Diagnostic yield of stereotactic needle-biopsies of sub-cubic centimeter intracranial lesions

Abstract

Background:

Stereotactic brain biopsies are widely used for establishing the diagnosis of intracranial lesions. Here we examine whether stereotactic biopsy of smaller brain lesions, defined for this study as being less than 1 cubic centimeter (1 cc) in volume, are associated with lowered diagnostic yield.

Methods:

We conducted a retrospective analysis of 267 consecutive patients who underwent stereotactic brain biopsy between 2007 and 2011. Lesion volumes were calculated and were stratified by <1 or >1 cc.

Results:

A total of 13 of 246 (5.2%) biopsies for lesions >1 cc resulted in nondiagnostic tissue or an incorrect diagnosis. In contrast, 5 of 21 (23.8%) biopsies for <1 cc lesions yielded nondiagnostic or incorrect diagnosis. Posthoc review of tissue from the <1 cc lesions suggests the neuropathologist's expertise in the handling and analysis of limited specimen as a critical parameter of successful diagnosis. The operative morbidities were low for both the <1 and >1 cc biopsies (0% and 1%, respectively).

Conclusion:

This study demonstrates that stereotactic cerebral biopsy of lesions less than a cubic centimeter in volume results in a lower diagnostic yield versus larger lesions (76.2% versus 94.8%). While auxiliary measures may be taken to improve diagnostic yield, these patients may be best managed in a specialized center with experienced stereotactic neurosurgeons and neuropathologists.

INTRODUCTION

Stereotactic-guided needle biopsy is a well-accepted method for obtaining tissue diagnosis for intracranial lesions that are not amenable to surgical resection. The accuracy, safety, and diagnostic yield of stereotactic needle-biopsies have been well established in the neurosurgical literature.[2,4,9,12,20,22,23] In terms of mechanical precision, stereotactically guided techniques have demonstrated an accuracy ranging from 1.2 to 2.8 mm.[4,24] In terms of safety, complications related to stereotactic biopsy ranged 1-8%.[1,2,5,8,10,21,23] In terms of diagnostic yield, several large series suggest that tissue diagnosis can be attained in >90% of the biopsies.[2,5,8,9,21,23]

Despite this extended literature, there has not been a study that looked specifically at how the size of the target influenced diagnostic yield. There are many reasons to expect lowered diagnostic yield for the smaller lesions. First, a smaller target will magnify the effect of any degree of mechanical deviation, however slight, within the stereotactic system. Second, tissues attained in the biopsy of smaller lesions are typically more limited in quantity, and the diagnostic yield directly correlates with the quantity of pathologic specimen secured and the expertise of the neuropathologist.[2,9,14,23] Finally, the limited specimens may be more demanding in terms of the pathologist's expertise.[15] Here we report our experience in stereotactic brain biopsies of the smaller lesions (defined as < 1 cm³) and compared the diagnostic yield of such biopsies to the larger lesions (defined as > 1 cm³).

METHODS

Study population

Electronic records from 267 consecutive patients who underwent stereotactic needle-biopsy from 2007 to 2011 by PW and CC were retrospectively reviewed. Information was collected regarding final pathology, morbidity, and indications. Magnetic resonance (MR) images for each case were imported into the Inomed system (Stereoplan Plus, Germany) for volumetric calculation. Based on this calculation, lesions were stratified into <1 cm³ or >1 cm³. For each case, postoperative computed tomography (CT) was evaluated for evidence of new hyper-density at the biopsy site. For clinical follow-up, transient neurologic deficit was defined as an altered postoperative neurologic examination that resolved within a month of surgery. This study was approved by the institutional review boards under IRB#2010-P-000134.

Surgical procedure

After induction of anesthesia, a Riechert/Mundinger stereotactic head frame (Inomed GmbH, Emmendingen, Germany) was secured onto the patient's cranium. A 1.25 mm slice-thickness contrast-enhanced CT imaging of the head was subsequently acquired using either an intraoperative scanner (Ceretom, Neurologica, MA) or a conventional CT scanner. CT images were processed to yield three-dimensional reconstructions using software by Inomed (Stereoplan Plus, Germany). Using these reconstructions, an optimal trajectory through the lesion was planned to avoid intersection of vasculature or sulci. In select cases, a positron emission tomographic (PET) scan was performed to define the hyper-metabolic area as a target. Image fusion of the CT, MRI, and PET where available was then performed.

After mounting the aiming bow to the head frame, a scalp incision (approximately 3 cm in length) and burr hole were placed at the planned entry site. Biopsy forceps (1.0 or 1.4 mm) were then advanced to the lesion under the guidance of the aiming bow. Serial biopsy through the entire lesion was performed. Depending on the size of the lesion, 2 to >10 biopsies were taken in 1 mm intervals.

Posthoc pathology review

For the <1 cm³ lesions that were biopsied and yielded nondiagnostic tissue or misdiagnosis, we retrieved the original slides as well as slides from the repeat biopsy or resection. The slides were reviewed with an independent neuropathologist (HR) who was not involved in the initial diagnosis. “Limited tissue” was used to designate situations where the pathologist felt that the tissue obtained during the first biopsy was lesional but definitive diagnosis was not possible due to tissue limitation.

RESULTS

Of the 267 patients, 21 (7.9%) had lesions that were <1 cm³. A summary of the final pathologic diagnosis for the 267 biopsied cases as stratified by <1 cm³ and >1 cm³ volume is presented in Table 1. A total of 246 (92%) of the biopsies were performed for lesions >1 cm³ and 21 (8%) were for lesions <1 cm³. Irrespective of lesion volume, the most frequently encountered category of diagnosis was tumor.

Table 1.

Final pathologic diagnosis of biopsied cases

Of the >1 cm³ biopsies, 5.2% (13/246) resulted in nondiagnostic tissue or incorrect diagnosis [Table 2]. All lesions were contrast enhancing on MRI. Of these, six biopsies (6/246 or 2.4%) yielded a diagnosis that was revised upon resection of the lesion, and seven biopsies (7/246 or 2.8%) yielded nondiagnostic tissues. Cases of misdiagnosis most often involved mis-staged glial tumors. The remaining nondiagnostic biopsies were performed for presumptive diagnoses related to inflammatory or infectious diseases [Table 2].

Table 2.

Incorrect diagnosis

All 21 lesions <1 cm³ were also contrast enhancing. Figure 1a–i illustrates representative radiographs from cases with a definitive diagnosis for lesions <1 cm³. The indications for biopsy of sub-cubic centimeter lesions are listed in Table 3. A total of 23.8% (5/21) of the <1 cm³ biopsies resulted in nondiagnostic or incorrect diagnosis. Of these cases, one biopsy (1/21 or 4.7%) yielded the diagnosis of anaplastic astrocytoma that was revised to sub-ependymoma upon resection (patient 21 in Table 3, Figure 2a). Four biopsies (4/21 or 19%) yielded nondiagnostic tissues. In one case, a second biopsy was performed to yield the diagnosis of a B cell lymphoma (patient 17 in Table 3, Figure 2b). In two other cases, surgical resection was performed due to enlargement of the lesions on serial MRI. Definitive diagnosis was made using the resected tissues (radiation necrosis and pilocytic astrocytoma, patients 8 and 12 in Table 3, Figure 2c and d). In the last case, the patient with the nondiagnostic biopsy was followed by serial MRIs. The patient initially presented with seizure and was treated with antiseizure medications without further neurologic events. The contrast enhancement resolved spontaneously at the 3-year follow-up (patient 13 in Table 3, Figure 2e).

Figure 1.

Illustrative cases with definitive diagnosis (a) Grade III oligoastrocytoma (b) Anaplastic astrocytoma (c) Grade III oligoastrocytoma (d) Anaplastic astrocytoma (e) Metastatic squamous cell carcinoma (f) Anaplastic astrocytoma (g) Metastatic carcinoma (h) Hodgkin's lymphoma (i) Metastasis

Table 3.

Indications for the sub-centimeter biopsies

Figure 2.

Cases of misdiagnosis (a) Misdiagnosed subependymoma (initial biopsy showed anaplastic astrocytoma, subsequent resection showed subependymoma. (b) Initial biopsy nondiagnostic. Repeat biopsy showed lymphoma (c) Radiation necrosis (d) Nondiagnostic (resection showed pilocytic astrocytoma) (e) Nondiagnostic

In terms of achieving a definitive diagnosis, the gross diagnostic yield was 94.8% and 76.2% for the >1 cm³ and the <1 cm³ lesions, respectively. The difference in diagnostic yield was statistically significant (P = 0.0081, Fisher's Exact Test).

Posthoc analysis of the tissue specimens from these lesions suggests that the primary cause for diagnostic failure was due to the limited amount of tissue [Table 4]. For all four cases, lesional tissue was obtained on the initial biopsies, but the limited specimens led to incorrect or indefinite diagnosis (patients 8, 12, 17, and 21 in Tables 3 and 4). For instance, the initial pathology report for patient 8 [Table 4] was nondiagnostic, but included pilocytic astrocytoma as a potential diagnosis. This diagnosis was confirmed based on the larger specimens secured through an open resection. Similar situations were encountered for cases 12 and 17. Case 21, where a subependymoma was initially misdiagnosed as an anaplastic astrocytoma, warrants additional comment. The initial diagnosis was made based on focal hypercellular regions thought to represent anaplastic astrocytoma. On subsequent resection, the lesion showed pathologic findings classic for subependymoma, without evidence of hypercellularity. The pathologic findings of the original specimen were subsequently attributed to artifact related to sample processing.

Table 4.

Reasons for mis-diagnosis in the biopsy of <1 cc lesions

Morbidities associated with the biopsies did not significantly differ based on the lesion size [Table 5]. No patients suffered from permanent neurologic deficit or death in either group. A total of 3 (0.8%) of 246 patients with >1 cm³ lesions had a transient neurologic deficit postbiopsy that resolved completely by 1 month follow-up. All three of these lesions were located in eloquent cortex.

Table 5.

Morbidities associated with stereotactic biopsies of lesions >1 cc or <1 cc in size

Postoperative CT demonstrated hyper-density at the site of biopsy in 15% (37/246) of the patients with >1 cm³ lesion and 9.5% (2/21) of the patients with <1 cm³ lesion (P = 0.7485, Fisher's Exact test). With the exception of the above noted cases, the hyper-density did not contribute to clinically detectable changes in neurologic examination.

DISCUSSION

This study represents the first to our knowledge to evaluate the diagnostic yield of stereotactic brain biopsy of small lesions (<1 cm³) relative to larger lesions. In our series, diagnostic yield from biopsies of lesions >1 cm³ (94.8%) was comparable to previously published rates of 90-96% for stereotactic brain biopsies without size stratification.[5,8,21] Biopsies of the <1 cm³ lesions, in contrast, were associated with a lowered diagnostic yield relative to the >1 cm³ lesions (76.2%, P = 0.0081).

Our study contributes to the field of neurosurgical oncology in two ways. First, our study suggest that the risk of a nondiagnostic biopsy for <1 cm³ lesions are significantly higher than most biopsies in general.[5,8,21] This is an issue that should be carefully reviewed during the informed consent for biopsy of <1 cm³ intracranial lesions. Second, given the inherent technical challenges associated with such biopsies, referral to centers where experienced stereotactic neurosurgeons and neuro-pathologists are available warrants consideration.

Posthoc review of the tissue specimen for the mis-diagnostic or nondiagnostic cases offered an opportunity to investigate the underlying cause. The observations that lesional tissue was obtained during the first biopsy and that the differential diagnosis generated using these specimens included the final diagnosis support the accuracy of the frame-based stereotactic biopsy. In this context, the major determinant of diagnosis for the <1 cm³ lesion appeared to be the familiarity of the pathologist in the handling and analysis of limited specimens. A smaller lesion necessarily translated into less tissue for examination, and the published literature suggested that diagnostic yield is a function of the amount of available specimen.[14] Further, sub-optimal processing of limited specimens may lead to artifacts prohibitive of definitive diagnosis.[15] Overall, our results suggest the neuropathologist's expertise in the handling and analysis of limited specimen as a critical determinant of diagnosis.

There are several limitations to this study. First and foremost, this study included patients treated by two surgeons (CC and PW) at a single institution. As such, the results presented here may not be broadly applicable. Second, the size criteria for small versus large lesions using a 1 cm³ volumetric cut off was somewhat arbitrary. Third, the number of <1 cm³ biopsies performed are fairly limited, constituting only 7.9% of all biopsies performed. Realizing the small sample size, we nevertheless performed this analysis of the patients accumulated over the 3-year interval with the goal of assessing the diagnostic yield in a timely manner. Finally, the retrospective design means that patient selection bias cannot be entirely excluded as the cause of the differential diagnostic yield. Despite these limitations, the implications of our findings contribute to the management of patients with a small intracranial lesion.

Many adjuncts to the biopsy technique and recent advances in technology may improve diagnostic yield in the biopsy of small lesions.[3,6,7,11,12,16–20] Notably, intraoperative MRI-based biopsies afford the opportunity to directly visualize the region of biopsy relative to the lesion as well as ascertainment of potential hemorrhage at the biopsy site.[3] Such information may be helpful in redirecting biopsy sites or may enable a greater number of biopsy specimens and thereby improve the diagnostic yield. Diagnosis through molecular analysis of limited specimen may also afford opportunities to enhance diagnostic yield.[13] The application of such technology to augment the diagnostic yield of challenging intracranial targets awaits investigation.

CONCLUSIONS

This study demonstrates that stereotactic cerebral biopsy of < 1 cm³ intracranial lesions is associated with a lower diagnostic yield relative to the >1 cm³ lesions (76.2% versus 94.8%, respectively). Morbidities associated with biopsy in both groups are comparable at approximately 1%. Our findings identify the neuropathologist's expertise in the handling and analysis of limited specimen as a critical determinant of tissue diagnosis. Our findings also support the accuracy of frame-based stereotactic biopsy in tissue acquisition. Given the technical challenges associated with these biopsies, consideration should be given for referral of patients with such lesions to centers where experienced stereotactic neurosurgeons and neuropathologists are available.