Brain MRI volumetry in a single patient with mild traumatic brain injury

Abstract

This letter to the editor describes the case of a 42 year old man with mild traumatic brain injury and multiple neuropsychiatric symptoms which persisted for a few years after the injury. Initial CT scans and MRI scans of the brain showed no signs of atrophy. Brain volume was measured using NeuroQuant®, an FDA-approved, commercially available software method. Volumetric cross-sectional (one point in time) analysis also showed no atrophy. However, volumetric longitudinal (two points in time) analysis showed progressive atrophy in several brain regions. This case illustrated in a single patient the principle discovered in multiple previous group studies, namely that the longitudinal design is more powerful than the cross-sectional design for finding atrophy in patients with traumatic brain injury.

To the Editor: This report describes the case of a 42 year-old man who was in a motor vehicle accident in May 2008, leading to mild traumatic brain injury (TBI) and other injuries. His magnetic resonance imaging (MRI) brain data were analysed with NeuroQuant® (an FDA-approved computer-automated method for measuring MRI brain volume) and susceptibility-weighted imaging (SWI) to assess for signs of old bleeding.

Background

Until recently, perhaps 5 years ago, most patients with mild TBI had little or no objective signs of injury. This situation has improved with the development of better techniques for assessing brain injury, including techniques for measuring brain volume and detecting signs of old bleeding. This case exemplifies the use of these techniques in a patient with mild TBI.

History

Prior to the accident, since age 20, the patient had obsessive compulsive disorder which was treated effectively with sertraline. He worked full-time as a professor at a local university. Otherwise he had no pre-accident neurological or psychiatric disorders.

In May 2008, he had slowed down for a yellow light when a pickup truck hit his small sedan from behind. He hit his head, incurring an abrasion and a laceration to the occipital scalp. He had witnessed loss of consciousness lasting a few minutes. He reported post-traumatic amnesia lasting ∼90 minutes. His Glasgow Coma Scale score was 14 out of 15. Based on these data, he satisfied the diagnostic criteria for traumatic brain injury [1] and the severity of his TBI was mild.

In the months following the accident, he experienced multiple persistent neuropsychiatric symptoms, including impaired concentration, distractibility, bradyphrenia, irritability, sleep impairment and fatigue. These symptoms proved to be debilitating to his work and social life.

Despite his classical clinical picture for TBI and post-traumatic symptoms, he initially had no objective findings of injury. A CT scan was normal. In December 2010, an MRI/SWI sequence showed signal loss in the right parietal and left frontal lobes, indicating hemosiderin deposition due to prior haemorrhage. The attending radiologist observed no signs of atrophy.

MRI scans performed in April 2011 and May 2012 were interpreted by the attending radiologist as having no change from the December 2010 MRI. These two MRIs were analysed using NeuroQuant®. For each MRI, the volume was measured of 11 brain regions (forebrain parenchyma, cortical gray matter, lateral ventricles, inferior lateral ventricle, hippocampus, amygdala, caudate, putamen, pallidum, thalamus and cerebellum) using previously published methods [2]. An additional four brain regions were included (whole brain parenchymal volume total CSF, cerebral white matter and brainstem) for analysis based on their high test–re-test reliability [2] and theoretical interest. Atrophy was defined as parenchymal volume less than the fifth normative percentile and ventricular enlargement was defined as volume greater than the 95th normative percentile. The results of the cross-sectional analyses on these two MRIs revealed no regions consistent with atrophy or ventricular enlargement.

Next, a longitudinal analysis was performed using previously published methods [3]. This analysis was conducted by subtracting, for each region, the volume at scan 2 from the volume at scan 1, dividing the result by the volume at scan 1 and expressing the resulting proportion as a percentage change; the annual rate of volume change was calculated by dividing the percentage change by the duration between scans (measured in years). Significant progressive atrophy was defined as volume loss below the fifth normative percentile and significant ventricular enlargement was defined as volume increase above the 95th normative percentile. The longitudinal analysis revealed progressive atrophy in the right forebrain parenchyma (volume change = −1.7% per year, 3.8 normative percentile) and right cerebral white matter (volume change = −3.0% per year, 4.6 normative percentile). Also, there was progressive enlargement of the left lateral ventricle (volume change = 10.1%, 99.3 normative percentile) and left total CSF region (volume change = 8.7%, 98.8 normative percentile).

Discussion

Neither the CT scan performed on the date of the accident nor the NeuroQuant® cross-sectional volumetric analyses showed abnormalities. However, the longitudinal volumetric analysis revealed significant atrophy of parenchymal regions and significant increase in CSF regions consistent with atrophy of the surrounding parenchyma. In other words, if the longitudinal analysis had not been done, no signs of atrophy would have been detected in this patient. This case exemplifies the principle discovered in previous group studies (for a review see [4], for a more recent study see [3]) that the longitudinal design is more powerful than the cross-sectional design for detecting atrophy in patients with TBI. To the authors’ knowledge, there is only one other case report regarding the use of a commercially available software program (NeuroQuant®) to assess atrophy in individual TBI patients [5]. This case report extends that literature because the previous study was limited to a cross-sectional design.

The patient had a pre-accident history of obsessive compulsive disorder (OCD), which has been associated with brain structural abnormalities. However, the pattern of brain regions commonly affected in patients with OCD (the orbitofrontal cortex, the anterior cingulate cortex, the basal ganglia and the thalamus) did not match the pattern found in this patient. Furthermore, the persistent and relatively rapid rate of volume decrease (e.g. 3% per year in the right cerebral white matter) was much more consistent with the effects of recent trauma than with the effects of OCD which had existed for 22 years. For example, if his white matter had decreased at a rate of 3% per year for 22 years, he would have lost most of his white matter and, if he survived, his brain structure would have been severely abnormal on post-accident testing, which it was not.

The finding of signs of old bleeding on SWI added important objective confirmatory evidence that this patient suffered a TBI and reduced the chance that the progressive atrophy found was due to some non-accident-related source (including OCD). However, the signs of old bleeding did not indicate an active brain problem, because the bleeding had long since resolved; in contrast, the progressive atrophy did indicate an active problem, which was concerning, in part because progressive atrophy has been found to correlate with poor outcome, including decreased ability to return to work or normal family relationships [3],[4]. Furthermore, it was particularly concerning that this patient with mild TBI showed progressive atrophy a relatively long time (4 years) after the accident; a previous group study of patients with mild TBI similarly found progressive atrophy up to a few years after the accident. This finding again raises questions, such as when does atrophy stop progressing in patients with TBI?

Declaration of interest: The authors report no conflicts of interest, including no financial interests or investments with CorTechs Labs®, maker of NeuroQuant®.