- Email: [email protected]
Reliable? The Value of Early Postoperative Magnetic Resonance Imaging after Cerebral Cavernous Malformation Surgery
Accepted Manuscript Reliable? The Value of Early Postoperative Magnetic Resonance Imaging after CCM Surgery Bixia Chen, Sophia Göricke, Karsten Wrede, Ramazan Jabbarli, Max Jaegersberg, Ulrich Sure, Philipp Dammann PII:
S1878-8750(17)30458-8
DOI:
10.1016/j.wneu.2017.03.135
Reference:
WNEU 5504
To appear in:
World Neurosurgery
Received Date: 20 December 2016 Revised Date:
27 March 2017
Accepted Date: 28 March 2017
Please cite this article as: Chen B, Göricke S, Wrede K, Jabbarli R, Jaegersberg M, Sure U, Dammann P, Reliable? The Value of Early Postoperative Magnetic Resonance Imaging after CCM Surgery, World Neurosurgery (2017), doi: 10.1016/j.wneu.2017.03.135. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Reliable? The Value of Early Postoperative Magnetic Resonance Imaging after CCM Surgery Bixia Chen1, Sophia Göricke2, Karsten Wrede1, Ramazan Jabbarli1, Max Jaegersberg3, Ulrich Sure1 and Philipp Dammann1,3 1
Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Essen,
RI PT
2
Department of Neurosurgery, University Hospital Essen, Essen, Germany
Germany
Department of Neurosurgery, University Hospital of Geneva, Geneva, Switzerland
M AN U
Corresponding author’s name and complete mailing address: Bixia Chen, MD University Hospital Essen Department of Neurosurgery Hufelandstrasse 55 45147 Essen, Germany
+49 201 723 1220
Email: [email protected]
TE D
Phone: +49 201 723 83165 Fax:
SC
3
Cavernous malformation, MRI, SWI, CCM, Cavernoma
Abstract:
257 words
Manuscript:
EP
Keywords:
words 1
Tables:
2
AC C
Figures:
References:
13
ACCEPTED MANUSCRIPT Abstract Background: Cerebral cavernous malformations (CCM) can cause intracerebral hemorrhage. The lesions themselves are frequently associated with perifocal hemosiderin deposits due to repetitive microhemorrhages. Main indications for a surgical treatment are recurrent symptomatic hemorrhages or cavernoma-related epilepsy (CRE). After surgical resection, follow-up MR-imaging is usually performed to confirm a) the complete
RI PT
resection of the CCM and, especially in cases of CRE, b) the complete resection of the hemosiderin deposits. Material and Methods: This prospective study evaluates the value of early postoperative MRI (within 72 hours) regarding the detection of CCM or hemosiderin remnants compared to a “standard” 3-6 months postoperative
SC
MRI control in 61 CCM cases.
Results: Sensitivity of early postoperative MRI for CCM remnant detection was 66.67% (95% CI 9.43% 99.16%), specificity was 76.74% (95% CI 61.37% - 88.24%), positive predictive value was 16.67% (95% CI
M AN U
2.09% – 48.41%) and negative predictive value was 97.06% (95% CI 84.67% - 99.93%). Due to the high number of patients that could not be evaluated due to imaging artifacts, sensitivity and specificity analysis was not performed for early postoperative MRI using T2*/SWI to assess hemosiderin remnants. Sensitivity of early postoperative MRI for hemosiderin remnant detection using T2 weighted sequences was 85.71% (95% CI 63.66% - 96.95%), specificity was 66.67% (95% CI 44.68% - 84.37%), positive predictive value was 69.23%
TE D
(95% CI 55.45% – 80.27%) and negative predictive value was 84.21% (95% CI 64.31% - 94.04%). Conclusion: Our data suggests that early postoperative MRI after CCM surgery is often hampered by imaging artefacts creating false positive results and therefore ineligible for a resection control. However, reliability of a
AC C
remnants.
EP
negative result on early postoperative T2 weighted MRI is relatively high regarding both CCM and hemosiderin
Introduction
Cerebral cavernous malformations (CCM) are low-flow vascular lesions that can cause intracerebral hemorrhage. The lesions mainly become symptomatic with epileptic seizures or focal neurological deficits [1]. In surgically treated patients, a complete resection of the CCM is mandatory to prevent future hemorrhage events[7]. Especially in patients with cavernoma related epilepsy (CRE) [10], additionally the complete resection of the peri-lesional hemosiderin rim is crucial as these patients are thought to have a better chance to stay seizure-free after surgery[3,11]. In clinical practice, resection extent is usually assessed using follow up
ACCEPTED MANUSCRIPT MRI. Immediately after CCM surgery postoperative changes like hemorrhagic products, edema, intracranial air inclusions and materials used for hemostasis may limit the assessment of CCM and hemosiderin remnants. However, confirming the complete resection of CCM and hemosiderin deposits directly after surgery may appear advantageous for both the surgeon and the patient. Furthermore, early postoperative MRI within 72h has been utilized in clinical series to assess CCM remnants [2]. Until now, the value and reliability of such early
RI PT
postoperative MRI has not yet been systematically examined. This prospective study aims to evaluate early postoperative MRI after CCM surgery regarding the resection extent of the CCM and hemosiderin deposits as well as the assessment of potential perioperative complications in comparison to a “gold standard” of a 3-6
SC
months follow-up MRI examination.
M AN U
Materials and Methods Study Design and Population
Sixty-one consecutive patients undergoing surgery for cranial or spinal CCM in the Department of Neurosurgery, University Hospital Essen, Germany were included in the study group. Each subject underwent
TE D
three MRI examinations: preoperative, early postoperative during hospital admission (within 3 days after surgery) and after a minimum of three - six months (“control”). An additional early postoperative computed tomography (CT) examination was routinely performed within 24 hours after surgery in cerebral CCM cases.
EP
The clinical status was assessed using the modified Rankin Scale (mRS) at admission, at the time of hospital discharge, and after 6 months. The study was conducted according to the principles expressed in the Declaration
AC C
of Helsinki and was approved by the local university institutional review board as part of a superordinate clinical study on CCM patients (review board identification 14-5751-BO) which assesses functional outcome and quality of life after CCM surgery.
Imaging All imaging was performed on whole-body 1.5 Tesla MRI systems. Preoperative MRI protocols included T1 (with and without contrast agent), T2 and T2*/SWI sequences, early postoperative MRI protocols included T1 with and without contrast agent (Gadolinium), T2, Diffusion and SWI/T2* sequences. Follow-up MRI protocols
ACCEPTED MANUSCRIPT were mainly performed in external departments and the protocols slightly varied (all including minimum T2, T1 (with / without contrast agent), T2* sequences).
RI PT
Image Analysis All preoperative and follow-up MRI and CT data were analyzed by raters experienced in CCM MRI (neuroradiologist (SG) / neurosurgeon (PD, US)). Postoperative data was assessed in consensus rating. In the preoperative MRI examinations, CCM size (max. diameter [mm]) and localization, presence of a hemosiderin
SC
rim and developmental venous anomaly were assessed. Early postoperative MR images were assessed focusing on presence of (1) CCM remnant, (2) hemosiderin rim remnant (on T2 or T2* and/or SWI), 3) integrity of
M AN U
developmental venous anomaly (DVA), (4) hemorrhage, and (5) diffusion impairment. CT images were analyzed for signs of hemorrhage or stroke. In the follow up MRI, presence of CCM remnant, hemosiderin rim remnant (on T2, T2* and/or SWI) and integrity of developmental venous anomaly were assessed. All ratings included the options: yes, no, not assessable.
TE D
Statistical Analysis
Statistical analysis was performed using IBM SPSS Statistics 22 (SPSS Inc., IBM Corp., North Castle, New
EP
York, USA). Interval- scaled data were expressed as mean and standard deviations, and nominal data were expressed as absolute numbers and valid percent. Data were tested for normal distribution by conducting a Shapiro-Wilk test, in addition to histograms and Q-Q plots. We used parametric statistics for between-group
AC C
comparison for normally distributed data and nonparametric statistics for non- normally distributed data. For categorical variables, c2 or Fisher exact tests (expected frequencies less than 5) were applied. Sensitivity and specificity calculations of the early postoperative MRI were performed, compared to the 3 months postoperative MRI as reference results.
Results In all 61 patients, CCM diagnosis was confirmed by histopathology. Mean age was 43.7 years (range: 2 – 73 years), 33 patients were female, 28 were male. CCM localizations included brain stem (n=40), frontal lobe
ACCEPTED MANUSCRIPT (n=5), temporal lobe (n=5), cerebellum (n=4), basal ganglia (n=3), spinal cord (n=3), and parietal lobe (n=1). Hemosiderin rims were preoperatively found in 48 subjects (79%), 22 patients had associated DVAs (36%). For baseline characteristics of the cohort see Table 1.
RI PT
Imaging results CCM remnants
CCM remnants on early postoperative MRI (using T1 and T2 sequences) were diagnosed in 12 patients (20%),
SC
34 (56%) were assessed as completely resected, and 15 patients (25%) could not be evaluated due to imaging artifacts and postoperative changes. In the 3-months follow up MRI rating, 3 patients (5%) were classified as
M AN U
incompletely (one patient (2%) with a clear remnant and 2 patients (3%) with suspicion of minimal remnant) and 58 patients (95%) as completely resected regarding CCM remnants. Consequently, sensitivity of early postoperative MRI for CCM remnant detection was 66.67% (95% CI 9.43% - 99.16%), specificity was 76.74% (95% CI 61.37% - 88.24%), positive likelihood ratio was 2.87 (95% CI 1.09 – 7.54), negative likelihood ratio was 0.43 (95% CI 0.09 – 2.17), positive predictive value was 16.67% (95% CI 2.09% – 48.41%) and negative
Hemosiderin remnants
TE D
predictive value was 97.06% (95% CI 84.67% - 99.93%).
EP
Hemosiderin remnants on early postoperative MRI using T2*/SWI were diagnosed in 2 patients (4%). In 3 patients (6%) a complete resection was diagnosed and 43 patients (90%) could not be evaluated due to imaging artifacts. Hemosiderin remnants on early postoperative MRI using T2 weighted sequences were diagnosed in 26
AC C
patients (52%). In 19 patients (40%) a complete resection was diagnosed and 3 patients (6%) could not be evaluated due to imaging artifacts. On follow-up postoperative MRI (using T2, T2*/SWI), incomplete resection was diagnosed in 21 patients (44%), complete resection in 27 patients (56%). Due to the high number of patients that could not be evaluated due to imaging artifacts, sensitivity and specificity analysis was not performed for early postoperative MRI using T2*/SWI to assess hemosiderin remnants. Sensitivity of early postoperative MRI for hemosiderin remnant detection using T2 weighted sequences was 85.71% (95% CI 63.66% - 96.95%), specificity was 66.67% (95% CI 44.68% - 84.37%), positive likelihood ratio was 2.57 (95% CI 1.42 – 4.65), negative likelihood ratio was 0.21 (95% CI 0.07 – 0.63), positive predictive value was 69.23% (95% CI 55.45% – 80.27%) and negative predictive value was 84.21% (95% CI 64.31% - 94.04%).
ACCEPTED MANUSCRIPT
Imaging rating results and results of diagnostic testing can be found in table 2 and 3. Early and follow up MRI readings on developmental venous anomalies were identical: DVAs were completely intact in 18 patients, partially intact in 2 patients, and resected in 2 patients. Hemorrhage was postoperatively
RI PT
detected in 9 early MRI examinations and in 6 CT examinations. Two patients underwent revision surgery due to hemorrhages with mass effect, which were identified on both CT and MRI. None of the additionally MRI diagnosed hemorrhages led to reoperation. Peri-lesional diffusion impairment was seen in 10 patients with early MRI. In a univariate analysis, it was associated with the probability of a functional deterioration on mRS at 6
SC
months control, compared to preoperative condition (Odds ratio 6.1333 (95% CI 1.28 – 29.36, p=0.0232)). Hypertrophic olivary degeneration was diagnosed in 4 patients (7%) in the follow-up MRI scan after 3 months.
M AN U
A typical MR imaging finding is presented in Figure 1.
Functional outcome
On admission, two patients were symptom free (mRS = 0), 25 patients had mild symptoms without significant
TE D
disability (mRS =1), 25 patients had a slight disability (mRS = 2), 5 patients suffered from moderate (mRS = 3) and 4 patients from moderately severe disability (mRS = 4). At time of discharge, 3 patients’ mRS improved, 39 patients remained at the same mRS level, 19 patients deteriorated. After six months, improvement was observed
EP
in 3 patients, the same status as preoperatively in 49 patients, and deterioration in 9 patients. 6 months
AC C
postoperative morbidity was consequently 15%.
Revision surgery for CCM remnants One patient (2%) with a clear CCM remnant on follow-up MRI underwent subsequent revision surgery. The two cases with suspicion of minimal remnant underwent further MRI follow-up (36 months). In both cases no rehemorrhage or other signal changes within the lesions occurred. See also Figure 2.
ACCEPTED MANUSCRIPT Discussion While preoperative MRI of CCM is clinically well established [8] and the subject of various studies[4-6,9,12], postoperative MRI of CCM has not been systematically analyzed yet and only few case series regarding this subject exist[2,7]. Nonetheless, postoperative MRI is a crucial technique to verify the achievement of treatment objectives such as complete resection of CCM and hemosiderin deposits (especially in CRE cases[11]). Kikuta et
RI PT
al.[7] analyzed a series of 10 patients with brainstem CCM which all underwent postoperative MRI within 2 months postoperatively. They found a correlation of incompletely resected CCM with a risk of a subsequent rehemorrhage. However, the method of image assessment and evaluation was not further specified. Cenzato et al.
SC
[2] presented a series of 30 brainstem CCM which were assessed by postoperative MRI within 72h after surgery. Remnants of CCM were described in 3 cases and these underwent subsequent revision surgery. Anyhow, results
M AN U
of MRI diagnosis were not further analyzed or compared with a second (subsequent) MRI examination. In our opinion, it seems clinically more reasonable to verify a complete resection of a CCM just after several weeks when all surgical artifacts (air, blood, hemostyptics) that may interfere with diagnosis of remnants are resorbed. This hypothesis was tested in this series, while potential benefits of an early MRI were tried to assess. Our results suggest that, using the three to six months postoperative MRI as a “gold standard” reference, the
TE D
value of early postoperative MRI is low for detecting/excluding both CCM (2%-5% prevalence) and hemosiderin remnants (44% prevalence). With an acceptable negative predictive value of 97.06% a diagnosis of a complete resection of the CCM is in fact very likely to be confirmed in the 3 months postoperative MRI, however, a positive predictive value of only 16.07% implies that the diagnosis of a CCM remnant on early
EP
postoperative MRI (potentially leading to revision surgery) is quite unreliable. Early T2*/SWI imaging to detect possible hemosiderin remnants proved to be completely inefficient with a drop out rate (no distinct diagnosis
AC C
possible due to artifacts) of 90%. T2 weighted imaging in this regard showed average positive and negative predictive values of 69.23% and 84.21%, respectively, making it also more reliable especially to exclude postoperative hemosiderin deposits.
It should be noted that the relatively high incidence of hemosiderin
remnants (44%) in our series can be attributed to the high proportion of brainstem cases in which the resection of peri-lesional hemosiderin is normally not intended. Apart from CCM or hemosiderin remnants, early postoperative MRI proved to be reliable in the diagnosis of associated DVA integrity, which was identified correctly in all cases on early MR images when compared to 3 months postoperative results. Compared to findings from routine postoperative CT scan, findings of early MRI led to no treatment changes in the acute management of our patients and showed no obvious additional value in
ACCEPTED MANUSCRIPT their acute management. However, regarding prognosis for functional outcome (e.g. in cases of brainstem CCM), early postoperative MRI may provide additional information.
RI PT
Study limitations One limitation of this study is of general nature. It may be quite hard to distinguish between very small remnants of a CCM (e.g. parts of a cavern) and simple postoperative changes, even on the 3-6 months postoperative images. 2 of our 3 CCM remnant diagnoses are of such kind (see figure 2). Furthermore, the risk for these
SC
“potential” remnants to cause (re-) hemorrhage is in our opinion negligible. Only further follow-up MR imaging may reveal a certain “activity” within the lesion that can confirm the suspicion of a real CCM remnant (which
M AN U
was not the case in this study). Another study limitation was that no standardized in-house 3-6 months postoperative MRI protocol was followed. This led to slight technical differences, e.g. in-plane resolution, slice thickness, MRI system, which may have hampered image analysis.
TE D
Conclusion
In this study, early postoperative MRI was not reliable in detecting postoperative CCM or hemosiderin remnant. Especially the suspicion of a remnant presence oftentimes proved to be false positive. It delivered no additional
EP
value regarding acute postoperative management when compared to postoperative CT. Thus, early postoperative MRI, in addition resource demanding and cost-intensive, cannot be recommended as a routine procedure. However, in cases of unclear neurological impairment after CCM surgery (especially in highly eloquent lesions
AC C
such as brainstem CCM), early MRI can deliver addition information that may be of relevance, especially regarding functional outcome.
References
1. Aiba T, Tanaka R, Koike T, Kameyama S, Takeda N, Komata T (1995) Natural history of intracranial cavernous malformations. J Neurosurg 83:56-59. doi:10.3171/jns.1995.83.1.0056 2. Cenzato M, Stefini R, Ambrosi C, Giovanelli M (2008) Post-operative remnants of brainstem cavernomas: incidence, risk factors and management. Acta Neurochir (Wien) 150:879-886; discussion 887. doi:10.1007/s00701-008-0008-4
ACCEPTED MANUSCRIPT 3. Dammann P, Schaller C, Sure U (2016) Should we resect peri-lesional hemosiderin deposits when performing lesionectomy in patients with cavernoma-related epilepsy (CRE)? Neurosurg Rev. doi:10.1007/s10143-016-0797-5 4. Dammann P, Wrede KH, Maderwald S, El Hindy N, Mueller O, Chen B, Zhu Y, Hutter BO, Ladd ME, Schlamann M, Sandalcioglu IE, Sure U (2013) The venous angioarchitecture of sporadic cerebral cavernous malformations: a susceptibility weighted imaging study at 7 T MRI. J
RI PT
Neurol Neurosurg Psychiatry 84:194-200. doi:10.1136/jnnp-2012-302599
5. de Champfleur NM, Langlois C, Ankenbrandt WJ, Le Bars E, Leroy MA, Duffau H, Bonafe A, Jaffe J, Awad IA, Labauge P (2011) Magnetic resonance imaging evaluation of cerebral
cavernous malformations with susceptibility-weighted imaging. Neurosurgery 68:641-647;
SC
discussion 647-648. doi:10.1227/NEU.0b013e31820773cf
6. Hart BL, Taheri S, Rosenberg GA, Morrison LA (2013) Dynamic contrast-enhanced MRI
doi:10.1007/s12975-013-0285-y
M AN U
evaluation of cerebral cavernous malformations. Transl Stroke Res 4:500-506.
7. Kikuta K, Nozaki K, Takahashi JA, Miyamoto S, Kikuchi H, Hashimoto N (2004) Postoperative evaluation of microsurgical resection for cavernous malformations of the brainstem. J Neurosurg 101:607-612. doi:10.3171/jns.2004.101.4.0607
8. Klostranec JM, Krings T (2015) Neuroimaging of cerebral cavernous malformations. J Neurosurg Sci 59:221-235
TE D
9. Mikati AG, Khanna O, Zhang L, Girard R, Shenkar R, Guo X, Shah A, Larsson HB, Tan H, Li L, Wishnoff MS, Shi C, Christoforidis GA, Awad IA (2015) Vascular permeability in cerebral cavernous malformations. J Cereb Blood Flow Metab 35:1632-1639. doi:10.1038/jcbfm.2015.98
EP
10. Rosenow F, Alonso-Vanegas MA, Baumgartner C, Blumcke I, Carreno M, Gizewski ER, Hamer HM, Knake S, Kahane P, Luders HO, Mathern GW, Menzler K, Miller J, Otsuki T, Ozkara C, Pitkanen A, Roper SN, Sakamoto AC, Sure U, Walker MC, Steinhoff BJ, Surgical Task Force
AC C
CoTSotI (2013) Cavernoma-related epilepsy: review and recommendations for management-report of the Surgical Task Force of the ILAE Commission on Therapeutic Strategies.
Epilepsia 54:2025-2035. doi:10.1111/epi.12402
11. Ruan D, Yu XB, Shrestha S, Wang L, Chen G (2015) The Role of Hemosiderin Excision in Seizure Outcome in Cerebral Cavernous Malformation Surgery: A Systematic Review and Meta-Analysis. PLoS One 10:e0136619. doi:10.1371/journal.pone.0136619 12. Tan H, Liu T, Wu Y, Thacker J, Shenkar R, Mikati AG, Shi C, Dykstra C, Wang Y, Prasad PV, Edelman RR, Awad IA (2014) Evaluation of iron content in human cerebral cavernous malformation using quantitative susceptibility mapping. Invest Radiol 49:498-504. doi:10.1097/RLI.0000000000000043
ACCEPTED MANUSCRIPT
Figure Legends Figure 1: Illustrative case of a false positive early postoperative finding of a CCM remnant. (A) and (B) showing T2
RI PT
weighted preoperative images of a pontine CCM. (C) showing suspected CCM remnant in the superior parts of the resection cavity on early postoperative MRI. (D) presenting the late postoperative MRI without suspicion of
SC
CCM remnant but minor residual peri-lesional hemosiderin deposits.
Figure 2:
Illustrative cases of diagnosed CCM remnants on 3-6 months follow-up MRI. (A) and (B) showing T2 sequences
M AN U
of a large “infiltrative” cerebellar CCM preoperatively (A) and a clear remnant (arrows) on the upper medial aspect on postoperative (B) imaging. The patient underwent revision surgery. (D) shows postoperative T2 sequences demonstrating signal changes at the bottom of the resection cavity of a large pontine CCM (C) which
AC C
EP
TE D
were also interpreted as a potential CCM remnant. The patient underwent follow-up MRIs.
MANUSCRIPT Table 1. Baseline characteristics of theACCEPTED cohort
AC C
EP
TE D
M AN U
SC
RI PT
Age [y] 43.7 (2- 73) Female [n] 33 (54%) Localization [n] supratentorial 14 (23%) infratentorial 47 (77%) Hemosiderin deposits* 48 (79%) Associated DVA 22 (36%) Max. diameter CCM [mm] 17.4 (±7.7) Postoperative functional outcome at 6 months† stable/improved 52 (85% ) deteriorated 9 (15%) * on preoperative MRI † on modified Rankin Scale
Table 2. Rating of postoperative MRI ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Early MRI* Control MRI † CCM remnant present 12 (20%) 3 (5%) not present 34 (56%) 57 (95%) not assessable/unclear 15 (25%) 0 (0%) Hemosiderin remnant (T2*/SWI) present 2 (4%) 21 (44%) not present 3 (6%) 27 (56%) not assessable/unclear 43 (90%) 0 (0%) Hemosiderin remnant (T2) present 26 (55%) 21 (44%) not present 9 (19%) 27 (56%) not assessable/unclear 12 (26%) 0 (0%) DVA intact 18 (82%) 18 (82%) partially intact 2 (9%) 2 (9%) resected 2 (9%) 2 (9%) * within 72h postoperatively † within 3-6 months postoperatively, hemosiderin remnants were assessed using T2*, SWI and T2 combined
ACCEPTED MANUSCRIPT Table 3 Diagnostic testing results CCM remnant diagnostic test
negative n=34
false negative (fn) n=1
true negative (tn) n=33
sensitivity [tp/ (tp+fn)] 66.67%
specificity [tn/(fp+tn)] 76.74%
not evaluable n=15
hemosiderin remnant diagnostic test
positive n=26 negative n=19
false negative (fn) n=3
true negative (tn) n=16
sensitivity [tp/ (tp+fn)] 85.71%
specificity [tn/(fp+tn)] 66.67%
not evaluable n=3
using T1,T2 sequences combined, 3using T2 sequences, 4using T2,T2* and SWI (not in all cases available) sequences combined
AC C
EP
TE D
1,2
positive predictive value [tp /(tp+fp)] 69.23% negative predictive value [tn/(fn+tn)] 84.21%
M AN U
early postoperative MRI3
follow-up MRI4 positive negative true positive (tp) false positive (fp) n=18 n=8
positive predictive value [tp /(tp+fp)] 16.67% negative predictive value [tn/(fn+tn)] 97.06%
RI PT
positive n=12
SC
early postoperative MRI1
follow-up MRI2 positive negative true positive (tp) false positive (fp) n=2 n=10
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Highlights:
RI PT SC M AN U TE D
• •
EP
•
Early postoperative compared to follow-up MRI after cavernous malformation surgery Early MRI is not reliable in detection of cavernoma or hemosiderin remnants. It is often hampered by artifacts and ineligible for resection control Early MRI delivers no additional value to acute postoperative management
AC C
•
ACCEPTED MANUSCRIPT Abbreviations list
EP
TE D
M AN U
SC
RI PT
Cerebral Cavernous Malformation Cavernoma-Related Epilepsy Magnetic Resonance Imaging Computed Tomography modified Ranking Scale
AC C
CCM CRE MRI CT mRS
S1878-8750(17)30458-8
DOI:
10.1016/j.wneu.2017.03.135
Reference:
WNEU 5504
To appear in:
World Neurosurgery
Received Date: 20 December 2016 Revised Date:
27 March 2017
Accepted Date: 28 March 2017
Please cite this article as: Chen B, Göricke S, Wrede K, Jabbarli R, Jaegersberg M, Sure U, Dammann P, Reliable? The Value of Early Postoperative Magnetic Resonance Imaging after CCM Surgery, World Neurosurgery (2017), doi: 10.1016/j.wneu.2017.03.135. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Reliable? The Value of Early Postoperative Magnetic Resonance Imaging after CCM Surgery Bixia Chen1, Sophia Göricke2, Karsten Wrede1, Ramazan Jabbarli1, Max Jaegersberg3, Ulrich Sure1 and Philipp Dammann1,3 1
Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Essen,
RI PT
2
Department of Neurosurgery, University Hospital Essen, Essen, Germany
Germany
Department of Neurosurgery, University Hospital of Geneva, Geneva, Switzerland
M AN U
Corresponding author’s name and complete mailing address: Bixia Chen, MD University Hospital Essen Department of Neurosurgery Hufelandstrasse 55 45147 Essen, Germany
+49 201 723 1220
Email: [email protected]
TE D
Phone: +49 201 723 83165 Fax:
SC
3
Cavernous malformation, MRI, SWI, CCM, Cavernoma
Abstract:
257 words
Manuscript:
EP
Keywords:
words 1
Tables:
2
AC C
Figures:
References:
13
ACCEPTED MANUSCRIPT Abstract Background: Cerebral cavernous malformations (CCM) can cause intracerebral hemorrhage. The lesions themselves are frequently associated with perifocal hemosiderin deposits due to repetitive microhemorrhages. Main indications for a surgical treatment are recurrent symptomatic hemorrhages or cavernoma-related epilepsy (CRE). After surgical resection, follow-up MR-imaging is usually performed to confirm a) the complete
RI PT
resection of the CCM and, especially in cases of CRE, b) the complete resection of the hemosiderin deposits. Material and Methods: This prospective study evaluates the value of early postoperative MRI (within 72 hours) regarding the detection of CCM or hemosiderin remnants compared to a “standard” 3-6 months postoperative
SC
MRI control in 61 CCM cases.
Results: Sensitivity of early postoperative MRI for CCM remnant detection was 66.67% (95% CI 9.43% 99.16%), specificity was 76.74% (95% CI 61.37% - 88.24%), positive predictive value was 16.67% (95% CI
M AN U
2.09% – 48.41%) and negative predictive value was 97.06% (95% CI 84.67% - 99.93%). Due to the high number of patients that could not be evaluated due to imaging artifacts, sensitivity and specificity analysis was not performed for early postoperative MRI using T2*/SWI to assess hemosiderin remnants. Sensitivity of early postoperative MRI for hemosiderin remnant detection using T2 weighted sequences was 85.71% (95% CI 63.66% - 96.95%), specificity was 66.67% (95% CI 44.68% - 84.37%), positive predictive value was 69.23%
TE D
(95% CI 55.45% – 80.27%) and negative predictive value was 84.21% (95% CI 64.31% - 94.04%). Conclusion: Our data suggests that early postoperative MRI after CCM surgery is often hampered by imaging artefacts creating false positive results and therefore ineligible for a resection control. However, reliability of a
AC C
remnants.
EP
negative result on early postoperative T2 weighted MRI is relatively high regarding both CCM and hemosiderin
Introduction
Cerebral cavernous malformations (CCM) are low-flow vascular lesions that can cause intracerebral hemorrhage. The lesions mainly become symptomatic with epileptic seizures or focal neurological deficits [1]. In surgically treated patients, a complete resection of the CCM is mandatory to prevent future hemorrhage events[7]. Especially in patients with cavernoma related epilepsy (CRE) [10], additionally the complete resection of the peri-lesional hemosiderin rim is crucial as these patients are thought to have a better chance to stay seizure-free after surgery[3,11]. In clinical practice, resection extent is usually assessed using follow up
ACCEPTED MANUSCRIPT MRI. Immediately after CCM surgery postoperative changes like hemorrhagic products, edema, intracranial air inclusions and materials used for hemostasis may limit the assessment of CCM and hemosiderin remnants. However, confirming the complete resection of CCM and hemosiderin deposits directly after surgery may appear advantageous for both the surgeon and the patient. Furthermore, early postoperative MRI within 72h has been utilized in clinical series to assess CCM remnants [2]. Until now, the value and reliability of such early
RI PT
postoperative MRI has not yet been systematically examined. This prospective study aims to evaluate early postoperative MRI after CCM surgery regarding the resection extent of the CCM and hemosiderin deposits as well as the assessment of potential perioperative complications in comparison to a “gold standard” of a 3-6
SC
months follow-up MRI examination.
M AN U
Materials and Methods Study Design and Population
Sixty-one consecutive patients undergoing surgery for cranial or spinal CCM in the Department of Neurosurgery, University Hospital Essen, Germany were included in the study group. Each subject underwent
TE D
three MRI examinations: preoperative, early postoperative during hospital admission (within 3 days after surgery) and after a minimum of three - six months (“control”). An additional early postoperative computed tomography (CT) examination was routinely performed within 24 hours after surgery in cerebral CCM cases.
EP
The clinical status was assessed using the modified Rankin Scale (mRS) at admission, at the time of hospital discharge, and after 6 months. The study was conducted according to the principles expressed in the Declaration
AC C
of Helsinki and was approved by the local university institutional review board as part of a superordinate clinical study on CCM patients (review board identification 14-5751-BO) which assesses functional outcome and quality of life after CCM surgery.
Imaging All imaging was performed on whole-body 1.5 Tesla MRI systems. Preoperative MRI protocols included T1 (with and without contrast agent), T2 and T2*/SWI sequences, early postoperative MRI protocols included T1 with and without contrast agent (Gadolinium), T2, Diffusion and SWI/T2* sequences. Follow-up MRI protocols
ACCEPTED MANUSCRIPT were mainly performed in external departments and the protocols slightly varied (all including minimum T2, T1 (with / without contrast agent), T2* sequences).
RI PT
Image Analysis All preoperative and follow-up MRI and CT data were analyzed by raters experienced in CCM MRI (neuroradiologist (SG) / neurosurgeon (PD, US)). Postoperative data was assessed in consensus rating. In the preoperative MRI examinations, CCM size (max. diameter [mm]) and localization, presence of a hemosiderin
SC
rim and developmental venous anomaly were assessed. Early postoperative MR images were assessed focusing on presence of (1) CCM remnant, (2) hemosiderin rim remnant (on T2 or T2* and/or SWI), 3) integrity of
M AN U
developmental venous anomaly (DVA), (4) hemorrhage, and (5) diffusion impairment. CT images were analyzed for signs of hemorrhage or stroke. In the follow up MRI, presence of CCM remnant, hemosiderin rim remnant (on T2, T2* and/or SWI) and integrity of developmental venous anomaly were assessed. All ratings included the options: yes, no, not assessable.
TE D
Statistical Analysis
Statistical analysis was performed using IBM SPSS Statistics 22 (SPSS Inc., IBM Corp., North Castle, New
EP
York, USA). Interval- scaled data were expressed as mean and standard deviations, and nominal data were expressed as absolute numbers and valid percent. Data were tested for normal distribution by conducting a Shapiro-Wilk test, in addition to histograms and Q-Q plots. We used parametric statistics for between-group
AC C
comparison for normally distributed data and nonparametric statistics for non- normally distributed data. For categorical variables, c2 or Fisher exact tests (expected frequencies less than 5) were applied. Sensitivity and specificity calculations of the early postoperative MRI were performed, compared to the 3 months postoperative MRI as reference results.
Results In all 61 patients, CCM diagnosis was confirmed by histopathology. Mean age was 43.7 years (range: 2 – 73 years), 33 patients were female, 28 were male. CCM localizations included brain stem (n=40), frontal lobe
ACCEPTED MANUSCRIPT (n=5), temporal lobe (n=5), cerebellum (n=4), basal ganglia (n=3), spinal cord (n=3), and parietal lobe (n=1). Hemosiderin rims were preoperatively found in 48 subjects (79%), 22 patients had associated DVAs (36%). For baseline characteristics of the cohort see Table 1.
RI PT
Imaging results CCM remnants
CCM remnants on early postoperative MRI (using T1 and T2 sequences) were diagnosed in 12 patients (20%),
SC
34 (56%) were assessed as completely resected, and 15 patients (25%) could not be evaluated due to imaging artifacts and postoperative changes. In the 3-months follow up MRI rating, 3 patients (5%) were classified as
M AN U
incompletely (one patient (2%) with a clear remnant and 2 patients (3%) with suspicion of minimal remnant) and 58 patients (95%) as completely resected regarding CCM remnants. Consequently, sensitivity of early postoperative MRI for CCM remnant detection was 66.67% (95% CI 9.43% - 99.16%), specificity was 76.74% (95% CI 61.37% - 88.24%), positive likelihood ratio was 2.87 (95% CI 1.09 – 7.54), negative likelihood ratio was 0.43 (95% CI 0.09 – 2.17), positive predictive value was 16.67% (95% CI 2.09% – 48.41%) and negative
Hemosiderin remnants
TE D
predictive value was 97.06% (95% CI 84.67% - 99.93%).
EP
Hemosiderin remnants on early postoperative MRI using T2*/SWI were diagnosed in 2 patients (4%). In 3 patients (6%) a complete resection was diagnosed and 43 patients (90%) could not be evaluated due to imaging artifacts. Hemosiderin remnants on early postoperative MRI using T2 weighted sequences were diagnosed in 26
AC C
patients (52%). In 19 patients (40%) a complete resection was diagnosed and 3 patients (6%) could not be evaluated due to imaging artifacts. On follow-up postoperative MRI (using T2, T2*/SWI), incomplete resection was diagnosed in 21 patients (44%), complete resection in 27 patients (56%). Due to the high number of patients that could not be evaluated due to imaging artifacts, sensitivity and specificity analysis was not performed for early postoperative MRI using T2*/SWI to assess hemosiderin remnants. Sensitivity of early postoperative MRI for hemosiderin remnant detection using T2 weighted sequences was 85.71% (95% CI 63.66% - 96.95%), specificity was 66.67% (95% CI 44.68% - 84.37%), positive likelihood ratio was 2.57 (95% CI 1.42 – 4.65), negative likelihood ratio was 0.21 (95% CI 0.07 – 0.63), positive predictive value was 69.23% (95% CI 55.45% – 80.27%) and negative predictive value was 84.21% (95% CI 64.31% - 94.04%).
ACCEPTED MANUSCRIPT
Imaging rating results and results of diagnostic testing can be found in table 2 and 3. Early and follow up MRI readings on developmental venous anomalies were identical: DVAs were completely intact in 18 patients, partially intact in 2 patients, and resected in 2 patients. Hemorrhage was postoperatively
RI PT
detected in 9 early MRI examinations and in 6 CT examinations. Two patients underwent revision surgery due to hemorrhages with mass effect, which were identified on both CT and MRI. None of the additionally MRI diagnosed hemorrhages led to reoperation. Peri-lesional diffusion impairment was seen in 10 patients with early MRI. In a univariate analysis, it was associated with the probability of a functional deterioration on mRS at 6
SC
months control, compared to preoperative condition (Odds ratio 6.1333 (95% CI 1.28 – 29.36, p=0.0232)). Hypertrophic olivary degeneration was diagnosed in 4 patients (7%) in the follow-up MRI scan after 3 months.
M AN U
A typical MR imaging finding is presented in Figure 1.
Functional outcome
On admission, two patients were symptom free (mRS = 0), 25 patients had mild symptoms without significant
TE D
disability (mRS =1), 25 patients had a slight disability (mRS = 2), 5 patients suffered from moderate (mRS = 3) and 4 patients from moderately severe disability (mRS = 4). At time of discharge, 3 patients’ mRS improved, 39 patients remained at the same mRS level, 19 patients deteriorated. After six months, improvement was observed
EP
in 3 patients, the same status as preoperatively in 49 patients, and deterioration in 9 patients. 6 months
AC C
postoperative morbidity was consequently 15%.
Revision surgery for CCM remnants One patient (2%) with a clear CCM remnant on follow-up MRI underwent subsequent revision surgery. The two cases with suspicion of minimal remnant underwent further MRI follow-up (36 months). In both cases no rehemorrhage or other signal changes within the lesions occurred. See also Figure 2.
ACCEPTED MANUSCRIPT Discussion While preoperative MRI of CCM is clinically well established [8] and the subject of various studies[4-6,9,12], postoperative MRI of CCM has not been systematically analyzed yet and only few case series regarding this subject exist[2,7]. Nonetheless, postoperative MRI is a crucial technique to verify the achievement of treatment objectives such as complete resection of CCM and hemosiderin deposits (especially in CRE cases[11]). Kikuta et
RI PT
al.[7] analyzed a series of 10 patients with brainstem CCM which all underwent postoperative MRI within 2 months postoperatively. They found a correlation of incompletely resected CCM with a risk of a subsequent rehemorrhage. However, the method of image assessment and evaluation was not further specified. Cenzato et al.
SC
[2] presented a series of 30 brainstem CCM which were assessed by postoperative MRI within 72h after surgery. Remnants of CCM were described in 3 cases and these underwent subsequent revision surgery. Anyhow, results
M AN U
of MRI diagnosis were not further analyzed or compared with a second (subsequent) MRI examination. In our opinion, it seems clinically more reasonable to verify a complete resection of a CCM just after several weeks when all surgical artifacts (air, blood, hemostyptics) that may interfere with diagnosis of remnants are resorbed. This hypothesis was tested in this series, while potential benefits of an early MRI were tried to assess. Our results suggest that, using the three to six months postoperative MRI as a “gold standard” reference, the
TE D
value of early postoperative MRI is low for detecting/excluding both CCM (2%-5% prevalence) and hemosiderin remnants (44% prevalence). With an acceptable negative predictive value of 97.06% a diagnosis of a complete resection of the CCM is in fact very likely to be confirmed in the 3 months postoperative MRI, however, a positive predictive value of only 16.07% implies that the diagnosis of a CCM remnant on early
EP
postoperative MRI (potentially leading to revision surgery) is quite unreliable. Early T2*/SWI imaging to detect possible hemosiderin remnants proved to be completely inefficient with a drop out rate (no distinct diagnosis
AC C
possible due to artifacts) of 90%. T2 weighted imaging in this regard showed average positive and negative predictive values of 69.23% and 84.21%, respectively, making it also more reliable especially to exclude postoperative hemosiderin deposits.
It should be noted that the relatively high incidence of hemosiderin
remnants (44%) in our series can be attributed to the high proportion of brainstem cases in which the resection of peri-lesional hemosiderin is normally not intended. Apart from CCM or hemosiderin remnants, early postoperative MRI proved to be reliable in the diagnosis of associated DVA integrity, which was identified correctly in all cases on early MR images when compared to 3 months postoperative results. Compared to findings from routine postoperative CT scan, findings of early MRI led to no treatment changes in the acute management of our patients and showed no obvious additional value in
ACCEPTED MANUSCRIPT their acute management. However, regarding prognosis for functional outcome (e.g. in cases of brainstem CCM), early postoperative MRI may provide additional information.
RI PT
Study limitations One limitation of this study is of general nature. It may be quite hard to distinguish between very small remnants of a CCM (e.g. parts of a cavern) and simple postoperative changes, even on the 3-6 months postoperative images. 2 of our 3 CCM remnant diagnoses are of such kind (see figure 2). Furthermore, the risk for these
SC
“potential” remnants to cause (re-) hemorrhage is in our opinion negligible. Only further follow-up MR imaging may reveal a certain “activity” within the lesion that can confirm the suspicion of a real CCM remnant (which
M AN U
was not the case in this study). Another study limitation was that no standardized in-house 3-6 months postoperative MRI protocol was followed. This led to slight technical differences, e.g. in-plane resolution, slice thickness, MRI system, which may have hampered image analysis.
TE D
Conclusion
In this study, early postoperative MRI was not reliable in detecting postoperative CCM or hemosiderin remnant. Especially the suspicion of a remnant presence oftentimes proved to be false positive. It delivered no additional
EP
value regarding acute postoperative management when compared to postoperative CT. Thus, early postoperative MRI, in addition resource demanding and cost-intensive, cannot be recommended as a routine procedure. However, in cases of unclear neurological impairment after CCM surgery (especially in highly eloquent lesions
AC C
such as brainstem CCM), early MRI can deliver addition information that may be of relevance, especially regarding functional outcome.
References
1. Aiba T, Tanaka R, Koike T, Kameyama S, Takeda N, Komata T (1995) Natural history of intracranial cavernous malformations. J Neurosurg 83:56-59. doi:10.3171/jns.1995.83.1.0056 2. Cenzato M, Stefini R, Ambrosi C, Giovanelli M (2008) Post-operative remnants of brainstem cavernomas: incidence, risk factors and management. Acta Neurochir (Wien) 150:879-886; discussion 887. doi:10.1007/s00701-008-0008-4
ACCEPTED MANUSCRIPT 3. Dammann P, Schaller C, Sure U (2016) Should we resect peri-lesional hemosiderin deposits when performing lesionectomy in patients with cavernoma-related epilepsy (CRE)? Neurosurg Rev. doi:10.1007/s10143-016-0797-5 4. Dammann P, Wrede KH, Maderwald S, El Hindy N, Mueller O, Chen B, Zhu Y, Hutter BO, Ladd ME, Schlamann M, Sandalcioglu IE, Sure U (2013) The venous angioarchitecture of sporadic cerebral cavernous malformations: a susceptibility weighted imaging study at 7 T MRI. J
RI PT
Neurol Neurosurg Psychiatry 84:194-200. doi:10.1136/jnnp-2012-302599
5. de Champfleur NM, Langlois C, Ankenbrandt WJ, Le Bars E, Leroy MA, Duffau H, Bonafe A, Jaffe J, Awad IA, Labauge P (2011) Magnetic resonance imaging evaluation of cerebral
cavernous malformations with susceptibility-weighted imaging. Neurosurgery 68:641-647;
SC
discussion 647-648. doi:10.1227/NEU.0b013e31820773cf
6. Hart BL, Taheri S, Rosenberg GA, Morrison LA (2013) Dynamic contrast-enhanced MRI
doi:10.1007/s12975-013-0285-y
M AN U
evaluation of cerebral cavernous malformations. Transl Stroke Res 4:500-506.
7. Kikuta K, Nozaki K, Takahashi JA, Miyamoto S, Kikuchi H, Hashimoto N (2004) Postoperative evaluation of microsurgical resection for cavernous malformations of the brainstem. J Neurosurg 101:607-612. doi:10.3171/jns.2004.101.4.0607
8. Klostranec JM, Krings T (2015) Neuroimaging of cerebral cavernous malformations. J Neurosurg Sci 59:221-235
TE D
9. Mikati AG, Khanna O, Zhang L, Girard R, Shenkar R, Guo X, Shah A, Larsson HB, Tan H, Li L, Wishnoff MS, Shi C, Christoforidis GA, Awad IA (2015) Vascular permeability in cerebral cavernous malformations. J Cereb Blood Flow Metab 35:1632-1639. doi:10.1038/jcbfm.2015.98
EP
10. Rosenow F, Alonso-Vanegas MA, Baumgartner C, Blumcke I, Carreno M, Gizewski ER, Hamer HM, Knake S, Kahane P, Luders HO, Mathern GW, Menzler K, Miller J, Otsuki T, Ozkara C, Pitkanen A, Roper SN, Sakamoto AC, Sure U, Walker MC, Steinhoff BJ, Surgical Task Force
AC C
CoTSotI (2013) Cavernoma-related epilepsy: review and recommendations for management-report of the Surgical Task Force of the ILAE Commission on Therapeutic Strategies.
Epilepsia 54:2025-2035. doi:10.1111/epi.12402
11. Ruan D, Yu XB, Shrestha S, Wang L, Chen G (2015) The Role of Hemosiderin Excision in Seizure Outcome in Cerebral Cavernous Malformation Surgery: A Systematic Review and Meta-Analysis. PLoS One 10:e0136619. doi:10.1371/journal.pone.0136619 12. Tan H, Liu T, Wu Y, Thacker J, Shenkar R, Mikati AG, Shi C, Dykstra C, Wang Y, Prasad PV, Edelman RR, Awad IA (2014) Evaluation of iron content in human cerebral cavernous malformation using quantitative susceptibility mapping. Invest Radiol 49:498-504. doi:10.1097/RLI.0000000000000043
ACCEPTED MANUSCRIPT
Figure Legends Figure 1: Illustrative case of a false positive early postoperative finding of a CCM remnant. (A) and (B) showing T2
RI PT
weighted preoperative images of a pontine CCM. (C) showing suspected CCM remnant in the superior parts of the resection cavity on early postoperative MRI. (D) presenting the late postoperative MRI without suspicion of
SC
CCM remnant but minor residual peri-lesional hemosiderin deposits.
Figure 2:
Illustrative cases of diagnosed CCM remnants on 3-6 months follow-up MRI. (A) and (B) showing T2 sequences
M AN U
of a large “infiltrative” cerebellar CCM preoperatively (A) and a clear remnant (arrows) on the upper medial aspect on postoperative (B) imaging. The patient underwent revision surgery. (D) shows postoperative T2 sequences demonstrating signal changes at the bottom of the resection cavity of a large pontine CCM (C) which
AC C
EP
TE D
were also interpreted as a potential CCM remnant. The patient underwent follow-up MRIs.
MANUSCRIPT Table 1. Baseline characteristics of theACCEPTED cohort
AC C
EP
TE D
M AN U
SC
RI PT
Age [y] 43.7 (2- 73) Female [n] 33 (54%) Localization [n] supratentorial 14 (23%) infratentorial 47 (77%) Hemosiderin deposits* 48 (79%) Associated DVA 22 (36%) Max. diameter CCM [mm] 17.4 (±7.7) Postoperative functional outcome at 6 months† stable/improved 52 (85% ) deteriorated 9 (15%) * on preoperative MRI † on modified Rankin Scale
Table 2. Rating of postoperative MRI ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Early MRI* Control MRI † CCM remnant present 12 (20%) 3 (5%) not present 34 (56%) 57 (95%) not assessable/unclear 15 (25%) 0 (0%) Hemosiderin remnant (T2*/SWI) present 2 (4%) 21 (44%) not present 3 (6%) 27 (56%) not assessable/unclear 43 (90%) 0 (0%) Hemosiderin remnant (T2) present 26 (55%) 21 (44%) not present 9 (19%) 27 (56%) not assessable/unclear 12 (26%) 0 (0%) DVA intact 18 (82%) 18 (82%) partially intact 2 (9%) 2 (9%) resected 2 (9%) 2 (9%) * within 72h postoperatively † within 3-6 months postoperatively, hemosiderin remnants were assessed using T2*, SWI and T2 combined
ACCEPTED MANUSCRIPT Table 3 Diagnostic testing results CCM remnant diagnostic test
negative n=34
false negative (fn) n=1
true negative (tn) n=33
sensitivity [tp/ (tp+fn)] 66.67%
specificity [tn/(fp+tn)] 76.74%
not evaluable n=15
hemosiderin remnant diagnostic test
positive n=26 negative n=19
false negative (fn) n=3
true negative (tn) n=16
sensitivity [tp/ (tp+fn)] 85.71%
specificity [tn/(fp+tn)] 66.67%
not evaluable n=3
using T1,T2 sequences combined, 3using T2 sequences, 4using T2,T2* and SWI (not in all cases available) sequences combined
AC C
EP
TE D
1,2
positive predictive value [tp /(tp+fp)] 69.23% negative predictive value [tn/(fn+tn)] 84.21%
M AN U
early postoperative MRI3
follow-up MRI4 positive negative true positive (tp) false positive (fp) n=18 n=8
positive predictive value [tp /(tp+fp)] 16.67% negative predictive value [tn/(fn+tn)] 97.06%
RI PT
positive n=12
SC
early postoperative MRI1
follow-up MRI2 positive negative true positive (tp) false positive (fp) n=2 n=10
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Highlights:
RI PT SC M AN U TE D
• •
EP
•
Early postoperative compared to follow-up MRI after cavernous malformation surgery Early MRI is not reliable in detection of cavernoma or hemosiderin remnants. It is often hampered by artifacts and ineligible for resection control Early MRI delivers no additional value to acute postoperative management
AC C
•
ACCEPTED MANUSCRIPT Abbreviations list
EP
TE D
M AN U
SC
RI PT
Cerebral Cavernous Malformation Cavernoma-Related Epilepsy Magnetic Resonance Imaging Computed Tomography modified Ranking Scale
AC C
CCM CRE MRI CT mRS