How Often Are MRI Abnormalities Found in Asymptomatic People?

Abstract & Commenatry

By Matthew E. Fink, MD, Vice Chairman, Professor of Clinical Neurology, Weill Medical College, Chief of Division of Stroke and Critical Care Neurology, NewYork-Presbyterian Hospital. Dr. Fink reports no financial relationship relevant to this field of study.

Synopsis: In an adult population older than age 45, brain MRI reveled that 1.8% had asymptomatic aneurysms, 1.6% had benign tumors, and 7.2% had silent brain infarcts.

Source: Vernooij MW, Ikram MA, Tanghe HL, et al. Incidental findings on brain MRI in the general population. N Engl J Med 2007;357:1821-1828.

The use of brain MRI by general physicians has increased dramatically for a variety of non-specific complaints such as headaches, dizziness, and minor head trauma. As a result, neurologists frequently are consulted to explain unexpected abnormalities that are found on brain imaging studies, and these findings often result in considerable anxiety and worry by referring physicians and their patients. Therefore, the study by Vernooij and colleagues from Rotterdam, the Netherlands, provides us with important information regarding the background frequency of brain pathology in an asymptomatic general population.

The subjects included 2000 people (mean age, 63.3 years; range, 45.7-96.7; 52.4% were women) from a population-based Rotterdam study that was designed to look at healthy aging. All scans were obtained with a 1.5-T GE scanner with an eight-channel head coil. The MRI protocol included four axial sequences: 3-D, T1- weighted sequence; 2-D proton-weighted sequence; 2-D FLAIR sequence; and 3-D T2-weighted gradient-echo (GRE) sequence. Slice thickness was 1.6 mm for all sequences except the FLAIR, which was 2.5 mm. Magnetic resonance arteriography (MRA) was not performed. Two trained reviewers, a resident in radiology and a resident in neurology, interpreted all scans. The reviewers were unaware of any clinical information about the subjects. Two experienced neuroradiologists reviewed the reported abnormalities, as well as an additional random sample of 230 scans to ensure accuracy and reliability of the interpretations.

Asymptomatic brain infarcts were present in 145 persons (7.2%); 112 were diagnosed as lacunar infarcts and 41 as cortical infarcts. The asymptomatic brain infarcts were found at increasing frequency with increasing age as evidenced by the following: 45 to 59 years (4.0%); 60 to 74 years (6.8%); and 75 to 97 years (18.3%).

The study identified 35 aneurysms (1.8%) and 31 (1.6%) primary benign tumors (18 meningiomas, 4 vestibular schwannomas, 2 lipomas, 1 trigeminal schwannoma, and 6 pituitary adenomas). There was one possible malignant tumor (low-grade glioma). Other findings included 7 cavernous angiomas, 1 metastases, 1 subdural hematoma, 22 arachnoid cysts, 18 Type-I Chiari malformations, 9 major vessel stenoses, 1 dermoid cyst, and 1 fibrous dysplasia. Regarding the brain aneurysms, all except 2 were in the anterior circulation and all except 3 were less than 7 mm in diameter. Four aneurysms were intracavernous. None of the persons with incidental brain findings reported any symptoms, except for 2 with hearing loss (vestibular schwannoma and transvestibular lipoma). The only urgent finding was a large, chronic subdural hematoma in an asymptomatic person who, in retrospect, had a minor head injury 4 weeks before the MRI scan was performed. The only other person who was referred for treatment had a 12 mm middle cerebral artery aneurysm.

Commentary

It has been the common experience of neurologists that incidental findings on MRI that have no relation to a patient's complaints are common, and this study bears that out. The frequency of incidental small aneurysms and benign tumors is consistent with other literature, and does not raise our concern. Only one aneurysm was large enough (12 mm) to warrant referral for possible treatment, and none of the benign tumors required treatment. It is possible that the addition of MRA to these studies may have identified some additional small aneurysms, but they would not have been clinically significant. The asymptomatic subdural hematoma occurred in a setting of minor head trauma, and the brain metastases occurred in a person with known lung cancer.

The authors' report of asymptomatic brain infarcts in 7.2% of persons is a more difficult issue to address. We were not given any information about risk factors for vascular disease in this population, and it would be important to correlate the MRI findings with rates of hypertension, cardiac disease, smoking, etc. However, the authors have not explained why they diagnosed all of these lesions as infarcts. Are all parenchymatous lesions that have the signal characteristics of cerebrospinal fluid (CSF) infarcts? In adults older than age 45, how many of the white matter lesions are non-ischemic demyelinating lesions? The authors have followed an arbitrary rule used by many neuroradiologists that diagnose the cause of "white matter hyperintensities" according to age; if the patient is older than age 50, the lesions are "infarcts." We need more objective imaging criteria for differentiating white matter hyperintensities to better understand cause and effect, and to make accurate diagnoses of our patients. In a middle-aged asymptomatic person, I believe that we still do not know what these white matter lesions represent pathologically.