The Impact of Gliomas on Cognition and Capacity

Lee and Hall

The personal and emotional responses that an individual and the family have to the condition may also make it difficult to opine on windows of capacity, especially in a postmortem situation. Are assessments of functioning, by lay people or even by doctors, potentially biased because they need to justify decisions made (e.g., being aggressive with temporarily prolonging palliative treatment versus deciding not to continue to pursue additional intervention)? Will individuals who think that palliative treatments may lead to the individual surviving cancer only focus on the positive aspects of an individual's functioning (e.g., moving better) and therefore not truly appreciate the strengths, weaknesses, or lack of change a patient may have in other domains (e.g., cognition)? If potential over- or underestimation of an individual's functioning with GBM is occurring, is this bias caused by emotional factors, such as not being able to admit someone is dying, or by self-interest?

Although aspects of all of these questions can be seen when assessing capacity in any medical condition, as this article will discuss, GBM does present some unique challenges due to the nature of treatments, the areas of the brain affected, the aggressive nature of such tumors, the speed of symptom progression, and limitations of the medical literature.

Impact on Neurocognitive Functioning

GBM is a fast-growing tumor with undifferentiated cell morphology and microscopic finger-like projections that invade surrounding areas of the brain. Healthy parenchymal tissue surrounding a GBM is often rapidly compromised, resulting in relatively swift symptom progression and cognitive decline greater than what would be expected from neuroimaging results alone. The growth rate of GBM is exceptionally quick, with the size of the tumor doubling in a median of seven weeks' time if untreated or unresponsive to treatment.⁶ GBM is particularly challenging to treat due to intratumor heterogeneity, whereby different regions within a single tumor are composed of different subgroups of tumor cells with different molecular characteristics, hence the subname “multiforme.”^7,8 This complexity often results in the need for varying therapeutic strategies that can result in different or compounding side-effect profiles, especially on cognition. Even after surgical resection and adjunct treatment with radiation or chemotherapeutic treatments, GBM typically recurs within an average of 6.9 months.⁹ In part, this resistance to standard treatment is why the five-year survival rate of GBM is less than 5 percent (Table 1).^10,11 Depending on multiple factors, such as the type or grade of the tumor, the brain areas affected, the length of time with the disease, and confounding variables (e.g., depression, sleep disruptions), many individuals with gliomas will have their cognitive functioning affected.^19,-,22

Research studies have reported that brain tumors, either primary or metastatic, commonly affect NCF, with as many as 75 to 80 percent of adults who undergo in-depth neurocognitive assessment at the time of treatment presentation showing some degree of cognitive impairment.^23,–,27 For example, Dwan and colleagues²⁵ studied 44 subjects with benign-to- low-grade (stage I-II) and malignant (stage III-IV) primary brain tumors and reported that 86 to 89 percent of subjects were classified as impaired at a testspecific level, 61 to 73 percent were classified as impaired at a domain-specific level, and 32 to 50 percent were classified as impaired at a global level. In the study, participants with brain tumors performed significantly below matched controls on tests of neuropsychological functioning, including executive function (p =.001) and memory (p <.001), but not for attention and processing speed (p >.05). This led to a recommendation that performance be measured using a multifaceted neuropsychological approach, otherwise specific deficits would not be appreciated.²⁵ Similar results have also been reported after surgical intervention, where there may be an improvement in one cognitive domain but a worsening in another, which means multiple domains must be tested to appreciate overall functioning.²⁸

The location of a tumor can also have a profound impact on deficits observed (e.g., damage in Broca's area results in an expressive aphasia), which often have ramifications for NCF and capacity. Patients with frontal lobe tumors not only show decreased inhibitory control, but also poorer executive functioning than those with non-frontal lobe tumors.²⁹ Noll and colleagues recently reported that patients with left temporal lobe gliomas of varying grades had frequent impairments in verbal learning and memory, executive functioning, language, and attention.²⁸ Other researchers have also found similar results when comparing right versus left temporal lobe location of primary tumors, in that left-sided tumors resulted in greater deficits in general.^26,30 Few of these studies tried to differentiate whether the effects corresponded with premorbid temporal dominance, however, which may have influenced outcomes.

Noll and colleagues also reported that, as grading worsened, the effects of impairment significantly increased independent of lesion size, seizure status, and medication treatment (e.g., antiepileptics, steroids), a finding that supports the concept of lesion momen- tum.²⁸ Unfortunately, there are a limited number of studies comparing the impact of low-grade gliomas to high-grade gliomas.^31,32 An additional concern for much of the literature on gliomas is that the studies' population sizes are often small and are usually limited to a single inpatient treatment location, which may increase the risk of sampling bias or procedural bias in the studies.

Although gliomas and other tumors may primarily present in a specific region of the brain, the overall impact on functioning may be difficult to predict due to the global effects. For example, a brain tumor may result in displacement of adjacent brain tissue, which can impede blood flow or cause increased intracranial pressure, endocrine disruptions, swelling, or edema.^25,33 These more chronic, diffuse effects of brain tumors are, in part, the reason that brain tumor lesions have been shown to impair cognition more significantly than other forms of brain damage, such as stroke or mechanical trauma, where effects from edema tend to be more temporary.^26,34

Cognitive Functioning Post-Treatment

Neurosurgical treatment is a common, if not primary, treatment method for many solid brain tu- mors.^2,35 The effects of surgery are difficult to predict, however, given the individual characteristics of each case, such as the varying areas of the brain that may be affected, the size or grading of the tumor, the surgical approach required, the degree of margin resection, and the historical lack of sufficient pre- and postoperative assessment studies.^35,–,37 While resection of a glioma is an inherently risky surgical procedure, most studies suggest that their removal (with or without clean margins) is possible without major long-term impairment of cognitive functioning due specifically to the surgery.^23,38 The effect size and duration of improvement varies from case to case and often depends on how well the glioma responds to adjunct treatment and how rapidly the glioma returns.²³

A recent Austrian study, for example, reported that, while the majority of 196 subjects with malignant gliomas, brain metastases, and meningiomas showed significant deficits in various neurocognitive domains presurgery, most either improved or did not decline further in terms of NCF in the short term after resection (i.e., 3 to 4 months).²⁴ This finding may be of particular importance to forensic examiners trying to determine, in a postmortem review, when an individual may have had the best cognitive functioning, or to advise when cognitive functioning may be at its best when it comes to creating or modifying legal documents.

The results of the Austrian study are similar to results found in a study by Habets and colleagues, which analyzed high-grade gliomas specifically.³⁹ They reported that 79 percent of the 62 subjects demonstrated cognitive impairment in at least one cognitive domain (e.g., verbal and working memory, attention, executive functioning, psychomotor function, information processing speed, or visual-spatial abilities) prior to surgery. At an average of five weeks after surgery, 39 subjects (63% of the original participants) were again tested, and 59 percent of that subgroup still showed deficits in at least one cognitive domain. Almost half (49%) of the surgically treated population showed improvement in cognitive function from presurgical baselines, whereas 23 percent showed further decline from presurgical baselines.³⁹ Another study of 29 subjects with gliomas reported that 55 percent of patients were unchanged after surgery, 21 percent improved, and 24 percent worsened in functioning in cognitive domains such as intelligence, executive functions, memory, and language.³⁷ The degree of worsening is similar for the two studies, but the degree of improvement was much larger for the first study. It should be noted, however, that many participants in the first study (37%) refused or were unable to repeat the testing postsurgery for various reasons (e.g., the testing was seen as too bothersome, or the subjects elected to not pursue additional treatment). Therefore, a potential limitation of several of the treatment studies is that they may reflect an optimistic or best-case scenario because the sickest patients or ones with limited responses may have dropped out.^37,39

The timing of a cognitive assessment can also affect the perceived benefit of the intervention because some negative impacts from surgery may resolve with time, such as edema. As an example, Wolf and colleagues noted worsening in cognitive functioning of surgical patients at the time of discharge from the hospital, with degrees of improvement occurring by the three-month mark.⁴⁰

In contrast to tumor resection, radiation therapy, a second modality for treating brain cancer, has been reported to more frequently and markedly impair NCF, depending on the delivery method (e.g., beam radiation versus whole brain), dose, and duration of treatment.^{34,36,41,–,44} In addition to general neurophysiological problems, patients who undergo radiotherapy also commonly exhibit cognitive or functional decline related to secondary factors, such as anxiety, depression, fatigue, and sleep distur- bances.²³ Research on subjects who received radiation therapy for brain tumors showed increased topological modularity(i.e., lesscommunication across cortical subnetworks), a finding that is associated with working memory deficits.^42,45 The effects of radiation therapy and the time for these effects to become evident may also depend on the dose (e.g., less than 2 Gy for low-grade gliomas has fewer effects and these effects take longer to manifest themselves than effects seen in patients receiving 60 Gy for highgrade gliomas).³⁶

Glioma patients typically are prescribed various pharmacological treatments, often including multiple classes of medications such as cortical steroids, anti-seizure medications, chemotherapy, and palliative medications (e.g., antiemetics), anyofwhichcan also affect NCF.^20,36 For example, topiramate has been shown, both in healthy patients and in patients with other central nervous system disorders such as epilepsy, to have effects on NCF, with impaired domains including visual-spatial memory, concentration, word fluency, verbal learning, and executive functioning.^46,–,49 Although the impact of topiramate alone may not be enough to cause one to lose capacity, it may have a more significant impact for an individual already dealing with cognitive deficits caused by other factors. A study looking at capacity to provide consent for research in patients with malignant gliomas noted that subjects on cortical steroid treatments and anticonvulsants were found to have poorer performance on the Capacity to Consent to Research Instrument⁵⁰ than those not prescribed the medications. Given that this study found that “phonemic and semantic word fluency predicted performance on the consent standards” (Ref. 50, p 3884), it is not surprising that anticonvulsants, some of which can affect word fluency, were associated with decreased capacity. It should also be noted that the studies reporting deficits with certain pharmacological treatments may serve more as a marker for the severity of the disease state or of a particular subtype of patient who is more willing to pursue aggressive treatment than the direct effects of the medication.

Traditional chemotherapeutic agents may also affect cognition in multiple ways, including neurotoxic effects (e.g., affects healthy progenitor cells as well as cancer cells) and side effects such as fatigue, which may further affect domains such as attention and concentration²¹.

The research literature does not specifically address how potential treatment combinations may interact or affect cognition. This may present a challenge for the forensic examiner, especially if there is a Daubert challenge, who must seek peer-reviewed literature supporting the confounding effects of treatment. In a peer-reviewed study that addresses this concern, Nolen and colleagues investigated hippocampal volumes in 13 subjects with high-grade gliomas.⁵¹ The subjects received surgical, radiation, and chemotherapy treatment (the primary agent was temozolomide, often followed with a course of bev- acizumab). Subjects were all affected by hippocampus loss as well as learning and memory deficits, but those who were exposed to more treatment modalities (e.g., radiation and both chemotherapy drugs) for longer periods of time had greater deterioration of hippocampal volume and more deficits on tasks as- sociated with hippocampal functioning. The authors focused on the hippocampus because that is an area of the brain also affected and studied in Alzheimer's dementia. They noted that once hippocampal change began, the loss of volume ensued at a rate faster than that typically seen with Alzheimer's dementia. A significant decrease in hippocampal volume was usually evident on magnetic resonance imaging after six months of treatment with bevaci- zumab, with continued change noted for up to a three-year time period. Although this study had a limited sample size, it does provide a good model to show the complex nature and interactions of varying treatment modalities and how there may be similarities as well as differences in the pathophysiology and progression of other disease states.

Capacity

Although cognitive functioning and capacity are not necessarily synonymous, a person's ability to consent to medical treatment or make important end-of-life decisions does require adequate NCF in domains such as memory, executive functioning, and language. Previous studies analyzing decision-making capacity in patients diagnosed with high-grade gliomas reported that nearly 90 percent ofthese patients lacked decision-making capacity in their final days before death, and some lacked decision-making capacity soon after being diagnosed.^33,52

Grisso and Applebaum originally proposed four criteria for decision-making capacity, which include the ability to communicate a choice, understand the relevant information, appreciate the situation and its consequences, and reason about treatment options.^53,–,55 The concepts are assessed in The MacArthur Competence Assessment Tool for Clinical Research, which was designed to offer a structured method for rating a person's abilities relevant to consenting to treatment.^55,56 In addition, thereareother assessment tools available for capacity, such as the Capacity to Consent to Treatment Instrument (CCTI), which uses hypothetical clinical vignettes in a structured interview to assess capacity.⁵⁰

In a study using the CCTI in individuals with newly diagnosed brain metastases, it was reported that simple attention, verbal fluency, verbal memory, processing speed, and executive functioning were all associated with understanding.²⁷ Verbal memory (e.g., episodic memory) and phonemic fluency were found to be the two best cognitive domains in which to assess initial deficits ifan individual lacked capacity for medical decision-making at the time of diagnosis.²⁷ When comparing the 41 subjects to healthy controls, it was found that 60 percent of the subjects with cancer demonstrated impaired capacity to consent (defined as 1.5 SD below the control group on capacity measures), with limitations being present shortly after diagnosis.²⁷ In a study of 26 subjects with malignant gliomas, capacity for medical decision-making was again assessed with the CCTI, with specificdomain breakdowns being reported.³³ More than 50 percent of those with malignant gliomas demonstrated capacity compromise (i.e., a combination of marginally capable and incapable outcomes) in understanding, 35 percent in reasoning, and 23 percent in appreciation.³³

Although assessment tools can be helpful, especially in a research setting, they often are not utilized in clinical treatment settings.^57,–,59 In addition, assessments for capacity for treatment may not be helpful for all determinations that need to be made.⁶⁰ For example, some capacity situations, such as testamentary capacity (described in Table 2), are not necessarily included on standard screens of capacity to consent for treatment.⁶¹ In addition, other formal mental health screenings, such as delirium screenings, or clinical observations that an individual is “lucid” or “oriented times three,”maynotnecessarily provide useful information regarding a patient's general higher-level thought process, comprehension, or delusional state.⁶² This general concept was highlighted in the testamentary capacity case of Miami Rescue Mission v. Roberts (2006), where the Appellate Court ruled:

We also find that the trial court correctly interpreted the legal effect of the evidence presented to it. Appellant cites to evidence in the record that Mrs. Manucy was found to be alert by the nurses treating her throughout her hospitalization. Dr. Ubeda, however, testified that he disagreed with the nurses' classification, and Dr. Currier explained that certain forms of delirium are often mistaken as cooperativeness in a patient. Further, a nurse's observation that a patient is alert does not equate to a legal finding that the patient possesses testamentary capacity We find that the trial court correctly determined that Mrs. Manucy was suffering from an insane delusion regarding Roberts and that she executed her 2005 will based on this delusion [Ref. 63, p 276].

With this being said, screens for general cognitive ability can be used to help a clinician determine capacity in conjunction with other information, assess the need for more formalized inquiry, or document degree of change over time.

It is often easier to objectively document or find documentation of neuromuscular changes or declines related to tumors (e.g., changes in gait, changes in handwriting) than cognitive changes in patient records. But motor strip damage or preserved motor functioning may not directly correlate with executive functioning or general cognitive abilities.^28,34,64 For example, a study of caregivers of patients with brain tumors noted that impairments of memory, reasoning, problem-solving, and judgment were the most frequently cited problems, with concerns regarding balance, walking, and speaking being less frequently identified.³⁴ Another example is a study of a mixed group of subjects with either malignant glioma or metastasis, which reported that medical decisionmaking capacity and cognitive functioning were correlated with scores from the Karnofsky Performance Status Scale (KPS),^{65, 66} although the KPS scores did not necessarily predict capacity.⁶⁷ The KPS (Table 3) is a frequently used scale in oncology that rates patients on the basis of their degree of functional impairment, such as the ability to work or to care for oneself, analogous to a Global Assessment of Functioning score in psychiatry. The KPS score tends to be heavily influenced more by physical functioning (e.g., ability to ambulate) than by upper-level cognitive abilities.³⁴ Hence, 46 percent of participants with a KPS score of 90-100 achieved capable classifications across all CCTI standards (of note, every subject that was rated at 100 passed), while only 23 percent of participants with KPS ratings of 70-80 and 0 percent of participants with KPS ratings of 50-60 were able to achieve the same bench- marks.⁶⁷ This concept is important for forensic examiners to understand because it is much more likely for KPS scores to be seen in oncologist charts than traditional cognitive screens or testing. There is also the potential for this scale to be used in a court situation to imply that a relatively high KPS (e.g., 7080, cares for self) equates to capacity or stable intellectual functioning over time, but this is not necessarily the case.

Treating physicians often tend to overestimate the degree of functioning a patient has, or the amount of improvement that a treatment intervention will provide. For example, one research study of 299 subjects who underwent intracranial tumor resection reported that the surgeon overestimated the functional level 62percent of the time at 30 days postsurgery, a tendency which “may have implications for clinical decision-making and for the accuracy of patient information” [Ref. 68, p 1173]. The possibility that treaters overestimate the degree of improvement may be important for forensic evaluators to consider when conducting postmortem forensic reviews.^20,54,69 It may help explain why family members, and even treatment providers, may subjectively believe the patient is doing better and has more capacity than is accurate, especially when the totality of the medical records are not supportive (e.g., records from treaters who did not perform the surgery).