REVIEW

REVIEW

Diagnostic Value and Safety of Brain Biopsy in

Patients With Cryptogenic Neurological Disease: A
Systematic Review and Meta-analysis of 831 Cases
Harrison Xiao Bai, MD*‡ BACKGROUND: The role of brain biopsy in patients with cryptogenic neurological
Yingjie Zou, MS*§ disease is uncertain.
Ashley M. Lee, MD‡ OBJECTIVE: To determine the risks and benefits of diagnostic brain biopsy for non-
neoplastic indications in immunocompetent patients.
Eric Lancaster, MD, PhD¶
METHODS: Appropriate studies were identified by searching electronic databases.
Li Yang, MD, PhD§
RESULTS: We screened 3645 abstracts and included 20 studies with a total of 831
‡Department of Neurology, The Second patients. Indications for biopsy were: (1a) severe neurological disease of unknown eti-
Xiangya Hospital, Central South Univer- ology in adults (n = 7) and (1b) in children (n = 2); (2) suspected primary angiitis of the
sity, Changsha, Hunan, China; §Depart- central nervous system (PACNS) (n = 3); (3) chronic meningitis of unknown cause (n = 3);
ment of Radiology, Hospital of the
University of Pennsylvania, Philadelphia; (4) atypical dementia (n = 4); and (5) nonneoplastic disease (n = 1). Diagnostic success
¶Department of Neurology, Hospital rates calculated for subgroups were 51.3% (34.5-68.1) for 1a, 53.8% (42.9-64.5) for 1b,
of the University of Pennsylvania, 74.7% (64.0-84.1) for 2, 30.3% (17.2-45.4) for 3, and 60.8% (41.2-78.8) for 4. Clinical impact
Philadelphia, Pennsylvania
rates were 30.5% (13.6-50.6) for 1a (n = 6), 67.1% (42.8-87.3) for 1b (n = 2), 8.3% (2.3-20.0)
*These authors contributed equally to for 3 (n = 1), and 14.2% (6.5-24.3) for 4 (n = 2). Lymphoma (n = 32) and Creutzfeldt-Jakob
this article. disease (n = 30) were the most common diagnoses on the final histopathology reports of
positive brain biopsies in 1a. In 1b, encephalitis (n = 7), PACNS (n = 6), and demyelination
Correspondence:
Li Yang, MD, PhD,
(n = 6) were the most common. The odds ratio for achieving a diagnostic biopsy when
Department of Neurology, there was a radiological target was 3.70 (P = .014, 95% confidence interval, 1.31-10.42).
The Second Xiangya Hospital, CONCLUSION: Brain biopsy in cryptogenic neurological disease was associated with
Central South University,
No. 139 Middle Renmin Rd,
the highest diagnostic yield in patients with suspected PACNS. The greatest clinical
Changsha, Hunan, impact was seen in children with cryptogenic neurological disease. The presence of
410011, P.R. China. a radiological target was associated with a higher diagnostic yield.
E-mail: yangli762@gmail.com
KEY WORDS: Atypical dementia, Brain biopsy, Cerebral vasculitis, Chronic meningitis, Cryptogenic neuro-
Received, November 24, 2014. logical disease, Diagnostic success
Accepted, February 23, 2015.
Published Online, April 7, 2015. Neurosurgery 77:283–295, 2015 DOI: 10.1227/NEU.0000000000000756 www.neurosurgery-online.com

B
Copyright © 2015 by the
Congress of Neurological Surgeons. rain biopsy is considered an invasive diag- are suspicious for opportunistic infections.4-6
nostic modality of last resort. The need for However, the risks and benefits of brain biopsy
brain biopsy has decreased in the era of in patients with cryptogenic neurological disease
magnetic resonance imaging.1 The role of brain have yet to be evaluated.7 This category com-
biopsy is well established in certain patient prises primarily acute or chronic neurological
populations, including those with suspected deterioration of unknown etiology, central ner-
neoplastic lesions where diagnostic yield ap- vous system vasculitis, chronic meningitis of
proaches 95%2,3 and immunocompromised pa- unknown etiology, and atypical dementia. This
tients with focal or multifocal brain lesions that is a particularly challenging subgroup of patients
in whom noninvasive diagnostic options are
frequently unrevealing. Therefore, it is often
ABBREVIATIONS: CI, confidence interval; CJD, imperative to obtain histological evidence before
Creutzfeldt-Jakob disease; CNS, central nervous initiating potentially toxic, yet life-saving
system; PACNS, primary angiitis of the central
therapies.
nervous system; RCVS, reversible cerebral vaso-
constriction syndrome The perceived risks and benefits of brain
biopsy vary widely among neurological centers

NEUROSURGERY VOLUME 77 | NUMBER 2 | AUGUST 2015 | 283

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

BAI ET AL

FIGURE 1. Flowchart for search strategy and study selection. AANS, American Association of Neurological Surgeons; CJD, Creutzfeldt-
Jakob disease.

around the world.7 Published evidence of diagnostic brain biopsy METHODS

in patients with cryptogenic neurological disease comprises
mostly small, retrospective, single-center analyses.8-27 These Search Strategy and Study Identification
studies were performed over long periods of time and consisted This study was conducted in accordance with Preferred Reporting
of nonhomogeneous patient groups.8-27 While some authors Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.28
argued that the diagnostic yield of the procedure was too low and Electronic databases that were searched included PubMed, Embase, Web
the results had limited impact on patient management,11-13,23 of Science, Cochrane Library, and the website of the American
Association of the Neurological Surgeons. Search terms included
others advocated for the procedure and proposed that it should
“biopsy” or “biopsies” in combination with “brain” or “intracranial”
no longer be considered as one of last resort.9,10,14,15,19,20,22,24,25 or “cerebral.” Searches were limited to human studies that were
The aim of the present review was to summarize the published published in English from 1985 to 2014. Reference lists from
evidence for the risks and benefits of diagnostic brain biopsy in publications retrieved were examined to identify additional eligible
patients without known HIV and neoplasm. studies.

284 | VOLUME 77 | NUMBER 2 | AUGUST 2015 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

 BRAIN BIOPSY FOR CRYPTOGENIC NEUROLOGICAL DISEASE: SYSTEMATIC REVIEW AND META-ANALYSIS

TABLE 1. Characteristics of the 20 Studies Included in the Final Analysisa

No. of
Study Year Country Patients Preop Imaging Stereotactic Indication Conclusion
8
Waltregny et al 1990 Belgium 28 NA Yes Presenile dementia Neutral
Hulette et al9 1992 USA 14 NA NA Atypical dementia Favor
Cheng et al10 1994 USA 37 NA No Chronic meningitis of unknown cause Favor
Smith et al26 1994 USA 21 CT, MRI, DSA NA Chronic idiopathic meningitis Neutral
Anderson et al11 1995 New Zealand 25 CT NA Chronic meningitis of unknown cause Against
Duna et al12 1995 USA 15 CT, MRI, DSA 47% PACNS Against
Javedan et al13 1997 USA 48 CT, MRI, DSA, 12% Progressive neurodegenerative Against
PET, fMRI disorders of unclear origin
Chu et al14 1998 USA 30 MRI, DSA NA PACNS Favor
Alrawi et al15 1999 USA 61 CT, MRI, MRA, 75% PACNS Favor
DSA
Warren et al16 2005 UK 90 MRI No Rapidly progressing dementia Neutral
Josephson et al17 2007 USA 64 NA NA Rapidly deteriorating neurological Neutral
condition/dementia
Pulhorn et al18 2008 UK 39 CT, MRI 22% Nonneoplastic undiagnosed Consider
neurological disorders
Venkateswaran 2008 UK 66 CT, MRI, DSA, NA Acute or chronic neurological Favor
et al19 SPECT deterioration
Burns et al20 2009 USA 42 MRI, MRA, DSA No Severe neurological disease of Favor
unknown etiology
Schott et al21 2010 UK 19 MRI NA Rapidly progressing dementia Neutral
Wong et al22 2010 UK 64 NA Yes Nonneoplastic disease Favor
Schuette et al23 2010 USA 51 MRI, DSA No Acute progressive neurological Against
decline
Rice et al24 2011 UK 52 MRI, MRA, DSA 2% Cryptogenic neurological disease Favor
Singh et al27 2014 UK 12 NA NA Undiagnosed neurological disorder Consider
Magaki et al25 2014 USA 53 NA NA Neurological disease of unknown Favor
etiology

a
CT, computed tomography; DSA, digital subtraction angiography; fMRI, functional magnetic resonance imaging; MRA, magnetic resonance angiogram; MRI, magnetic
resonance imaging; NA, not available; PACNS, primary angiitis of the central nervous system; PET, positron emission tomography; Preop, preoperative; SPECT, single-photon
emission computerized tomography.

Selection Criteria extracted included year of publication, country in which the study was
Studies were included if they reported (1) original data on patients conducted, duration of the study, cohort size, sex, age, indication for
undergoing brain biopsies, (2) diagnostic success rates, (3) the impact on biopsy, location of biopsy, imaging modality that was used for biopsy
patient management, and (4) rate of procedure-related complications (ie, planning, number of patients with a radiologically identifiable target,
morbidity and mortality). Studies were excluded if they (1) included diagnostic yield, detailed histopathological results, number of biopsies
patients with suspected neoplastic lesions or known immunocompromised that altered patient management or prognosis, and procedure-related
status, (2) included patients who had not undergone extensive investigations morbidity and mortality rates. Diagnostic success was defined as tissue
with less invasive diagnostic methods prior to biopsy, (3) focused on obtained via biopsy that was either diagnostic or suggestive of a specific
correlating other diagnostic modalities (eg, cytology, imaging, etc) with disease process. Nondiagnostic biopsies were defined as nonspecific
biopsy results, (4) reported a patient population of less than 10, (5) abnormalities (typically reactive gliosis) or normal tissue. The biopsy
presented a reanalysis of subpopulations that had already been included in was considered to have had clinical impact if the biopsy result guided
other studies, or (6) were commentaries, technical notes, or review articles clinical management or improved prognostic accuracy. In each study, the
summarizing the results of previous studies. Two authors (HXB and YZ) number of biopsies that were performed annually by each center was
examined the titles, abstracts, and full texts (when necessary) of all eligible estimated by dividing the number of cases by the duration of the study.
articles to identify those that met the inclusion criteria. When necessary, We used the middle year of the study duration as a rough estimate of the
a third author (LY) reviewed articles to reconcile any differences. time interval during which the study was conducted.

Data Extraction Statistical Analysis

For articles that were selected for data analysis, 2 or more authors read Study-specific proportions of outcome measures (diagnostic success,
the article in full text and performed data extraction. Parameters that were clinical impact, morbidity) were transformed by using the Freeman-

NEUROSURGERY VOLUME 77 | NUMBER 2 | AUGUST 2015 | 285

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

BAI ET AL

Tukey variant of arcsine square to attain the symmetry of confidence were published between 1990 and 2014. Most studies were
intervals that were necessary for statistical combination. Transformed conducted in the United States (n = 11)9,10,12-15,17,20,23,25,26 or
proportions were then combined by using the random-effects model. the United Kingdom (n = 7).16,18,19,21,22,24,27 Cohort size ranged
Consistency of findings across studies was assessed with I2 statistics. from 12 to 90 patients with a median of 40.5 patients per study. In
Heterogeneity was assessed with subgroup analyses by comparing results
77% of the studies (10/13), more than 1 imaging modality was
from studies that were grouped according to study-level characteristics.
Potential publication bias was assessed with funnel plots, Begg test, and
used for prebiopsy planning.12-15,18-20,23,24,26 The method of
Egger test. Sensitivity analysis was conducted by repeatedly calculating biopsy was stereotactic in 2 studies8,22 and open in 4 stud-
the effect size with 1 study being omitted per iteration. All analyses were ies.10,16,20,23 A combination of open and stereotactic biopsies was
performed with Comprehensive Meta-analysis Version 2 (Biostat, performed in 5 studies with the percentage of patients undergoing
Englewood, New Jersey). Statistical tests were 2-sided. Results were stereotactic biopsies ranging from 2% to 75%.12,13,15,18,24 The
considered to be statistically significant if P , .05. following conclusions were reached by the individual studies
regarding brain biopsy in patients with cryptogenic neurological
RESULTS disease: in favor (n = 9),9,10,14,15,19,20,22,24,25 may consider (n =
2),18,27 neutral (n = 5),8,16,17,21,26 and against (n = 4).11-13,23
Search Results
Overall, our search strategy retrieved 3645 potentially eligible Meta-analysis
publications (Figure 1). After screening titles and abstracts, 3558 The weighted proportions for diagnostic success, clinical
articles did not meet the criteria for inclusion in the meta-analysis. impact, and morbidity across the studies are shown in Figure 2.
Full texts were retrieved for 87 studies. Upon reviewing these studies, In detail, the weighted-average proportion per random-effects
70 were eliminated for not meeting inclusion criteria for the meta- modeling was 54.0% (95% confidence interval [CI]: 45.1%-
analysis. Three additional studies were identified by examining the 62.8%) for diagnostic success, 32.3% (95% CI: 20.8%-45.0%)
reference lists from publications retrieved. Twenty studies were for clinical impact, and 9.0% (95% CI: 6.5%-11.9%) for
included in the final analysis (Table 1). The indications for biopsy morbidity. After 8 studies were excluded given that preoperative
were severe neurological disease of unknown etiology (n = 9; 7 adult imaging was either not available (“NA”) or computed tomogra-
and 2 pediatric studies), primary angiitis of the central nervous phy (“CT”) alone (Table 1), the weighted-average proportion was
system (PACNS) (n = 3), chronic meningitis of unknown cause (n = 53.3% (95% CI: 41.3%-65.1%) for diagnostic success, 32.7%
3), atypical dementia (n = 4), and nonneoplastic disease (n = 1). We (95% CI: 18.0%-49.5%) for clinical impact, and 10.3% (95%
did not include studies where the indication for biopsy was suspected CI: 7.6%-13.3%) for morbidity.
herpes encephalitis (n = 2), tuberculoma (n = 2), or Creutzfeldt- Sensitivity analysis for random-effects models that calculated
Jakob disease (CJD) (n = 1) because either alternative methods of pooled proportions upon exclusion of single studies in turn showed
diagnosis have been well-established,29,30 or the management has similar results, indicating that overall estimates were not driven by
been shown to be unaffected by the results of the biopsies.31 the findings of single studies. Between-study heterogeneity was mild
for morbidity (I2 = 38%, 95% CI: 0%-66%) and high for
Study Population diagnostic success (I2 = 85%, 95% CI: 78%-90%) and clinical
Study and patient characteristics are summarized in Table 1. impact (I2 = 91%, 95% CI: 86%-94%). Funnel plots of study-
Our meta-analysis included a total of 831 patients. All studies specific proportions were approximately symmetrical and most

FIGURE 2. Forrest plots of outcome measures (A, diagnostic success; B, clinical impact; and C, morbidity) assessed in the present meta-analysis. Squares and horizontal bars
indicate point estimates and 95% confidence intervals of proportions in the individual studies. Diamonds indicate summary estimates calculated with random-effects
meta-analyses.

286 | VOLUME 77 | NUMBER 2 | AUGUST 2015 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

 BRAIN BIOPSY FOR CRYPTOGENIC NEUROLOGICAL DISEASE: SYSTEMATIC REVIEW AND META-ANALYSIS

data points were within the funnel area for diagnostic success and chronic meningitis (8.3%). Clinical impact rate was not reported
clinical impact, indicating low evidence for presence of publication in the PACNS studies. Lymphoma (n = 32) and CJD (n = 30)
bias. There was no statistical evidence of publication bias among were the most common diagnoses on the final histopathology
studies per the Begg test and Egger test for diagnostic success reports of positive brain biopsies in adult patients with
(P = .70 and .62, respectively) and clinical impact (P = .22 and .10, cryptogenic neurological disease. In children with cryptogenic
respectively). neurological disease, encephalitis (n = 7), PACNS (n = 6), and
demyelination (n = 6) were the most common. As expected,
Meta-regression PACNS (n = 36), chronic inflammation (n = 16) and Alzheimer
Meta-regression demonstrated a strong correlation between disease (n = 29) were the most common diagnoses when the
diagnostic success rates and the time interval during which the indications were PACNS, chronic meningitis, and atypical
study was performed (P = .007) and the presence of a radiological dementia, respectively.
target (P , .001) (Figure 3). Diagnostic success rates were not
correlated with patient age (P = .26), sex (P = .27), and the Target vs No Target
number of cases performed per year in each center (P = .75). Eight studies reported separate diagnostic success rates for brain
biopsies with an identified target on neuroimaging and those
Subgroup Analyses without.10,11,13,16,18,20,22,24 In cases without a target for biopsy, 4
The weighted proportions for diagnostic success and clinical studies sampled the nondominant frontal lobe13,16,18,22 and 1
impact for each of the subgroups were summarized in Figure 4. study sampled the frontal or temporal lobe8; it was the surgeon’s
The diagnostic success rates per indication were listed in preference in 1 study20; the other 2 studies did not report on the
descending order as follows: PACNS (74.7%), atypical dementia location chosen for biopsy in such cases.10,24 The odds ratio for
(60.8%), cryptogenic neurological disease in children (53.8%), achieving a diagnostic biopsy when there was a radiological target
cryptogenic neurological disease in adults (51.3%), and chronic was 3.63 (P = .003, 95% CI: 1.56%-8.41%) (Figure 5).
meningitis (30.3%) (Table 2). The rate of clinical impact per Between-study heterogeneity for this effect was moderate (I2 =
indication was listed in descending order as follows: cryptogenic 54%, 95% CI: 0%-79%). After 8 studies were excluded given
neurological disease in children (67.1%), cryptogenic neurolog- that preoperative imaging was either “NA” or “CT” alone, the
ical disease in adults (30.5%), atypical dementia (14.2%), and odds ratio was 3.28 (P = .012, 95% CI: 1.30%-8.24%).

FIGURE 3. Meta-regression of cases per year (A), middle year of the study period (B), sex (C), target (D), and age (E) on diagnostic success. Sizes of circles are proportional to
the inverse variance of the Freeman-Tukey transformed proportions.

NEUROSURGERY VOLUME 77 | NUMBER 2 | AUGUST 2015 | 287

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

BAI ET AL

FIGURE 4. Forrest plots of diagnostic success rate for each of the subgroups A, cryptogenic neurological disease in adults. B, cryptogenic neurological disease in children.
C, PACNS. D, chronic meningitis. E, atypical dementia. CI, confidence interval; PACNS, primary angiitis of the central nervous system.

Morbidity and Mortality Outside of this subset of patients, the sensitivity of brain
Overall, 63 procedure-related complications were reported in biopsies varies significantly from study to study. In 1979,
these studies. Details are shown in Table 2. Intraparenchymal Kaufman and Catalano reviewed 10 series of brain biopsy in
hemorrhage was the most common complication reported patients with unexplained neurological disorders that were
overall: 4 cases in adult cryptogenic neurological disease, 3 published from 1963 to 1979.33 They observed diagnostic
cases in PACNS, and 4 cases in atypical dementia. Wound yields as high as 43% and concluded that in carefully selected
infection was the most common complication in pediatric patients and in centers where tissue obtained could be
cryptogenic neurological disease and atypical dementia. There exhaustively examined, brain biopsy played a useful role. 33
were 9 cases of seizures and it was reported for all indications Since then, advances in neuroimaging and CSF/serological
except PACNS. Temporary hemiparesis occurred in 5 patients testing have significantly altered the indication and risk/
after biopsy. Three individual studies reported permanent benefit assessment of this procedure. Nonetheless, the
morbidity for 1 patient each.8,9,19 There was only 1 case of probability of achieving a definite diagnosis with potential
mortality reported, where the patient developed subdural to guide clinical management must be weighed against the
empyema and an aspiration pneumonia from which she risks of the procedure and the alternative option of empirical
ultimately died.22 In another study, 4 patients died within 12 therapy. Currently, limited evidence exists to guide clinicians
days of the biopsy; however, it was unclear if the causes of death in the difficult decision as to proceed to brain biopsy. Ideally,
were procedure related.11 one would want to know the odds of a biopsy having utility
for each specific preoperative clinical scenario, but the
DISCUSSION number of patients reported for any single indication is quite
small. In this review, we aimed to summarize the current
When definite diagnosis cannot be reached by less invasive literature on brain biopsy in patients with cryptogenic
means, brain biopsy is the last remaining method by which to neurological disease in the era of modern imaging. Based
make a premortem diagnosis. Its role in the diagnosis and on our results, we created a clinical algorithm for neuro-
management of neoplastic and infectious lesions in severely surgeons when consulted on patients with cryptogenic
immunocompromised patients has been well-established.2,4,32 neurological disease (Figure 6).

288 | VOLUME 77 | NUMBER 2 | AUGUST 2015 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

NEUROSURGERY

TABLE 2. Subgroup Analyses for the Selected Studiesa

Adult Cryptogenic Pediatric Cryptogenic PACNS Chronic Meningitis Atypical Dementia
No. of studies 7 2 3 3 4
No. of patients 349 78 106 83 151
Diagnostic success, % (95% CI) 51.3 (34.5-68.1) 53.8 (42.9-64.5) 74.7 (64.0-84.1) 30.3 (17.2-45.4) 60.8 (41.2-78.8)
Clinical impact, % (95% CI) 30.5 (13.6-50.6)b 67.1 (42.8-87.3) — 8.3 (2.3-20.0)c 14.2 (6.5-24.3)d
Final pathologye Lymphoma: 32 Encephalitis: 7 PACNS: 36 Chronic inflammation: 16 Alzheimer: 29
CJD: 30 PACNS: 6 Lymphoma: 10 Neoplasm: 11 CJD: 27
Demyelination: 21 Demyelination: 6 Alzheimer: 5 Sarcoid: 5 Pick: 6
Encephalitis: 20 TB: 4 Hypertensive changes: 5 Lymphoma: 2 Chronic encephalitis: 4
Alzheimer: 17 Rasmussen: 3 Abscess: 3 Wegener: 2 KUFS: 3
PACNS: 17 Infarct: 2 AVM: 2 PACNS: 3

BRAIN BIOPSY FOR CRYPTOGENIC NEUROLOGICAL DISEASE: SYSTEMATIC REVIEW AND META-ANALYSIS
Infarct: 10 SSPE: 2 Demyelination: 2 Lewy body: 2
Sarcoid: 6 Neoplasm: 2 Infarct: 2 MS: 2
Abscess: 3
Granulomatous disease: 3
PML: 3
Neoplasm: 2
Rheumatoid meningitis: 2
Complication IPH: 4 Infection: 2 IPH: 3 Seizure: 2 Wound infection: 5
Hemiparesis:3 Stroke: 2 Leg weakness: 1 Seizure: 4
SAH: 2 Seizure: 1 IPH: 4
Delirium: 2 Meningitis: 1 Bleeding: 3
Seizure: 2 Torticollis: 1 Abscess: 2
CSF leak: 1 Subdural effusion: 1 Hemiparesis: 1
Abscess: 1
VOLUME 77 | NUMBER 2 | AUGUST 2015 | 289

a
PACNS, primary angiitis of central nervous system; CJD, Creutzfeldt-Jakob Disease; IPH, intraparenchymal hemorrhage; CSF, cerebrospinal fluid; SAH, subarachnoid hemorrhage; PML, progressive multifocal
leukoencephalopathy; AVM, arteriovenous malformation; SSPE, subacute sclerosing panencephalitis; MS, multiple sclerosis; TB, tuberculosis.
b
Six studies.
c
One study.
d
Two studies.
e
Only diagnoses with .2 cases are included.

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

BAI ET AL

FIGURE 5. Odds ratio of target vs no target. CI, confidence interval.

Summary of Findings showed that the presence of a target on imaging was associated with
Diagnostic Success higher diagnostic success rates. This association was further
confirmed by pooling diagnostic success rates from studies that
We found an average diagnostic success rate of 54% across studies
reported success rates separately for patients with a biopsy target vs
in this meta-analysis. The wide range of diagnostic success rates
those without. The odds of achieving a diagnostic biopsy was 3.6
reported (from 20% to 83%) reflects the heterogeneity that exists
times higher with the presence of a radiological target than those
among studies. Through meta-regression analysis, we found that
without. Therefore, clinicians should expect a higher likelihood of
some of this heterogeneity can be explained by the time interval
obtaining diagnostic biopsy when a radiologically identified lesion
during which the study was performed and the presence or absence
can be successfully targeted.
of a radiological target for biopsy. Age, sex, and the number of
procedures performed per year had no impact on the heterogeneity.
Schott et al compared their experience of brain biopsy in patients Clinical Impact
with rapidly progressive dementia between 2 time intervals (1989- The value of diagnostic biopsies is debatable, if the identified
2003 and 2004-2009).21,33 They found that brain biopsy was less disease is not treatable. A nondiagnostic biopsy can still be of clinical
frequently required in the latter period, but still provided useful relevance. The exclusion of certain conditions may permit with-
diagnostic information in difficult cases.33 The diagnostic success drawal of potentially useless treatments and/or give clinicians the
rate was much higher in the latter period (74%) in comparison confidence to proceed with treatments that are otherwise contra-
with the earlier period (43%). Similarly, Venkateswaran et al19 indicated. For example, if the prebiopsy picture is consistent with
showed an improvement in the diagnostic yield of biopsies over PACNS, the patient should be treated with steroid and cyclophos-
their study period (1988-2003). In our meta-regression analysis, phamide.34 However, if the subsequent brain biopsy is non-
there was a clear trend of increasing diagnostic success rates over diagnostic, then the probability of a PACNS diagnosis is lowered.
the years. This may be explained by the improved surgical This will affect the decision to initiate immunosuppressive
techniques or the fact that fewer, but more selective patients were medications and the duration of treatment because the side effects
sent for brain biopsy because of improved diagnostic techniques. are significant. Another example is CJD. If the clinical presentation
The justification for nontargeted, open biopsy is usually based is consistent with CJD but the biopsy is nondiagnostic, the
on the assumption that the disease process is widespread, pro- probability of the patient having CJD is again lowered. In a study
gressive, and unremitting. In addition, there should be a reasonable of 26 patients with suspected CJD, 1 patient with a nondiagnostic
chance of obtaining diagnostic tissue. Some studies suggested that biopsy was later found to have Niemann-Pick type C, and another
diagnostic yield was not affected by the presence of discrete patients with a nondiagnostic biopsy was later found to have
imaging abnormalities.15,20 However, our meta-regression analysis amyotrophic lateral sclerosis and frontotemporal dementia.35

290 | VOLUME 77 | NUMBER 2 | AUGUST 2015 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

 BRAIN BIOPSY FOR CRYPTOGENIC NEUROLOGICAL DISEASE: SYSTEMATIC REVIEW AND META-ANALYSIS

FIGURE 6. A clinical algorithm for neurosurgeons consulting on a patient with cryptogenic neurological disease. CJD, Creutzfeldt-Jakob disease; CSF, cerebrospinal fluid;
DSA, digital subtraction angiography; PACNS, primary angiitis of the central nervous system.

To assess the overall benefit provided by brain biopsy beyond its uncommon. We found a morbidity rate of 9% across the studies.
ability to reveal a specific diagnosis, we defined the clinical impact This compares favorably with the complication rate reported for
of a biopsy as a meaningful change in treatment or an improved biopsy of suspected malignancy, where the incidence of post-
prognostic accuracy based on its results. We found an average operative complications may be as high as 12% (including
clinical impact rate of 32% through random-effects modeling. The a symptomatic hemorrhage rate of 3.8%),2,36 permanent mor-
reported clinical impact rate varied from 8% to 83% in individual bidity 5%,37 and mortality 1%.24 This likely reflects the
studies. We did not attempt to investigate the cause of this increased vascularity of tumors (particularly glioblastoma), thus
heterogeneity because the clinical usefulness of a diagnostic biopsy making hemorrhage more likely during biopsy.24 Because the
was often subjectively interpreted and the standards that were used number of complications was far and scarce in each study, only
to measure impact were different among studies. Even in patients a small number of studies had commented on the clinical finding
with a definite diagnosis but no impact on clinical management, that may have been associated with such complications. In a large
they may have gained emotional comfort simply with information study of 270 patients who underwent image-guided stereotactic
from the biopsy. Despite these limitations, our results suggested brain biopsy, McGirt et al38 found that diabetes mellitus and
that brain biopsy can be of clinical value. Clinicians should take thalamic and basal ganglia locations were independent risk factors
individual clinical scenario into consideration when discussing the for procedure-related complications. In another study of 225
risks and benefits of brain biopsy with their patients. CT-assisted stereotactic brain biopsies, Sawin et al39 reported an
increased risk of morbidity with preoperative use of antiplatelet
agents and chronic corticosteroids. In addition, deep-seated
Morbidity lesions, malignant gliomas, and a greater number of biopsy
In our meta-analysis, procedure-related mortality was zero. attempts were also associated with complications.39 From our
Permanent neurological deficit as a result of the biopsy was breakdown of morbidities by indication for biopsy, we observed

NEUROSURGERY VOLUME 77 | NUMBER 2 | AUGUST 2015 | 291

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

BAI ET AL

that some were more frequently encountered for certain Wilson40 conducted a retrospective review of pediatric brain
indications. For example, intraparenchymal hemorrhage ac- biopsies involving 45 children with neurological deterioration
counted for 75% of the complications in PACNS. This may between 1968 and 1974. The diagnostic yield in their study was
be due to the increased vascularity of these lesions. only 13%.40 Furthermore, most of the diagnoses could have been
determined with noninvasive methods today.40 A retrospective
Preoperative Imaging study by MacGregor et al41 looked at children who presented to
Among the 20 studies included in the meta-analysis, 8 studies a neurology service with history of developmental regression
had preoperative imaging as either “NA” or “CT.”8-11,17,22,25,27 between the years 1964 and 1976. Their diagnostic or prognostic
One study used CT alone for prebiopsy imaging,11 whereas the yield was 39%.41 However, if current state-of-the-art diagnostic
other 7 studies did not provide any information on prebiopsy techniques were available at the time of the study, the yield would
imaging. Three of the 8 studies that did not provide this decrease to 15% to 20%.41 In our subgroup analysis, the
information were from 1990 to 19948-10; 2 studies were reported diagnostic success rate of pediatric patients with cryptogenic
in abstract format only.20,21 Our calculations indicate that even neurological disease was 53.8%. This was slightly higher than
after excluding these 8 studies, diagnostic success, clinical impact, that for adult patients with cryptogenic neurological disease, but
and morbidity rates remained similar. After excluding the 8 lower than that of PACNS. However, the clinical impact rate
studies where preoperative imaging was either “NA” or “CT” (67.1%) was higher than that for any other subgroup. In the
alone, the odds ratio of achieving a diagnostic biopsy in the series of 66 children who underwent brain biopsies by
presence of radiological targets was slightly lower, albeit still Venkateswaran et al,34 the most frequently diagnosed disease
statistically significant. This indicates that a radiological target was vasculitis (18.2%), which is amenable to treatment with
seen on primitive imaging modality (ie, CT alone) is more likely immunosuppressants. Because pharmaceutical treatment of
to result in a diagnostic biopsy in comparison with lesions that central nervous system vasculitis is not benign, a definite
were identified on more than 1 imaging modality, if we assume diagnosis is necessary before starting a patient on treatment.
that those studies from the earlier years used CT alone. The Of the 6 patients who were diagnosed with vasculitis by brain
reason for this is unclear but may be related to selection bias. biopsy, 5 improved with treatment. Another common diagnosis
Further studies are needed to examine the effect of preoperative was infection with a specific agent (4 tuberculosis, 1 measles, 1
imaging on target identification and diagnostic success rates in aspergillosis, 1 cysticercosis, 1 toxoplasmosis). All of these have
patients with cryptogenic neurological disease who are referred highly effective treatment options.42 Similarly, inflammatory
brain biopsy. demyelination (n = 3) is responsive to steroids or disease-
modifying therapies.43 Even among the 34 nondiagnostic
biopsies, 12 (18.2%) were considered useful and necessary.19
Cryptogenic Neurological Disease in Adults
Consequently, high clinical impact of brain biopsy in children
The diagnostic success rate (51%) was similar to the overall could be attributed to both their physiological resilience to
diagnostic rate, but the heterogeneity was still high (I2 = 91%). recover from severe brain disease and the curability of the diseases
The study with the highest diagnostic yield (83%) included being diagnosed.
patients with rapidly deteriorating neurological conditions or
dementia so the inclusion criteria could have been broader in this Primary Angiitis of the Central Nervous System
study.17 In addition, the high proportion of biopsies with PACNS is an extremely rare and challenging disease to diagnose.
histopathological findings consistent with CJD may reflect Currently, there is no consensus on diagnostic criteria. The
referral bias and suggested that biopsy was used only as diagnostic criteria proposed by Calabrese and colleagues in 1988
a confirmatory test. In the study with the lowest diagnostic included preexisting neurological deficit that remained unex-
success rate (27%), more stringent inclusion criteria could have plained after complete evaluation, evidence of vascular damage
been applied because a few patients were reported to have within the central nervous system (CNS) detected by angiography
undergone more sophisticated imaging studies such as positron or biopsy, and exclusion of other differential diagnoses.44 Although
emission tomography, single-photon emission CT, and func- these criteria were not yet validated by large clinical trials, they
tional magnetic resonance imaging (MRI) before biopsy.13 The had been used widely for clinical reporting and investigation.
clinical impact rate of 30.5% was also similar to the overall Currently, there is no diagnostic laboratory study for PACNS.
impact rate. The most common histopathological diagnoses were MRI has a sensitivity of 90% to 100%, but poor specificity;
lymphoma and CJD. Other diagnoses such as demyelination, magnetic resonance angiography and computed tomography
encephalitis, Alzheimer disease, and PACNS were also common. angiography are noninvasive, but their resolution is often
inadequate for detecting the vasculitic changes in PACNS.45
Cryptogenic Neurological Disease in Children Angiography has a very low diagnostic yield for PACNS because
A limited number of studies have examined the diagnostic yield the classically described segmental stenoses are commonly seen in
of brain biopsy in children with cryptogenic neurological disease, nonvasculitic disorders such as atherosclerosis, radiation vasculop-
because biopsy is rarely done in children. Boltshauser and athy, infection, and vasospasm.34 The typical angiographic finding

292 | VOLUME 77 | NUMBER 2 | AUGUST 2015 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

 BRAIN BIOPSY FOR CRYPTOGENIC NEUROLOGICAL DISEASE: SYSTEMATIC REVIEW AND META-ANALYSIS

of PACNS consists of alternating areas of stenosis and dilation, of patients presenting with chronic meningitis, the etiology is not
referred to as beading.34 The sensitivity of angiography for the identified.11 In our meta-analysis, brain biopsy was associated
diagnosis of PACNS ranges from 40% to 90% and the specificity with the lowest diagnostic yield (30.3%) when the indication was
can be as low as 30%.34,45 In addition, PACNS typically affects to evaluate for chronic meningitis. Such low yields were likely
small cerebral arteries, which cannot be seen properly even with the because improved serological and CSF testing for herpes simplex
resolution of conventional angiography.34 Some patients with normal virus and other viruses had precluded the need for brain biopsy in
angiography were later found to have PACNS on brain biopsy.46 In these patients. Similarly, autoimmune and paraneoplastic causes
a French multicenter cohort study of patients with PACNS, 10 of the of encephalitis, such as anti-NMDAR, anti-LGI1/VGKC, and
15 patients with biopsy-proven PACNS had normal angiography anti-Hu, have well-validated diagnostic antibody tests in either
findings.46 In a correlative study by Kadhlodayan et al,47 none of serum or CSF.53,54 Brain biopsy is not recommended unless
the 14 patients with typical angiographic findings of vasculitis had thorough antibody testing and cancer screening are unreveal-
primary angiitis of the CNS on brain biopsy. Reversible cerebral ing.53,54 The clinical impact rate was also low (8.3%). Frequently,
vasoconstriction syndrome (RCVS) is the most common mim- a definite diagnosis can be achieved by sampling the meninges
icker of PACNS because of its similar angiographic appearance. alone.10 Furthermore, meningeal enhancement on MRI has a good
RCVS is differentiated from PACNS by clinical presentation predictive value. In the study by Cheng et al,10 80% of biopsies
(thunderclap headache is typical), normal CSF, and reversibility of obtained from enhancing lesions were diagnostic, whereas only 9%
the angiographic abnormalities shown by follow-up vascular of biopsies with negative imaging studies yielded diagnostic results.
imaging.34 RCVS is also frequently associated with vasoconstric- The second most common diagnosis on histopathology was
tive drug exposure or recent childbirth.48 It is crucial to neoplasm. Based on these results, we do not recommend
differentiate PACNS from RCVS because immunosuppressive intracranial biopsy in patients with chronic meningitis of unknown
therapy is not indicated in the latter. etiology.
Brain biopsy is considered the gold standard for diagnosis of
PACNS. Biopsy-positive PACNS differs from angiographically Atypical Dementia
positive PACNS in vessel size and clinical presentation.34,45,46,49 In
In most cases, the etiology of dementia can be determined
the 3 studies that were included in our meta-analysis, 2 studies
accurately by clinical criteria alone, and brain biopsy is unneces-
used brain biopsy “to rule out vasculitis”12,14; the indication was
sary. In patients with atypical presentations, the utility of tissue
“multifocal neurological deficits and signal abnormalities on MRI
diagnosis remains controversial. Brain biopsy may be used to rule
or CT of the brain” in the third study.15 We found that brain
out treatable causes of dementia such as encephalitis or chronic
biopsy had the highest diagnostic yield (74.7%) when the
meningitis.7 As early as 1979, Torack et al55 reviewed the role of
indication was to evaluate for suspected PACNS. However, brain
brain biopsy in the management of dementia, and observed that
biopsy may have false negatives because some patients were later
biopsy results rarely altered patient management or disease
diagnosed as PACNS based on a combination of clinical course
outcomes. In fact, one of the major criticisms of brain biopsy
and radiological evidence.12 In these false-negative brain biopsies,
in patients with dementia is that the diagnoses made are
lesions could have irregular involvement or were inaccessible.
frequently of diseases without any current treatment options.
Although these 3 studies did not report clinical impact, the impact
In agreement with this sentiment, we found a diagnostic success
was presumably high because PACNS frequently responds to
rate of 60.8% in patients with atypical dementia, but the clinical
aggressive immunosuppressive therapy. The morbidity rate was
impact rate was only 14.2%. In our analysis, the most common
3.8% for biopsy of suspected PACNS in our study, in comparison
diagnoses on histopathology were Alzheimer disease and CJD.
with 0.8% for angiography.50 Morbidity associated with aggressive
Consequently, we do not recommend biopsy in this subset of
immunosuppression has been shown to be significantly greater
patients at this time. This may change in the future should
than those with cerebral angiography or brain biopsy.14 The
disease-modifying therapies become available for neurodegener-
second most common diagnosis was lymphoma. Previous studies
ative disorders such as Alzheimer disease.
have shown that CNS lymphoma and intravascular lymphoma can
be difficult to distinguish from PACNS based on clinical
presentation and imaging characteristics alone.51 Based on these Limitations
results, we recommend that, in patients with suspected PACNS, Our study was limited owing to the retrospective nature of the
accurate diagnosis should not rely on any single study. Diagnosis included studies and the small number of patients in each. We
should only be made after careful correlations among clinical, were unable to control for subtle differences in clinical indications
radiographical, and pathological findings. that prompted biopsy among studies (selection bias), differences in
the population studied (institutional referral bias), the quality,
Chronic Meningitis of Unknown Etiology extent, and amount of tissue that were obtained at biopsy, and the
Chronic meningitis can be caused by a variety of infectious and extent of the pathological analysis. Moreover, the advance in
noninfectious processes. It is arbitrarily defined as meningitis that medical imaging and surgical techniques over time may have
lasts for at least 4 weeks without any improvement.52 In one-third affected the indications for brain biopsy, as well as its safety and

NEUROSURGERY VOLUME 77 | NUMBER 2 | AUGUST 2015 | 293

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

BAI ET AL

utility. Incomplete and/or inconsistent recording of data in the 8. Waltregny A, Maula AA, Brucher JM. Contribution of stereotactic brain biopsies
to the diagnosis of presenile dementia. Stereotact Funct Neurosurg. 1990;54-55:
individual studies may have limited the validity of the performed
409-412.
meta-regression analysis. The small number of studies in certain 9. Hulette CM, Earl NL, Crain BJ. Evaluation of cerebral biopsies for the diagnosis of
subgroups (eg, pediatric patients with cryptogenic neurological dementia. Arch Neurol. 1992;49(1):28-31.
disease) predisposed our meta-analysis to publication bias. Because 10. Cheng TM, O’Neill BP, Scheithauer BW, Piepgras DG. Chronic meningitis: the
role of meningeal or cortical biopsy. Neurosurgery. 1994;34(4):590-595; discussion
larger prospective studies are impractical for these groups of
596.
patients, our meta-analysis provided important data for neurolo- 11. Anderson NE, Willoughby EW, Synek BJ. Leptomeningeal and brain biopsy in
gists and neurosurgeons alike. We hope that our findings will lead chronic meningitis. Aust N Z J Med. 1995;25(6):703-706.
to further research in the optimal use of brain biopsy in clinical 12. Duna GF, Calabrese LH. Limitations of invasive modalities in the diagnosis of
primary angiitis of the central nervous system. J Rheumatol. 1995;22(4):662-667.
practice. 13. Javedan SP, Tamargo RJ. Diagnostic yield of brain biopsy in neurodegenerative
disorders. Neurosurgery. 1997;41(4):823-828; discussion 828-830.
14. Chu CT, Gray L, Goldstein LB, Hulette CM. Diagnosis of intracranial vasculitis:
CONCLUSION a multi-disciplinary approach. J Neuropathol Exp Neurol. 1998;57(1):30-38.
15. Alrawi A, Trobe JD, Blaivas M, Musch DC. Brain biopsy in primary angiitis of the
Despite other diagnostic advances, brain biopsy remains a useful central nervous system. Neurology. 1999;53(4):858-860.
tool in establishing a definite diagnosis in patients with crypto- 16. Warren JD, Schott JM, Fox NC, et al. Brain biopsy in dementia. Brain. 2005;128
genic neurological disease. In our meta-analysis, the indication to (pt 9):2016-2025.
17. Josephson SA, Papanastassiou AM, Berger MS, et al. The diagnostic utility of brain
evaluate for suspected PACNS was associated with the highest biopsy procedures in patients with rapidly deteriorating neurological conditions or
diagnostic yield. Brain biopsy had the greatest clinical impact in dementia. J Neurosurg. 2007;106(1):72-75.
pediatric patients with cryptogenic neurological disease, but the 18. Pulhorn H, Quigley DG, Bosma JJ, Kirollos R, du Plessis DG, Jenkinson MD. Impact
impact was limited in patients with atypical dementia or chronic of brain biopsy on the management of patients with nonneoplastic undiagnosed
neurological disorders. Neurosurgery. 2008;62(4):833-837; discussion 837-838.
meningitis of unknown etiology. Brain biopsy should be avoided 19. Venkateswaran S, Hawkins C, Wassmer E. Diagnostic yield of brain biopsies in
in patients whose clinical presentation is consistent with CJD or children presenting to neurology. J Child Neurol. 2008;23(3):253-258.
Alzheimer disease, given that even a diagnostic biopsy will unlikely 20. Burns JD, Cadigan RO, Russell JA. Evaluation of brain biopsy in the diagnosis of
change patient management. Additionally, brain biopsies with severe neurologic disease of unknown etiology. Clin Neurol Neurosurg. 2009;111
(3):235-239.
radiological targets were associated with higher diagnostic yields 21. Schott JM, Reiniger L, Thom M, et al. Brain biopsy in dementia: clinical
than those without. indications and diagnostic approach. Acta Neuropathol. 2010;120(3):327-341.
22. Wong SH, Jenkinson MD, Faragher B, Thomas S, Crooks D, Solomon T. Brain
biopsy in the management of neurology patients. Eur Neurol. 2010;64(1):42-45.
Disclosures 23. Schuette AJ, Taub JS, Hadjipanayis CG, Olson JJ. Open biopsy in patients with
This work was supported by the Natural Science Foundation of China (grant acute progressive neurologic decline and absence of mass lesion. Neurology. 2010;
81301988) to Dr Yang, and China Ministry of Education Doctoral Program Spot 75(5):419-424.
Foundation (grant 20130162120061) to Dr Yang. Dr Lancaster received 24. Rice CM, Gilkes CE, Teare E, Hardie RJ, Scolding NJ, Edwards RJ. Brain biopsy
in cryptogenic neurological disease. Br J Neurosurg. 2011;25(5):614-620.
compensation for teaching courses for Grifols Inc, and has provided paid expert
25. Magaki S, Yong WH, Khanlou N, Tung S, Vinters HV. Comorbidity in dementia:
opinions for the National Vaccine Injury Compensation Program (VICP). The update of an ongoing autopsy study. J Am Geriatr Soc. 2014;62(9):1722-1728.
authors have no personal, financial, or institutional interest in any of the drugs, 26. Smith JE, Aksamit AJ Jr. Outcome of chronic idiopathic meningitis. Mayo Clin
materials, or devices described in this article. Proc. 1994;69(6):548-556.
27. Singh A, Iyer A, Burn S, et al. Brain biopsy in children being investigated for
REFERENCES undiagnosed neurological conditions. Arch Dis Child. 2014;99(suppl 1):A1-A212.
28. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items
1. Callovini GM. Is it appropriate to redefine the indication for stereotactic brain for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med.
biopsy in the MRI Era? Correlation with final histological diagnosis in 2009;6(7):e1000097.
supratentorial gliomas. Minim Invasive Neurosurg. 2008;51(2):109-113. 29. Lakeman FD, Whitley RJ. Diagnosis of herpes simplex encephalitis: application of
2. Livermore LJ, Ma R, Bojanic S, Pereira EA. Yield and complications of frame- polymerase chain reaction to cerebrospinal fluid from brain-biopsied patients and
based and frameless stereotactic brain biopsy—the value of intra-operative correlation with disease. National Institute of Allergy and Infectious Diseases
histological analysis. Br J Neurosurg. 2014;28(5):637-644. Collaborative Antiviral Study Group. J Infect Dis. 1995;171(4):857-863.
3. Krieger MD, Chandrasoma PT, Zee CS, Apuzzo ML. Role of stereotactic biopsy 30. Stoeck K, Sanchez-Juan P, Gawinecka J, et al. Cerebrospinal fluid biomarker
in the diagnosis and management of brain tumors. Semin Surg Oncol. 1998;14(1): supported diagnosis of Creutzfeldt-Jakob disease and rapid dementias: a longitu-
13-25. dinal multicentre study over 10 years. Brain. 2012;135(pt 10):3051-3061.
4. Skolasky RL, Dal Pan GJ, Olivi A, Lenz FA, Abrams RA, McArthur JC. HIV- 31. Wasay M, Farooq S, Khowaja ZA, et al. Cerebral infarction and tuberculoma in
associated primary CNS lymorbidity and utility of brain biopsy. J Neurol Sci. 1999; central nervous system tuberculosis: frequency and prognostic implications.
163(1):32-38. J Neurol Neurosurg Psychiatry. 2014;85(11):1260-1264.
5. Antinori A, Ammassari A, Luzzati R, et al. Role of brain biopsy in the management 32. Hornef MW, Iten A, Maeder P, Villemure JG, Regli L. Brain biopsy in patients
of focal brain lesions in HIV-infected patients. Gruppo Italiano Cooperativo AIDS with acquired immunodeficiency syndrome: diagnostic value, clinical perfor-
& Tumori. Neurology. 2000;54(4):993-997. mance, and survival time. Arch Intern Med. 1999;159(21):2590-2596.
6. Rosenow JM, Hirschfeld A. Utility of brain biopsy in patients with acquired 33. Kaufman HH, Catalano LW Jr. Diagnostic brain biopsy: a series of 50 cases and
immunodeficiency syndrome before and after introduction of highly active a review. Neurosurgery. 1979;4(2):129-136.
antiretroviral therapy. Neurosurgery. 2007;61(1):130-140; discussion 34. Hajj-Ali RA, Singhal AB, Benseler S, Molloy E, Calabrese LH. Primary angiitis of
140-141. the CNS. Lancet Neurol. 2011;10(6):561-572.
7. Gilkes CE, Love S, Hardie RJ, Edwards RJ, Scolding NJ, Rice CM. Brain biopsy 35. Heinemann U, Krasnianski A, Meissner B, et al. Brain biopsy in patients with
in benign neurological disease. J Neurol. 2012;259(5):995-1000. suspected Creutzfeldt-Jakob disease. J Neurosurg. 2008;109(4):735-741.

294 | VOLUME 77 | NUMBER 2 | AUGUST 2015 www.neurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited

 BRAIN BIOPSY FOR CRYPTOGENIC NEUROLOGICAL DISEASE: SYSTEMATIC REVIEW AND META-ANALYSIS

36. Field M, Witham TF, Flickinger JC, Kondziolka D, Lunsford LD. Comprehensive 51. Song DK, Boulis NM, McKeever PE, Quint DJ. Angiotropic large cell lymphoma
assessment of hemorrhage risks and outcomes after stereotactic brain biopsy. with imaging characteristics of CNS vasculitis. AJNR Am J Neuroradiol. 2002;23
J Neurosurg. 2001;94(4):545-551. (2):239-242.
37. McGirt MJ, Woodworth GF, Coon AL, et al. Independent predictors of morbidity 52. Helbok R, Broessner G, Pfausler B, Schmutzhard E. Chronic meningitis.
after image-guided stereotactic brain biopsy: a risk assessment of 270 cases. Neurology. 2009;256(2):168-175.
J Neurosurg. 2005;102(5):897-901. 53. Lancaster E, Dalmau J. Neuronal autoantigens—pathogenesis, associated disorders
38. McGirt MJ, Woodworth GF, Coon AL, et al. Independent predictors of morbidity and antibody testing. Nat Rev Neurol. 2012;8(7):380-390.
after image-guided stereotactic brain biopsy: a risk assessment of 270 cases. 54. Lancaster E, Martinez-Hernandez E, Dalmau J. Encephalitis and antibodies to
J Neurosurg. 2005;102(5):897-901. synaptic and neuronal cell surface proteins. Neurology. 2011;77(2):179-189.
39. Sawin PD, Hitchon PW, Follett KA, Torner JC. Computed imaging-assisted 55. Torack RM. Adult dementia: history, biopsy, pathology. Neurosurgery. 1979;4(5):
stereotactic brain biopsy: a risk analysis of 225 consecutive cases. Surg Neurol. 434-442.
1998;49(6):640-649.
40. Boltshauser E, Wilson J. Value of brain biopsy in neurodegenerative disease in
childhood. Arch Dis Child. 1976;51(4):264-268. COMMENT
41. MacGregor DL, Humphrey RP, Armstrong DL, Becker LE. Brain biopsies for
neurodegenerative disease in children. J Pediatr. 1978;92(6):903-905.
42. Schoeman JF, Van Zyl LE, Laubscher JA, Donald PR. Effect of corticosteroids
on intracranial pressure, computed tomographic findings, and clinical
I n this article, the authors present an interesting meta-analysis and literature
review examining the risk and benefits of diagnostic brain biopsy in pa-
tients with cryptogenic neurological disease. Twenty studies were included in
outcome in young children with tuberculous meningitis. Pediatrics. 1997;
99(2):226-231.
the final analysis, with a total of 831 patients. The indications for biopsy were
43. Pohl D, Tenembaum S. Treatment of acute disseminated encephalomyelitis. Curr severe neurological disease of unknown etiology, primary angiitis of the CNS,
Treat Options Neurol. 2012;14(3):264-275. chronic meningitis, atypical dementia, and nonneoplastic disease. Diagnostic
44. Calabrese LH, Mallek JA. Primary angiitis of the central nervous system. Report of success and clinical impact rates were then evaluated and calculated for each
8 new cases, review of the literature, and proposal for diagnostic criteria. Medicine. of these subgroups. The authors found that brain biopsy in cryptogenic
1988;67(1):20-39. neurological disease was associated with the highest diagnostic yield (74.7%)
45. Rodriguez-Pla A, Monach PA. Primary angiitis of the central nervous system in
adults and children. Rheum Dis Clin North Am. 2015;41(1):47-62, viii.
in patients with suspected PACNS. Moreover, the greatest clinical impact
46. de Boysson H, Zuber M, Naggara O, et al. Primary angiitis of the central nervous rate was seen in children with cryptogenic neurological disease (67.1%),
system: description of the first fifty-two adults enrolled in the French cohort of which can likely be attributed to both the curability of the diseases being
patients with primary vasculitis of the central nervous system. Arthritis Rheumatol. diagnosed with immunosuppressants, as well as the general physiological
2014;66(5):1315-1326. resilience associated with this patient population. Finally, the authors noted
47. Kadkhodayan Y, Alreshaid A, Moran CJ, Cross DT III, Powers WJ, Derdeyn CP. that the presence of a radiological target was associated with a higher diag-
Primary angiitis of the central nervous system at conventional angiography.
Radiology. 2004;233(3):878-882.
nostic yield (OR = 3.70, P = .014). Although the latter finding is somewhat
48. Yang L, Bai HX, Zhao X, Xiao Y, Tan L. Postpartum cerebral angiopathy intuitive, it still provides value more so in those patients without a radio-
presenting with non-aneurysmal subarachnoid hemorrhage and interval develop- graphic target. While the clinical utility of biopsy appears highest in the
ment of neurological deficits: a case report and review of literature. Neurol India. pediatric population, perhaps regardless of age we should be less inclined to
2013;61(5):517-522. offer biopsy for those patients without a clear radiographic target.
49. Giannini C, Salvarani C, Hunder G, Brown RD. Primary central nervous system
vasculitis: pathology and mechanisms. Acta Neuropathol. 2012;123(6):759-772.
Aria Jamshidi
50. Hellmann DB, Roubenoff R, Healy RA, Wang H. Central nervous system
angiography: safety and predictors of a positive result in 125 consecutive patients Jonathan H. Sherman
evaluated for possible vasculitis. J Rheumatol. 1992;19(4):568-572. Washington, DC

NEUROSURGERY VOLUME 77 | NUMBER 2 | AUGUST 2015 | 295

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited