James Provenzale, MD
Medscape coverage of: American Roentgen Ray Society 103rd Annual Meeting
The 103rd Annual Meeting of the American Roentgen Ray Society (ARRS), May 4-9, San Diego, California, showcased neuroradiologic topics in 3 major venues: scientific presentations and posters, instructional courses, and a 2-day review course that provided a broad review of important clinical topics. The scientific presentations covered a wide variety of topics, including head and neck radiology, cerebral angiography, and advanced MRI techniques.
Rise of Novel Contrast Agents
One topic that is currently of great interest to neuroradiologists and was covered in the meeting is the utility of novel contrast agents to better depict CNS disease processes, such as neoplasms and cerebral infarction. For instance, in the field of tumor imaging, there is a need for high-molecular-weight contrast agents that do not leak across an interrupted blood-brain barrier and can serve as blood-pool imaging agents. Another area of research lies in the development of contrast agents that are specific for tumors but are not taken up by normal tissue, which offers a high contrast-to-background ratio. An analogous method is the use of contrast agents that are taken up by normal tissues but not by tumors. Both developments could become milestones in the utility of contrast agents to improve diagnostic sensitivity and specificity.
With this issue in mind, one of the more interesting areas of research in the field of head and neck imaging discussed at ARRS 2003 was presented by S.C. Rankin, MD, Guy's Hospital, London. England. Dr. Rankin and colleagues examined the effects of a novel contrast agent in distinguishing between benign and metastatic cervical lymph nodes in patients who had head and neck cancer. One of the well-known imaging pitfalls of CT and MRI in the diagnosis of metastatic spread of head and neck cancers to cervical lymph nodes is the primary reliance on size criteria, which results in limitations in imaging sensitivity and specificity.
For instance, sensitivity is lowered in this setting because small lymph nodes frequently harbor metastases; ie, they are negative by imaging criteria but positive by pathologic criteria. The use of superparamagnetic iron-oxide (SPIO) particles as a contrast agent is an attempt to overcome these limitations. SPIO particles are taken up in substantial quantities by normal lymph nodes but not by lymph nodes that are infiltrated by tumor. Because of the paramagnetic effects of such particles, normal lymph nodes exhibit a marked decrease in signal intensity on T2*-weighted images. However, lymph nodes infiltrated by tumor do not exhibit such a decrease in signal intensity because of the low or absent uptake of SPIO. In the study by Dr. Rankin and colleagues, 18 patients underwent SPIO imaging prior to neck dissection, and 739 lymph nodes underwent pathologic evaluation. Twelve patients had positive neck dissections and 7 patients had negative dissections. SPIO imaging correctly predicted nodal metastases in 10 patients (83% sensitivity) and predicted the absence of metastases in 7 patients (100% specificity). A total of 54 lynch nodes were found to be infiltrated at pathologic evaluation. Although SPIO-enhanced MRI detected only half of these lesions, 62% of the false-negative lymph nodes detected by imaging were micrometastases in which tumor size was < 3 mm. The consensus of the discussion surrounding this presentation was that future work in a larger series of patients is necessary for this promising form of imaging to be validated, with a concentration on a reduction in the number of false-positive lymph nodes.
Diffusion-Weighted MRI: Expanded Utility
Over the past few years, diffusion-weighted MRI has evolved from a technique that is primarily used to evaluate acute cerebral infarction to one that is employed to study a variety of central nervous system (CNS) diseases. For instance, diffusion-weighted imaging has been used to study rates of water motion within regions of white matter that are affected by cranial trauma (so-called diffuse axonal injury), tumors (which typically have elevated water diffusibility, except for tumors that have high cellularity, such as lymphoma), and abscesses (which usually show restricted water diffusibility). A number of scientific presentations and lectures at ARRS 2003 explored new ways in which diffusion-weighted imaging can enhance clinical and research practice.
D.L. Camacho, MD, and colleagues from the University of North Carolina, Chapel Hill, reported on the use of diffusion-weighted MRI to distinguish between toxoplasmosis abscess and primary CNS lymphoma in HIV-positive patients. The distinction between these 2 entities is problematic using conventional MRI techniques. In the past, clinicians have either depended on the effects of a trial period of antibiotic therapy (with shrinkage of lesions serving as presumptive evidence of toxoplasmosis infection) or an additional nuclear medicine imaging study, such as positron emission tomography, to clinch the diagnosis. In this retrospective study, investigators assessed whether diffusion imaging could distinguish toxoplasmosis from lymphoma in patients with HIV. The study population consisted of 13 patients who had toxoplasmosis abscesses and 8 patients who had primary CNS lymphoma. Apparent diffusion coefficient (ADC) values were measured within lesions and expressed as a ratio of values within lesions compared with values within normal white matter in the brain. Mean ADC ratios within toxoplasmosis abscesses (1.63) were significantly higher than ratios seen within lymphoma lesions (1.14). However, considerable overlap was observed between the distribution of ratios, which limited the feasibility of discriminating between the 2 types of lesions based on ADC ratios. However, it appears that very high ADC ratios may be predictive of toxoplasmosis.
Mark Mullins, MD, PhD, Massachusetts General Hospital, Boston, Massachusetts, reported on the performance profiles of diffusion-weighted imaging in the assessment of the acute stroke patient. The study population consisted of 122 patients with an admitting diagnosis of acute stroke in whom diffusion-weighted imaging was performed within 24 hours after a well-documented ictus. This study, which was an extension of multiple previous studies, showed that the sensitivity, specificity, accuracy, and positive predictive value of diffusion-weighted imaging were all within the 90% to 100% range, but that the negative predictive value of the study was more on the order of 55%. This study validates the belief that diffusion-weighted MRI is an important technique for the accurate assessment of the acute stroke patient. Nonetheless, a small subset of patients exists in whom a false-negative MRI diffusion imaging study will be seen.
The previously mentioned studies used isotropic (3-direction) diffusion-weighted technique, which facilitates the measurement of the rate of microscopic water diffusion. However, tensor diffusion-weighted imaging is an MRI technique that employs diffusion gradients in at least 6 orthogonal directions and allows calculation of the diffusion tensor. This technique aids in the calculation of both the rate and directionality of microscopic water diffusion. Tensor diffusion imaging measures the tendency for water motion to occur in 1 direction as opposed to all other directions (ie, anisotropy) and, therefore, offers the possibility of depicting white matter tracts.
A number of the presentations in the 2-day Neuroradiology Review course offered insights into the means by which tensor diffusion imaging may eventually change clinical practice patterns. P. Ellen Grant, MD, Massachusetts General Hospital, Boston, outlined the work being undertaken at that institution in the areas of evaluating congenital CNS disorders and pediatric white matter disorders. In both of these presentations, the utility of measuring anisotropy values to monitor development of white matter tracts was highlighted.
In another study on the imaging evaluation of adult white matter disorders, presented by the author of this review, the role of tensor diffusion imaging in assessment of multiple sclerosis was explored. This study demonstrated that regions of normal-appearing white matter that are adjacent to plaques are, in many cases, markedly decreased compared with white matter regions in subjects who do not have multiple sclerosis. In fact, one can redefine the size of multiple sclerosis plaques by including the region that is abnormal on T2-weighted images and the adjacent region that has the same degree of anisotropy decrease as the plaque. The resultant plaque as defined on tensor diffusion-weighted images is often on the order of 250% of the size of the abnormality seen on T2-weighted images. These findings indicate that the conventional method of assessing lesion burden on T2-weighted images substantially underestimates the degree of white matter involvement seen in multiple sclerosis.
Focus on General Neuroradiology
W.K. Erly, MD, University of Arizona, Tucson, presented a study that explored the rates of agreement between readings of emergency head CT scans that were interpreted by general radiologists and neuroradiologists in an academic setting. A total of 716 consecutive CT scans were interpreted by 1 of 15 general radiologists using teleradiology during off-hour coverage, and a preliminary report was rendered. Scans were reinterpreted the next day by 1 of 5 neuroradiologists at a university-based medical center; the degree of agreement between the preliminary and final interpretations was compared using the following choices: agreement, insignificant disagreement, and significant disagreement. Agreement was found in 95% of cases, insignificant disagreement in 3%, and significant disagreements in 2% (16 cases, of which 11 were false-negative and 5 were false-positive). Forty-seven cases exhibited important or active pathology, and 11 of the disagreements were seen in these cases (of which 3 cases were false-negative interpretations of CT scans that showed pituitary masses). Although the rates of disagreement were relatively low, most of the disagreements were seen in the abnormal cases.
The ARRS 2003 meeting offered practitioners a window on the future of
neuroradiology, as well as a comprehensive review of key issues in this
subspecialty of imaging. The refinement of techniques outlined in the scientific
sessions and review courses is likely to reach widespread use in the community
practice setting over the next couple of years and will advance the utility
of neuroimaging in the management of disease entities, such as stroke,
brain tumors, and head and neck cancers.
Copyright © 2003, Medscape