Impact of acute lymphoblastic leukemia therapy on attention and working memory in children

Keywords::

• acute lymphoblastic leukemia
• cognitive late effects
• magnetic resonance imaging
• neurotoxicity
• quality of life

Imagine by the time you finish reading this sentence you cannot remember how it started, by the time you get to the top of the stairs you cannot remember what you were retrieving or by the time you get to the phone you cannot remember the number you were about to dial. Now imagine that you are not showing early signs of aging but rather you are in the prime of your childhood years. You have yet to face the challenges of adolescence, including tracking alternating high school schedules, juggling the multiple steps required for college acceptance, automatizing the micro-routines involved in learning to drive and monitoring the nuances of teenage social interactions. These are precisely the challenges faced by an ever-growing group of survivors following modern-day therapy for acute lymphoblastic leukemia (ALL). While these cognitive late effects are now well documented, we are just beginning to discover the underlying mechanisms that are responsible – a crucial step towards developing interventions that can meaningfully impact quality of life.

Acute lymphoblastic leukemia is the most common malignancy of childhood and adolescence. Over the past few decades, there have been dramatic improvements in treatment outcomes, resulting in 5-year event-free survival rates of nearly 90% [1]. Increased cure rates have led to increased focus on the quality of life of these survivors. Adverse side effects, including cognitive late effects, were substantially reduced with the elimination of radiation therapy as a component of most ALL therapies [2]. However, neurotoxicity is still a relevant issue and ALL survivors remain subject to increased cognitive impairments secondary to disease and treatment [3].

In order to identify neurologic processes that underlie cognitive late effects and develop targeted cognitive interventions, it is crucial that research moves beyond assessing global abilities to unveil specific neuropsychological deficits. Historically, cognitive outcomes have been evaluated using measures assessing global abilities such as intellectual functioning (IQ) and academic achievement, where healthy controls typically perform better than children treated for ALL [4]. Recent research has begun to examine more specific and functional cognitive processes. An emerging consensus holds that a decline in IQ is secondary to one or more central processing deficits involving attention [5], processing speed [6] or executive functions such as working memory [7].

Secondary to these core processing deficits, ALL survivors may experience decreased academic achievement, reduced occupational success and poorer overall quality of life [8]. ALL survivors treated with chemotherapy alone process information slower than age-matched controls, especially when multiple pieces of information have to be processed or attention has to be focused [9]. Adequate attention is required, but not sufficient, for good working memory performance, which requires additional mental manipulation of information. Deficits in attention and working memory affect a large portion of ALL survivors, with a recent study reporting working memory deficits in almost 70% of survivors after 2 years off therapy [10]. These abilities appear especially sensitive to treatment-related changes, detecting difficulties potentially missed by global intelligence measures. Findings from this study are consistent with a recent meta-analysis of 28 studies of cognitive late effects of childhood ALL treatment, which identified clinically significant deficits in attention, speed of information processing and executive functioning [11].

Pathophysiology of damage

The mechanisms of ALL treatment-induced neurotoxicity remain unclear. The elimination of irradiation from standard treatment regimens was made possible through more intensive CNS chemotherapy administered intravenously or intrathecally. Antimetabolite chemotherapeutic agents including methotrexate (MTX) and cytarabine are commonly used for CNS treatment, with MTX being the primary focus of recent late-effect studies. Damage to two components of the CNS, such as oligodendrocytes and the vasculature, may contribute to neurotoxicity, resulting in neurocognitive late effects in childhood ALL survivors. Iron contained in oligodendrocytes plays an important role in myelogenesis, and damage to oligodendrocytes could result in demyelination. This mechanism is consistent with the destruction of oligodendrocytes observed over 30 years ago in a postmortem study of patients with leukemia treated with MTX [12].

Methotrexate can inhibit the turnover of myelin lipids and proteins, resulting in intramyelinic splits and intralamellar formation of vacuoles in which interstitial fluid can accumulate. The main mechanism of action of MTX is inhibition of dihydrofolate reductase (DHFR), which is primarily responsible for the conversion of dihydrofolates to their active form of folate. Methylenetetrahydrofolate reductase (MTHFR) plays a key role in determining the available levels of folate. Alterations in the function of folate pathway enzymes can lead to reduced folate, thereby increasing toxic levels of homocysteine, and may induce ischemic side effects [13].

Glucocorticoids are also major components of ALL therapy that affect metabolism and have anti-inflammatory and immunosuppressive effects. Their inhibition of glucose utilization by neurons and glia increases the concentration of glutamate, particularly in the hippocampus, which can potentially cause excitotoxic neuronal death as a result of overstimulation. Dexamethasone has been shown to be a potent cytotoxic agent, with higher rates of CNS penetrance than prednisone and prednisolone.

Identifying those at risk

Genetic polymorphisms

Although several chemotherapeutic agents in ALL therapy may contribute to neurotoxicity and negative neurocognitive outcomes, MTX has been the primary focus of many studies. The dose, cumulative exposure and infusion rate of MTX have been associated with the degree of neurocognitive deficit [14]. To fully explore this relationship between MTX and cognitive late effects, one must consider the entire folate pathway and the associated genetic polymorphisms. For example, MTHFR has been linked with attention deficits in ALL survivors [15]. Alterations in the function of folate pathway enzymes may alter CNS folate levels, chemotherapy-related toxic effects or both. We believe that neurocognitive decline can occur secondary to elevated homocysteine, thus having a primary impact on neuroanatomical structure and function, which can be directly assessed through advanced neuroimaging techniques.

The neurotransmitter dopamine is critically important for functions mediated by the prefrontal cortex, including attention and working memory [16]. The catechol O-methyltransferase (COMT) gene codes for the COMT enzyme, which plays a key role in the degradation of synaptic dopamine. The COMT gene has two polymorphisms, methionine (Met) and valine (Val); the Met polymorphism results in slower dopamine degradation and thus greater dopamine availability. In healthy adults, the Met polymorphism is associated with better prefrontal cortex function. For example, adults homozygous for the Met polymorphism have been shown to perform better on a frontally mediated task, the Wisconsin Card Sorting Test [17]. Children homozygous for the Met polymorphism have been shown to perform significantly better on a working memory task dependent on the dorsolateral prefrontal cortex [18]. It may be that genes regulating dopamine circuitry, such as COMT, serve as resiliency factors against the emergence of cognitive late effects in children treated for ALL, thus offering a mechanism for individual variability among those with similar treatment histories. If these relationships are demonstrated, cancer treatment regimens could be tailored based on the risk of later emerging cognitive issues.

Advanced neuroimaging of damage

High-dose methotrexate given intravenously has a substantial toxic effect on the CNS and can lead to severe neurologic morbidity. Leukoencephalopathy (LE), which is seen as white-matter hyperintensity on T2-weighted MRI, is the most common manifestation, and may be either persistent or transient [19]. The frequency and severity of LE depends on MTX dose, cumulative exposure or other clinical variables, and is more prevalent than previously believed, with some MRI studies demonstrating rates of LE as high as 86% after seven courses of high-dose MTX, with a reduction to approximately 40% prevalence by the end of therapy [20]. The duration of these changes and their relationship to long-term pathology are still unclear [21]. However, transient changes in white matter may be the result of demyelination, which could lead to decreased axonal density and brain atrophy. Smaller white matter volumes, particularly in the frontal lobes, have been observed in ALL survivors and were associated with declines in attention, intelligence and academic achievement [22,23].

Anatomical localization of LE appears to preferentially impact the white matter of the frontal lobes and long association fibers throughout the corona radiata, including the longitudinal fasciculus [24]. LE within the frontal lobes during active therapy has been associated with deficits in attention processes and executive functioning, such as working memory, at therapy completion [10]. The frontal lobes of the brain are the last to fully develop with respect to cerebral white matter, with myelination extending into the third decade of life. Given this protracted development, cognitive processes supported by the frontal lobes may be particularly vulnerable to neurotoxicity associated with CNS-directed therapy.

Diffusion tensor imaging, an advanced MRI method based on local microstructural characteristics of water diffusion, has been shown to be a more sensitive measure of myelin integrity than simple white matter volume changes. Fractional anisotropy (FA) in the frontal lobes in normal subjects demonstrates a normal developmental trajectory associated with attention, processing speed and working memory [25], which could be changed as a result of insult. Unfortunately, studies of diffusion metrics in patients treated for ALL are fairly rare and limited to cross-sectional studies of survivors. These few studies have demonstrated reductions in FA within the frontal white matter and along the fronto-occipital fasciculus relative to controls, and FA was also associated with intelligence and processing speed [26,27].

There are two additional ancillary measurements of myelin degradation that could potentially further elucidate the impact of ALL therapy on neuroanatomy: quantitative magnetization transfer (MT) imaging and quantitative multiple exponential T2 (qT2) measurements. MT imaging is a relatively lengthy procedure but has demonstrated some success in one small study of patients during active ALL therapy. A whole-brain histogram analysis of magnetization transfer ratio before and after intravenous and intrathecal MTX demonstrated a significant difference in the peak height but no significant shift in the peak [28]. Multiple exponential fitting of the qT2 relaxation components in white matter can be used to generate myelin water fraction maps (<40–50 ms), which have been shown to correlate strongly with histological staining for myelin. An additional advantage of adding the quantitative multiple exponential T2 measures would be the ability to carefully characterize the long T2 components (200–800 ms), which are indicative of inflammation, extensive destruction with increased extracellular water and severe axonal damage [29]. Unfortunately, the qT2 method is more difficult to implement, limited in the amount of brain imaged, and requires extensive post-processing, making it more suited for research than routine clinical care.

The advent of more refined neuroimaging techniques combined with higher field strengths have provided a unique opportunity to more carefully examine the impact of therapy for ALL on the normal developing brain. Limited pilot studies substantially support our hypotheses that patients treated for ALL will exhibit disrupted myelin integrity relative to age- and sex-matched controls in superior long association fibers early in therapy. We believe that genetic polymorphisms in the folate pathway will result in this neuroimaging phenotype. Correlating these advanced imaging measures with therapy, genetic factors and sensitive neurocognitive testing of specific domains holds great potential for understanding the impact therapy on individual patients, thus facilitating the design and evaluation of behavioral or pharmacologic interventions when needed.

Prevention & remediation of cognitive late effects

Refining treatment delivery

The evolution of ALL-directed therapy has reflected the goal of modifying treatment to obtain high survival rates while minimizing toxicities, including cognitive late effects. The elimination of radiation therapy from most ALL treatment regimens has had the greatest impact in this regard. Risk-adapted therapies that reduce the dose of intravenous and/or intrathecal chemotherapy for those children with better disease prognosis, based on presenting clinical and biological factors, are also serving to optimize cognitive outcomes [10]. It is likely that a plateau will be reached in the ability to limit toxicity while maintaining high survival rates, emphasizing the need to develop interventions to prevent or remediate cognitive late effects that may not be entirely avoidable. Given the prominence of attention and working memory difficulties among cognitive late effects in ALL survivors, initial intervention approaches have borrowed heavily from the existing attention-deficit/hyperactivity disorder (ADHD) literature.

Pharmacologic interventions

Stimulant medications have been used for decades to successfully and safely treat children diagnosed with ADHD. The most commonly prescribed medication for ADHD is methylphenidate. A placebo-controlled, double-blind crossover study including ALL survivors revealed significant improvements in parent and teacher ratings of attention, and teacher ratings of social skills following treatment with methylphenidate [30]. More recently, it has been demonstrated that these attention and behavioral benefits are sustained by maintenance doses of methylphenidate over the course of a year and benefits are observable across home and school environments [31]. While this study failed to reveal significant improvements on standardized measures of academic skills, there were anecdotal reports of improved academic grades secondary to better planning and organizational abilities. Newer stimulant medications such as modafanil and nonstimulant medications such as donepezil, typically used to treat Alzheimer’s dementia, are beginning to emerge in oncology literature and may also hold promise for addressing common cognitive late effects of ALL survivors [32,33]. As there are survivors for whom these medications will not be a viable option due to medical contraindications, parental preferences, medication intolerance or poor response, the development of nonpharmacologic interventions is a necessity.

Nonpharmacologic interventions

Butler, Copeland et al.’s tripartite model, the cognitive remediation program (CRP), uses techniques from brain injury rehabilitation (e.g., massed practice on attention tasks), special education (e.g., training in metacognitive strategies such as task preparedness and monitoring) and clinical psychology (e.g., cognitive behavioral strategies such as reframing of cognitive struggles) [34]. CRP includes 20, 2-h sessions, completed one-on-one with a therapist. A multicenter, randomized controlled trial of CRP in 167 childhood cancer survivors revealed significant improvement on a parent report measure of attention, an increase in academic skills and use of more metacognitive strategies in problem solving [34]. Similar behavioral interventions are being developed that vary program elements, such as increased teaching of strategies that can be implemented by parents and teachers in their respective environments [35]. These studies offer initial encouragement, but the personnel time and financial requirements are great for generally modest benefits. Investigations including less expensive, less therapist-driven, nonpharmacologic interventions are underway that may increase the availability of options. While lagging with respect to research investigation, school interventions including accommodations (e.g., preferred seating or additional time) and modifications (e.g., shortened assignments) are an integral part of any comprehensive intervention plan [36]. Accordingly, close monitoring of cognitive abilities and proactive involvement of the school are imperative to optimize academic and vocational success.

Future directions

Similar to findings in the ADHD literature, it may be advantageous to approach cognitive late effects with a combination of pharmacologic and nonpharmacologic interventions. Furthermore, prophylactic approaches that seek to prevent rather than remediate cognitive late effects may prove most promising. These approaches could include pharmacologic, nonpharmacologic and/or alternative (e.g., diet or exercise) interventions that would be administered during or shortly after treatment, prior to the emergence of cognitive or behavioral problems. Furthermore, coupling intervention studies with neuroimaging technology may lead to the identification of a neurophysiological profile that provides early (e.g., end of treatment) guidance regarding the selection of alternative interventions, thus reducing the need to wait for significant problems to develop or to try multiple interventions before identifying the best match for a particular child.

What this means for those treating patients

When a child is treated for ALL, curing the cancer should not be the only goal. A significant proportion of children surviving ALL therapy develop clinically meaningful deficits in attention, information processing speed and executive functions such as working memory. These deficits can have a detrimental impact on survivors’ quality of life. Failure to closely assess and monitor these survivors may lead to confusion or misattribution surrounding cognitive struggles.

Detecting specific neurocognitive deficits requires focused, serial testing beyond assessment of global abilities. Identifying those patients at risk of developing these deficits may require genetic testing for polymorphisms associated with folate metabolism or dopamine regulation. Neuroimaging assessments, soon after completion of intravenous MTX, can identify those patients with extensive LE who are more likely to have persistent irreversible damage in white matter tracts essential for efficient frontal lobe functioning.

Curing the cancer is only the beginning for these patients. Survivors of childhood ALL with deficits in attention, processing speed or working memory will probably experience increased difficulties with the greater academic and social demands throughout adolescence and, ultimately will have to face the added neurologic burden associated with normal aging. Once patients at risk for deficits in attention and working memory have been identified, either prospectively early in therapy or when the deficits are made manifest at end of therapy, there are pharmacologic or nonpharmacologic interventions with emerging empirical support that can be tailored to the individual patient to mitigate the impact of these deficits.

Financial & competing interests disclosure

This work was supported in part by NIH grants R01-CA90246 and R21-CA131616, and by Cancer Center Support grant P30-CA21765 from the National Cancer Institute and by the American Lebanese Syrian Associated Charities. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.