- Email: [email protected]
Craniocervical Junction and Posterior Fossa Dimensions can Affect Need for Decompressive Craniectomy in Posterior Cranial Fossa Hemorrhage
Original Article
Craniocervical Junction and Posterior Fossa Dimensions can Affect Need for Decompressive Craniectomy in Posterior Cranial Fossa Hemorrhage Ethan A. Neufeld1, Sarah T. Menacho2, Lubdha M. Shah1
BACKGROUND: Posterior fossa hemorrhage (PFH) of the cerebellum is managed by decompressive craniectomy when there is clinical deterioration. There is no current consensus on an objective imaging method to determine which patients need surgery before clinical deterioration. We developed an imaging scoring tool by assessing initial hemorrhage diameter and posterior fossa (PF) measurements to determine which patients will benefit from early surgical intervention.
-
METHODS: For this caseecontrol study, we reviewed the electronic medical record to identify adults who presented with PFH over a 10-year period at our institution. Chart review for clinical findings and inciting factors were documented. The average diameter of PFH and the surrogate PF volume on initial imaging studies were measured. These measurements were correlated with surgical intervention. A scoring tool was developed based on radiographic and clinical data.
CONCLUSIONS: Patients presenting with PFH with smaller PF volumes may be more likely to require surgery as determined by clinical standards. The proposed scoring system based on simple measurements on initial computed tomography and magnetic resonance imaging may help surgeons consider early surgical intervention in those patients with PFH with smaller PF volumes.
-
-
RESULTS: Fifty-one patients met the inclusion criteria. The average hemorrhage diameter and the surrogate PF volume measurements were statistically different between surgical and nonsurgical cases (P < 0.001 and P [ 0.019, respectively). The scoring system was created by dividing average hemorrhage diameter by surrogate PF volume and multiplying by 1000. The median score of nonsurgical patients was 9.1, and the median score of surgical patients was 15.6.
-
Key words Cerebellum - Cerebrovascular disorders - Posterior cranial fossa - Stroke -
Abbreviations and Acronyms CCJ: Craniocervical junction CT: Computed tomography EMR: Electronic medical record GCS: Glasgow Coma Scale MRI: Magnetic resonance imaging MPRAGE: magnetization-prepared rapid acquisition gradient echo PF: Posterior fossa PFH: Posterior fossa hemorrhage
e570
www.SCIENCEDIRECT.com
INTRODUCTION
S
pontaneous intraparenchymal posterior fossa hemorrhage (PFH) accounts for w10% of intracranial hemorrhages and has a crude annual incidence rate of w4 per 100,000 in the United States.1,2 Risk factors for PFH include age, male sex, use of anticoagulation, and hypertension. Hypertension is the greatest risk factor, with 60%e90% of spontaneous PFH cases occurring in hypertensive patients.3,4 Patients have variable symptoms depending on the site of hemorrhage, but unanimously present with headache and impaired consciousness.5 Primarily cerebellar hemorrhages feature varying degrees of ataxia. Primarily pontomedullary hemorrhages are far more devastating, presenting with lower cranial nerve dysfunction, respiratory failure, or coma.6 Mainstays of management of patients with PFH include blood pressure management, reversal of anticoagulation, and potentially interventions to reduce cerebral edema.7 In comparison with the more common supratentorial intraparenchymal hemorrhage syndromes, PFH of the
PFHS: Posterior fossa hemorrhage score TDPF: Transverse diameter of the posterior fossa From the Departments of 1Neuroradiology and 2Neurosurgery, University of Utah, Salt Lake City, Utah, USA To whom correspondence should be addressed: Ethan A. Neufeld, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2019) 127:e570-e577. https://doi.org/10.1016/j.wneu.2019.03.208 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
cerebellum is relatively unique in that surgical management is more frequently employed for treatment.4,8 There is some controversy in the literature regarding the appropriate surgical management of patients with cerebellar PFH, with much of the literature discussing surgical management of patients with infarction rather than hemorrhage. Although some centers advocate for craniotomy with duraplasty with or without resecting of the posterior arch of C1, others advocate for craniectomy with or without duraplasty.4 The decision to proceed with surgical intervention is based on clinical deterioration in these patients, with decreasing Glasgow Coma Scale (GCS) score and progression of imaging findings including increasing hemorrhage size and regional mass effect.9 The surgical criteria in the literature are controversial; some authors have proposed strict cutoffs of hemorrhage diameters such that hematomas >3e4 cm should undergo decompression regardless of clinical findings, whereas others have suggested grading systems based on regional mass effect, including degree of fourth ventricular effacement or compression of the quadrigeminal cistern.4,10-12 Even with surgical intervention, mortality remains high, with 30-day mortality rates in the range of 25%e30% in patients who undergo surgical intervention.9 What has yet to be explored is the influence of a patient’s anatomy on the need for surgical decompression in the setting of cerebellar PFH. Many measurement techniques exist to evaluate the craniocervical junction (CCJ) and the posterior fossa (PF) on midline sagittal imaging.13 Although there are clearly genetic conditions such as achondroplasia and Chiari malformations that have a diminutive PF as part of their pathology, the spectrum of sizes of the PF throughout the normal population may have an influence on the need for decompression. Prior research involving patients with Chiari I malformation has evaluated the influence of PF measurements on symptomatology.14,15 It stands to reason that patients with smaller PF may have less tolerance for PFH. We hypothesize that patients with smaller PF/CCJ dimensions may be more likely to require surgical decompression by current clinical standards. Our
Figure 1. Measurement techniques for average hemorrhage diameter on computed tomography (CT). Axial CT soft tissue algorithm CT (A) and coronal soft tissue algorithm reformat CT (B). Greatest diameter on axial sequence was obtained with
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
EARLY CRANIAL PFH INTERVENTION
goal was to develop an objective imaging method to determine which patients need surgery before clinical deterioration based on these anatomic measurements.
METHODS Institutional review board approval was obtained with a waiver of patient consent. A retrospective review of the electronic medical record (EMR) system at our institution was performed to identify all computed tomography (CT) or magnetic resonance imaging (MRI) studies performed between January 1, 2007, and January 1, 2017, on patients >18 years of age with “cerebellar” and/or “posterior fossa hemorrhage.” Imaging studies were excluded if there was evidence of a mass lesion associated with the site of hemorrhage, either by imaging appearance or knowledge of preexisting mass at that site; if there was contemporaneous supratentorial mass effect from territorial infarction, hemorrhage, or intra-axial mass (the presence of punctate embolic infarction or microhemorrhage <5 mm in the supratentorial brain were not used as exclusion criteria), or if motion degradation precluded accurate measurements. Imaging was reviewed by a neuroradiology fellow, independent of chart review (without knowledge of patient outcome or surgical intervention). Studies were included if the site of hemorrhage was in the cerebellum, solely pontomedullary hemorrhages were not included. On CT soft tissue algorithm images, the area of hemorrhage was defined as the area of intra-axial hyperattenuation demarcated by either normal brain parenchyma or a rim of hypoattenuation. The diameter of the area of hemorrhage was measured as the greatest axial diameter with corresponding perpendicular dimension and greatest craniocaudal dimension on coronal reconstruction image (Figure 1). On MRI, the diameter of hemorrhage on the axial T2-weighted sequence was defined as the region of predominantly mass-like T2 hyperintensity demarcated by either normal brain parenchyma or a rim of vasogenic edema (T2 hyperintensity) and was measured as the greatest axial diameter with corresponding perpendicular dimension (Figure 2).
corresponding perpendicular dimension on the same slice as denoted by yellow lines in A. Greatest craniocaudal diameter was obtained on coronal reformat sequence as denoted by yellow line in B.
www.journals.elsevier.com/world-neurosurgery
e571
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Figure 2. Measurement techniques for average hemorrhage diameter on magnetic resonance imaging. Axial T2 sequence (A) and coronal magnetization-prepared rapid acquisition gradient echo sequence (B). Greatest diameter on axial sequence
The greatest craniocaudal dimension of the hemorrhage was measured on coronal reformat of T1-weighted magnetizationprepared rapid acquisition gradient echo (MPRAGE) sequence as the heterogeneous mass-like T1 signal demarcated by normal brain parenchyma. The 3 measurements obtained on each case were averaged to determine the mean PFH diameter. For the cases that went to surgery, any imaging performed immediately before surgery was evaluated and the aforementioned measurements were also recorded.
Figure 3. Sagittal measurements of the posterior fossa and craniocervical junction. Midline sagittal computed tomography (A) or sagittal magnetization-prepared rapid acquisition gradient echo magnetic resonance (B) sequences were used. Twining’s line: distance between the dorsum sella and the torcula/internal occipital protuberance. Clival length: distance between the dorsum sella and the tip of the basion. McRae’s
e572
www.SCIENCEDIRECT.com
was obtained with corresponding perpendicular dimension on the same slice as denoted by yellow lines in A. Greatest craniocaudal diameter was obtained on coronal sequence as denoted by yellow line in B.
Measurements of the PF and CCJ were also performed on the initial CT/MRI examinations per standardized measurement techniques.13-16 On CT, anatomic measurements were performed at the midline on sagittal soft tissue algorithm reconstruction. On MRI, measurements were performed at the midline on sagittal T1 MPRAGE sequence. Measurements included Twining’s line, clival length, McRae’s line, interval between the internal occipital protuberance and opisthion, and the anteroposterior diameter of the tentorium (Figure 3). To determine the greatest transverse
line: distance between the tip of the basion and tip of the opisthion. Tentorium: anteroposterior distance from the insertion of the tentorium at the internal occipital protuberance to its most visualized anterior extent. IOP-OP: distance between the tip of the internal occipital protuberance to the tip of the opisthion.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Table 1. Demographics of the Patient Population at Presentation Total Patients (n [ 51)
Nonsurgical Management (n [ 29)
Surgical Management (n [ 22)
P Value
Mean age in years (range)
66.5 (28e89)
67.3 (44e89)
65.2 (28e89)
0.769
Sex, no. male
30
16
14
0.543
Anticoagulation use, no. (%)
17 (33.3)
10 (34.4)
7 (31.8)
0.842
Hypertensive (SBP >140)
39 (76.5)
22 (75.9)
17 (77.3)
0.906
Mean initial GCS score
12
13.2
10.3
<0.001
SBP, systolic blood pressure; GCS, Glasgow Coma Scale.
Figure 4. Measurement technique of the transverse diameter of the posterior fossa (TDPF) on coronal soft tissue algorithm reformat computed tomography (A) or coronal reformat of magnetization-prepared rapid acquisition gradient echo magnetic resonance imaging sequence (B). TDPF defined as the distance between the most lateral points of the superior aspect of the groove for the transverse-sigmoid sinus junction. This point was chosen because the apex of the groove is reproducible between measurements.
diameter of the posterior fossa (TDPF), a measurement between the most lateral points of the groove for the transverseesigmoid sinus junctions was performed. The TDPF measurement was performed on coronal soft tissue algorithm CT reconstruction (Figure 4) and on coronal reformat of the T1 MPRAGE sequence. From these measurements, a surrogate measurement for the PF volume was created. The surrogate measurement is the product of the clival length, Twining’s line, and TDPF. These measurements approximate the craniocaudal, anterior-posterior, and transverse dimensions, respectively. A scoring system—the posterior fossa hemorrhage score (PFHS)—was created by dividing the average hemorrhage diameter by the surrogate PF volume and multiplying by 1000. Prior literature has explored using average hemorrhage diameter as a tool to distinguish surgical from nonsurgical patients, but there is no consensus on cutoff and there is significant variability in the severity of a patient’s presentation for a given hemorrhage size. By combining hemorrhage diameter with patients’ PF dimensions the PFHS can account for a patient’s ability to accommodate for a given volume of hemorrhage.
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
The EMR was reviewed after all imaging measurements were performed. The primary outcome was the decision to proceed to surgical decompression. Other outcomes evaluated were in-hospital mortality and major neurologic deficits before discharge. Parameters documented at initial evaluation included use of therapeutic anticoagulation, blood pressure measurement at initial evaluation, GCS score, and major neurologic examination findings such as ataxia, decerebrate/decorticate posturing, focal weakness/numbness, or cranial nerve dysfunction. The cases were reviewed through their hospital course and grouped into 4 categories: discharge without surgical management; discharge after surgical management; patient death without surgical management; and patient death after surgical management. Major neurologic deficits at the time of discharge were documented. GCS score and major neurologic deficits of postsurgical patients were compared with their initial presentation. When available, the documented indication for surgical decompression was noted (i.e., declining GCS score, increasing compression of the fourth ventricle, et cetera). The surgical procedure itself was reviewed in every surgical case to determine the major surgical findings and details of the intervention(s) performed. Statistical analysis was performed to compare patients who had surgery and those who were successfully managed without surgery. The average hemorrhage diameter, each individual PF and CCJ measurement, and surrogate PF volume were compared between groups using the 2-tailed t test. The use of anticoagulation was compared between the groups using the c2 test analysis. A P value of <0.05 was prospectively determined to indicate a statistically significant difference. RESULTS Of the 98 cases initially identified, 23 cases were excluded as they were the result of hemorrhage into a preexisting intra-axial neoplasm or a metastatic lesion, 19 cases were excluded because they featured contemporaneous supratentorial hemorrhage (extraaxial or intraparenchymal), territorial infarction, or intra-axial mass with regional mass effect, and 5 cases were excluded
www.journals.elsevier.com/world-neurosurgery
e573
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Table 2. Details of Measurements Performed on Initial Computed Tomography/Magnetic Resonance Imaging Total Patients (n [ 51)
Nonsurgical Management (n [ 29)
Surgical Management (n [ 22)
P Value
Hemorrhage diameter (cm)
3.6
2.9
4.6
<0.001
Twining’s line (cm)
8.00
8.11
7.85
0.051
Clival length (cm)
3.74
3.80
3.64
0.161
McRae’s line (cm)
3.52
3.58
3.42
0.052
Tentorial diameter (cm)
4.46
4.49
4.44
0.829
IOP-OP interval (cm)
4.19
4.21
4.17
0.754
TDPF (cm)
10.50
10.61
10.38
0.148
Surrogate posterior fossa volume (cm3)
315.71
328.66
297
0.019
Posterior fossa hemorrhage score
11.61
8.91
15.55
<0.001
IOP-OP, internal occipital protuberance and opisthion; TDPF, transverse diameter of the posterior fossa.
because the imaging had motion degradation. The remaining 51 cases met the inclusion criteria. This included 29 nonsurgical cases and 22 surgical cases. Of note, only 19 of the 22 surgical cases actually proceeded to surgery. The remaining 3 cases were patients with marked clinical deterioration at presentation for whom surgery was planned but died prior to surgery. As these patients met the clinical standards for surgical decompression, they were included in the statistical analysis. The demographics of the patients in each group are outlined in Table 1. Thirty-nine patients (76.5%) were hypertensive at the time of presentation with systolic blood pressure measurements >140 mmHg (value chosen based on prior literature). There was no statistically significant difference in the incidence of hypertension at presentation between surgical and nonsurgical patients via the c2 test analysis (P ¼ 0.906). Among the 51 cases, 17 of the patients were on therapeutic anticoagulation at the time of presentation, including 7/22 surgical patients and 10/29 nonsurgical patients. Most of these patients (16) were on therapeutic warfarin and 1 patient was on rivaroxaban. There was no correlation between anticoagulation use at time of presentation and need for surgery via the c2 test analysis (P ¼ 0.842). There was no correlation between use of anticoagulation at time of surgery and average hemorrhage diameter using the 2-tailed t test (P ¼ 0.555). The mean initial GCS score at presentation for the surgical patients was 10.3 compared with 13.2 for the patients treated nonsurgically (P < 0.001). The indication for surgery in all 22 of the surgical patients was documented as decreasing or persistently low GCS score; 2 patients had documentation of clinical signs of brainstem compression. Only 3 of the 22 surgical cases had clear
e574
www.SCIENCEDIRECT.com
Figure 5. Graph of average hemorrhage diameters comparing patients managed surgically versus those managed without surgery. Values of y-axis are in cm. Mean hemorrhage diameter in the surgical group ranged from 3e6 cm with a mean value of 4.6 cm, whereas the mean hemorrhage diameter in the nonsurgical group ranged from 2e4.4 cm, with a mean value of 2.9 cm. This is a statistically significant difference with a P value of <0.001.
EMR documentation of imaging findings that were an impetus to proceed to surgery. Two surgical cases documented increasing fourth ventricle compression, whereas 1 surgical case documented effacement of the quadrigeminal cistern. On imaging review, all 22 surgical cases had 1 or both of these findings on the CT examination immediately prior to surgery. The 19 patients who actually proceeded to surgery underwent suboccipital craniectomy with duraplasty. Seventeen of these patients had a midline decompression and 2 patients had off-midline decompression. Imaging parameters of the groups are outlined in Table 2. The mean diameter of hemorrhage of surgical patients was 4.6 cm compared with 2.9 cm for nonsurgical patients (P < 0.001) (Figure 5). There was no statistically significant difference for any single PF or CCJ measurement between the surgical and nonsurgical groups (Table 2), but there was a statistically
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
significant difference in the surrogate PF volume between the surgical and nonsurgical groups (P ¼ 0.019) (Figure 6). The mean PFHS (method for calculation described in Methods section) for the nonsurgical group was 8.91 with a range of 5.3e13.1, and the mean for the surgical group was 15.55 with a range of 11.7e20 (Figure 7). No surgical patient scored below 11.7 and no nonsurgical patient scored above 13.1. DISCUSSION Spontaneous PFH is potentially devastating neurologic event that has a high rate of morbidity and mortality. There is evidence that suboccipital decompressive craniectomy can be a life-saving measure for patients with PFH by alleviating regional mass effect and thereby potentially preventing life-threatening hydrocephalus and/or herniation syndromes.9,17 Although there is a general consensus that only a subset of patients who present with more severe cases of PFH need to undergo decompression, there remains no consensus regarding what the selection criteria to determine the need for surgery should be. Close neurologic assessment is performed until patients’ clinical findings either deteriorate or improve before a decision about surgical decompression is made.9 Prior literature has examined factors such as average hematoma/hemorrhage diameter and both qualitative and quantitative evaluation of the degree of PF mass effect. Taneda et al.18 and subsequently van Loon et al.10 described a grading system of imaging severity based on the degree of compression of the quadrigeminal cistern. This finding was thought to be useful because it not only accounted for mass effect by the hemorrhage itself but also by vasogenic edema and hydrocephalus. Kirollos et al.11 alternatively evaluated compression and shift of the fourth ventricle as an imaging marker of severity, with greater degrees of shift and larger hematoma size correlating with need for surgery and poor outcomes. However, these imaging findings do not necessarily correlate with a patient’s clinical status, and hemorrhage morphology and location can limit the applicability of the findings. In a retrospective review, Kobayashi et al.12 developed a protocol that categorized patients into surgical management or nonsurgical management using a hematoma cutoff size of 4 cm and a GCS score cutoff of 14. An alternative approach that has been proposed is to use volume of hematoma as an indication for intervention via open decompression or aspiration. Thirty cc has been proposed as a cutoff value of hematoma volume for intervention.19 Our study takes a different approach in patients presenting with spontaneous PFH, focusing on how a given patient’s PF and CCJ measurements may contribute to the severity of his or her symptoms, which may necessitate surgery. Patients with relatively smaller PF may not be able to tolerate PFH as well as patients with larger PF and may succumb to complications with relatively lower PFH volumes. If we apply preexisting approaches to determine the indication for surgery via imaging measurements on the patients in our study, we get results that do not correlate as well with the severity of the patients’ clinical presentation or need for surgery. If we were to use a conservation cutoff of average hemorrhage diameter of 3 cm or greater as an indication for surgery, then 15 patients would have been managed conservatively and 36 patients
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
Figure 6. Graph of posterior fossa volume surrogate measurements comparing patients managed surgically versus those managed without surgery. Values of y-axis are in cm3. The volume surrogate measurement in the surgical group ranged from 229.47e374.33 with a mean value of 297, whereas the volume surrogate measurement in the nonsurgical group ranged from 267.54e427.62 with a mean value of 328.66. This is a statistically significant difference with a P value of 0.019.
would have been considered surgical candidates. Using this cutoff, 14 patients would have proceeded to surgery who were successfully managed conservatively with minimal neurologic deficits in our study. If we were to use a cutoff of average hemorrhage diameter of 4 cm, then 31 patients would have been managed conservatively and 20 patients would have been considered surgical candidates. Using this cutoff, 2 patients would have proceeded to surgery who were successfully managed conservatively with minimal neurologic deficits and 3 patients would have been managed conservatively who—in our study—had small PF and presented with severe clinical deterioration. Finally, if we used hematoma volume of 30 cc as a cutoff to determine need for surgical management, then 19 patients would have been managed conservatively and 32 patients would have been considered surgical candidates. Using this cutoff,
www.journals.elsevier.com/world-neurosurgery
e575
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Figure 7. Graph of posterior fossa hemorrhage score (PFHS) values comparing patients managed surgically versus those managed without surgery. The PFHS in the surgical group ranged from 11.7e20 with a mean value of 15.55, whereas the PFHS in the nonsurgical group ranged from 5. 3e13.1 with a mean value of 8.91. This is a statistically significant difference with a P value of <0.001.
11 patients would have proceeded to surgery who were successfully managed conservatively with minimal neurologic deficits, and 1 patient would have been managed conservatively who had a small PF and presented with severe clinical deterioration. It would appear then that measurements of hemorrhage size alone as discussed in prior literature do not tell the whole story. The combined measurement of the hemorrhage size and the individual patient’s PF anatomic constraints (as encapsulated by the PFHS) could help guide early surgical intervention before potentially devastating clinical decompensation. This objective measurement would be most useful in patients with spontaneous PFH who are “borderline.” Such patients would be those with relatively large hemorrhage diameters (>3e4 cm) but with minimal symptoms or those with relatively smaller hemorrhage diameters but with depressed GCS scores. By incorporating PF
e576
www.SCIENCEDIRECT.com
measurements such as the PFHS, such patients may be able to be classified into surgical or nonsurgical cases. Based on the distribution of PFHS in both groups, patients who score <11 could be considered nonsurgical patients, whereas those who score >14 could be considered surgical patients. Patients who score in the 11e14 range could be considered indeterminate and management would likely depend more heavily on the clinical status of the patient. A surgeon could adopt a more conservative policy in which any patient scoring >11 on the PFHS would be taken for surgical decompression or could take a more liberal policy in which any patient scoring <14 on the PFHS would be managed conservatively. Although the surrogate PF volume is not a true volumetric assessment as can be assessed by 3-dimensional volumetric analysis of the PF, we elected to use a product of 2-dimensional measurements that can be performed easily on any workstation. Although our method seems to indicate that anatomic criteria observed on imaging can influence necessity for suboccipital decompression, ultimately the decision will rely on the neurosurgical expertise and clinical judgement. An important technical consideration is the applicability of these techniques. Our CT protocol uses 4-mm width axial slice acquisition with 2-mm width coronal and sagittal reconstructions. Many imaging centers may not have CT protocols or scanners that allow for such slice dimensions, and this could limit the reliability of measurements. Furthermore, at our institution, patients who had their heads rotated or angled in the CT gantry at the time of acquisition will routinely have their axial sequences reformatted into more anatomic positioning and have their coronal and sagittal reconstructions performed from the repositioned sequence. This likewise may not be an available option for many imaging centers, and thus obtaining a true midline slice on sagittal imaging may be challenging. Also, MRI measurements were made on a sagittal T1 MPRAGE sequence, a 3-dimensional T1-weighted gradient recall echo sequence that allows for reformatting in multiple planes and enables measurement of the TDPF quickly. Although we do not anticipate that measurements performed on a more traditional sagittal T1 spin-echo sequence would be substantially different, this should be considered when applying measurement values. This study does have some limitations. With our relatively restrictive exclusion criteria, the patient group comprised only 51 patients, limiting the strength of the study. We only evaluated whether patients were or were not using therapeutic anticoagulation, and the use of other potentially relevant medications such as aspirin was not considered. Chart review also had its limitations as only the patients’ GCS score and major neurologic examination findings were consistently documented. For patients who had surgery, it was not always clear what imaging findings were considered for those patients. Specifically, it was not clear whether findings that have been evaluated in the neurosurgical literature, such as increasing fourth ventricle shift/compression or quadrigeminal cistern effacement, were considered in the management of the surgical patients. On review of the imaging prior to surgery, all patients who were taken to surgery had increasing effacement of the fourth ventricle and/or effacement of the quadrigeminal cistern. Whether these findings were a direct impetus to proceed to surgery was only documented in 3 of the surgical cases. Finally, underlying cerebellar parenchymal volume
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
loss could be a factor that would allow patients to tolerate greater volumes of hemorrhage without clinical decompensation. Accurate volumetric assessment of the cerebellum is able to be performed via quantitative analysis of a volumetric MRI sequence, but this is not routinely available and cannot be performed with CT. Furthermore, the presence of hemorrhage at the time of examination would preclude accurate volumetric analysis. We advise that if a patient has profound preexisting cerebellar volume loss that the surgeon consider this as a factor that may allow a patient to tolerate conservative management with relatively larger hemorrhage volumes. CONCLUSIONS Our study is concordant with the hypothesis that larger hemorrhage volumes in spontaneous PFH at initial presentation are
REFERENCES 1. Luney MS, English SW, Longworth A, et al. Acute posterior cranial fossa hemorrhage: is surgical decompression better than expectant medical management? Neurocrit Care. 2016;25:365-370. 2. Sacco S, Marini C, Toni D, Olivieri L, Carolei A. Incidence and 10-year survival of intracerebral hemorrhage in a population-based registry. Stroke. 2009;40:394-399. 3. Salvetti M, Paini A, Bertacchini F, et al. Therapeutic approach to hypertensive emergencies: hemorrhagic stroke. High Blood Press Cardiovasc Prev. 2018;25:191-195. 4. Amar AP. Controversies in the neurosurgical management of cerebellar hemorrhage and infarction. Neurosurg Focus. 2012;32:E1. 5. Dinsdale HB. Spontaneous hemorrhage in the posterior fossa. A study of primary cerebellar and pontine hemorrhages with observations on their pathogenesis. Arch Neurol. 1964;10:200-217. 6. Karadan U, Supreeth RN, Manappallil RG, Jayakrishnan C. Twenty-four syndrome: an untold presentation of pontine hemorrhage. J Stroke Cerebrovasc Dis. 2018;27:e73-e74. 7. Dastur CK, Yu W. Current management of spontaneous intracerebral haemorrhage. Stroke Vasc Neurol. 2017;2:21-29. 8. Mathew P, Teasdale G, Bannan A, OluochOlunya D. Neurosurgical management of cerebellar haematoma and infarct. J Neurol Neurosurg Psychiatry. 1995;59:287-292.
more likely to require surgical decompression by clinical standards. Our results also demonstrate a statistically significant difference in the volume of PF between surgical and nonsurgical patients as determined by current clinical standards via our surrogate measurement. These findings suggest that patients with smaller PF volumes are more likely to require surgical decompression and may require decompression with relatively smaller volumes of hemorrhage. We suggest that these 2 factors are both predictors of the need for surgical decompression and should both be considered when initially evaluating a patient presenting with primary cerebellar PFH. Our derived PFHS—which can be performed using simple workstation measurements—incorporates these 2 factors and accentuates the differences between the 2 groups. The PFHS may be able to determine which patients will require early surgical decompression at initial presentation.
9. Arnone GD, Esfahani DR, Wonais M, et al. Surgery for cerebellar hemorrhage: a national surgical quality improvement program database analysis of patient outcomes and factors associated with 30-vday mortality and prolonged ventilation. World Neurosurg. 2017;106:543-550. 10. van Loon J, Van Calenbergh F, Goffin J, Plets C. Controversies in the management of spontaneous cerebellar haemorrhage. A consecutive series of 49 cases and review of the literature. Acta Neurochir (Wien). 1993;122:187-193. 11. Kirollos RW, Tyagi AK, Ross SA, van Hille PT, Marks PV. Management of spontaneous cerebellar hematomas: a prospective treatment protocol. Neurosurgery. 2001;49:1378-1386 [discussion: 13861377]. 12. Kobayashi S, Sato A, Kageyama Y, Nakamura H, Watanabe Y, Yamaura A. Treatment of hypertensive cerebellar hemorrhage–surgical or conservative management? Neurosurgery. 1994;34:246-250 [discussion: 250-241]. 13. Smoker WR. Craniovertebral junction: normal anatomy, craniometry, and congenital anomalies. Radiographics. 1994;14:255-277. 14. Sekula RF Jr, Jannetta PJ, Casey KF, Marchan EM, Sekula LK, McCrady CS. Dimensions of the posterior fossa in patients symptomatic for Chiari I malformation but without cerebellar tonsillar descent. Cerebrospinal Fluid Res. 2005;2:11. 15. Alperin N, Loftus JR, Oliu CJ, et al. Magnetic resonance imaging measures of posterior cranial fossa morphology and cerebrospinal fluid physiology in Chiari malformation type I. Neurosurgery. 2014;75:515-522 [discussion: 522].
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
16. Rojas CA, Bertozzi JC, Martinez CR, Whitlow J. Reassessment of the craniocervical junction: normal values on CT. AJNR Am J Neuroradiol. 2007; 28:1819-1823. 17. Puffer RC, Graffeo C, Rabinstein A, Van Gompel JJ. Mortality rates after emergent posterior fossa decompression for ischemic or hemorrhagic stroke in older patients. World Neurosurg. 2016;92: 166-170. 18. Taneda M, Hayakawa T, Mogami H. Primary cerebellar hemorrhage. Quadrigeminal cistern obliteration on CT scans as a predictor of outcome. J Neurosurg. 1987;67:545-552. 19. Aguilar MI, Brott TG. Update in intracerebral hemorrhage. Neurohospitalist. 2011;1:148-159.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Contents of this submission were presented at the 2018 Annual Meeting of the American Society of Neuroradiology in Vancouver, British Columbia, Canada on June 6, 2018. Received 7 December 2018; accepted 20 March 2019 Citation: World Neurosurg. (2019) 127:e570-e577. https://doi.org/10.1016/j.wneu.2019.03.208 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e577
Craniocervical Junction and Posterior Fossa Dimensions can Affect Need for Decompressive Craniectomy in Posterior Cranial Fossa Hemorrhage Ethan A. Neufeld1, Sarah T. Menacho2, Lubdha M. Shah1
BACKGROUND: Posterior fossa hemorrhage (PFH) of the cerebellum is managed by decompressive craniectomy when there is clinical deterioration. There is no current consensus on an objective imaging method to determine which patients need surgery before clinical deterioration. We developed an imaging scoring tool by assessing initial hemorrhage diameter and posterior fossa (PF) measurements to determine which patients will benefit from early surgical intervention.
-
METHODS: For this caseecontrol study, we reviewed the electronic medical record to identify adults who presented with PFH over a 10-year period at our institution. Chart review for clinical findings and inciting factors were documented. The average diameter of PFH and the surrogate PF volume on initial imaging studies were measured. These measurements were correlated with surgical intervention. A scoring tool was developed based on radiographic and clinical data.
CONCLUSIONS: Patients presenting with PFH with smaller PF volumes may be more likely to require surgery as determined by clinical standards. The proposed scoring system based on simple measurements on initial computed tomography and magnetic resonance imaging may help surgeons consider early surgical intervention in those patients with PFH with smaller PF volumes.
-
-
RESULTS: Fifty-one patients met the inclusion criteria. The average hemorrhage diameter and the surrogate PF volume measurements were statistically different between surgical and nonsurgical cases (P < 0.001 and P [ 0.019, respectively). The scoring system was created by dividing average hemorrhage diameter by surrogate PF volume and multiplying by 1000. The median score of nonsurgical patients was 9.1, and the median score of surgical patients was 15.6.
-
Key words Cerebellum - Cerebrovascular disorders - Posterior cranial fossa - Stroke -
Abbreviations and Acronyms CCJ: Craniocervical junction CT: Computed tomography EMR: Electronic medical record GCS: Glasgow Coma Scale MRI: Magnetic resonance imaging MPRAGE: magnetization-prepared rapid acquisition gradient echo PF: Posterior fossa PFH: Posterior fossa hemorrhage
e570
www.SCIENCEDIRECT.com
INTRODUCTION
S
pontaneous intraparenchymal posterior fossa hemorrhage (PFH) accounts for w10% of intracranial hemorrhages and has a crude annual incidence rate of w4 per 100,000 in the United States.1,2 Risk factors for PFH include age, male sex, use of anticoagulation, and hypertension. Hypertension is the greatest risk factor, with 60%e90% of spontaneous PFH cases occurring in hypertensive patients.3,4 Patients have variable symptoms depending on the site of hemorrhage, but unanimously present with headache and impaired consciousness.5 Primarily cerebellar hemorrhages feature varying degrees of ataxia. Primarily pontomedullary hemorrhages are far more devastating, presenting with lower cranial nerve dysfunction, respiratory failure, or coma.6 Mainstays of management of patients with PFH include blood pressure management, reversal of anticoagulation, and potentially interventions to reduce cerebral edema.7 In comparison with the more common supratentorial intraparenchymal hemorrhage syndromes, PFH of the
PFHS: Posterior fossa hemorrhage score TDPF: Transverse diameter of the posterior fossa From the Departments of 1Neuroradiology and 2Neurosurgery, University of Utah, Salt Lake City, Utah, USA To whom correspondence should be addressed: Ethan A. Neufeld, M.D. [E-mail: [email protected]] Citation: World Neurosurg. (2019) 127:e570-e577. https://doi.org/10.1016/j.wneu.2019.03.208 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
cerebellum is relatively unique in that surgical management is more frequently employed for treatment.4,8 There is some controversy in the literature regarding the appropriate surgical management of patients with cerebellar PFH, with much of the literature discussing surgical management of patients with infarction rather than hemorrhage. Although some centers advocate for craniotomy with duraplasty with or without resecting of the posterior arch of C1, others advocate for craniectomy with or without duraplasty.4 The decision to proceed with surgical intervention is based on clinical deterioration in these patients, with decreasing Glasgow Coma Scale (GCS) score and progression of imaging findings including increasing hemorrhage size and regional mass effect.9 The surgical criteria in the literature are controversial; some authors have proposed strict cutoffs of hemorrhage diameters such that hematomas >3e4 cm should undergo decompression regardless of clinical findings, whereas others have suggested grading systems based on regional mass effect, including degree of fourth ventricular effacement or compression of the quadrigeminal cistern.4,10-12 Even with surgical intervention, mortality remains high, with 30-day mortality rates in the range of 25%e30% in patients who undergo surgical intervention.9 What has yet to be explored is the influence of a patient’s anatomy on the need for surgical decompression in the setting of cerebellar PFH. Many measurement techniques exist to evaluate the craniocervical junction (CCJ) and the posterior fossa (PF) on midline sagittal imaging.13 Although there are clearly genetic conditions such as achondroplasia and Chiari malformations that have a diminutive PF as part of their pathology, the spectrum of sizes of the PF throughout the normal population may have an influence on the need for decompression. Prior research involving patients with Chiari I malformation has evaluated the influence of PF measurements on symptomatology.14,15 It stands to reason that patients with smaller PF may have less tolerance for PFH. We hypothesize that patients with smaller PF/CCJ dimensions may be more likely to require surgical decompression by current clinical standards. Our
Figure 1. Measurement techniques for average hemorrhage diameter on computed tomography (CT). Axial CT soft tissue algorithm CT (A) and coronal soft tissue algorithm reformat CT (B). Greatest diameter on axial sequence was obtained with
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
EARLY CRANIAL PFH INTERVENTION
goal was to develop an objective imaging method to determine which patients need surgery before clinical deterioration based on these anatomic measurements.
METHODS Institutional review board approval was obtained with a waiver of patient consent. A retrospective review of the electronic medical record (EMR) system at our institution was performed to identify all computed tomography (CT) or magnetic resonance imaging (MRI) studies performed between January 1, 2007, and January 1, 2017, on patients >18 years of age with “cerebellar” and/or “posterior fossa hemorrhage.” Imaging studies were excluded if there was evidence of a mass lesion associated with the site of hemorrhage, either by imaging appearance or knowledge of preexisting mass at that site; if there was contemporaneous supratentorial mass effect from territorial infarction, hemorrhage, or intra-axial mass (the presence of punctate embolic infarction or microhemorrhage <5 mm in the supratentorial brain were not used as exclusion criteria), or if motion degradation precluded accurate measurements. Imaging was reviewed by a neuroradiology fellow, independent of chart review (without knowledge of patient outcome or surgical intervention). Studies were included if the site of hemorrhage was in the cerebellum, solely pontomedullary hemorrhages were not included. On CT soft tissue algorithm images, the area of hemorrhage was defined as the area of intra-axial hyperattenuation demarcated by either normal brain parenchyma or a rim of hypoattenuation. The diameter of the area of hemorrhage was measured as the greatest axial diameter with corresponding perpendicular dimension and greatest craniocaudal dimension on coronal reconstruction image (Figure 1). On MRI, the diameter of hemorrhage on the axial T2-weighted sequence was defined as the region of predominantly mass-like T2 hyperintensity demarcated by either normal brain parenchyma or a rim of vasogenic edema (T2 hyperintensity) and was measured as the greatest axial diameter with corresponding perpendicular dimension (Figure 2).
corresponding perpendicular dimension on the same slice as denoted by yellow lines in A. Greatest craniocaudal diameter was obtained on coronal reformat sequence as denoted by yellow line in B.
www.journals.elsevier.com/world-neurosurgery
e571
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Figure 2. Measurement techniques for average hemorrhage diameter on magnetic resonance imaging. Axial T2 sequence (A) and coronal magnetization-prepared rapid acquisition gradient echo sequence (B). Greatest diameter on axial sequence
The greatest craniocaudal dimension of the hemorrhage was measured on coronal reformat of T1-weighted magnetizationprepared rapid acquisition gradient echo (MPRAGE) sequence as the heterogeneous mass-like T1 signal demarcated by normal brain parenchyma. The 3 measurements obtained on each case were averaged to determine the mean PFH diameter. For the cases that went to surgery, any imaging performed immediately before surgery was evaluated and the aforementioned measurements were also recorded.
Figure 3. Sagittal measurements of the posterior fossa and craniocervical junction. Midline sagittal computed tomography (A) or sagittal magnetization-prepared rapid acquisition gradient echo magnetic resonance (B) sequences were used. Twining’s line: distance between the dorsum sella and the torcula/internal occipital protuberance. Clival length: distance between the dorsum sella and the tip of the basion. McRae’s
e572
www.SCIENCEDIRECT.com
was obtained with corresponding perpendicular dimension on the same slice as denoted by yellow lines in A. Greatest craniocaudal diameter was obtained on coronal sequence as denoted by yellow line in B.
Measurements of the PF and CCJ were also performed on the initial CT/MRI examinations per standardized measurement techniques.13-16 On CT, anatomic measurements were performed at the midline on sagittal soft tissue algorithm reconstruction. On MRI, measurements were performed at the midline on sagittal T1 MPRAGE sequence. Measurements included Twining’s line, clival length, McRae’s line, interval between the internal occipital protuberance and opisthion, and the anteroposterior diameter of the tentorium (Figure 3). To determine the greatest transverse
line: distance between the tip of the basion and tip of the opisthion. Tentorium: anteroposterior distance from the insertion of the tentorium at the internal occipital protuberance to its most visualized anterior extent. IOP-OP: distance between the tip of the internal occipital protuberance to the tip of the opisthion.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Table 1. Demographics of the Patient Population at Presentation Total Patients (n [ 51)
Nonsurgical Management (n [ 29)
Surgical Management (n [ 22)
P Value
Mean age in years (range)
66.5 (28e89)
67.3 (44e89)
65.2 (28e89)
0.769
Sex, no. male
30
16
14
0.543
Anticoagulation use, no. (%)
17 (33.3)
10 (34.4)
7 (31.8)
0.842
Hypertensive (SBP >140)
39 (76.5)
22 (75.9)
17 (77.3)
0.906
Mean initial GCS score
12
13.2
10.3
<0.001
SBP, systolic blood pressure; GCS, Glasgow Coma Scale.
Figure 4. Measurement technique of the transverse diameter of the posterior fossa (TDPF) on coronal soft tissue algorithm reformat computed tomography (A) or coronal reformat of magnetization-prepared rapid acquisition gradient echo magnetic resonance imaging sequence (B). TDPF defined as the distance between the most lateral points of the superior aspect of the groove for the transverse-sigmoid sinus junction. This point was chosen because the apex of the groove is reproducible between measurements.
diameter of the posterior fossa (TDPF), a measurement between the most lateral points of the groove for the transverseesigmoid sinus junctions was performed. The TDPF measurement was performed on coronal soft tissue algorithm CT reconstruction (Figure 4) and on coronal reformat of the T1 MPRAGE sequence. From these measurements, a surrogate measurement for the PF volume was created. The surrogate measurement is the product of the clival length, Twining’s line, and TDPF. These measurements approximate the craniocaudal, anterior-posterior, and transverse dimensions, respectively. A scoring system—the posterior fossa hemorrhage score (PFHS)—was created by dividing the average hemorrhage diameter by the surrogate PF volume and multiplying by 1000. Prior literature has explored using average hemorrhage diameter as a tool to distinguish surgical from nonsurgical patients, but there is no consensus on cutoff and there is significant variability in the severity of a patient’s presentation for a given hemorrhage size. By combining hemorrhage diameter with patients’ PF dimensions the PFHS can account for a patient’s ability to accommodate for a given volume of hemorrhage.
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
The EMR was reviewed after all imaging measurements were performed. The primary outcome was the decision to proceed to surgical decompression. Other outcomes evaluated were in-hospital mortality and major neurologic deficits before discharge. Parameters documented at initial evaluation included use of therapeutic anticoagulation, blood pressure measurement at initial evaluation, GCS score, and major neurologic examination findings such as ataxia, decerebrate/decorticate posturing, focal weakness/numbness, or cranial nerve dysfunction. The cases were reviewed through their hospital course and grouped into 4 categories: discharge without surgical management; discharge after surgical management; patient death without surgical management; and patient death after surgical management. Major neurologic deficits at the time of discharge were documented. GCS score and major neurologic deficits of postsurgical patients were compared with their initial presentation. When available, the documented indication for surgical decompression was noted (i.e., declining GCS score, increasing compression of the fourth ventricle, et cetera). The surgical procedure itself was reviewed in every surgical case to determine the major surgical findings and details of the intervention(s) performed. Statistical analysis was performed to compare patients who had surgery and those who were successfully managed without surgery. The average hemorrhage diameter, each individual PF and CCJ measurement, and surrogate PF volume were compared between groups using the 2-tailed t test. The use of anticoagulation was compared between the groups using the c2 test analysis. A P value of <0.05 was prospectively determined to indicate a statistically significant difference. RESULTS Of the 98 cases initially identified, 23 cases were excluded as they were the result of hemorrhage into a preexisting intra-axial neoplasm or a metastatic lesion, 19 cases were excluded because they featured contemporaneous supratentorial hemorrhage (extraaxial or intraparenchymal), territorial infarction, or intra-axial mass with regional mass effect, and 5 cases were excluded
www.journals.elsevier.com/world-neurosurgery
e573
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Table 2. Details of Measurements Performed on Initial Computed Tomography/Magnetic Resonance Imaging Total Patients (n [ 51)
Nonsurgical Management (n [ 29)
Surgical Management (n [ 22)
P Value
Hemorrhage diameter (cm)
3.6
2.9
4.6
<0.001
Twining’s line (cm)
8.00
8.11
7.85
0.051
Clival length (cm)
3.74
3.80
3.64
0.161
McRae’s line (cm)
3.52
3.58
3.42
0.052
Tentorial diameter (cm)
4.46
4.49
4.44
0.829
IOP-OP interval (cm)
4.19
4.21
4.17
0.754
TDPF (cm)
10.50
10.61
10.38
0.148
Surrogate posterior fossa volume (cm3)
315.71
328.66
297
0.019
Posterior fossa hemorrhage score
11.61
8.91
15.55
<0.001
IOP-OP, internal occipital protuberance and opisthion; TDPF, transverse diameter of the posterior fossa.
because the imaging had motion degradation. The remaining 51 cases met the inclusion criteria. This included 29 nonsurgical cases and 22 surgical cases. Of note, only 19 of the 22 surgical cases actually proceeded to surgery. The remaining 3 cases were patients with marked clinical deterioration at presentation for whom surgery was planned but died prior to surgery. As these patients met the clinical standards for surgical decompression, they were included in the statistical analysis. The demographics of the patients in each group are outlined in Table 1. Thirty-nine patients (76.5%) were hypertensive at the time of presentation with systolic blood pressure measurements >140 mmHg (value chosen based on prior literature). There was no statistically significant difference in the incidence of hypertension at presentation between surgical and nonsurgical patients via the c2 test analysis (P ¼ 0.906). Among the 51 cases, 17 of the patients were on therapeutic anticoagulation at the time of presentation, including 7/22 surgical patients and 10/29 nonsurgical patients. Most of these patients (16) were on therapeutic warfarin and 1 patient was on rivaroxaban. There was no correlation between anticoagulation use at time of presentation and need for surgery via the c2 test analysis (P ¼ 0.842). There was no correlation between use of anticoagulation at time of surgery and average hemorrhage diameter using the 2-tailed t test (P ¼ 0.555). The mean initial GCS score at presentation for the surgical patients was 10.3 compared with 13.2 for the patients treated nonsurgically (P < 0.001). The indication for surgery in all 22 of the surgical patients was documented as decreasing or persistently low GCS score; 2 patients had documentation of clinical signs of brainstem compression. Only 3 of the 22 surgical cases had clear
e574
www.SCIENCEDIRECT.com
Figure 5. Graph of average hemorrhage diameters comparing patients managed surgically versus those managed without surgery. Values of y-axis are in cm. Mean hemorrhage diameter in the surgical group ranged from 3e6 cm with a mean value of 4.6 cm, whereas the mean hemorrhage diameter in the nonsurgical group ranged from 2e4.4 cm, with a mean value of 2.9 cm. This is a statistically significant difference with a P value of <0.001.
EMR documentation of imaging findings that were an impetus to proceed to surgery. Two surgical cases documented increasing fourth ventricle compression, whereas 1 surgical case documented effacement of the quadrigeminal cistern. On imaging review, all 22 surgical cases had 1 or both of these findings on the CT examination immediately prior to surgery. The 19 patients who actually proceeded to surgery underwent suboccipital craniectomy with duraplasty. Seventeen of these patients had a midline decompression and 2 patients had off-midline decompression. Imaging parameters of the groups are outlined in Table 2. The mean diameter of hemorrhage of surgical patients was 4.6 cm compared with 2.9 cm for nonsurgical patients (P < 0.001) (Figure 5). There was no statistically significant difference for any single PF or CCJ measurement between the surgical and nonsurgical groups (Table 2), but there was a statistically
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
significant difference in the surrogate PF volume between the surgical and nonsurgical groups (P ¼ 0.019) (Figure 6). The mean PFHS (method for calculation described in Methods section) for the nonsurgical group was 8.91 with a range of 5.3e13.1, and the mean for the surgical group was 15.55 with a range of 11.7e20 (Figure 7). No surgical patient scored below 11.7 and no nonsurgical patient scored above 13.1. DISCUSSION Spontaneous PFH is potentially devastating neurologic event that has a high rate of morbidity and mortality. There is evidence that suboccipital decompressive craniectomy can be a life-saving measure for patients with PFH by alleviating regional mass effect and thereby potentially preventing life-threatening hydrocephalus and/or herniation syndromes.9,17 Although there is a general consensus that only a subset of patients who present with more severe cases of PFH need to undergo decompression, there remains no consensus regarding what the selection criteria to determine the need for surgery should be. Close neurologic assessment is performed until patients’ clinical findings either deteriorate or improve before a decision about surgical decompression is made.9 Prior literature has examined factors such as average hematoma/hemorrhage diameter and both qualitative and quantitative evaluation of the degree of PF mass effect. Taneda et al.18 and subsequently van Loon et al.10 described a grading system of imaging severity based on the degree of compression of the quadrigeminal cistern. This finding was thought to be useful because it not only accounted for mass effect by the hemorrhage itself but also by vasogenic edema and hydrocephalus. Kirollos et al.11 alternatively evaluated compression and shift of the fourth ventricle as an imaging marker of severity, with greater degrees of shift and larger hematoma size correlating with need for surgery and poor outcomes. However, these imaging findings do not necessarily correlate with a patient’s clinical status, and hemorrhage morphology and location can limit the applicability of the findings. In a retrospective review, Kobayashi et al.12 developed a protocol that categorized patients into surgical management or nonsurgical management using a hematoma cutoff size of 4 cm and a GCS score cutoff of 14. An alternative approach that has been proposed is to use volume of hematoma as an indication for intervention via open decompression or aspiration. Thirty cc has been proposed as a cutoff value of hematoma volume for intervention.19 Our study takes a different approach in patients presenting with spontaneous PFH, focusing on how a given patient’s PF and CCJ measurements may contribute to the severity of his or her symptoms, which may necessitate surgery. Patients with relatively smaller PF may not be able to tolerate PFH as well as patients with larger PF and may succumb to complications with relatively lower PFH volumes. If we apply preexisting approaches to determine the indication for surgery via imaging measurements on the patients in our study, we get results that do not correlate as well with the severity of the patients’ clinical presentation or need for surgery. If we were to use a conservation cutoff of average hemorrhage diameter of 3 cm or greater as an indication for surgery, then 15 patients would have been managed conservatively and 36 patients
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
Figure 6. Graph of posterior fossa volume surrogate measurements comparing patients managed surgically versus those managed without surgery. Values of y-axis are in cm3. The volume surrogate measurement in the surgical group ranged from 229.47e374.33 with a mean value of 297, whereas the volume surrogate measurement in the nonsurgical group ranged from 267.54e427.62 with a mean value of 328.66. This is a statistically significant difference with a P value of 0.019.
would have been considered surgical candidates. Using this cutoff, 14 patients would have proceeded to surgery who were successfully managed conservatively with minimal neurologic deficits in our study. If we were to use a cutoff of average hemorrhage diameter of 4 cm, then 31 patients would have been managed conservatively and 20 patients would have been considered surgical candidates. Using this cutoff, 2 patients would have proceeded to surgery who were successfully managed conservatively with minimal neurologic deficits and 3 patients would have been managed conservatively who—in our study—had small PF and presented with severe clinical deterioration. Finally, if we used hematoma volume of 30 cc as a cutoff to determine need for surgical management, then 19 patients would have been managed conservatively and 32 patients would have been considered surgical candidates. Using this cutoff,
www.journals.elsevier.com/world-neurosurgery
e575
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
Figure 7. Graph of posterior fossa hemorrhage score (PFHS) values comparing patients managed surgically versus those managed without surgery. The PFHS in the surgical group ranged from 11.7e20 with a mean value of 15.55, whereas the PFHS in the nonsurgical group ranged from 5. 3e13.1 with a mean value of 8.91. This is a statistically significant difference with a P value of <0.001.
11 patients would have proceeded to surgery who were successfully managed conservatively with minimal neurologic deficits, and 1 patient would have been managed conservatively who had a small PF and presented with severe clinical deterioration. It would appear then that measurements of hemorrhage size alone as discussed in prior literature do not tell the whole story. The combined measurement of the hemorrhage size and the individual patient’s PF anatomic constraints (as encapsulated by the PFHS) could help guide early surgical intervention before potentially devastating clinical decompensation. This objective measurement would be most useful in patients with spontaneous PFH who are “borderline.” Such patients would be those with relatively large hemorrhage diameters (>3e4 cm) but with minimal symptoms or those with relatively smaller hemorrhage diameters but with depressed GCS scores. By incorporating PF
e576
www.SCIENCEDIRECT.com
measurements such as the PFHS, such patients may be able to be classified into surgical or nonsurgical cases. Based on the distribution of PFHS in both groups, patients who score <11 could be considered nonsurgical patients, whereas those who score >14 could be considered surgical patients. Patients who score in the 11e14 range could be considered indeterminate and management would likely depend more heavily on the clinical status of the patient. A surgeon could adopt a more conservative policy in which any patient scoring >11 on the PFHS would be taken for surgical decompression or could take a more liberal policy in which any patient scoring <14 on the PFHS would be managed conservatively. Although the surrogate PF volume is not a true volumetric assessment as can be assessed by 3-dimensional volumetric analysis of the PF, we elected to use a product of 2-dimensional measurements that can be performed easily on any workstation. Although our method seems to indicate that anatomic criteria observed on imaging can influence necessity for suboccipital decompression, ultimately the decision will rely on the neurosurgical expertise and clinical judgement. An important technical consideration is the applicability of these techniques. Our CT protocol uses 4-mm width axial slice acquisition with 2-mm width coronal and sagittal reconstructions. Many imaging centers may not have CT protocols or scanners that allow for such slice dimensions, and this could limit the reliability of measurements. Furthermore, at our institution, patients who had their heads rotated or angled in the CT gantry at the time of acquisition will routinely have their axial sequences reformatted into more anatomic positioning and have their coronal and sagittal reconstructions performed from the repositioned sequence. This likewise may not be an available option for many imaging centers, and thus obtaining a true midline slice on sagittal imaging may be challenging. Also, MRI measurements were made on a sagittal T1 MPRAGE sequence, a 3-dimensional T1-weighted gradient recall echo sequence that allows for reformatting in multiple planes and enables measurement of the TDPF quickly. Although we do not anticipate that measurements performed on a more traditional sagittal T1 spin-echo sequence would be substantially different, this should be considered when applying measurement values. This study does have some limitations. With our relatively restrictive exclusion criteria, the patient group comprised only 51 patients, limiting the strength of the study. We only evaluated whether patients were or were not using therapeutic anticoagulation, and the use of other potentially relevant medications such as aspirin was not considered. Chart review also had its limitations as only the patients’ GCS score and major neurologic examination findings were consistently documented. For patients who had surgery, it was not always clear what imaging findings were considered for those patients. Specifically, it was not clear whether findings that have been evaluated in the neurosurgical literature, such as increasing fourth ventricle shift/compression or quadrigeminal cistern effacement, were considered in the management of the surgical patients. On review of the imaging prior to surgery, all patients who were taken to surgery had increasing effacement of the fourth ventricle and/or effacement of the quadrigeminal cistern. Whether these findings were a direct impetus to proceed to surgery was only documented in 3 of the surgical cases. Finally, underlying cerebellar parenchymal volume
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.03.208
ORIGINAL ARTICLE ETHAN A. NEUFELD ET AL.
EARLY CRANIAL PFH INTERVENTION
loss could be a factor that would allow patients to tolerate greater volumes of hemorrhage without clinical decompensation. Accurate volumetric assessment of the cerebellum is able to be performed via quantitative analysis of a volumetric MRI sequence, but this is not routinely available and cannot be performed with CT. Furthermore, the presence of hemorrhage at the time of examination would preclude accurate volumetric analysis. We advise that if a patient has profound preexisting cerebellar volume loss that the surgeon consider this as a factor that may allow a patient to tolerate conservative management with relatively larger hemorrhage volumes. CONCLUSIONS Our study is concordant with the hypothesis that larger hemorrhage volumes in spontaneous PFH at initial presentation are
REFERENCES 1. Luney MS, English SW, Longworth A, et al. Acute posterior cranial fossa hemorrhage: is surgical decompression better than expectant medical management? Neurocrit Care. 2016;25:365-370. 2. Sacco S, Marini C, Toni D, Olivieri L, Carolei A. Incidence and 10-year survival of intracerebral hemorrhage in a population-based registry. Stroke. 2009;40:394-399. 3. Salvetti M, Paini A, Bertacchini F, et al. Therapeutic approach to hypertensive emergencies: hemorrhagic stroke. High Blood Press Cardiovasc Prev. 2018;25:191-195. 4. Amar AP. Controversies in the neurosurgical management of cerebellar hemorrhage and infarction. Neurosurg Focus. 2012;32:E1. 5. Dinsdale HB. Spontaneous hemorrhage in the posterior fossa. A study of primary cerebellar and pontine hemorrhages with observations on their pathogenesis. Arch Neurol. 1964;10:200-217. 6. Karadan U, Supreeth RN, Manappallil RG, Jayakrishnan C. Twenty-four syndrome: an untold presentation of pontine hemorrhage. J Stroke Cerebrovasc Dis. 2018;27:e73-e74. 7. Dastur CK, Yu W. Current management of spontaneous intracerebral haemorrhage. Stroke Vasc Neurol. 2017;2:21-29. 8. Mathew P, Teasdale G, Bannan A, OluochOlunya D. Neurosurgical management of cerebellar haematoma and infarct. J Neurol Neurosurg Psychiatry. 1995;59:287-292.
more likely to require surgical decompression by clinical standards. Our results also demonstrate a statistically significant difference in the volume of PF between surgical and nonsurgical patients as determined by current clinical standards via our surrogate measurement. These findings suggest that patients with smaller PF volumes are more likely to require surgical decompression and may require decompression with relatively smaller volumes of hemorrhage. We suggest that these 2 factors are both predictors of the need for surgical decompression and should both be considered when initially evaluating a patient presenting with primary cerebellar PFH. Our derived PFHS—which can be performed using simple workstation measurements—incorporates these 2 factors and accentuates the differences between the 2 groups. The PFHS may be able to determine which patients will require early surgical decompression at initial presentation.
9. Arnone GD, Esfahani DR, Wonais M, et al. Surgery for cerebellar hemorrhage: a national surgical quality improvement program database analysis of patient outcomes and factors associated with 30-vday mortality and prolonged ventilation. World Neurosurg. 2017;106:543-550. 10. van Loon J, Van Calenbergh F, Goffin J, Plets C. Controversies in the management of spontaneous cerebellar haemorrhage. A consecutive series of 49 cases and review of the literature. Acta Neurochir (Wien). 1993;122:187-193. 11. Kirollos RW, Tyagi AK, Ross SA, van Hille PT, Marks PV. Management of spontaneous cerebellar hematomas: a prospective treatment protocol. Neurosurgery. 2001;49:1378-1386 [discussion: 13861377]. 12. Kobayashi S, Sato A, Kageyama Y, Nakamura H, Watanabe Y, Yamaura A. Treatment of hypertensive cerebellar hemorrhage–surgical or conservative management? Neurosurgery. 1994;34:246-250 [discussion: 250-241]. 13. Smoker WR. Craniovertebral junction: normal anatomy, craniometry, and congenital anomalies. Radiographics. 1994;14:255-277. 14. Sekula RF Jr, Jannetta PJ, Casey KF, Marchan EM, Sekula LK, McCrady CS. Dimensions of the posterior fossa in patients symptomatic for Chiari I malformation but without cerebellar tonsillar descent. Cerebrospinal Fluid Res. 2005;2:11. 15. Alperin N, Loftus JR, Oliu CJ, et al. Magnetic resonance imaging measures of posterior cranial fossa morphology and cerebrospinal fluid physiology in Chiari malformation type I. Neurosurgery. 2014;75:515-522 [discussion: 522].
WORLD NEUROSURGERY 127: e570-e577, JULY 2019
16. Rojas CA, Bertozzi JC, Martinez CR, Whitlow J. Reassessment of the craniocervical junction: normal values on CT. AJNR Am J Neuroradiol. 2007; 28:1819-1823. 17. Puffer RC, Graffeo C, Rabinstein A, Van Gompel JJ. Mortality rates after emergent posterior fossa decompression for ischemic or hemorrhagic stroke in older patients. World Neurosurg. 2016;92: 166-170. 18. Taneda M, Hayakawa T, Mogami H. Primary cerebellar hemorrhage. Quadrigeminal cistern obliteration on CT scans as a predictor of outcome. J Neurosurg. 1987;67:545-552. 19. Aguilar MI, Brott TG. Update in intracerebral hemorrhage. Neurohospitalist. 2011;1:148-159.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Contents of this submission were presented at the 2018 Annual Meeting of the American Society of Neuroradiology in Vancouver, British Columbia, Canada on June 6, 2018. Received 7 December 2018; accepted 20 March 2019 Citation: World Neurosurg. (2019) 127:e570-e577. https://doi.org/10.1016/j.wneu.2019.03.208 Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.
www.journals.elsevier.com/world-neurosurgery
e577