Abstract
Background:
A goal of medical research is to advance knowledge of the molecular biology underlying human brain function. Yet, few studies of human brain biology have been performed using brain tissue from living people. This is due to the lack of safe approaches to sampling the living human brain for rigorous scientific inquiry.
Objectives:
The Living Brain Project (LBP) developed a method to biopsy a small volume of prefrontal cortex (PFC) tissue during deep brain stimulation (DBS) lead placement procedures. The objective of this LBP report is to establish the safety of the PFC biopsy approach.
Methods:
Acute adverse events (i.e., infection, intracranial hemorrhage [ICH], and seizures) were tracked following 1,152 DBS procedures performed on 590 patients. A PFC biopsy was obtained in 652 procedures (“biopsy group”) and no biopsy was obtained in 500 procedures (“non-biopsy group”). Cognitive health was assessed at baseline and one year after DBS surgery for 144 of the patients. Rates of acute adverse events and changes in cognitive health were compared between the biopsy and non-biopsy groups.
Results:
No infections occurred in either group. No statistically significant difference in ICH rate was observed between groups (1.7% biopsy group vs. 1.4% non-biopsy group; Chi-square test p-value = 0.88), and this observation held regardless of the anatomical location or the clinical severity of the ICH. No statistically significant difference in seizure rate was observed between groups (0.2% biopsy group vs. 0.4% non-biopsy group; p-value = 0.82). No statistically significant associations were observed between number of biopsies and changes in cognitive health over time.
Conclusions:
DBS procedures involving PFC biopsies for the LBP demonstrate a safety profile comparable to DBS procedures without biopsies.
Keywords: Deep brain stimulation, Brain biopsy, Surgical complications, Neuropsychological outcomes, Movement disorders, Safety analysis, Living brain tissue
Introduction
Certain fundamental questions about the molecular basis of human brain function can only be addressed using brain tissue samples from living human participants (e.g., “What genes are active when a person experiences a particular emotion?”). Therefore, approaches are needed for obtaining such samples in a safe and scalable manner. Generally, such approaches are lacking.
Deep brain stimulation (DBS) is an elective neurosurgical treatment for neurological and mental illnesses. A common technique for safely implanting the DBS electrode involves cauterizing a small volume of the prefrontal cortex (PFC) prior to delivery cannula penetration. While most individuals who undergo a DBS procedure do not experience an adverse event in the acute period (i.e., hours to months) following the procedure, potentially dangerous types of acute adverse events following DBS procedures are intracranial hemorrhage (ICH), seizures, and infections. According to one meta-analysis that included over 25,000 DBS procedures, ICH occur following 1.6%, seizures occur following 1.5%, and infections occur following 2.4% of procedures1. According to another meta-analysis, approximately half of ICH following DBS procedures are asymptomatic, and the anatomical locations of these ICH include the entry point of the DBS electrode into the brain (16% of ICH following DBS), the track of the electrode through the brain (31%), and the target of the electrode deep in the brain (7%)2. A potential long-term adverse event following DBS procedures is post-operative decline in cognitive function. Studies suggest there is a statistically significant increased risk of decline in function across multiple domains of cognition following DBS procedures, through the clinical significance of these statistical observations is unclear3-5.
The Living Brain Project (LBP) was designed to address the unmet need of collecting living human brain tissue to advance medical research. For the LBP, a technique was developed to biopsy the region of PFC that would otherwise be cauterized during DBS procedures. By utilizing this technique, the LBP has already made several contributions to medical research, including (1) a comprehensive characterization of living human brain biology6-9, (2) technology to study human brain circuits at the synapse-level10, (3) technology to deliver gene therapy to the brain11, and (4) insights into the pathogenesis of Alzheimer’s disease12. Here, analyses establishing the safety of the LBP PFC biopsy procedure are presented of data from 1,152 DBS procedures.
Methods
Study design and cohort
Data was analyzed retrospectively from clinical and research records of 1,152 DBS procedures performed at a single medical center between 2013 and 2024 (Table 1). The 1,152 procedures were performed on 590 individuals (“the full study cohort”). A computed tomography (CT) scan was performed within hours following each procedure. Patient medical records were reviewed for acute ICH, seizure, and infection from the period of time spanning the procedure date to 90 days post-procedure. A subset of patients was followed for approximately one year to assess cognitive outcomes.
Table 1.
Demographic and Clinical Characteristics of Procedures in the Biopsy and Non-Biopsy Groups
| Characteristic | Biopsy Group (n=652) | Non-biopsy Group (n=500) |
|---|---|---|
| Demographics | ||
| Mean age (years) | 61 | 62 |
| Male sex (%) | 65% | 67% |
| DBS Target | ||
| STN (%) | 63% | 65% |
| GPi (%) | 19% | 23% |
| VIM (%) | 7% | 8% |
| Clinical Indication | ||
| Parkinson's disease (%) | 75% | 79% |
| Dystonia (%) | 8% | 10% |
| Essential tremor (%) | 7% | 7% |
| OCD (%) | 6% | 2% |
| MDD (%) | 4% | 2% |
| Device Manufacturer | ||
| Abbott (%) | 56% | 60% |
| Medtronic (%) | 44% | 40% |
Abbreviations: DBS, deep brain stimulation; STN, subthalamic nucleus; GPi, globus pallidus internus; VIM, ventral intermediate nucleus; OCD, obsessive-compulsive disorder; MDD, major depressive disorder.
Most individuals in the full study cohort (N = 532) underwent two procedures, but 45 individuals underwent one procedure and 13 individuals underwent three or more procedures. Most of the individuals in the full study cohort (N = 409) elected to enroll in the LBP. A PFC biopsy for the LBP was obtained in 652 of the 792 procedures performed on these 409 individuals. These 652 procedures will be referred to as the “biopsy group” and the 500 procedures where a PFC biopsy was not obtained for the LBP will be referred to as the “non-biopsy group.”
The procedures in the biopsy group were performed on 384 unique individuals, including 268 individuals with two biopsies taken (i.e., one biopsy from each hemisphere) and 116 individuals with one biopsy taken (74 left hemisphere biopsies and 42 right hemisphere biopsies). The procedures in the non-biopsy group were performed on 304 unique individuals, including 98 individuals who were also amongst the 384 individuals represented in the biopsy group (i.e., individuals who underwent two of more procedures and a biopsy was obtained in one of, but not all of, the procedures) and 206 individuals from who zero biopsies were obtained.
Additional details of the study cohort are presented in Table 1 and the Supplementary Information, including data showing that demographic, clinical, and device-related characteristics were similar between the biopsy and non-biopsy groups.
Ethics statement
The study was carried out with the approval of the Human Research Protection Program at the Icahn School of Medicine at Mount Sinai. DBS patients in the study included LBP participants (N = 409) and individuals who did not participate in the LBP (N = 181). The LBP participants provided informed consent that covered clinical data extraction from medical records. Clinical data extraction from medical records of individuals who were not LBP participants was performed for clinical quality improvement purposes. All patients undergoing DBS procedures at Mount Sinai were eligible for the study. LBP eligibility was limited to English-speaking individuals over 18 years old.
Surgical procedure and biopsy collection
All DBS procedures were performed by a single neurosurgeon using standard stereotactic techniques. The vast majority of the procedures were implantation procedures (i.e., the placement of a DBS electrode into the brain of an individual for the first time). Implantation procedures were categorized into the biopsy group (i.e., procedures where a PFC biopsy was obtained) and the non-biopsy group (i.e., procedures where a PFC biopsy was not obtained). The non-biopsy group also included a small number of revision procedures (i.e., the adjustment or replacement of a previously implanted DBS electrode). All procedures involved a standard frontal burr hole and stereotactic cannula placement, with cortical surface preparation differing only by the inclusion of a PFC biopsy in the biopsy group (obtained prior to cauterization and using a standardized punch tool), as detailed in the Supplementary Information.
Assessment of adverse events
Acute adverse events were defined as an adverse event that occurred within 90 days of the DBS procedure. Three types of acute adverse events were considered: ICH, seizure, and infection. Review of the electronic medical record for each of the 1,152 procedures was performed to identify occurrences of these three types of acute adverse events. For each DBS procedure, a non-clinician chart review was performed to flag potential acute adverse events, followed by a clinician chart review if flagged. In the non-clinician step, a trained research staff member (non-clinician) manually reviewed the electronic medical record from the three months post-procedure following a standardized protocol. Five staff members (three from the LBP team and two from an unrelated program) conducted these reviews. Procedures were flagged if keywords (e.g., blood, hematoma, seizure, infection) appeared in relevant records (e.g., post-procedure CT radiology reports). The clinician review was performed on all flagged procedures by the neurosurgeon (BHK) and a neurologist staff member with no previous involvement in the LBP (SA).
Assessment of post-operative cognitive function
For 144 of the 590 individuals in the full study cohort, two neuropsychological testing sessions (“NP sessions”) were conducted approximately one year apart. These 144 individuals will be referred to as “the NP cohort.” NP sessions were performed at part of routine clinical care of DBS patients. At each NP session a battery of neuropsychological tests was administered characterizing the general intelligence, language abilities, visuospatial abilities, attention, processing speed, executive function, memory, and affective state of the individual.
The NP cohort was comprised of 92 individuals who had at least one biopsy taken (“the NP biopsy cohort”) and 52 individuals who had zero biopsies taken (“the NP non-biopsy cohort”). The NP biopsy cohort was further comprised of 30 individuals who had one biopsy taken (“the NP unilateral biopsy cohort”) and 62 individuals who had two biopsies taken (“the NP bilateral biopsy cohort”).
Data from NP sessions was extracted from the electronic medical record by trained staff using a standardized protocol. When appropriate, raw scores were normalized by the administering neuropsychologist (who was not a part of the research team) using normative data. Analyses were restricted to the 26 tests that were administered in both NP sessions for a sufficient proportion of the NP cohort. For each of 26 tests administered, the following three analyses were performed to examine the relationship between the change in the test scores over time and the PFC biopsy:
test score changes over time were correlated with the number of biopsies obtained in the NP cohort
test score changes over time in the NP biopsy cohort were compared to test score changes over time in the NP non-biopsy cohort
test scores that significantly worsened over time in the NP cohort were further examined to verify that the worsening was observed regardless of the number of biopsies obtained.
Further details on the NP cohort and NP sessions are provided in the Supplementary Information, including a list of the 26 tests analyzed.
Statistical analysis
Statistical analyses were conducted in R (version 4.2.0) primarily using the base stats R package. Chi-squared tests were implemented using the chisq.test() function. Correlation tests were implemented using the cor.test() function with the “method” parameter set to either “pearson” or “kendall” as indicated in the main text. Student’s t tests were implemented using the t.test() function with the “paired” parameter set to “TRUE” when appropriate. Adjustment of p-values for multiple tests was performed using the p.adjust() function with the “method” parameter set to “fdr” (i.e., the Benjamini-Hochberg method). Statistical significance was defined as a p-value less than 0.05 (or, when appropriate, an adjusted p-value less than 0.05).
Results
Biopsy sizes
Biopsy sizes were measured for 231 biopsies. The mean biopsy volume measured was 0.04 cm3 (equal to 40 mm3) and the median volume was 0.03 cm3 (equal to 30 mm3). Visual inspection of the distribution of biopsy volumes (Figure 1) showed one extreme outlier (0.64 cm3 = 640 mm3). This one measurement is suspected to be the result of a measuring error since the approximate maximum biopsy volume that is possible given the dimensions of the surgical instrument used to obtain the biopsies for the LBP (0.075 cm3 = 75 mm3; see Supplementary Information). The volumes of the remaining 230 biopsies were tightly distributed around the mean (standard deviation of these 230 biopsies = 0.02 cm3 = 20 mm3), as expected given the surgical instrument dimensions.
Figure 1.
Histogram showing the distribution of PFC biopsy volumes obtained for the Living Brain Project.
Acute adverse events
ICH occurred following 18 (1.6%) of the procedures, seizure occurred following 3 (0.3%) of the procedures, and infection did not occur following any of the procedures. No difference in the ICH rate was observed between the biopsy group and the non-biopsy group (11 ICH in the biopsy group [1.7% of the biopsy group]; 7 ICH in the non-biopsy group [1.4% of the non-biopsy group]; Chi-square test p-value = 0.88). No difference in seizure rate was observed between the biopsy group and the non-biopsy group (1 seizure in the biopsy group [0.2% of the biopsy group]; 2 seizures in the non-biopsy group [0.4% of the non-biopsy group]; Chi-square test p-value = 0.82) (Figure 2).
Figure 2.
Bar plot showing the rates of acute adverse events do not differ between the biopsy group (pink bars) and the non-biopsy group (blue bars). The vertical axis shows the acute adverse event type and the horizontal axis shows the acute adverse event rate.
ICHs were further categorized in three ways: (1) based on the anatomical location of the ICH (11 ICH near the electrode entry point [“cortical ICH”] and 7 ICH along the electrode track [“subcortical ICH”]); (2) based on whether the individual exhibited clinical symptoms in the setting of the ICH (7 ICH where the individual exhibited symptoms [“symptomatic ICH”] and 11 ICH where the individual did not exhibit symptoms [“asymptomatic ICH”]); (3) based on whether the ICH resulted in the individual remaining in the hospital for longer than is standard following a DBS procedure (11 ICH resulting in a longer hospitalization [“extended-stay ICH”] and 7 ICH not resulting in a longer hospitalization [“standard-stay ICH”]). No difference was observed between the biopsy group and the non-biopsy group with respect to the rates of cortical ICH, subcortical ICH, symptomatic ICH, asymptomatic ICH, extended-stay ICH, or standard-stay ICH (all Chi-square test p-values > 0.05; Figure 2 and Supplementary Table 1).
Post-operative cognitive function
The correlation between cognitive test score changes over time and the number of biopsies obtained was examined using Kendall’s rank correlation coefficient (τ). After adjusting for 26 tests, no statistically significant correlations were observed (Figure 3). A nominally significant correlation (i.e., a p-value less than 0.05 prior to multiple test adjustment) was observed for two tests (the PDQ-8 communication score and the PDQ-8 total score), and for both tests a higher number of biopsies was associated with clinical improvement over time (Supplementary Table 2).
Figure 3.
Box plot showing there is no relationship between the number of biopsies performed (indicated by the blue, pink, and purple colors) and changes in test scores over time for 26 tests performed during neuropsychological testing sessions. The vertical axis in both plots shows the score change over time. The horizontal axis in both plots shows the test name. Tests are grouped into the top and bottom plots based on whether a positive score change implies clinical improvement over time (top plot) or clinical deterioration over time (bottom plot).
Test score changes over time in the NP biopsy cohort were compared to test score changes over time in the NP non-biopsy cohort using Student’s t test. After adjusting for 26 tests, no statistically significant differences between the NP biopsy cohort and NP non-biopsy cohort were observed (Supplementary Table 3). A nominally significant difference (i.e., a p-value less than 0.05 prior to multiple test adjustment) was observed between the NP biopsy cohort and the NP non-biopsy cohort for two tests (the PDQ-8 activities of daily living score and the PDQ-8 total score), and in both instances the NP biopsy cohort had a better clinical outcome than the NP non-biopsy cohort (Supplementary Table 3).
The worsening of test scores over time was assessed in the NP cohort using a paired Student’s t test. After adjusting for 26 tests, a statistically significant worsening of test scores was observed for two tests of verbal fluency (the FAS test and the animal naming test) and one test of apathy (Starkstein rating scale) (Supplementary Table 4). These three findings were observed irrespective of the number of biopsies obtained (Supplementary Table 5).
Discussion
This report examines the safety of obtaining PFC biopsies for research purposes during DBS procedures using the technique developed for the LBP by presenting the results of analyses of data from 1,152 DBS procedures (652 in the biopsy group and 500 in the non-biopsy group). The rates of acute adverse events observed following these 1,152 procedures were comparable to rates previously reported in large meta-analyses, and no associations were observed between the biopsy procedure and acute adverse event rates. Trajectories of cognitive function following DBS observed in the NP cohort were comparable to trajectories previously reported in DBS patients4,13,14 – namely, declines in verbal fluency and apathy – and no associations were observed between the biopsy procedure and cognitive trajectory. Altogether, these findings empirically establish the safety of obtaining PFC biopsies for research purposes during DBS procedures using the technique developed for the LBP. Additionally, the tight distribution of biopsy sizes demonstrate that the LBP PFC biopsy procedure can be highly standardized, reproducible, and precise. The strengths of this study design include a large sample size, systematic documentation of acute adverse events, and longitudinal follow-up.
Several limitations of the study should also be noted. First, assignment to the biopsy group versus the non-biopsy group was non-random, which may have introduced selection bias. Second, the single-center nature of the study may limit generalizability. Third, complete NP testing data (i.e., two time points) was available only for only 144 of the 590 patients, and this could potentially affect the generalizability of the findings. Fourth, there is a subjective element to making the determination of whether an ICH has occurred following a neurosurgical procedure. When reviewing CT scans performed immediately after a DBS procedure, radiologists often note the presence of blood. In most cases, this blood is not an ICH, it is simply the expected state of the brain in the immediate aftermath of a neurosurgical procedure. Determining what constitutes an ICH in this setting therefore requires the expertise of a neurosurgeon who specializes in DBS. In the ideal world, this individual would be an entirely external, unbiased reviewer (e.g., a DBS surgeon from an external institution with no opinion on the LBP). However, the amount of work required (i.e., manual chart review of outcomes for over 1,000 DBS procedures) and multiple regulatory barriers make this approach infeasible in practice. Fifth, due to the small number of acute adverse events, the statistical power to investigate the relationship between acute adverse event rates and indication for DBS is limited and therefore not included in the report.
The findings of this report were anticipated since PFC biopsies for the LBP are obtained from tissue that would otherwise be cauterized during the standard DBS procedure. As such, regardless of whether a PFC biopsy is obtained for the LBP or not, approximately the same amount of PFC tissue volume will be lost during the DBS procedure. While subtle differences in tissue volume loss between approaches (cauterization alone versus biopsy followed by cauterization) may exist, they are not measurable with current methods and appear clinically inconsequential. The establishment of this safe and standardized biopsy protocol opens new avenues for human brain research. an important step forward for human brain research.
Conclusions
The findings of this report suggest that obtaining brain biopsies during DBS surgery using the standardized protocol developed for the LBP targeting non-eloquent cortex, does not significantly increase the risk of acute adverse events of long-term cognitive decline. Studies of living human brain tissue enabled by this protocol can help advance understanding of the molecular origins of brain health and illness.
Supplementary Material
Supplementary Table 3. Results of tests of the difference in neuropsychological test score changes over time in the biopsy group and non-biopsy group.
Supplementary Table 2. Results of tests of the correlation between the number of biopsies obtained and changes in neuropsychological test scores over time.
Supplementary Table 1. Results of tests of association between the number of biopsies obtained and ICH rates.
Supplementary Table 4. Results of tests of the significance of changes in neuropsychological test scores over time in the NP cohort.
Supplementary Table 5. Results of tests of the significance of changes in neuropsychological test scores over time in subsets of the NP cohort.
Data Access Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1.Rasiah NP, Maheshwary R, Kwon CS, Bloomstein JD & Girgis F Complications of Deep Brain Stimulation for Parkinson Disease and Relationship between Micro-electrode tracks and hemorrhage: Systematic Review and Meta-Analysis. World Neurosurg 171, e8–e23 (2023). 10.1016/j.wneu.2022.10.034 [DOI] [PubMed] [Google Scholar]
- 2.Cheyuo C. et al. Comprehensive characterization of intracranial hemorrhage in deep brain stimulation: a systematic review of literature from 1987 to 2023. J Neurosurg 141, 381–393 (2024). 10.3171/2024.1.JNS232385 [DOI] [PubMed] [Google Scholar]
- 3.Xie Y, Meng X, Xiao J, Zhang J & Zhang J Cognitive Changes following Bilateral Deep Brain Stimulation of Subthalamic Nucleus in Parkinson's Disease: A Meta-Analysis. Biomed Res Int 2016, 3596415 (2016). 10.1155/2016/3596415 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Combs HL et al. Cognition and Depression Following Deep Brain Stimulation of the Subthalamic Nucleus and Globus Pallidus Pars Internus in Parkinson's Disease: A Meta-Analysis. Neuropsychology review 25, 439–454 (2015). 10.1007/s11065-015-9302-0 [DOI] [PubMed] [Google Scholar]
- 5.Bucur M & Papagno C Deep Brain Stimulation in Parkinson Disease: A Meta-analysis of the Long-term Neuropsychological Outcomes. Neuropsychology review 33, 307–346 (2023). 10.1007/s11065-022-09540-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Rodriguez de Los Santos M. et al. Divergent landscapes of A-to-I editing in postmortem and living human brain. Nat Commun 15, 5366 (2024). 10.1038/s41467-024-49268-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Vornholt E. et al. Characterizing cell type specific transcriptional differences between the living and postmortem human brain. medRxiv (2024). 10.1101/2024.05.01.24306590 [DOI] [Google Scholar]
- 8.Liharska LE et al. A study of gene expression in the living human brain. medRxiv (2023). 10.1101/2023.04.21.23288916 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kopell BH et al. Multiomic foundations of human prefrontal cortex tissue function. medRxiv (2024). 10.1101/2024.05.17.24307537 [DOI] [Google Scholar]
- 10.Karlupia N. et al. Immersion Fixation and Staining of Multicubic Millimeter Volumes for Electron Microscopy-Based Connectomics of Human Brain Biopsies. Biol Psychiatry 94, 352–360 (2023). 10.1016/j.biopsych.2023.01.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wang C. et al. Intravenous administration of blood-brain barrier-crossing conjugates facilitate biomacromolecule transport into central nervous system. Nature biotechnology (2024). 10.1038/s41587-024-02487-7 [DOI] [PubMed] [Google Scholar]
- 12.Kosoy R. et al. Genetics of the human microglia regulome refines Alzheimer's disease risk loci. Nat Genet 54, 1145–1154 (2022). 10.1038/s41588-022-01149-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Greif TR, Askari A, Cook Maher A, Patil PG & Persad C Anterior lead location predicts verbal fluency decline following STN-DBS in Parkinson's disease. Parkinsonism Relat Disord 92, 36–40 (2021). 10.1016/j.parkreldis.2021.10.012 [DOI] [PubMed] [Google Scholar]
- 14.Tan ZG, Zhou Q, Huang T & Jiang Y Efficacies of globus pallidus stimulation and subthalamic nucleus stimulation for advanced Parkinson's disease: a meta-analysis of randomized controlled trials. Clin Interv Aging 11, 777–786 (2016). 10.2147/CIA.S105505 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 3. Results of tests of the difference in neuropsychological test score changes over time in the biopsy group and non-biopsy group.
Supplementary Table 2. Results of tests of the correlation between the number of biopsies obtained and changes in neuropsychological test scores over time.
Supplementary Table 1. Results of tests of association between the number of biopsies obtained and ICH rates.
Supplementary Table 4. Results of tests of the significance of changes in neuropsychological test scores over time in the NP cohort.
Supplementary Table 5. Results of tests of the significance of changes in neuropsychological test scores over time in subsets of the NP cohort.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.