Neuroimages As Proof of Toxic Exposure Injuries: Not Always What They Appear To Be
(This paper was presented at the West Virginia State Bar's Annual Litigation Seminar, February 6, 2004)
Like many of you, I defend my clients against toxic exposure claims. In our practices, we are on the front lines of an important war in which the laws of evidence, designed to shield the jury from unreliable technical and scientific opinions, are pitted against an army of plaintiffs who are continuously probing the law’s defenses with the latest technological wizardry designed to arouse the senses and excite the jury’s passions. Perhaps nowhere do we take the war more seriously or wage it with more ferocity than over the use of neuroimagery to prove injury and causation in brain injury cases.
Brain injuries have always been one of the most expensive (or, for plaintiffs, lucrative) class of injuries in personal injury and products liability litigation. While the incentive to prevail on both sides has remained the same, the ways in which plaintiffs attempt to prove brain injuries are changing. In the past, plaintiffs’ alleged brain injuries were typically far more obvious than those we see today. The plaintiff had suffered a catastrophic accident or acute toxic exposure, and it was self-apparent that some degree of harm had occurred. More and more, though, the injuries that are claimed are far less obvious and require much more subtlety to detect than ever before. This new generation of claimed injuries has spawned such conditions as “mild traumatic brain injury” (MTBI), “post-concussive syndrome,” “multiple chemical sensitivity syndrome” (MCS), and “toxic encephalopathy.”
In order to prove their case, plaintiffs are often on the “bleeding edge” of technology, in techno-speak, seeking to become the first in line to use a new technique to illustrate their claims. In the world of brain injury litigation, plaintiffs proffer medical experts armed with the latest neuroimaging techniques, which have not yet stood the test of time and acceptance necessary to be reliable to support their claims. These experts attempt to use the fruits of this technology, the vibrant imagery, to depict brain activity and to pass it off as reliable, objective evidence of both brain injury and causation.
This use of neuroimagery to demonstrate cause and effect is, at best, premature. Before such evidence is permitted, there must be accepted scientific evaluation to establish supportable patterns of abnormalities associated with specific conditions and their causes. This requires reliable, controlled studies of patterns and conditions to establish a database of consistent information to substantiate the relationships. Currently, sufficient data simply does not exist to support the findings and conclusions offered by plaintiffs' experts on the issues of causation and cognitive deficit.
This is not to say that neuroimaging technologies are not valuable tools. To the contrary, there is no question they are useful medical tools with a broad range of applications. For example, it is well recognized that PET and SPECT scans can be used in the management of patients with stroke, epilepsy, brain tumors, and Alzheimer’s disease. Scan abnormalities have been recognized in patients with an array of conditions such as depression, schizophrenia, and cerebral vascular disease. These techniques certainly can be used to identify abnormalities of the brain. However, there is not, at this time, adequate evidence to support the confident use of PET and SPECT scans to establish causal relationships.
When plaintiffs attempt to take this technology from the examining room to the courtroom, the problem becomes clear. As Judge Posner has noted generally, is that “the courtroom is not the place for scientific guesswork, even of the inspired sort” for “law lags science; it does not lead it.” Rosen v. Ciba-Geigy Corp., 78 F.3d 316, 319 (7th Cir. 1996). Our rules of evidence, whether following Daubert, Frye or some other standard, reflect this maxim and demand that scientific or technical evidence be reliable before it is allowed to go to the jury for consideration.
This is the point on which our battle turns, for, until the science has developed to allow neuroimaging to reliably depict either causation or injury, the laws of evidence must lag technological advances and continue to exclude these highly persuasive and potentially prejudicial images.
As discussed below, the development of this technology and attempts to use it in litigation are ever-increasing. Appropriately, the courts have generally remained vigilant. In nearly every case, the courts have properly performed their gate-keeping role by excluding neuroimagery where the purposes for which it is offered are not yet supported by science.
II. Neuroimaging Technologies
There are currently five key neuroimaging technologies in use that are likely to be proffered as evidence in brain injury litigation. The most common of these are Positron Emission Tomography (PET) and Single Photon Emission Computerized Tomography (SPECT), which are used, respectively, to reflect the body’s metabolism and perfusion, or blood flow. Additionally, while not yet as widely-used in litigation, Magnetoencephalography (MEG), Electroencephalography (EEG) and Functional Magnetic Resonance Imaging (fMRI) are also becoming common neuroimaging modalities. MEG and EEG generally reflect electrical activity, while fMRI reflects the flow, volume and oxygenation of blood within bodily tissue. 
PET is most commonly used to reflect brain function or metabolism. The basic procedure for PET scans has remained virtually unchanged in recent years. In a PET scan of the brain, a patient is injected with glucose labeled with a positron-emitting radionuclide or “tracer.” The brain burns glucose as fuel, so the glucose travels to where the brain is working. The tracers are radioactive and give off gamma ray signals.
As the patient lies on a table passing through a circular gamma ray scanner, PET measures the amount of physiological function in terms of glucose metabolism, and a computer reassembles the signals into color-coded images. If those parts of the brain are healthy and/or functional, they will consume a large quantity of the radioactively tagged glucose and appear as a bright orange or red color. Those parts of the brain that are damaged or simply inactive at the time of the scan will absorb little, if any, glucose, and will show up blue or purple on the finished scan. Yellows and greens are between these extremes. The colors have no intrinsic significance, but are merely a way of creating visual contrast to distinguish varying levels of metabolic activity in various parts of the brain.
SPECT is similar in many ways to PET, but less precise. SPECT essentially measures blood flow. A radioactive chemical is administered intravenously to the patient, but the radioactive chemical remains in the bloodstream and does not enter the brain tissue. As a result, SPECT measures brain activity indirectly by gauging cerebral “perfusion” or the brain’s vascular supply. Because damaged brain tissue normally shuts down its own blood supply, focal vascular defects on a SPECT scan may be used by a plaintiff’s attorney as circumstantial evidence of brain damage. The advantages of SPECT are that they are readily available and relatively inexpensive as compared to PET scans (the cost of a PET study typically is between $2,000 and $4,000, depending on the complexity of the study, as compared to a cost of approximately $750 to $1,500 for a SPECT scan).
There is no question that the use of PET and SPECT in the medical community has soared within recent years. A 1997 survey, for example, found that there were 73 PET scanning sites in the United States, concentrated primarily in the East and Midwest. Today, there are nearly 250 PET scan centers spread across the nation. While the cost of a PET scanner (between $1 and $3 million, depending on application-specific accessories) usually means that only the larger medical centers can afford them, they can be trailer mounted and easily transported to smaller facilities. SPECT scanners still far outnumber PET scanners.
The clinical applications for PET and SPECT have continued to expand within the fields of neurology and psychiatry. PET and SPECT imaging is used to diagnose and treat epilepsy, dementias and movement disorders (such as Parkinson’s Disease, Huntington’s Disease and Tourette Syndrome). These technologies are used in a variety of brain research applications, as in, for example, pinpointing the area of the brain responsible for ADD/ADHD in children. As well, all of the major pharmaceutical manufacturers use them when applying for FDA approval of a new drug application, to show how the drug affects the metabolism and function of the human brain.
Another advance in the use of these technologies, particularly PET, has been the extension into the areas of cardiology and oncology. By measuring blood flow and metabolic rate within the heart, for example, PET scans can identify the areas of decreased blood flow caused by blockages and differentiate muscle damage from living muscle. Because PET scanning can also detect functional differences or changes in cancerous cells and observe structural changes which occur in the tissue, PET scanning has proved to be an invaluable tool in the early detection of cancer. Oncologists rely on PET scanning also to determine the spread of cancer, whether a tumor is responding to a given course of treatment, and if a patient is free of cancer following the treatment.
There continue to be improvements and refinements to the technologies. There are better cameras and better image processing software. Practitioners and researchers are refining the methodology for administering these tests and interpreting the results. New systems are emerging. For example, GE Medical Systems now manufactures a combined PET and CT scanner which shows the anatomical location from CT along with the metabolic activity of PET in a single study. Scientists are also beginning to tout ultrasound-based images as a cheaper, faster alternative to these other types of imaging. Ultrasound imaging has not yet played much of a role in brain injury litigation, but we should remain aware that it may soon find its way into the courtroom. Seemingly, no end is in sight for the medical miracles of these technologies.
III. Neuroimagery Under Daubert
Whether in the context of a toxic tort case or closed head injury, a plaintiff typically attempts to use neuroimagery both to prove the existence of a brain abnormality and to establish a causal link between the abnormality and the exposure or trauma. In a plaintiff’s ideal world, the images alone would establish proof of diagnosis – i.e., the plaintiff would use the scans to demonstrate that the abnormality resulted from whatever wrongful act spurred the lawsuit. The more cautious approach, and the one a plaintiff is most likely to take, is to have an expert opine that the abnormalities on brain scan are consistent with a diagnosis of toxic exposure or head trauma. Here, the plaintiff’s expert can appear to exclude all other potential causes and support his or her differential diagnosis with impressive high-tech imagery.
From the defense perspective, testimony about consistencies alone is unacceptable, especially given the extremely persuasive nature of neuroimagery and the effect it can have on the jury. As such, it is imperative that the courts screen this evidence through Daubert review (or whatever review is provided by the jurisdiction’s rules of evidence).
All too often in this day and age, a plaintiff’s lawyer whose expert faces exclusion on reliability grounds attempts to meet Daubert’s criteria with generalities. A plaintiff seeking to admit neuroimagery to prove toxic brain damage, for example, would have the judge ignore whether this technology has been tested, subjected to peer review and accepted as reliable in this specific context. Rather, the plaintiff might attempt to mask those shortcomings by highlighting all of the many medical advances in neuroimaging in recent years. For example, a plaintiff can show that PET has been an established scientific methodology within the relevant medical community for now well over a decade. An abundance of published medical and scientific literature now exists on the subject of all of the neuroimaging modalities, both directly and tangentially in connection with research into disease processes and their treatments.
This limited history is insufficient. In challenging this evidence, it is crucial to have the court focus on the specific purpose for which the plaintiff is seeking to introduce this evidence and to judge its admissibility in the proper context. None of the neuroimaging techniques have met the acceptable levels of methodology and criteria to diagnose cognitive dysfunction due to toxic exposure. There are no articles in the published medical literature that would support this use neuroimagery. It is important to highlight to the court the numerous ways in which this evidence does not and cannot meet the Daubert standard:
No standard exists for normal or abnormal. In order to attribute a specific abnormality to a specific exposure, there must be accepted scientific evaluation to establish supportable patterns of abnormalities associated with specific conditions and their causes. Not only must there be a “baseline” for normal, but there must be a database of consistent information from controlled studies showing patterns of conditions and their causes to allow a given neuroimage to be classified as “normal” or “abnormal.” No such database exists.
There is no reliable database of images acquired under the same protocol. Despite significant advances in the technology and underlying methodologies, there are significant inconsistencies from individual to individual; from scan to scan; from scanner to scanner; from interpreter to interpreter, etc. Many factors affect the quality of the image and prevent a standardized quantitative analysis on which to base any empirical study.
No theory or technique to identify specific disease processes and their causes has been identified or carried out. Again, the published scientific and medical literature to date has not advanced any theories or techniques (much less, tests or studies of any particular methodologies) that would support any conclusions regarding the effect on the brain of exposure to toxic substances. As there is no methodology in place to identify these disease processes, the potential rate of error is unknown.
No theory or technique has been subjected to peer review, nor has the scientific community accepted such a theory or technique.
Again, the scientific developments in neuroimaging and its medical use have been nothing short of breathtaking. However, those developments have been designed for use in the medical arena, not the courtroom. As such, science has not yet seen fit to study neuroimaging from an evidentiary or admissibility perspective. Simply put, the technological breakthroughs are designed to aid patients medically, not win cases. While medical science has made tremendous progress, it simply has not been in the direction of establishing cause and effect in toxic exposure cases. Until it reliably does so, neuroimagery cannot meet the Daubert standard.
IV. The Courts’ Handling of Neuroimaging Evidence
Perhaps surprisingly, there are not an overwhelming number of published decisions addressing the admissibility of neuroimaging evidence. Those that have addressed the issue have been limited almost exclusively to PET and SPECT scan imagery.
One of the earliest, and most notable, cases was Guilbeau v. W.W. Henry Co., 85 F.3d 1149 (5th Cir. 1996). The plaintiff in that case allegedly suffered from chronic toxic encephalopathy resulting from a short-term exposure to chemicals contained in a carpet adhesive. Plaintiff offered the testimony of Dr. Thomas Callender, who testified as an expert in internal medicine, neurotoxicology and occupational medicine. Dr. Callender’s testimony was based in large part on the results of a 1990 SPECT scan which purported to show abnormal areas of decreased circulation. Despite a rigorous cross-examination exposing the lack of foundation for his testimony and the unreliability of the SPECT scan evidence, the jury returned a verdict of $2.9 million. The verdict subsequently was overturned on the ground that the product was not defective, and, consequently, we have little indication of how the Fifth Circuit may have treated the admissibility of SPECT scan evidence. Nevertheless, it is worth noting that Judge Reynaldo Garza, in his dissenting opinion, expressed his belief that the plaintiff had carried his burden in proving causation through the use of SPECT scan evidence.
In another significant case, Hose v. Chicago Northwestern Transportation Co., 70 F.3d 968 (8th Cir. 1995), the trial court allowed PET scan evidence for the limited use of eliminating alternative theories of injury. Mr. Hose allegedly suffered toxic encephalopathy resulting from occupational exposure to manganese fumes. In his FELA action against his employer, the court permitted Mr. Hose’s expert to testify on the basis of a PET scan. Specifically, his expert testified that the PET scan did not match the typical patterns for Alzheimer’s, alcoholism or stroke, but, rather, showed a pattern consistent with dementia and an “overall picture. . .very consistent with manganese encephalopathy and toxicity as the cause of this dementia.” The judge allowed this testimony despite his admission that the scientific literature did not explain how manganese toxicity would appear on a PET scan. The jury returned a $1.33 million verdict. On appeal, the Eighth Circuit upheld the verdict and held that the trial court did not abuse its discretion in allowing the PET scan evidence. The Eighth Circuit’s opinion relied to a great extent on the fact that the expert’s testimony was limited to showing consistency with, as opposed to diagnostic proof of, manganese encephalopathy.
In another FELA action alleging multiple chemical sensitivity syndrome, however, a trial court excluded an expert’s testimony which relied heavily on SPECT images. Summers v. Missouri Pacific Railroad System, 897 F. Supp. 533 (E.D. Okla. 1995). The court in that case noted the lack of reliable scientific and medical data to support the use of SPECT technology to diagnose neurotoxic exposure.
Since these cases were decided, there has been only one reported opinion squarely addressing this issue in the context of a claimed toxic exposure. In Rhilinger v. Jancics, et al., 1998 WL 1182058 (Mass. Super. 1998), a Massachusetts trial court allowed a plaintiff’s expert to testify as to the results of a SPECT test to prove injuries consistent with toxic encephalopathy. Ms. Rhilinger alleged that she developed toxic solvent encephalopathy from exposure to chemicals stored illegally in the basement of her apartment building. She asserted that the chemicals leaked and caused a migration of fumes into her residence which ultimately caused brain damage. After she experienced various physical and cognitive symptoms including fatigue, dizziness, memory loss, hoarseness and stuttering in her speech, she was seen by a Dr. Howard Hu at the Center for Occupational and Environmental Medicine at the Massachusetts Respiratory Hospital. Dr. Hu diagnosed Ms. Rhilinger with toxic encephalopathy.
The defendants brought a Daubert challenge to the admissibility of the SPECT scan evidence on three grounds, namely, that the use of SPECT scan imaging to diagnose brain injury due to chemical exposure (1) is not supported by valid scientific evidence; (2) has not been subjected to empirical testing or otherwise properly tested or studied; and (3) has not been generally accepted as a diagnostic tool for toxic encephalopathy in the relevant scientific and medical communities.
The court included a thorough discussion of SPECT scans, and noted that the medical community had used SPECT technology for at least 15 years. The court acknowledged that, as of that time, there was inadequate evidence to support the use of SPECT scans to establish the existence of a cause and effect relationship in toxic exposure cases. Indeed, the plaintiff’s own doctor testified that a SPECT scan could not be relied upon as definitive evidence to rule in or to rule out a diagnosis of TSE. Nevertheless, the court relied on Dr. Hu’s testimony that: (1) “individuals who have a history of exposure to chemicals are at a higher risk of having abnormal SPECT scan results”; (2) Ms. Rhilinger’s SPECT scan results were consistent with the results of those individuals; (3) Ms. Rhilinger had no other identifiable disease process or diagnosis to account for her abnormal SPECT scan results; and, (4) her MRI was normal, thus ruling out other possible origins of her cognitive difficulties.
In holding that the SPECT evidence was admissible, the court stated that:
There is no dispute that SPECT scans show abnormalities in brain function. Neither is there a dispute that SPECT scans cannot conclusively establish the existence or non-existence of TSE in a patient. Plaintiff’s experts do not opine that the SPECT scan does, in fact, establish the diagnosis. They merely assert that is one of a constellation of diagnostic tools which they used and considered consistent with their conclusion that Rhilinger suffers from TSE.
There also is no dispute that SPECT scanning is relevant to prove or disprove the other possible explanations for plaintiff’s condition. The scientific evidence submitted by both parties approves of the use of SPECT scans to identify other brain disorders such as epilepsy, stroke and dementia. Plaintiff’s proffered expert testimony does not go beyond what is considered scientifically defensible use of SPECT scan technology.
The judge held that SPECT evidence is “scientifically reliable” in such a toxic tort case. Thus, as in the Hose case, the plaintiff had her SPECT evidence admitted in a round-about way by using it to show a diagnosis “consistent with” toxic exposure.
The debate over the admissibility of PET/SPECT evidence has also been addressed in a number of other contexts during the last few years. Penney v. Praxair, 116 F.3d 330 (8th Cir. 1997), for example, involved a claim of brain injuries following an automobile accident. The 62-year-old plaintiff allegedly suffered a whiplash-type injury and claimed symptoms including headaches, dizziness, vertigo, ringing ears, a sore neck and various other maladies. He underwent both an MRI and a CT scan, neither of which showed any brain injury. The plaintiff then had a PET scan which, according to his expert, revealed “abnormalities which were consistent with a traumatic brain injury.” 116 F.3d at 332. The district court refused to admit the PET scan evidence, reasoning that it would “not be helpful to the jury in deciding the issues when compared with the likelihood that the jury would misapply the evidence.” Id. On appeal, the Eighth Circuit upheld the district court ruling. The appellate court’s analysis is instructive on some of the problems with the numerous variables affecting PET scan evidence. Specifically, plaintiff’s expert failed to account for the plaintiff’s age and medication taken days before the procedure, and the expert could not explain how plaintiff’s abnormalities compared to a control group of individuals with similar characteristics.
In re Air Crash at Little Rock Arkansas, 291 F.3d 503 (8th Cir. 2002) involved a dispute over whether post-traumatic stress disorder is in itself a “physical injury” for which recovery under the Warsaw Convention was allowed. The Eighth Circuit rejected as speculative the testimony of plaintiff’s expert to the effect that “people with chronic PTSD may have brain dysfunction, meaning that PTSD is both biological and physical.” Id. at 511. While this case did not deal with the admissibility of PET or SPECT evidence, it bears mentioning because the Eighth Circuit noted in its opinion that the plaintiff failed to have one of these tests performed to confirm the existence of a physical injury. Thus, the Eighth Circuit seemed to indicate that a PET or SPECT just might have proved the requisite physical injury from PTSD.
The admissibility of PET / SPECT scan evidence arises most frequently in the criminal context, typically as defendants attempt to introduce neuroimaging evidence to negate or mitigate criminal intent, or to show incompetence to stand trial. See, e.g., United States v. Mezvinsky, 206 F.Supp.2d 661 (E.D.Pa. 2002) (rejecting defendant’s attempt to prove “insanity” defense through PET scan evidence); Jackson v. Calderon, 211 F.2d 1148 (9th Cir. 2000) (rejecting habeas petitioner’s attempt to use PET scan evidence to show chronic PCP abuse at time of criminal act). While these cases are not directly on point, they are nevertheless instructive in showing how courts are addressing some of the evidentiary shortcomings found with PET/SPECT scan evidence.
In one of the more notorious cases, Vincent “The Chin” Gigante (reputed head of the Genovese crime family) repeatedly attempted to use PET scan images to show that he was incompetent to stand trial which the Southern District of New York rejected. In United States v. Gigante, 982 F.Supp. 140 (S.D.N.Y. 1997), the court refused to admit the testimony of Dr. Monte Buchsbaum of New York’s Mount Sinai Medical School. Dr. Buchsbaum opined that a PET scan would allow him to “measure where and how much brain activity occurred,” and further opined that Mr. Gigante’s PET scan revealed “organic brain dysfunction, possibly due to Alzheimer’s disease or multi-infarct dementia.” Id. at 147. The district court rejected this testimony, highlighting several recurring shortcomings with PET scan evidence:
No baseline – Dr. Buchsbaum had never reviewed earlier PET scans of the defendant, so he was unable to state reliably whether the recent PET scan revealed new dementia or dementia that had existed all along.
Control group – the doctor compared the defendant’s PET scan to a fairly small control group, so the comparison was not necessarily statistically-significant.
Variables / Medication – the defendant was taking medication when his PET scan was conducted, and the doctor failed to explain whether the PET scan results were affected by the medication, or whether people in the control group were also taking similar medication when their PET scans were conducted.
See also United States v. Gigante, 996 F.Supp. 194 (S.D.N.Y. 1998) (rejecting Mr. Gigante’s second attempt to introduce PET scan evidence). The shortcomings on which the trial court focused are essentially some of the same that arise in the toxic tort context.
In sum, no clear trend has emerged among the courts of accepting or rejecting PET and SPECT scans. In certain instances – even when confronted with the evidence that this technology has not met acceptable scientific levels of methodology and criteria to diagnose neurotoxic injuries – courts are letting it in, even if only for the limited purpose of showing a differential diagnosis and injuries “consistent with” toxic exposure.
The real danger in all this, of course, is that, to the average juror, a neuroimage is as “objective” as a photograph and suggests scientific reliability where there is none. For this reason, it can truly be said that neuroimages are not always what they appear to be.
V. What the Future Holds
The few reported cases might suggest that plaintiffs’ lawyers have not taken an interest in pursuing these technologies in toxic tort and personal injury litigation. As the recent proliferation of attorney websites shows, however, these technologies are fast becoming common weapons in the plaintiff’s arsenal. Numerous plaintiffs’ websites advertise their successful use of neuroimaging technology in pursing mild traumatic brain injury and toxic brain injury claims on behalf of clients. (See, e.g., www.braininjurylaw.com; www.headinjurylaw.com; www.themoldsource.com). The subject has been a hot topic in recent years at ATLA conventions and in other plaintiffs’ networks.
Undoubtedly, the plaintiffs’ bar will continue the push to make claimed diseases like MTBI and MCS household names. Plaintiffs will make inroads into altogether new areas of toxic tort and neuroimaging will continue to develop technologically and be used in new and yet unforeseen areas of medicine. We can no doubt expect to see more and more reliance on all of the latest in neuroimaging evidence to support these claims.
Even from a defense perspective, this is not to say that neuroimagery and neuroimaging technology are things we should thwart. They are certainly valuable tools in the medical field, and their advances no doubt are a benefit to society. However, as Judge Posner reminds us, the law must lag technology. While we may one day see this sector of medical science following the lead of the plaintiffs’ bar to provide the needed methodologies, controls, empirical evidence and so forth to enable neuroimaging modalities to prove neurotoxic injuries, that day has not yet arrived. Until then, neuroimagery simply has no place in a toxic tort trial to support claims of cognitive dysfunction.
 An in-depth discussion of all of these technologies would far exceed the scope of this paper, the focus of which will remain on PET and SPECT since they are the neuroimaging techniques most commonly used in litigation.
 The Second Circuit has mentioned EEG in the admissibility context in a toxic exposure case involving damage to the plaintiff’s nervous system. In Amorgianos v. Nat’l R.R. Passenger Corp., 303 F.3d 256 (2d Cir. 2002), the plaintiff alleged nervous system injuries due to exposure to paint fumes while painting in a confined area. The court stated that “neurological abnormalities may be observed on some neurophysiological or neuroradiological measures, such as CT and MRI scans, electromyography and electroencephalography.” 303 F.3d at 263. However, the court held that it was within the district court’s discretion to conclude that such studies, on which plaintiff’s experts based their opinion, were insufficient to support the expert opinions. 303 F.3d at 269.