Assessing Cognitive Function in Transplantation and CAR-T Recipients: Expert Recommendations from the Survivorship, Aging and Biobehavioral Special Interest Groups of ASTCT

Abstract

Cognitive impairment is a prevalent yet underexplored comorbidity and complication in hematopoietic stem cell transplantation (HCT) and chimeric antigen receptor (CAR) T-cell therapy. Affecting up to half of patients, cognitive impairment may include acute phases, manifesting as transplant-associated altered mentation and encephalopathy (TAME) or immune effector cell-associated neurotoxicity syndrome (ICANS), and may persist for years post-treatment as cancer-related cognitive impairment (CRCI). Such dysfunction undermines autonomy, healthcare management, work reintegration, and quality of life. This consensus review synthesizes current evidence on CRCI across the timeline of transplant and cellular therapy, organized into pre-, peri-, and post-therapy phases, with additional focus on specific populations, such as older adults and pediatric patients. It highlights gaps in understanding of cognitive impairment risks, trajectory, and impact, alongside the challenges of standardizing assessments in diverse practice settings. Key recommendations, endorsed by the American Society for Transplantation and Cellular Therapy (ASTCT) Aging, Biobehavioral Research, and Survivorship Special Interest Groups, advocate for cognitive assessment pre- and post-therapy using validated instruments, like the Montreal Cognitive Assessment (MoCA) or Blessed Orientation-Memory-Concentration Test (BOMC). We additionally recommend supplementing with patient-reported outcomes (PROs) measures for comprehensive evaluation. If cognitive impairment is identified, we recommend action items, including exclusion of alternative etiologies, reconsideration of therapy or caregiving plan, and referrals for additional evaluation and rehabilitation, among others. Practical guidance for implementation across clinical and research settings is provided, emphasizing the need for multidisciplinary strategies to address identified impairments. This work aims to establish a framework for systematic cognitive monitoring, improving patient outcomes and quality of life while guiding future research to address significant knowledge and implementation gaps.

Introduction

For a growing number of patients with malignant and non-malignant hematologic disorders, transplantation and cellular therapies (TCT) offer the best opportunity for long-term disease control and potential cure. In hematopoietic cell transplantation (HCT), increase in donor options, development of less toxic conditioning regimens, and improvements in supportive care have broadened transplant eligibility criteria and improved long-term outcomes. Similarly, chimeric antigen receptor (CAR) T-cell therapies have led to significant improvements in progression-free (PFS) and overall survival (OS) in B-cell malignancies, and improved supportive care strategies have facilitated increased use of these therapies.

As uptake and outcomes of TCT continue to improve, it is critical to understand associated toxicities. Compared with infection, graft-vs-host disease, and cytokine release syndrome, there is less research dedicated to cognitive impairment. Cancer-related cognitive impairment (CRCI) refers to a decline in neurocognitive functioning that is independent of normal aging and may include deficits in memory, attention, concentration, and/or information processing speed, and has been associated with a current or prior diagnosis of cancer [1,2]. CRCI is remarkably common, affecting 25 – 50% of allogeneic HCT candidates [3–6], and has been identified as one of the top five concerns of patients undergoing such therapies [7]. During TCT, many patients are affected by either transplant-associated alerted mentation and encephalopathy (TAME) or immune effector cell-associated neurotoxicity syndrome (ICANS). Shortly after TCT, a decrease from baseline cognition is seen in nearly half of patients, regardless of a prior diagnosis of TAME or ICANS [8]. In a subset of patients, prolonged CRCI can persist, lasting up to half a decade or longer[9,10].

The immediate and sustained impact on cognitive function can lead to difficulty in patients resuming autonomy, managing their own healthcare, returning to work, and achieving an acceptable quality of life[11,12]. For providers caring for TCT patients, it is critical to understand CRCI throughout the treatment course. However, there are currently few standards and guidance regarding which patients should receive testing for cognitive impairment, what cognitive screening measure to administer, how often cognitive function should be assessed, and what clinical steps should be taken when impairments are identified.

This paper summarizes the available data regarding CRCI across the timeline of TCT for both adult and pediatric patients. It is organized first by timepoint (pre-TCT, peri-TCT, and post-TCT) (Figure 1), with subsequent sections highlighting populations requiring additional consideration (older adults, pediatrics) and emerging data. We conclude by offering tiered recommendations regarding cognitive screening assessment and offer practical suggestions regarding implementation. These recommendations are sorted into Tier 1 recommendations , which are well supported by data and relatively simple to operationalize, and Tier 2 recommendations, which are either supported by limited data and/or aspirational to operationalize. These recommendations are intended to be helpful across multiple practice settings and have been endorsed by the American Society for Transplantation and Cellular Therapy (ASTCT) Aging, Biobehavioral Research, and Survivorship Special Interest Groups (SIGs). The goal of this paper is to standardize best practices for cognitive assessment in TCT recipients in both standard of care and research settings.

Figure 1. Cognitive assessment and response across the time course of TCT.

TCT, transplantation and cellular therapy; American Society for Transplantation and Cellular Therapy, ASTCT; Montreal Cognitive Assessment, MoCA; Blessed Orientation-Memory-Concentration Test, BOMC; Patient Reported Outcome Measurement Information, PROMIS; central nervous system, CNS; transplant-associated altered mentation and encephalopathy, TAME; immune effector. Cell-associated neurotoxicity syndrome, ICANS

How Do We Assess Cognitive Impairment?

Brief Screening Measures

Several screening measures have been used for clinical cognitive assessment across numerous CRCI studies. It is important to note the distinction between psychometrically valid brief screening instruments versus more comprehensive neuropsychological testing, consisting of full batteries of tests. Brief screening refers to standardized tests that can be administered and scored quickly to efficiently provide a quantitative estimate of general cognitive function. Because of their brevity, brief screening instruments can be implemented in busy clinical settings. However, brevity can reduce sensitivity in detecting CRCI and nuanced cognitive impairments that a formal neuropsychological testing can detect [13].

Currently, no formal recommendations exist with respect to which specific cognitive screening instruments to use when evaluating adult patients undergoing TCT. All commonly used screening tools were developed for evaluating dementia and have limited data regarding sensitivity, specificity, and other test properties in CRCI specifically. The National Comprehensive Cancer Network (NCCN) [14] and the Young International Society for Geriatric Oncology (Young SIOG) [15] recommend the use of various instruments, including the MiniCog [16], Mini-Mental State Examination (MMSE) [17], Montreal Cognitive Assessment (MoCA) [18], and Blessed Orientation-Memory-Concentration Test (BOMC) [19] across various malignancies. The domains assessed, time requirements, and practical considerations for each screening measure are summarized in Table 1 [20]. Additional recommendations for pediatric patients are reviewed in the “Pediatrics” section.

Table 1.

Cognitive assessment tools and associations with TCT outcomes

Assessment	Domains assessed	Completion Time [20].	Practical considerations	Associations with alloHCT outcomes
MoCA [18]. www.mocatest.org	Memory, visuospatial, orientation, attention, language, executive	10-15 minutes (15 – 30 questions depending on written, online, or telephone version)	Copyrighted but freely available for use after registration on the website Available in 6 languages Audio-only and telephone versions available for patients with visual impairment Good sensitivity for detecting mild cognitive impairment compared to other tools Training required Many different versions available to avoid biases from learning effects	• Baseline impairment associated with OS [3,6,22]. or NRM [6]. • Baseline impairment not associated with length of stay [22]. • Impairment at time of initial assessment predicted lower likelihood of proceeding to alloHCT [3].
BOMC [19]. https://img.thebody.com/legacyAssets/40/77/F3-6.pdf	Memory, orientation, attention	2-3 minutes (6 questions)	Copyrighted but freely available Only available in English No visual elements No online application	• Baseline impairment (score ≥7) associated with inferior 1-year OS and higher NRM in older adults [43]. • Baseline impairment (score ≥11) associated with inferior PFS [100].
MiniCog [16]. www.mini-cog.com	Memory, visuospatial, executive	3 minutes (4 questions)	Copyrighted but freely available for use Available in 16 languages Can be self-administered online	• Baseline impairment (score <3) not associated with post-alloHCT PFS, OS, or NRM [41].
MMSE [17].	Memory, visuospatial, orientation, attention, language, praxis, executive	10-15 minutes (23 questions)	Copyrighted Cannot be self-administered; requires provider to complete	• Baseline impairment not associated with PFS or OS [42].
PROMIS-CF8a [32].	Patient reported: Physical function, cognitive function (memory, attention), mental health, social wellbeing, quality of life	5-10 minutes (8 questions)	Copyrighted, but freely available for use Fixed form and computer-adapted versions available; digital administration allows for ease of EMR integration Available in 6 languages Can be self-administered online Correlates with MoCA, MMSE [34,35].	• Limited data on prognostic ability, especially with cognitive-specific domain
FACT-Cog3 [33].	Patient reported: Perceived cognitive impairment, perceived cognitive abilities, quality of life	10-15 minutes (37 questions)	Copyrighted, license required to use Available in 31 languages Cannot be self-administered online No data correlating results with objective screening measures	• Limited data on prognostic ability
EORTC QLQ [37].	Patient reported: Physical, social, emotional, and cognitive function; symptoms; quality of life;	10-15 minutes (30 questions)	Copyrighted but freely available for use Extensively translated (> 100 languages) Cannot be self-administered online No data correlating results with objective screening measures	• Limited data on prognostic ability, especially with cognitive-specific domain

Transplantation and cellular therapy, TCT; allogeneic hematopoietic cell transplantation; alloHCT; overall survival, OS; non-relapse mortality, NRM; progression-free survival, PFS; electronic medical record, EMR; Montreal Cognitive Assessment, MoCA; Blessed Orientation-Memory-Concentration Test, BOMC; Mini-Mental Status Examination, MMSE; Patient Reported Outcome Measurement Information System Cognitive Short Form 8-item-a, PROMIS-CF8a; Functional Assessment of Cancer Therapy-Cognitive, FACT-Cog3; European Organization for Research and Treatment of Cancer Core Quality of Life, EORTC QLQ

The Mini-Cog is a brief screen for cognitive impairment based on two components: delayed three-word recall and clock drawing [16]. Possible cognitive impairment is identified by either a delayed word recall score of 0 out of 3 or a delayed recall score of 1 or 2 and an abnormal clock drawing [16]. The MiniCog has the advantage of being quick (2-3 minutes), easy to administer, does not require special training, and is freely available [16]. Recently, the American Society of Clinical Oncology (ASCO) Guidelines endorsed the Mini-Cog as part of a “practical” geriatric assessment in older adults [21].

The Mini-Mental Status Examination (MMSE) is an 11-question measure testing 5 domains of cognitive function: orientation, registration, attention and calculation, recall, and language. A score of 23 out of 30 indicates cognitive impairment. It has been validated in clinical and research settings and can measure changes in cognitive function with repeated use [17]. Of note, since its original development, this tool has become proprietary, which may limit its use.

The Montreal Cognitive Assessment (MoCA) is more involved, with 30 questions, but is more sensitive to subtle deficits in executive function compared with other screening instruments, such as the MMSE [22,23]. MoCA can detect early impairments in memory and executive function, with 83% sensitivity to detect mild cognitive impairment and 94% to detect dementia [23]. For patients with hematological malignancies, a score of ≤25/30 is optimal to identify performance below normative levels in at least one domain, and a score of ≤22/30 identifies impaired performance in two or more domains when using the in-person version of the assessment [13, 24]. A score of <23/30 has been incorporated into a risk prediction model for 1-year non-relapse mortality (NRM) in older adults undergoing allogeneic HCT [25].

The Blessed Orientation-Memory-Concentration Test (BOMC) is a six-item cognitive screen with scores ranging from 0-28; scores ≥5 are suggestive of cognitive impairment [19]. The BOMC has been incorporated into cancer-specific geriatric assessment [26]. and is predictive of chemotherapy toxicity [27].

Patient Reported Outcomes

Unlike neuropsychological tests, which are performance-based and capture neurocognitive function at a fixed point in time, patient reported outcome (PRO) measures are patient or proxy administered and assess the subjective experience of cognitive difficulty in daily life. CRCI research has demonstrated inconsistent correlation between objective cognitive assessments and self-reported cognitive function [28]. Rather, cognitive PROs tend to correlate with measures of fatigue, emotional distress (anxiety, depressive symptoms), and/or sleep disturbance [29]. PROs can be administered in standard practice to alert care teams of developing complications, thus improving patient care and outcomes. To optimize the benefits of PROs in clinical practice, measures should be brief to minimize patient burden, have high sensitivity and specificity, and be highly interpretable by care teams to allow for actionable intervention.

The optimal PRO measure(s) to use in TCT research and clinical practice remains an open question. In a recent comprehensive review of self-reported cognitive function measures among adult transplant patients, 21 different measures across 56 studies were described [30]. Table 2 summarizes key studies that have used PRO measures to investigate neurocognitive function in TCT recipients.

Table 2:

Summary of Studies Investigating Patient-Reported Outcomes Addressing Neurocognitive Function in TCT recipients

Study title	Patient population (N, treatment type)	Median age	Neurocognitive PRO instrument	Assessment points	PRO study conclusions and comments
Long-term patient reported neurocognitive outcomes in adult survivors of hematopoietic cell transplant. (Wu et al.) [101].	N = 1861, Adult auto- and alloHCT recipients surviving ≥2 years from HCT	64.2 years (interquartile range, 56.8-70.5)	Neuro-QOL: 8 question self-reported measure addressing perceived difficulties in cognitive abilities and application to daily tasks reflecting cognitive quality of life. NCQ: 33 questions, 4 domains corresponding to emotional regulation, task efficiency, memory and organization.	Cross-sectional study (≥2 years post-HCT)	HCT respondents reported average Neuro-QOL scores similar to general population. On the NCQ, 43% reported impairment in 1 or more domain, higher than the general population. Hearing and sleep concerns were associated with impaired patient-reported neurocognitive scores. No correlative objective neuropsychological testing.
Longitudinal Patient Reported Outcomes with CAR-T Cell Therapy Versus Autologous and Allogeneic Stem Cell Transplant (Sidana et al.) [102].	N = 104, CAR-T patients	62 years (range 26-77)	Neuro-QOL	Pre-CAR-T, 2 weeks after CAR-T, monthly for 6 months.	No decline in self-reported cognitive function after CAR-T and overall trajectory was positive compared to baseline.
Change in Patients’ Perceived Cognition Following Chimeric Antigen Receptor T-Cell Therapy for Lymphoma (Barata et al.) [83].	N = 118, CAR-T patients	61 years (standard deviation 12)	ECog: 40 item questionnaire assessing 6 domains (memory, language, visuospatial abilities, planning, organization, and divided attention) resulting in a global perceived cognition score. Satisfaction with cognition was also evaluated.	Baseline, day 90, day 360.	Mean levels of perceived cognition did not change from baseline to day 90 but worsened from day 90 to day 360 in global cognition and in the domains of memory, language, organization, and divided attention (p values<0.05). Although statistically significant, changes were small. Greater baseline fatigue, apromis nxiety, and depression were associated with worse global cognition at day 90. Patients with more severe ICANS post-CART reported worse global cognition at day 360 (p<0.05), although there were no differences in perceived cognition by severity of CRS.
Self-endorsed cognitive problems versus objectively assessed cognitive impairment in blood or bone marrow transplantation recipients: a longitudinal study. (Murdaugh et al.) [103].	N = 378, auto- and alloHCT patients; N = 90, healthy controls	52.2 years (range (19 – 73)	NIS: 95 item questionnaire rated on 5-point scale with 7 subscales (critical items, cognitive efficiency, attention, memory, frustration, tolerance, learning verbal and academic skills).	Baseline, 6 months, 1, 2, 3 years post-HCT	HCT recipients reported more cognitive problems at all time points and rate of change in NIS scores was significantly greater in HCT patients. There was a modest, but statistically significant correlation between self-endorsed cognitive problems and objective cognitive impairment.
Patient-Reported Neuropsychiatric Outcomes of Long-Term Survivors after Chimeric Antigen Receptor T Cell Therapy (Ruark et al.)	N = 40; CAR-T recipients surviving ≥ 1 year.	54 years (range 22 – 74)	4 cognitive functioning questions- author developed questions assessing difficulty with concentration, finding words, memory, or solving problems since CART-.		19 (47.5%) reported at least one cognitive difficulty and/or depression/anxiety. Depression prior to CAR-T was significantly association with self-reported CAR-T difficulty.

Transplantation and cellular therapy, TCT; hematopoietic cell transplantation, HCT; chimeric antigen receptor T-cell therapy, CAR-T; patient reported outcomes, PRO; quality of life, QOL; immune effector cell neurotoxicity syndrome, ICANS; cytokine release syndrome, CRS; Neuro-Quality of Life Cognitive Function Short Form, Neuro-QOL; Childhood Cancer Survivor Study Neurocognitive Questionnaire, NCQ; Everyday Cognition Questionnaire, ECog; Neuropsychological impairment scale, NIS;

The following PRO measures are commonly used and validated:

Patient Reported Outcome Measurement Information System (PROMIS) is a set of measures that was developed by the National Institutes of Health for use in diverse clinical and research settings. PROMIS measures have demonstrated strong validity and reliability in various cancer populations, including HCT, have a high level of patient responsiveness, and are relatively easy to implement and score [31,32]. The PROMIS suite of measures provides T-score conversion tables, allowing for standardized interpretation relative to the general population, including direct numerical comparison with other PRO measures that may co-occur and influence cognition (e.g., fatigue, depression, sleep disturbance). The PROMIS Cognitive Function Short Form-8-item-a (CF8a) was noted to have high utility by the Cancer Neuroscience Initiative Working Group following critical appraisal of the 8 most common PRO instruments used to measure CRCI [33]. PROMIS-CF8a and PROMIS-Applied Cognition also correlate with MoCA and MMSE, respectively, in geriatric populations [34,35].

The Functional Assessment of Cancer Therapy-Cognitive Function, version 3 (FACT-Cog3), is widely used in both clinical and research settings [36]. It consists of 37 items and assesses cognitive function across 4 subscales: (1) perceived cognitive impairments (PCI); (2) perceived cognitive abilities; (3) comments from others; and (4) quality of life. The 4 scales are commonly scored separately and are, therefore, useful granular measures of cognitive experience. However, given the substantial overlap in items between the FACT-Cog PCI (which is comprised of 18 or 20 items) and the PROMIS-CF8a, as a measure of patient-reported cognitive impairment alone, the CF8a may offer more convenience due to its brevity yet high reliability.

The European Organization for Research and Treatment of Cancer Core Quality of Life questionnaire (EORTC QLQ-C30) is the most widely used PRO measure of cognitive function, both in transplant as well as in general cancer research and practice [30,37]. It is a 30-item quality of life questionnaire, although only 2 questions address cognition, and thus accuracy in CRCI detection may be compromised. In HCT, the EORTC QLQ-C30 is typically used alongside the Functional Assessment of Cancer Therapy subscale for Bone Marrow Transplantation (FACT-BMT) [38].

Cognitive Impairment Prior to TCT

Baseline Cognitive Impairment in TCT Patients

The prevalence of CRCI in TCT candidates is unknown. In the literature, rates range from 12-89%, although most studies describe CRCI in roughly half of pre-TCT patients [4–6,39].

This variability is frequently attributed to methodological differences, such as investigators evaluating cognition at different timepoints relative to prior chemotherapy. Differences in both neuropsychological test batteries and investigators’ operationalized definitions CRCI also contribute to variable prevalence estimates.

Impact of Pre-TCT Cognitive Impairment on Post-TCT Survival Outcomes

Evidence is mixed regarding the impact of baseline cognitive impairment on survival outcomes in TCT (Table 1). Available data is limited not only by the heterogeneity in screening instruments but also by the outcomes evaluated. In the transplantation setting, the MoCA has been applied pre-transplant in several single-center studies, which have largely found no association between baseline cognition and post-HCT survival [3,6,22]. One study showed that patients with cognitive impairment at the time of initial assessment were less likely to proceed to HCT [3]. Recently, BMT-CTN 1704, a large national study that prospectively applied a geriatric assessment (GA) in older adult alloHCT candidates, showed lower cognitive scores by MoCA independently added to prediction of 1-year NRM as a component of the Composite Health Assessment Risk Model (CHARM) score; CHARM further predicted for post-alloHCT disability, inferior patient-reported physical and mental function, and decline in post-transplant cognition by MoCA at day 100. [25,40]. The MiniCog (score <3) and MMSE (score < 27) have failed to predict survival outcomes after alloHCT [41,42]. Finally, a multi-institutional retrospective study of geriatric assessment (GA) in older adults found that of all components of a brief GA metrics collected pre-alloHCT, cognitive impairment as measured by BOMC (score ≥7) was the only risk factor for inferior OS and increased NRM [43].

In CAR-T, a single-institution prospective study demonstrated an association between even mild cognitive impairment pre-CAR-T (MoCA < 26/30) and inferior post-CAR-T overall survival [44].; however, larger and multi-center studies are lacking. Importantly, to date, in both HCT and CAR-T, most studies have focused on survival or NRM as outcome metrics. The impact of baseline cognitive functioning on other post-treatment outcomes, including healthcare utilization, risk of long-term cognitive impairment, and impact on other quality of life parameters is limited but should be included in future prospective studies.

Tiered Recommendations: Pre-TCT

Tier 1 Recommendations

• Type of assessment:

  • Adults: There is no gold standard assessment. We endorse the following as data-supported options:

    • MoCA has been established in prospective trials and risk assessment models [25]. and is associated with HCT outcomes.
    • BOMC is a shorter assessment with substantial supportive retrospective data [43] and has also been endorsed by ASTCT [45]
  • Pediatrics: Multiple instruments have been endorsed by the Children’s Oncology Group (COG), as reviewed in Table 4 of a prior publication [46].
  • In all assessments, consider the patient’s native language and English proficiency, literacy, possible hearing or visual impairment, and level of fatigue at the time of testing, and revise testing type and timing if necessary.
• Population to assess:

  • Adults: Patients > 50 years of age or any adult with concerns for frailty or functional impairment with known CNS malignancy or prior CNS-directed therapy or with clinical or caregiver concern.
  • Pediatrics: All children (and/or caregiver proxy) that meet standardized age requirements, which are dependent on the test used [46]

• Timeframe of assessment: Assess at initial transplant referral visit or time of TCT evaluation and comorbidity assessment

• Action if impairment identified: Exclusion of alternative etiologies for cognitive impairment (e.g., metabolic encephalopathy, infection, medication side effect; cerebrovascular; psychiatric); review of medications and de-prescribing as indicated; critical assessment of TCT therapy plan (e.g., treatment intensity, caregiver requirements)

Tier 2 Recommendations

• Type of assessment:

  • Incorporation of PRO assessment, such as PROMIS-CF8a or as part of PROMIS 29+2
  • PROMIS parent proxy measures for pediatric patients

• Population to assess: All patients

• Timeframe of assessment:

  • Second assessment prior to conditioning therapy, especially if patient has received substantial additional therapy since initial visit, has a change in clinical status, or has significant frailty and/or advanced age. Data describing cognitive changes (via MoCA) between initial and pre-HCT visit have been described [3,25]
  • If multiple timeframes are assessed, the same assessment should be used, with awareness of possible learning effects and need for version changing

• Action if impairment identified: Referrals for additional evaluation (e.g., geriatrics, neuropsychology); consider cognitive rehabilitation therapy (e.g., cognitive behavioral therapy memory or attention worksheets, smartphone-based memory applications) [6]; implement strategies for improving medication adherence; increased awareness of possible delirium in the acute care setting

Cognitive Impairment Peri-TCT

During conditioning chemotherapy, lymphodepleting chemotherapy, and/or radiation, as well as shortly after transplantation or cellular therapy, patients are particularly vulnerable to cognitive decline. This may be due to the therapy administered (e.g., ICANS after CAR T-cell therapy), a result of complications of therapy (e.g., transplant-associated thrombotic microangiopathy [TA-TMA], mucosal injury, infections), hospital-acquired delirium, medication side effects (e.g., calcineurin inhibitors, corticosteroids), GVHD, or a combination of the above.

Transplant-Associated Altered Mentation and Encephalopathy (TAME)

Up to 73% of adults and children experience delirium during the initial HCT course, and delirium has been associated with increased NRM, 1-year mortality, worse post-HCT neurocognitive functioning, and quality of life [47,48] This spectrum of altered consciousness following HCT prompted the ASTCT Committee on Practice Guidelines to propose a new diagnosis, TAME, to capture the range of acute neurocognitive changes that can occur early (≤100 days) and late (>100 days) post-HCT [45]. A diagnosis of TAME encompasses patients who fulfill the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria for delirium, as well as those experiencing acute neurocognitive changes that fall short of meeting all DSM-5 delirium criteria. Patients who experience TAME often face longer hospital stays, prolonged recovery time, and a higher risk of poor post-HCT outcomes [45].

The Confusion Assessment Method (CAM) is recommended for screening and diagnosis [45]. CAM is a clinical tool widely used to diagnose delirium in various clinical settings by focusing on four key features: acute onset and fluctuating course, inattention, disorganized thinking, and altered level of consciousness as evaluated by a clinician. To diagnose TAME using CAM, both acute onset with fluctuating course and inattention must be present, along with either disorganized thinking or altered level of consciousness [49].

Pre-HCT assessment of cognitive impairment may help identify patients at higher risk for TAME, guiding discussions on the risks and benefits of HCT, and allowing for adjustments to reduce that risk [45]. However, as TAME is a newly described entity, and screening measures are not yet widely adopted in standard clinical practice, further research is needed to understand the pre-HCT risk factors, treatment factors, and long-term implications of a TAME diagnosis.

Immune effect cell-associated neurotoxicity syndrome (ICANS)

The incidence of ICANS following CAR T-cell therapy varies widely due to differences in grading scales, CAR design and development, and clinical trial design, but is generally reported as 15-62% [50–52]. Symptoms of ICANS may include cognitive impairment, aphasia, altered consciousness, motor weakness, seizures, and cerebral edema. ICANS severity is graded using ASTCT Consensus Grading, which utilizes an immune effector cell-associated encephalopathy (ICE) score (0-10) assessing orientation, handwriting changes, level of consciousness, presence of seizures, and motor deficits, and is completed multiple times per day [53].

Predictive factors for ICANS are variable and include factors associated with disease (e.g., disease bulk, CNS involvement) and treatment (e.g., co-stimulatory domain) as well as patient-specific factors. The Endothelial Activation and Stress Index (EASIX) score is predictive for ICANS and incorporates disease and patient comorbidity factors, including MRI brain, lumbar puncture, and neurologic assessments [54]. A lower baseline MoCA is also associated with increased risk and severity of ICANS in CD19 and BCMA-targeting [55–57]. Interestingly, changes in MoCA scores may occur early in the course of ICANS, and there may be utility in adopting MoCA scoring in the post-infusion periods to monitor for ICANS or determine the need for preemptive therapy [58].

Tiered Recommendations: Peri-TCT

Tier 1 Recommendations

• Type of Assessment: Conduct CAM and ICE assessments as appropriate: CAM assessments twice daily on days 6 – 20 and daily through hospitalization; otherwise, ICE assessments at least twice daily through hospitalization. In pediatric patients, the Cornell Assessment of Pediatric Delirium (CAP-D) is an alternative to CAM [59].

• Population to Assess: All patients. Maintain a high level of awareness for TAME and ICANS, discuss risks with patients and caregivers prior to TCT.

• Action if Impairment Identified: Exclude alternative etiologies; start ICANS treatment per protocol. Increase support for the patient, both during hospitalization and upon discharge. This may involve heightened non-pharmacologic delirium prevention strategies and arranging for home health services, additional caregiving support, or other resources to ensure adequate care and monitoring.

Post-Treatment Cognitive Impairment, Late Effects, and Survivorship

Cognitive impairment is a significant issue for survivors of TCT. In HCT, neurocognitive function declines in 40% of adults during the first few months post-HCT and is characterized by deficits in memory, complex attention, verbal fluency, and executive functioning [9,60–62]. Identifying specific predictors for post-HCT neurocognitive impairment is difficult, but inpatient delirium and prolonged hospitalization have been reported [63] Studies utilizing longitudinal neuropsychological testing have demonstrated that cognitive function frequently returns to pre-HCT levels over time [9]. However, up to half of HCT recipients demonstrate mild global cognitive impairment up to 5 years post-HCT, with 12-28% of these individuals demonstrating moderate to severe impairment [9,39,61,64,65]. Declines in memory and executive function were present in patients of all ages, with additional declines in verbal fluency in older adults [66]. For long-term survivors with CRCI, quality of life and daily functioning are often adversely impacted. Survivors report difficulty with return to work, everyday tasks, social interactions and can experience frustration, feelings of inadequacy, and social isolation [67–69].

Less is known about long-term CRCI in CAR-T recipients. Patients may experience slight worsening in cognitive function from baseline to day 90 post-CAR-T, with slight improvement by 1 year [70]. Conversely, another prospective observational study showed no evidence for neurological or cognitive toxicity 6-12 months after CAR-T, even in patients who developed acute neurotoxicity [71]. At 3 years post-CAR-T, late cognitive difficulties, including difficulty concentrating and speaking, have been self-reported in up to 38% of patients [72].

Tiered Recommendations: Post-TCT

We acknowledge there is limited data and no established gold standard for post-TCT cognitive assessment as well as interventions to address cognitive decline in the post-HCT period; thus, our “Tier 1” recommendations are largely based on consensus recommendation rather than published data. This indicates a significant need for further research, especially regarding early intervention strategies.

Tier 1 Recommendations

• Assessment: Use the same assessment tool as in the pre-TCT assessment; maintain awareness for learning effects and the need for version changing.

• Timeframe of Assessment: Post-TCT assessment at the end of transplant/beginning of survivorship visit, corresponding to the day 100 timepoint centers are required to submit clinical outcomes for reporting.

Tier 2 Recommendations

• Timeframe of Assessment: Day 90 and 1-year post-TCT, again paralleling timepoints centers are required to submit clinical outcome data for registry reporting. Consider additional assessment at significant cognitive milestones (e.g., return to work or school)

• Action if Impairment Identified: Referrals for additional evaluation (e.g. geriatrics, neuropsychology); referral to rehabilitation, such as efforts that focus on balance and orientation to support both cognitive and physical recovery [73].; referral for formal neuropsychologic evaluation, especially if applying for disability or accommodations; incorporate cognitive-stimulating activities such as jigsaw puzzles and games, as suggested by research studies [74]., to help maintain or improve cognitive function in patients; recommendations for exercise or mindfulness programs

Special Populations and Considerations

Older Adults

CRCI in older adults has long been recognized as a source of morbidity and mortality. Studies dedicated to assessing CRCI in patients with hematologic malignancies have demonstrated an association with survival, function, and care utilization [75,76]., while studies evaluating treatment have shown a negative impact on cognition longitudinally, regardless of therapy intensity [77,78]. By contrast, other studies have shown little to no deleterious impact from therapy, likely related to heterogeneity in underlying diagnosis and assessment instrument used [79].

The impact of cognition on TCT is particularly relevant to older adults, in whom the prevalence of baseline cognitive impairment is higher due to normal aging processes. In older adults, data are mainly derived from single-institution studies of pre-transplant GA, which demonstrate baseline cognitive impairment rates of 12-24% without a consistent impact on NRM or OS [80]. However, in a multi-institutional retrospective study of HCT recipients aged 50+, pre-HCT cognitive impairment was shown to be a strong predictor of NRM and OS [43].; notably, rates of cognitive impairment were similar in patients ages 50-59, 60-69, and 70+, though this may reflect patient selection for transplant. In another study, baseline cognitive abilities were also similar between older and younger transplant recipients, and both were inferior to age-matched controls without hematologic malignancy, suggesting that cancer treatment received pre-transplant may impact baseline cognition [81]. Data from the recently completed prospective CHARM study (BMT CTN 1704) may provide additional data regarding cognitive trajectory in older adults after alloHCT [25].

In CAR-T cell therapy, data evaluating CRCI in older adults has shown that baseline cognitive status, as measured by a comprehensive GA or MoCA, is associated with length of hospitalization, need for intensive care, and inferior survival [49,82]. Studies specifically examining cognitive outcomes, however, have shown that while there is a transient drop in cognition post-CAR-T, there is generally either a return to pre-CAR-T baseline or even slight improvement [70,71]. Interestingly, studies that have shown neurocognitive decline have used PRO assessments [83]., which point towards possible limitations of relying on objective cognitive assessment tools alone in this population. Most studies in CAR-T cell therapy, however, do not focus on older adults. As use of these therapies in older adults increases, there is ongoing need for both geriatric-specific studies and the development of geriatric-specific tools and metrics [84].

Pediatrics

In pediatric TCT recipients, approximately 20-40% of patients have some degree of CRCI post-TCT, although these data are limited by retrospective studies with relatively small sample sizes, population heterogeneity, and non-standardized assessments [46,85]. Assessing CRCI in pediatric TCT recipients requires distinctive considerations. Age-based variations in development limit the use of a single standardized battery, and age at treatment may also impact risk, type, and extent of cognitive impairment. Further, many patients, especially those with acute lymphoblastic leukemia, are often pretreated with intrathecal chemotherapy and/or central nervous system (CNS)-directed radiation, which could significantly impact their pre-TCT neurocognitive function. Similarly, children often undergo HCT for non-malignant diseases such as sickle cell disease and inborn errors of metabolism, which may predispose them to have baseline neurocognitive deficits due to the underlying disease process. Therefore, it is important to collect prospective, age-appropriate, and disease-specific data starting from baseline pre-TCT and at periodic intervals post-TCT in the pediatric population.

Given the knowledge gaps regarding neurocognitive impairment in pediatric TCT survivors, in 2022, the Children’s Oncology Group (COG) Neurocognition in Cellular Therapies Task Force published guidelines for monitoring cognition post-TCT, including recommendations for specific assessments [46]. Based on a review of published evidence and expert opinions, the guidelines recommend monitoring neurocognitive function pre-HCT and at 1, 2, and 4-5 years post-HCT, with additional assessments during major academic milestones, such as transitioning from elementary to middle school.

The taskforce also acknowledged that limited personnel and lack of insurance coverage may limit access to formal neurocognitive assessment for many patients and recommended a risk-based approach where cognitive assessment is prioritized for patients exposed to CNS-directed or total body radiation at ≤6 years of age, a history of CNS pathology, or prior ICANS. In addition to domains routinely assessed in adults, pediatric cognitive assessment should also emphasize educational performance, including average grades, a history of failed or repeated grades, and school-based support.

Along with formal neurocognitive assessments, validated PRO measures include the PROMIS cognitive function scale and Behavior Rating Inventory of Executive Function (BRIEF) [86]. PROMIS self-report scale is validated for age >8 and parent proxy scale for age >5, although concurrent administration of both self- and parent-proxy report survey is recommended for optimal assessment [87]. The COG taskforce also recommends the use of Cogstate (www.cogstate.com) [88]., a comprehensive, validated computerized battery currently incorporated into several COG-sponsored trials.

Biomarkers, Imaging and Novel Metrics to Assess Cognition

To date, most studies of CRCI in TCT rely on either provider or patient-administered assessments. However, there are limited data on the use of additional objective measures, including biomarkers and imaging, in research settings. Following alloHCT, neurocognitive impairments correlate with an activated immune system in the CNS [67,89]. When compared to pre-HCT measurements, patients with elevated IL-6 and TNF-RII but decreased CRP at day 90 post-HCT have inferior cognitive performance [66]. Genetic predisposition, such as single nucleotide polymorphisms in the blood-brain barrier, telomere homeostasis, and DNA repair genes, have also been correlated with neurocognitive impairment post-HCT [90]. Multiple studies have also associated cytokines with ICANS development and decreased cognition, including increased peripheral levels of IL-6, IFN-γ, IL-10, and IL-15 [91]. Currently, there are no validated biomarkers for cognitive function.

Neuroimaging changes during HCT are common but imperfectly correlate with neurocognitive function [92]. In patients with ICANS, MRI changes are evident in approximately half of patients [93]. A specialized MRI technique, resting state functional connectivity (RSFC), may correlate with attention and recall [94]. While MRI is generally less helpful in assessing cognition, it is critical in ruling out alternative etiologies for cognitive change, including structural abnormalities, cerebrovascular complications, and posterior reversible encephalopathy syndrome (PRES). Another imaging technique, transcranial doppler, may provide a novel and sensitive method of ICANS assessment and has the benefit of being done at the bedside [95]. Visual EEG-based grading may also improve ICANS assessments [96].

Recommendations

We acknowledge that the available data and taskforce recommendations, summarized above and in Table 3, are often retrospective, incomplete, and/or heterogenous. Further, the logistics of implementing cognitive testing are unique to different TCT centers and patient populations. With these caveats, we offer some practical considerations for incorporating these recommendations into standard clinical practice.

Table 3:

Summary of Tiered Task Force Recommendations across the TCT Course

Type of Assessment	Population to Assess	Timeframe of Assessment	Action if Impairment Identified
Pre-TCT
Tier 1 Recommendations
• MoCA • BOMC • Pediatrics: Multiple instruments endorsed by COG	• Adults > 50 years of age • Significant frailty • Known CNS malignancy; prior CNS-directed therapy • All children	• Initial TCT referral or time of TCT evaluation and comorbidity assessment	• Exclude alternative etiologies • Medication review; de-prescribing as able • Critical assessment of TCT therapy plan
Tier 2 Recommendations
• PROMIS-CF8a • Pediatrics: PROMIS parent proxy measures	• All patients	• Additional pre-TCT timepoints, especially if additional therapy, change in clinical status, or high-risk population	• Referral for additional evaluations (geriatrics, neuropsychology) • Cognitive rehabilitation therapy • Medication adherence strategies
Peri-TCT
Tier 1 Recommendations
• CAM (Transplant) • ICE (Cellular Therapy)	• All patients	• CAM: twice daily days 6 – 20 post-transplant; daily through hospitalization • ICE: at least twice daily	• Exclude alternative etiologies • Start ICANs treatment per protocol
Tier 2 Recommendations
No additional recommendations	No additional recommendations	No additional recommendations	• Increase support for patient: arrange home health services additional caregiving support, etc.
Post-TCT
Tier 1 Recommendations
• Same as used in pre-TCT assessment • Maintain awareness for learning effects	• Adults > 50 years of age • Significant frailty • Known CNS malignancy; prior CNS-directed therapy • Significant cognitive concern during peri-TCT • All children	• Day 100 post-TCT (End of peri-TCT period/beginning of survivorship care)	No formal recommendations
Tier 2 Recommendations
• • No additional recommendations	• All patients	• Day 180 post-TCT • 1 year post-TCT • Any significant cognitive milestone (e.g., return to work or school)	• Referrals for additional evaluations; formal neuropsychologic evaluation • Referrals for physical, occupational, cognitive rehabilitation • Exercise or mindfulness programs • Incorporate jigsaw puzzles, games

Transplant and Cellular Therapy, TCT; Montreal Cognitive Assessment, MoCA; Blessed Orientation-Memory-Concentration Test, BOMC; Mini-Mental Status Examination, MMSE; Children’s Oncology Group, COG; Patient Reported Outcome Measurement Information System Cognitive Short Form 8-item-a, PROMIS-CF8a; Confusion Assessment Method, CAM; Immune Effector Cell Encephalopathy, ICE; Immune effector cell-associated neurotoxicity syndrome, ICANS

There are multiple options for conducting cognitive assessments as part of clinical workflow (Figure 2). While these can be done by a specialized geriatric clinic or a TCT clinic provider or nurse, it may also fall to the medical assistant as part of standard patient rooming and vital sign assessment [3]. MoCA does require specific training to administer, but this training is available through their website, allowing anyone to be certified to perform assessments. Of note, when using MoCA longitudinally, it is recommended to rotate through different versions to avoid training effects. Results from in-person assessments may be entered into the medical record (e.g., Epic flowsheet) or electronic database (e.g., REDCap [Vanderbilt University, Nashville]). Both MoCA [97]. and BOMC [98]. can also be administered as telephone versions, which can ease the logistical burden of longitudinal assessment or be a good option for visually impaired patients. Telephone or telehealth assessments may also not be equally appropriate for all populations. In a study of virtual cognitive assessments in older adults, older and frailer patients were less likely to complete assessments [99]. Finally, when conducting cognitive assessments, it is important to note the patient’s mental condition and avoid assessment if impaired (e.g., if the patient recently took a narcotic or is struggling with delirium).

Figure 2. Multidisciplinary Approach to Cognitive Impairment in TCT.

TCT, transplantation and cellular therapy; TAME, transplant-associated altered mentation and encephalopathy; ICANS, immune effector cell-associated neurotoxicity syndrome

For PROs, assessments can be sent to patients before appointments through the electronic medical record (e.g., Epic/MyChart) or by email from research databases (e.g., REDCap). This approach has the advantages of (1) allowing patients to complete assessments at their convenience, (2) allowing results to be reviewed prior to the appointment, and (3) results are directly entered into databases, saving time and avoiding errors with manual entry. If results are not available, staff may follow up with patients to complete assessments using electronic devices (e.g., tablets) or paper forms.

Conclusion

This position paper summarizes the current evidence in cognitive assessments along the TCT timeline and provides recommendations for implementation both clinically and in research settings. Measuring changes in cognition across the TCT timeline has been limited due to heterogeneity in assessments utilized, timepoints for baseline and post-treatment longitudinal evaluations, thresholds for impairment, and practical barriers to implementation in busy clinical settings. Screening tools and the use of PRO measures can help overcome some of the barriers, but these are limited due to the heterogeneity in tools utilized and in limited data on the impact of PRO measures of cognitive function on outcomes. In CAR-T, there is a dearth of data on the impact of cognitive impairment, as identified by screening tools, on treatment-related toxicities or outcomes. The recommendations in this paper were created to address these barriers.

In TCT recipients, we recommend clinical evaluation of cognition at a minimum of pre-TCT and ideally at multiple timepoints, using the same instrument across timepoints, institutions, and research studies. Standardization across timepoints would allow for improved monitoring of cognitive changes over time and with treatment. Early recognition of cognitive decline can allow for earlier investigations and interventions to, at minimum, stabilize the trajectory of cognitive decline and ideally reverse it. Standardization of assessments across institutions would enhance the ability to perform collaborative multi-institutional retrospective or observational studies. PRO measures should be included as part of cognitive evaluation in TCT recipients and can overcome the personnel and time barriers of implementation, although additional studies associating PRO cognitive assessments with objective assessments and clinical outcomes are needed. Current literature investigating pre-TCT cognition focuses on NRM and OS. Future studies are needed to investigate the impact of pre-TCT cognition, as measured by both objective assessment and PRO measures, on peri-TCT cognitive complications (TAME, ICANS), as well as quality of life, return-to-work, and other long-term sequelae. Finally, to better understand cognitive trajectories along the TCT timeline and impact on outcomes, using cognitive tools validated and with normative data in non-English speaking populations is imperative. Beyond language translations, standardized cognitive screening tools need to account for patient age, education level, and cultural context.

Standardizing cognitive assessment clinically and in research will allow our community to bridge the gap in understanding cognitive dynamics across TCT, better identify patients impacted by CRCI, and develop interventions. We have provided practical recommendations to address these gaps and barriers to implementation.

Highlights.

• Cognitive impairment is common pre-, peri-, and post-therapy for HCT and CAR-T (TCT)

• This paper offers consensus recommendations to standardize cognitive assessment.

• MoCA or BOMC are recommended pre-TCT and at transition to survivorship

• PRO assessments, like PROMIS-CF8a, should be incorporated as feasible

• Recommendations address assessment gaps in pediatric and geriatric populations