The high-field-strength curmudgeon. AJNR 2004;25:168-169.


 Wiggins GC, Triantafyllou C, Potthast A, et al. 32-channel 3 Tesla receive-only phased-array head coil with soccer-ball element geometry. Magn Reson Med 2006;56(1):216-223.


 Bradley WG, Waluch V. Blood flow: magnetic resonance imaging. Radiology 1985;154:443-450.


 Wong EC, Cronin M, Wu WC, et al. Velocity-selective arterial spin labeling. Magn Reson Med 2006;55(6):1334-1341.


 Knight-Scott J, Shanbhag DD, Dunham SA. A phase rotation scheme for achieving very short echo times with localized stimulated echo spectroscopy. Magn Reson Imaging 2005;23(8):871-876. Epub 2005 Oct 13.


 Gatehouse PD, Thomas RW, Robson MD, et al. Magnetic resonance imaging of the knee with ultrashort TE pulse sequences. Magn Reson Imaging 2004;22(8):1061-1067.


 Forbes KP, Pipe JG, Karis JP, Heiserman JE. Improved image quality and detection of acute cerebral infarction with PROPELLER diffusion-weighted MR imaging. Radiology 2002;225(2):551-555.


 Hynynen K et al. Noninvasive MR Image-guided focal opening of the blood brain barrier in rabbits. Radiology 2001;220:640-646.


 Black PM, Moriarty T, Alexander E 3rd, et al. Development and implementation of intraoperative magnetic resonance imaging and its neurosurgical applications. Neurosurgery 1997;41(4):831-842; discussion 842-845.


 Tuszynski MH, Thal L, Pay M, et al. A phase 1 clinical trial of nerve growth factor gene therapy for Alzheimer disease. Nat Med 2005;11(5):551-55. Epub 2005 Apr 24.


 Goldman M, Johannesson H, Axelsson O, Karlsson M. Hyperpolarization of 13C through order transfer from parahydrogen: a new contrast agent for MRI. Magn Reson Imaging 2005;Feb;23(2):153-157.

Like an expanding bubble, the number of MR applications continues to rise exponentially. Looking back over the last 27 years, I see several major MRI epochs: low- to midfield systems (late 1970s to mid-1980s), 1.5T with 10 mT/m gradients (mid-1980s to mid-1990s), and 1.5T with echo-planar gradients (mid-1990s to early 2000s). We entered a new epoch a few years ago: 3T with echo-planar gradients. Examining changes currently occurring, and understanding why they occur, can help us predict further changes to come over the next decade.

As era of 3T MR takes hold, new opportunities emerge

Since the introduction of cardiac catheterization in the 1940s, development and implementation of cardiovascular imaging techniques have been a collaborative effort among several specialties, particularly radiology and cardiology. Many pioneers in CV imaging have held joint appointments.

Because cardiologists have predominantly performed coronary angiography and cardiac echocardiography in recent years, many radiologists have had only minimal training in these areas. Most cardiologists, on the other hand, have not been trained in the physics and technology of MR, CT, and PET imaging. With the advent of clinically applicable advanced CV imaging such as cardiovascular MR (CMR) and CT (CCT) and wider dissemination of cardiac PET, patients will benefit from the combined expertise of cardiologists and radiologists.

Achieving a productive working relationship between these two specialties is generally not easy. In many institutions, radiologists and cardiologists compete over who is best qualified to read advanced cardiac imaging. At the University of California, San Diego, we believe both groups bring essential expertise to this enterprise, and we have formed a joint venture to maximize benefit to the patient. While radiologists could certainly learn to read cardiac anatomy and pathology as well as cardiologists, and cardiologists could learn the complicated physics of MR imaging and the technology and radiation exposure associated with CT and PET, it is unlikely that all but a few will be fully conversant with both disciplines.

The authors have learned not only from each other but from our colleagues in the discipline in which we did not train. When we first worked together at Long Beach Memorial Medical Center, Dr. Bradley taught MR physics and image optimization to his fellows and Dr. Grover. As we performed and read cases together, Dr. Grover taught the fellows and Dr. Bradley how to optimize the study to answer the clinician's question and how to interpret cardiac anatomy, function, and physiology to help the clinician care for the patient.

Introducing new tests into practice patterns takes time, for multifactorial reasons. Clinicians need to see data that convince them the new tests will provide information as good as or better than the old tests. Hearing about the advantages of a new test from a colleague in their own field, rather than from someone who will derive economic benefit from new referrals, is likely to be more persuasive.

Even in academic institutions, economics plays a role. When Dr. Bradley moved to UCSD and realized that the institution did not have a CMR specialist or program, he met with Dr. Anthony DeMaria, the chief of cardiology, and proposed hiring Dr. Grover as codirector, with a cardiothoracic radiologist, of advanced cardiac imaging. Radiology and cardiology agreed to split Dr. Grover's billings, as all cases are read out jointly with both radiology and cardiology trainees in attendance. When Dr. Grover is in San Diego, we read from various cardiac imaging workstations. When she is out of town, we read and teach using the Internet and VitalConnect (Vital Images), which allows 4D cardiac imaging and uses an arrow that can be seen on both the local and remote monitor.

CMR at 1.5T is fast becoming a clinical reality. How cardiologists and radiologists partner to provide the highest quality studies depends on local factors. With 3T CMR, however, radiologists, physicists, and cardiologists will have to work closely to optimize clinically relevant imaging.

We recently installed three 3T MR scanners at UCSD and are beginning to explore applications of 3T imaging for CMR. While it would seem logical that both temporal and spatial resolution could be improved at higher field strength, the physics have proven challenging. For example, the wavelength of a radiowave at 128 MHz (resonance frequency at 3T) is about the same dimension as an adult chest. The dielectric effect produces standing waves that result in signal variation and difficulties in refocusing phase in coherent bright blood techniques such as steady-state free precession pulse sequences (e.g., FIESTA and TrueFISP). T1 prolongation causes contrast changes at 3T, and sequences that are dependent on T1 must be modified. An example is delayed enhancement in which the myocardium is nulled based on its T1 value.

In this initial period of 3T cardiac imaging, people with expertise in MR physics and pulse sequences (usually physicists and radiologists) will likely play a larger role. Once the protocols and artifacts are worked out, people with knowledge of cardiovascular physiology and patient care will probably assume a larger role. The synergy of the partners in the joint venture may not lead to comparable involvement at all stages of new technology development and evolution.

CCT provides a good example of how collaboration improves test performance and interpretation. Radiologists can design protocols that optimize temporal and spatial resolution while minimizing radiation. Cardiologists can make decisions about and administer beta blockers to regularize and slow the heart rate. Cardiologists and radiologists can then interpret the cardiovascular and noncardiovascular findings and artifacts of the studies together.

When the UCSD Center for Molecular Imaging installed a PET scanner and cyclotron, our primary focus, like many other PET centers, was initially cancer detection. But myocardial stress perfusion imaging is an emerging area in which collaboration between radiologists and cardiologists is advantageous.

A myocardial perfusion study involves stress and rest scans. Our cyclotron can produce short-lived isotopes such as nitrogen-13 to make N-13 ammonia for myocardial perfusion imaging. Because N-13 is short-lived, it must be injected onsite.

Stress is achieved by an intravenous injection of the coronary vasodilator adenosine. Electrocardiograms are monitored before and during the stress study, and a crash cart is available. While a few cardiac radiologists are comfortable administering adenosine, interpreting the electrocardiographic response, and treating problems that may arise, most radiologists prefer that someone accustomed to these tasks perform them.

Our radiology-cardiology joint venture enables us to increase the types of studies that can be done in our PET facility, to our mutual benefit as well as that of patients.

The joint venture also provides cross-fertilization of ideas, leading to new areas of research that neither party might be able to persue individually. At UCSD, Dr. Graeme Bydder, known for his development of FLAIR and STIR, has been applying ultrashort TE (UTE) MR imaging to musculoskeletal, hepatic, and central nervous system structures and abnormalities with short T2s. Working with cardiologists like Dr. Grover, Dr. Bydder and his group hope to use UTE for imaging of calcification, which should allow them to differentiate stable, calcified plaque from noncalcified, and perhaps vulnerable, plaque.

Setting aside our turf battles and Manhattan Projects and working together fosters complementary expertise that can potentially maximize patient care. When we adopt this frame of mind and forget about who gets paid for what, collaboration becomes a powerful tool.

DR. BRADLEY is chair of radiology, and DR. GROVER-MCKAY is director of advanced cardiovascular imaging, both at the University of California, San Diego. Dr. Bradley has received grants/research support from GE Healthcare.

Convert turf battles into productive joint ventures

Achieving a productive working relationship between these two specialties is generally not easy. In many institutions, radiologists and cardiologists compete over who is best qualified to read advanced cardiac imaging. At UCSD, we believe both groups bring essential expertise to this enterprise, and we have formed a joint venture to maximize benefit to the patient. While radiologists could certainly learn to read cardiac anatomy and pathology as well as cardiologists, and cardiologists could learn the complicated physics of MR imaging and the technology and radiation exposure associated with CT and PET, it is unlikely that all but a few will be fully conversant with both disciplines until a truly integrated training program is developed.

Stress is achieved by an intravenous injection of the coronary vasodilator adenosine. Electrocardiograms (ECGs) are monitored before and during the stress study, and a crash cart is available. While a few cardiac radiologists are comfortable administering adenosine, interpreting the electrocardiographic response, and treating problems that may arise, most radiologists prefer that someone accustomed to these tasks perform them.

The joint venture also provides cross- fertilization of ideas, leading to new areas of research that neither party might be able to persue individually. At UCSD, Dr. Graeme Bydder, known for his development of FLAIR and STIR, has been applying ultrashort TE (UTE) MR imaging to musculoskeletal, hepatic, and central nervous system structures and abnormalities with short T2s. Working with cardiologists like Dr. Grover, Dr. Bydder and his group hope to use UTE for imaging of calcification, which should allow them to differentiate stable, calcified plaque from noncalcified, and perhaps vulnerable, plaque.

Dr. Bradley is a professor and chair of radiology, and Dr. Grover-McKay is director of advanced cardiovascular imaging, at the University of California, San Diego. Dr. Bradley has received grants/research support from GE Healthcare.

Convert imaging turf battles into productive joint ventures

During the two decades in which MR has been used clinically, progress has been more sporadic than steady. If one were to plot it out, MR advances would be represented by a series of steep climbs followed by plateaus, rather than by a slow, steady climb. The steep parts include the transition from resistive to superconducting magnets in the early 1980s, the transition from low-field superconducting magnets to 1.5T superconducting systems in the mid-1980s, the introduction of echo-planar systems in the mid-1990s, and the introduction of "cardiovascular systems" with even stronger, faster gradients in the late '90s.

It appears that we are now approaching another steep part of the curve, due to the synergy provided by the greater 3T field strength of the latest commercial MRI systems and to "parallel imaging," known more commonly as SENSE (Philips), ASSET (GE), and iPAT (Siemens). With apologies to GE and Siemens, this technique will henceforth be called SENSE, because Philips invented it. SENSE is a technique that trades signal-to-noise for speed. Just as S/N is reduced by the square root of 2 (40%) whenever the number of excitations is halved in a conventional spin-echo image, the same is true of SENSE: A SENSE factor of 2 halves the acquisition time and reduces the S/N by 40%. SENSE is therefore most useful when there is excess S/N, such as at 3T. It requires the use of phased-array coils, and the number of coil elements determines the theoretical limit on the SENSE factor. SENSE works by intentionally reducing the field-of-view and the number of phase-encoding steps, which in turn reduces the acquisition time. This would normally lead to wraparound artifact, or "aliasing," but because the local sensitivity of each coil in the phased array is known, the aliasing can be "unwrapped."

Phased-array coils also have the advantage of higher S/N than standard quadrature RF coils. The greater S/N of these coils and the higher field strength can be traded for speed or higher spatial resolution at the same acquisition time. This leads us to 1024 MR angiography.

In the circle of Willis, 1024 MRA has been a Holy Grail for neuroradiologists. Earlier MR systems either did not have the computer power to acquire a 1024 x 1024 image or didn't have the necessary S/N. Attempts to boost S/N with gadolinium were complicated by venous enhancement. As the acquisition time for the gradient-echo techniques used for 3D time-of-flight MRA is TR x Np x Ns, increasing the number of phase-encode steps (Np) to 1024 would either lead to an excessively long acquisition time or mandate a reduction in TR. At some point, a very low TR limits the time available for inflow of unsaturated blood and reduces flow-related enhancement, which is the basis for TOF MRA.

Increasing the matrix size from 512 to 1024 along one axis decreases S/N by 50% (assuming FOV is held constant). Halving pixel dimension along both the phase and frequency axes reduces S/N to 25% of what it was (holding acquisition time and slice thickness constant). Since the eye, like most organs, responds logarithmically to physiologic stimuli, a loss of 75% of the original S/N will be most noticeable at low S/N levels. The added S/N afforded by 3T magnets and the latest generation of eight-channel head coils allows this drop to go essentially unnoticed. With a rectangular FOV, a smaller number of phase steps can be used to achieve the same spatial resolution, reducing the acquisition time.

Figures 1, 2, and 3 show 1024 MRA of the circle of Willis. The images were acquired on a short-bore 3T GE TwinSpeed EXCITE with an eight-channel head coil, using a 1024 x 608 acquisition over a 16 x 12-cm FOV, yielding pixels of 160 x 200 microns. This compares with the typical 1024 digital subtraction angiogram, which has an in-plane spatial resolution of 250 microns, and a typical CT angiogram, which has spatial resolution on the order of 500 microns.

The slice thickness of the MRA is 0.8 mm (800 microns), which is zero interpolated to 400 microns. It was based on a spoiled GRASS sequence with a TR of 31 msec and TE of 6 (fractional echo). Although it took 14 minutes and 12 seconds to acquire, a SENSE factor of 2 would have brought the time down to seven minutes and six seconds.

Figure 1 is a full maximum intensity projection image of the data set in the transaxial plane. All the arteries of the circle of Willis appear larger than in lower resolution MRA. This is due to the fact that intravoxel dephasing has been minimized at the periphery of the vessel where the velocity changes (and the phase dispersion) are greatest.

This level of detail clearly will allow diagnosis of spasm from subarachnoid hemorrhage, vasculitis, and intracranial atherosclerosis, greatly extending the clinical applications of MRA in the brain. Although the spatial resolution of MRA can surpass that of DSA, detection of abnormalities is a function not only of spatial resolution but also of S/N. DSA also has the advantage of temporal information.

To use a technique like TRICKS to obtain temporal phases from an MR angiogram, contrast would have to be used and 1024 spatial resolution would probably need to be reduced to get clinically reasonable spatial resolution.

Still, the initial results are impressive. Figures 1 and 2 (MIP in the sagittal plane) show the ophthalmic and anterior choroidal arteries. Figure 3 is a MIP of this data that is set in the coronal plane. It shows the lenticulostriate arteries penetrating several centimeters into the basal ganglia. As parallel imaging and phased-array technology continue to advance, the results are only going to get better.

Dr. Bradley is a professor and chair of radiology at University of California, San Diego Healthcare in San Diego.

William G. Bradley Jr., MD, PhD

Articles

CT perfusion for stroke: Should you use it?

Contrast FLAIR tops FLAIR and contrast T1-weighted images

Contrast FLAIR tops FLAIR and contrast T1-weighted images

As era of 3T MR takes hold, new opportunities emerge

Convert turf battles into productive joint ventures

Convert imaging turf battles into productive joint ventures

Use of 1024 MRA at 3T enhances neuro scans