Earlier this year, news circulated through typical press release channels regarding U.S. Food and Drug Administration (FDA) approval of a rapid, handheld traumatic brain injury (TBI) blood test. These press releases conveyed the advancement as “revolutionary” and “a game changer.”

Does the substance match the hype?

The test in question is the Banyan Biomarkers’ Brain Trauma Indicator (BTI) running on the Abbott i-STAT Alinity platform. This BTI panel measures two biomarkers of cerebral injury: ubiquitin C-terminal hydrolase-L1 (UCH-L1) and glial fibrillary acidic protein (GFAP). GFAP is a cytoskeletal protein belonging to the class of intermediate filaments specific for astrocytes. UCH-L1 is a 24-kDa protein component of the ubiquitinylation pathway with specific expression in cerebral neuron cell bodies. GFAP peaks approximately 20 hours after injury and declines over 72 hours, while UCH-L1 peaks eight hours after injury and declines rapidly within 48 hours. These tests, when combined, produce area-under-the-curve estimates for predicting traumatic intracranial injury ranging between 0.72 and 0.93 within the first 12 hours from injury.¹ The BTI implements this combination at predetermined cutoff values with the intention of predicting traumatic intracranial injuries on CT scan, and it is this specific use that the FDA approved.²

The Data

The key trial reviewed by the FDA was ALERT-TBI, a multicenter trial sponsored on multiple levels by Banyan Biomarkers.³ In this prospective trial, which recruited between 2012 and 2014 and published in 2018, the study team enrolled 1,977 adults with nonpenetrating traumatic brain injury (TBI) and a Glasgow Coma Scale (GCS) score of 9–15. The average age of those enrolled was 49 years, and the bulk of these (93 percent) had a documented GCS of 15 on presentation. There were 42 percent with loss of consciousness, 33 percent with posttraumatic amnesia, and 21 percent intoxicated with drugs or alcohol. Of the 1,959 included in their final analysis, 125 had a traumatic injury identified on neuroimaging, with eight requiring neurosurgical intervention.

The topline results touted by the authors and spokespeople for the manufacturer are sensitivity and its cousin, negative predictive value. Of the 125 patients with traumatic injury identified on neuroimaging, three were missed by the BTI panel.

This provided a sensitivity of 97.6 percent, with the lower limit of the 95 percent confidence interval down to 93.1 percent. With 671 patients showing a negative BTI result, the negative predictive value was therefore 99.6 percent, with the lower confidence interval bound at 98.7 percent. The authors’ publication concluded with the suggestion that these results support the “role of this biomarker test for ruling out the need for a head CT scan among patients with TBI presenting at emergency departments in whom a head CT is felt to be clinically indicated.”

Unfortunately, this conclusion is best described as aspirational. There is no disputing these biomarkers rise and fall in relation to TBI, but the key missing feature for generalizing these data into practice is the inclusion criteria for the study. Patients became eligible for inclusion if CT of the head was planned for their emergency department evaluation, but the eligibility criteria did not provide definitions regarding appropriateness of that CT. Simply stated, it is a straw-man comparator to imply value for imaging reduction when the imaging being performed was of uncertain necessity.

Simple clinical decision instruments, such as the Canadian CT Head Rule, are readily available and cost nothing, in contrast to the BTI.⁴ Despite the introduction of this rule nearly 20 years ago, in community practice, it is estimated that unnecessary imaging occurs in one-third of ED visits for minor head injury.⁵ So perhaps before we clamor for a blood test, we ought to simply realize the preexisting low-hanging fruit for imaging reduction already available to us.

The other unfortunate aspect of these results is that the panel functions as a one-way exclusion instrument only. In ALERT-TBI, the BTI demonstrated a specificity estimate of 36.4 percent, leading to a positive predictive value of merely 9.5 percent. These test characteristics are hardly dissimilar to our richly flawed experience with D-dimer over the past few decades. As we’ve found time and time again, inappropriately selecting patients for evaluation with D-dimer causes a downstream increase in advanced imaging rather than a decrease. A negative BTI result may potentially prove helpful, but a positive result does not. Two-thirds of those enrolled in ALERT-TBI showed positive BTI results, as do almost a fifth of normal healthy controls in the general population. If the BTI assay were to be as widely deployed as its proponents might wish, the overall result would be more imaging, not less. This is indication creep.

Lastly, there remain false negatives with the BTI, with the upper bound of the 95 percent confidence interval reaching up to 7 percent of traumatic injuries. Because of this, in part, the FDA has required compliant labeling to explicitly state, “A negative result is generally associated with the absence of acute intracranial lesions. An appropriate neuroimaging method is required for diagnosis of acute intracranial lesions.” None of the missed cases ultimately required neurosurgical intervention, but this sample simply isn’t large enough to produce reliable estimates of the overall safety of this product.