- Email: [email protected]
Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test
Journal of Clinical Neuroscience xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test Markus Hoffermann a,⇑, Lukas Bruckmann a, Kariem Mahdy Ali a, Karla Zaar a, Alexander Avian b, Gord von Campe a a b
Department of Neurosurgery, Medical University of Graz, Graz, Austria Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Graz, Austria
a r t i c l e
i n f o
Article history: Received 18 August 2016 Accepted 15 October 2016 Available online xxxx Keywords: Neurocognitive function Neuropsychological assessment Brain tumor Brain surgery Brain tumor surgery NeuroCogFX
a b s t r a c t Neurocognitive assessment becomes increasingly important in neuro-oncology. The presence and degree of neurocognitive deficits in patients with brain tumors appear to be important not only as outcome measures but also in treatment planning and as possible prognostic markers for tumor-progression. Common screening methods for neurocognitive deficits are often insufficient in uncovering subtle changes or harbor the risk of being observer-dependent and time-consuming. We present data of brain tumor patients screened by a computer-based neurocognitive assessment tool before and after surgery. 196 patients with tumor resections were tested at our institution using the NeuroCog FxÒ software 2 days before and 3–4 months after surgery. Additionally to the test results, patient-related information, such as age, sex, handedness, level of education, pre- and postoperative neurological status, KPS, location and histopathological diagnosis were recorded. These prospectively collected results were correlated in the here presented retrospective study. The majority of patients with malignant gliomas, metastases and meningiomas showed significant deficits in various neurocognitive domains, most of them improved or did not decline in their postoperative neurocognitive performances. Interestingly, there was no significant correlation of neurocognitive deficits and brain tumor location. In future, standardized neuropsychological assessment should become an essential part of the management and care of patients with brain tumors to provide a more personalized and tailored treatment. Further studies will improve the understanding of the influence of various treatment modalities on neuro-cognition. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Brain tumors are heterogeneous diseases ranging from benign neoplasias, e.g. Meningioma WHO grade I, to very aggressive tumors like Glioblastoma WHO grade IV. Epileptic seizures, motor or sensory deficits, headaches, cranial nerve deficits or neuropsychological changes are possible initial symptoms. A neurosurgical intervention is usually the first crucial step in the treatment of brain tumor patients, at least to establish a histopathological diagnosis or to achieve tumor removal if it is safely possible. Tumor resection procedures are entailed by the risk to impinge on various functions of the brain, depending on the localization of the treated lesion. In addition to a surgical procedure, most brain tumor patients need adjuvant treatment like radio- and/or chemotherapy. The fact that possible side-effects of brain tumor specific therapies ⇑ Corresponding author at: Dept. of Neurosurgery, Medical University of Graz, Auenbruggerplatz 29, 8036 Graz, Austria. Fax: +43 316138513368. E-mail address: [email protected] (M. Hoffermann).
don’t only affect simply measureable neurological functions like motor function but also more complex neurocognitive functions seems obvious and has become an important part of brain tumor research [4,6,16,20,22,31,32,37–40]. It is now well accepted that brain tumors and related treatments can impair cognitive functions. Not only neurocognitive deficits themselves, but also their possible predictive and prognostic value [18,24] and possible treatment strategies [10–12,22] have gained growing interest in the past. Especially in malignant brain tumors, a decline in neurocognitive functions may even be a measurable sign for tumor progression [2,23]. Although deficits in neuropsychological functions of brain tumor patients have been investigated in a number of studies, the latter predominantly provide cross-sectional data from already treated patients and there is very little focus on the influence of surgery on these deficits [27,29]. We therefore decided to perform prospective longitudinal assessments of individual neurocognitive functions in our setting of patient care. In this retrospective study we share our initial experiences with the use of a computer-based neurocognitive screening tool. The
http://dx.doi.org/10.1016/j.jocn.2016.10.030 0967-5868/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
2
Case report / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
test was implemented at our institution in 2009 and meanwhile it has been shown by Kerrigan et al. that screening for neurocognitive deficits in brain tumor patients is not only of scientific interest but might also prevent subjective misperception of the patients’ mental capacity and thereby of their ability to provide informed consent [17]. Neuropsychological assessment is usually a time-consuming procedure that has to be carried out by specialized neuropsychologists and provides detailed information on various neuropsychological functions, e.g. language processing, memory, calculation or spatial orientation. Of course, not all of these functions may be of immediate clinical interest in the majority of brain tumor patients. This may contribute to the fact that testing for neuropsychological deficits mostly is not part of routine patient examination during brain tumor treatment. But, as shown before, these deficits seem to be of prognostic and predictive value and their revelation may contribute to an optimal treatment design [2,17,18]. Considering these facts, the department of Neurosurgery Graz has implemented a screening test battery for neurocognitive deficits in the routine pre-operative check-up and the first post-operative follow-up of patients with brain tumors requiring craniotomy procedures. Before the implementation of a neuropsychological screening tool, one has to make sure that the desired test is short (approx. 30 min.), repeatable, shows good psychometric properties (validity, reliability, and population norms) and is sensitive to changes in cognitive function, highly standardized and relatively simple to administer. It should also be completable by most patients, even those with considerable cognitive deficits, to avoid selection bias [22]. These features recommended by Meyers et al. in 2006 can be extended as proposed by Correa et al. in a review of 2007, which provided the basis for the development of NeuroCogFxÒ: [9] Assessment of cognitive domains sensitive to tumor and treatment effects, standardized training procedures and certification for individuals involved in test administration and availability in different languages [5]. In contrast to the Mini Mental State Examination [13,22,25], a dementia screening tool that is also widely used in brain tumor trials [3,33], we found that the neurocognitive screening software NeuroCogFXÒ meets the majority of these requirements to a high extent [8,9]. This software has originally been designed for neurocognitive screening in a variety of neurological disorders, especially epilepsy, at the Department of Epileptology, University of Bonn, Germany, and it has been standardized by testing 242 healthy individuals. Therefore, age standardization, critical differences to judge individual performances and an extensive assignment of subtests to neuropsychological domains are provided. Sensitivity and specificity for individual diagnosis of fractional neuropsychological dysfunction range from medium to high, but to a certain extent, training effects and only medium retest-reliability have to be taken into consideration for the interpretation of test results [8]. Recently, Fliessbach et al. have published data on practicability, retest reliability, practice effects, critical differences and validity for neurocognitive assessment with NeuroCogFxÒ in brain tumor patients enrolled in the German Glioma Network. In this study, the software was used parallel to a battery of established neuropsychological tests. Practicability was found to be good, retest reliability was medium-sized for most subtests in the control group (retest reliability r1 2 = 0.5–0.7) and brain tumor group (r1 2 = 0.6–0.8), but low in the 2-back test and simple reaction time in brain tumor patients (r1 2 = 0.18 and 0.33, respectively). Significant practice effects were seen in all subtests except the 2-back test and simple reaction time. These effects were not found when testing healthy individuals for a third or fourth time. The study revealed highly significant and strong correlations between NeuroCogFXÒ subtests and corresponding established tests (Pearson correlation r = 0.43–0.80). Regarding validity of the test, the authors were able to show in a factor analysis, that
the software subtests represent 5 important cognitive domains, most of which are typically altered in brain tumor patients: (I) psychomotor speed, (II) attention/executive functions and visual working memory, (III) verbal memory and word fluency, (IV) verbal short-term memory and (V) figural memory. However, the factors attention/executive functions and visual working memory may be underrepresented [9]. Mean test duration is 25 min and the software design allows standardized, repeatable and simple application, in principle even by trained non-academic staff. With eight different subtests, the domains verbal short time memory, working memory, reaction time, selective attention, susceptibility to interference or cognitive flexibility, verbal learning and recognition, and phonemic-literal verbal fluency are assessed. We provide only a short summary of the tested neurocognitive domains in Table 1, as detailed descriptions of the test properties have been published previously [8,9,14].
2. Materials and methods During routine preoperative checkup 1–2 days before undergoing craniotomy procedures and at their first postoperative followup exams after three to four months, a total of 196 patients (116 female, 80 male) with surgically treatable intracranial tumors underwent neurocognitive screening assessment using the NeuroCogFXÒ software. Patient’s mean age was 56.6 ± 13.9 years, ranging from 18.6 to 80.3 years (f: 56.6 ± 14.4, m: 56.7 ± 13.2). All tests were performed from 2009 to 2012. Patients treated for pituitary adenomas were not included. All patients were informed that their test results will have no impact on their treatment plan. Testing was conducted by the two main authors (M.H. and L.B.) or occasionally by medical students under their supervision. Clinical data regarding handedness, focal neurologic deficits, seizures and Karnofsky Performance Status were stored in a prospective database. The histopathological diagnoses were categorized into 9 subgroups (Meningioma, High Grade Glioma [= HGG = Glioma WHO III and IV] and Low Grade Glioma [=LGG = Glioma WHO I and II], Metastasis, Vestibular Schwannoma, Hemangioblastoma, Lymphoma and Epidermoid Cyst). In order to provide comparable tumor location data, categorization into 15 subgroups was performed (Supratentorial: frontal, rolandic, parieto-occipital and temporal in left and right hemisphere; frontal midline-dominant lesions; Infratentorial: cerebello-pontine angle or hemispheres, midline cerebellar lesions and brainstem lesions). The NeuroCogFXÒ subtest results and total scores with percentile ranks were automatically saved by the software in text form. These were transferred to IBM SPSS Statistics (Release 20.0.0. 2011. Chicago (IL), USA: SPSS Inc., an IBM Company). For interpretation of individual test results, the graduation system proposed by the test developers was used. Individual scores were given percentile ranks according to previously published normative data. Test results were categorized as follows: percentile rank (PR) = 0: very poor; PR < 3: poor; PR < 16: marginal; PR 16–84: normal; PR > 84: very good. The same classification was used for interpretation of subtests, overall score and overall test quality. The ‘‘overall score” of NeuroCogFXÒ is given in standard values, with a mean value of 100 and standard deviation of 10. Categories were merged for group comparisons (frontal vs. other hemispheric, left vs. right hemisphere, left temporal vs. left frontal) and analysis of changes in pre- and postoperative test as ‘‘impaired” (marginal, poor or very poor, PR < 16) and ‘‘not impaired” (normal and very good, PR P 16). Either Chi-square Test or Fisher’s exact Test were used for group comparisons. McNemar Test was used for analysis of changes between
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
3
M. Hoffermann et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx Table 1 Description of NeuroCogFXÒ subtests (modified from Hoppe C et al., 2009; Fliessbach K et al., 2010). Subtest
Function
Task
Measures
Validation test
Digit Span
Verbal short-term memory
Score: number of correct responses
Two Back
Working memory
Digit span forward (WMS-R) d2 (total score)
Simple Reaction
Alertness
Successive visual presentation of single digits (1/s) from a digit sequence with increasing length (3–9); two trials for each span; immediately recall the digit sequence by typing Continuous presentation of single digits (1/s); react (press spacebar) as fast as possible if present digit equals the second to the last digit React as fast as possible when a blue circle occurs (press spacebar)
Go/No Go
Selective attention
React (Go) as fast as possible if a blue circle occurs (press spacebar) but ignore yellow circles (No Go)
Reaction time: median of reaction times for correct reactions
Inverted Go/No Go
Selective attention and cognitive flexibility Verbal learning and recognition
React (Go) as fast as possible if a yellow circle occurs (press spacebar) but ignore blue circles (No Go)
Reaction time: median of reaction times for correct reactions
Three trials of same word list learning (12 words), subsequent yes/no recognition tests (test items: distractor items 1:2; spacebar press indicates ‘‘yes”) plus delayed recognition test Three trials of same figure list learning (7 checkerboard patterns with four delighted fields in a 3 3 matrix); subsequent yes/no recognition tests (spacebar press indicates ‘‘yes”) plus delayed recognition test Producing orally words with random first letter (L, P or S) within 1 min as fast as possible
Score: number of correctly recognized test items minus number of incorrect classifications/2 Score: number of correctly recognized test items minus number of incorrect classifications/2
AVLT (learning)
Score: number of correct words
Semantic word fluency
Verbal Memory
Figural Memory
Figural learning and recognition
Phonematic Fluency
Phonematic literal word fluency
Score: number of correct reactions minus number of incorrect reactions Reaction time: median of reaction times
TAP mean reaction times TAP mean reaction times TAP mean reaction times
ROCF (memory quotient)
Table 2 Performance of all patients in the eight different NeuroCogFXÒ subtests (n and %). Pre
Digit span 2-Back test Reaction test Go/no Go Inverse Go/no Go Verbal memory Figural memory Verbal fluency
Post
Impaired
Not impaired
No change
Improvement
Decline
55 (29.1%) 52 (35.9%) 67 (35.6%) 76 (40.9%) 60 (32.6%) 81 (45.0%) 60 (33.3%) 90 (49.2%)
134 (70.9%) 93 (64.1%) 121 (64.4%) 110 (59.1%) 124 (67.4%) 99 (55.0%) 120 (66.7%) 93 (50.8%)
139 (73.5%) 88 (60.7%) 136 (72.3%) 129 (69.4%) 123 (66.8%) 120 (66.7%) 120 (66.7%) 117 (63.9%)
31 (16.4%) 33 (22.8%) 23 (12.2%) 40 (21.5%) 31 (16.8%) 36 (20.0%) 47 (26.1%) 30 (16.4%)
19 (10.1%) 24 (16.6%) 29 (15.4%) 17 (9.1%) 30 (16.3%) 24 (13.3%) 13 (7.2%) 36 (19.7%)
pre- and postoperative results. No multiple testing p-value adjustment was performed since this was an exploratory analysis in a very heterogeneous population. Continuous data are presented as mean, ± SD and range. Categorical data are given as counts and percentages. A p-value of <0.05 was considered statistically significant. The conduction of this study was approved by the local ethics committee of the Medical University of Graz. 3. Results Most of the 196 patients were right-handed (n = 177, 90.3%), 11 were left-handed (5.6%) and 8 patients reported to have no dominant hand (4.1%). The predominant histopathological diagnosis was Meningioma, followed by HGG. Details on histopathological diagnosis are given in Table 3. The most common tumor location was left frontal. See Table 4 for detailed information on tumor locations.
The following educational level according to the International Standard Classification of Education was observed in the cohort: 19.9% reached level 2 (second stage of basic education), 62.2% level 3b (upper secondary education), 8.2% level 3a (upper secondary education, which provides direct access to tertiary education) and 7.7% level 5a (tertiary education). Four patients stated they had no education. Median KPS was 100 (min 60, max 100) preoperatively and 100 (min 30, max 100) postoperatively. Mean test duration was 24.7 ± 4.4 min preoperatively and 22.9 ± 3.6 postoperatively. 4. Test results of all patients The NeuroCogFXÒ overall score could not be calculated in 39 cases, when not all subtests were conducted correctly. Most comprehension problems occurred with the 2-back test. This test could not be completed correctly in 15.8% of preoperative tests and 17.2% of postoperative tests. In the preoperative setting, 6.5% of patients
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
4
Case report / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
Table 3 Histopathological diagnoses and Overall NeuroCogFXÒ Score (range) before and after surgery (AThe NeuroCogFXÒ overall score could not be calculated in 39 cases, when not all subtests were conducted correctly). Overall scoreA (n = 157)
Diagnosis
No. of patients (n = 196)
%
Meningioma High Grade Glioma Low Grade Glioma Metastasis Vestibular schwannoma Others (Lymphoma, Hemangioblastoma, Epidermoid)
76 55 23 27 12 3
38.8 28.1 11.7 13.8 6.1 1.5
Pre 93 91 97 91 96 96
(88–99) (84–97) (93–99) (80–102) (89–98) (95–97)
Post
Sign. change (p-value)
98 93 96 90 97 92
0.001 0.736 0.813 0.925 0.166
(90–102) (86–99) (90–100) (86–96) (94–103) (89–95)
Table 4 Tumor locations and Overall NeuroCogFXÒ Score (with standard deviation) before and after surgery. (AThe NeuroCogFXÒ overall score could not be calculated in 39 cases, when not all subtests were conducted correctly. BCPA = cerebellopontine angle). Diagnosis
No. of patients (n = 196)
%
Left frontal Left rolandic Left parieto-occipital Left temporal Right frontal Right rolandic Right parieto-occipital Right temporal Midline frontal Cerebellum left side Cerebellum right side Cerebellum midline CPA left B CPA right B Brainstem
44 14 17 18 25 8 14 11 10 6 7 6 10 5 1
22.4 7.1 8.7 9.2 12.8 4.1 7.1 5.6 5.1 3.1 3.6 3.1 5.1 2.6 0.5
also had trouble with comprehension of the figural memory test and 4% with the verbal memory test. During the follow-up tests, 11 patients (5.5%) were found to be in too poor clinical condition or to have neurologic deficits that would compromise test results significantly, so these were not tested again. The figural memory test could not be conducted correctly in 2.5% of postoperative patients and the verbal memory test in 2%. Two Patients (1%) refused re-testing. In preoperative tests, the most poor or very poor subtest results were seen in the verbal memory and the verbal fluency test. Significant improvements at follow-up performance were seen in the Go/no Go test (not impaired: preoperative 59% vs. follow up 72%; p = 0 .003) and the figural memory test (not impaired: preoperative 67% vs. follow up 86%; p < 0.001). No significant improvement or worsening was seen in the other subtests. For details, see Table 2. 5. Test results and histology 5.1. Subtest results in high grade gliomas Preoperative deficits (marginal, poor and very poor performance) were most commonly (61.4%) seen in the verbal memory test but also in 56.3% in the Go/no Go, 50.0% in the verbal fluency, 49.0% in the reaction, 40.6% in the 2-back, 36.4% in the figural memory, 34.8% in the inverse Go/no Go, and 32.7% in the digit span tests. A significant improvement was observed in the Go/no Go test (15 out of 27 impaired improved, five out of 21 not impaired declined, p = 0.041). A tendency toward an improvement was seen in the figural memory test (twelve out of 16 impaired improved, four out of 28 not impaired declined, p = 0.077). 5.2. Subtest results in low grade gliomas Preoperative deficits (marginal, poor and very poor performance) were most commonly (52.6%) seen in the 2-back test and
Overall scoreA (n = 157) Pre
Post
Sign. change (p-value)
92 95 92 90 90 93 94 92 95 92 91 97 99 98 99
91 (85–100) 97 (90–103) 94 (92–100) 90 (86–98) 94 (90–102) 97 (88–98) 97 (93–107) 95 (91–100) 99 (97–113) 93 (89–103) 94 (87–97) 90 (86–118) 101 (96–115) 97 (97–106) 101
0.703 0.685
(84–96) (88–105) (88–95) (88–97) (78–100) (89–95) (92–14) (89–96) (94–105) (84–98) (83–104) (97–102) (90–111) (94–98)
0.404 0.037
0.142
also in 43.5% in the digit span, 39.1 % in the verbal memory and figural memory, 30.4% in the verbal fluency, 26.1% in the Go/no Go and reaction tests and 21.7% in the inverse Go/no Go test. A tendency toward a decline was observed in the verbal fluency test (one out of seven impaired improved, seven out of 16 not impaired declined, p = 0.070) and a tendency toward an improvement was seen in the digit span test (five out of ten impaired improved, none out of 13 not impaired declined, p = 0.063). 5.3. Subtest results in meningiomas Preoperative deficits (marginal, poor and very poor performance) were most commonly (46.7%) seen in the verbal fluency test and also in 41.9% in the verbal memory, 38.7% in the Go/no Go, 36.0% in the inverse Go/no Go and reaction time, 31.1% in the figural memory and in the 2-back, and 26.3% in the digit span test. A significant improvement was observed in the figural memory test (20 out of 23 impaired improved, five out of 51 not impaired declined, p = 0.004). A tendency toward an improvement was seen in the Go/no Go (15 out of 29 impaired improved, six out of 46 not impaired declined, p = 0.078), 2-back test (13 out of 19 impaired improved, five out of 42 not impaired declined, p = 0.096) and verbal memory tests (16 out of 31 impaired improved, seven out of 43 not impaired declined, p = 0.093). 5.4. Subtest results in metastases Preoperative deficits (marginal, poor and very poor performance) were most commonly (62.5%) seen in the verbal fluency test and also in 47.4% in the 2-back, 41.7% in the verbal memory and in the figural memory, 36.0% in the Go/no Go and inverse Go/no Go, 30.8% in the digit span and the reaction tests. A significant improvement was observed in the figural memory test (nine out of ten impaired improved, one out of 14 not impaired declined, p = 0.021).
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
M. Hoffermann et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
5.5. Test results and tumor location Comparisons between subtest results (marginal, very poor and poor vs. normal and very good) in several different tumor localizations groups (frontal vs. other hemispheric, left vs. right hemisphere, left temporal vs. left frontal) did mostly not depict significant differences except for the preoperative Go/no Go test, in which patients with left temporal tumors showed more marginal, poor or very poor results (p = 0.022) and the postoperative figural memory test, in which patients with tumors in the right hemisphere showed more marginal, poor or very poor results (p = 0.011). 6. Discussion In our series, patients with malignant brain tumors (HGG and Metastases) scored significantly lower in the preoperative screening than patients with more favorable diagnoses. In the follow-up tests, there was no significant rise in impaired overall results. Talacchi et al. have shown that postoperative cognitive decline in glioma patients is likely to be linked to significant MR imaging alterations, e.g. brain edema, and that HGG patients are more likely to experience overall improvement of cognitive functions after surgery [30]. The glioma patients in our study did not seem to improve to a great extent, unlike patients with Meningiomas. This observation is most likely related to the infiltrating and by itself destroying growth pattern of gliomas as compared to pure mass effect by extra-axial lesions or the often fast-recovering tumor edema that can be observed in Meningiomas. Our results do not allow a statistically significant correlation between tumor histology and impaired cognitive domains, but alertness, selective attention and verbal memory seemed to be more affected in HGG patients than in most other groups, especially LGG. Interestingly, the LGG patients performed worse in verbal short-term and working memory tests. It has also been shown by Raysi Dehcordi et al. that the pattern of pre- and postoperative deficits of LGG patients differs from those of HGG patients, who showed marked impairment in working memory, fluency function and processing speed both pre- and postoperatively [26]. In our study, we were able to show an overall post-operative improvement or at least no significant decline in most cognitive domains in all histology groups, although preoperative deficits were evident in a high number of patients with HGG, Metastasis and Meningioma. Together with the fact that our follow-up tests were performed at three months, before radiation or chemotherapy effects are to be expected, we conclude that surgery by itself is unlikely to provoke neurocognitive disturbances in most brain tumor patients. Similar findings have previously been reported by some of the few groups who have investigated pre- and postoperative deficits in patients with LGG, HGG and Meningioma [1,15,21,26,34]. Of course, since the study cohort consists of patients who were eligible for tumor surgery, it can be assumed that there is high selection bias. The natural course of the tumor diseases was not part of this investigation, but one would expect further decline in cognitive function without surgery or other tumor specific treatment. Although Meningioma is a common type of benign brain tumors, little is known about cognitive function in patients suffering from this condition. In contrast to what clinicians might presume, van Nieuwenhuizen et al. have shown in a study on Meningioma patients, that patients with this disease show significant impairment of neurocognitive function [36]. Tucha et al. conducted tests of cognitive function preoperatively and at a 4–9-
5
month postoperative follow-up on patients who underwent frontal meningioma resection. Except in the case of working memory, comparisons between preoperative and postoperative assessments of cognition revealed. No significant differences in memory, visuoconstructive abilities, or executive function were found, but a postoperative improvement of attention function was observed along with slight improvement in working memory. The authors concluded that frontal Meningioma resection does not negatively affect cognition [35]. Our study supports these findings, since the meningioma group scored significantly better in the postoperative overall results. Furthermore, we also found the domain of working memory significantly improved in these patients, but independent of tumor location. We were not able to depict significant differences in the impairment of certain cognitive domains when comparing various tumor locations independent of histology. There was only a trend towards a higher degree of improved postoperative results in right hemisphere lesions when compared to left hemisphere tumors. It can be assumed that neurocognitive changes in brain tumor patients are more of a global phenomenon, supporting network theories rather than a location-oriented point of view. Scheibel et al. have shown that patients with left-hemispheric tumors (glioblastomas and other histological entities) showed worse performance in verbal tasks and patients with right-sided lesions scored lower in facial recognition, for example. These laterality effects were similar to those found in focal neurological disorders, such as cerebrovascular disease [28]. In contrast to these findings, our own results are supported by Klein et al., who have shown for diffuse infiltrating gliomas that brain tumor-related cognitive deficits are not restricted to any single domain, and that even in cases where deficits were related to tumor localization, patients tended to exhibit global deficits [19]. Rather than reflecting a local effect, these and our own findings indicate that whole brain cooperation is essential for intact cognitive performance. Douw et al. also conducted a much longer follow-up study of survivors among the patients in the aforementioned study by Klein et al., at a mean of 12 years after initial diagnosis. Attentional dysfunction was found to have deteriorated significantly in patients who received radiotherapy. This deterioration was independent of tumor lateralization, fraction dose, size, age, extent of resection and antiepileptic drug use. A progressive decline in attentional function was observed even among patients who received fraction doses that were considered to be safe (62 Gy) [7]. The results of these studies suggest that radiotherapy has a high impact on the cognitive status on a long-term basis. A possible drawback of our study is that the effects of radiotherapy, chemotherapy and anti-epileptic drugs are not reflected in our final analysis. These factors have been part of our initial analyses but failed to show any significant effect because of low power. This data is not shown in order to avoid over-interpretation of possible findings. Another limitation can be seen in the follow-up timeframe. Since this is a retrospective study, the intention was to identify a high number of comparable cases. Therefore, only the test results before surgery and at the first follow-up exam were used because in the majority of cases, the test was not conducted more than once post-operatively. In our opinion, repetitive testing is still advisable especially in glioma patients, since a decline in function can be indicative of tumor progression. It is also important to be aware of possible training effects in the re-test scenario, which have not been systematically assessed in this study population. The lack of additional Quality of life, depression and anxiety assessment is another drawback of this study, since these factors can influence neurocognitive performance per se. Also, it has to be taken into account that the software is not designed to detect deficits in executive functions.
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
6
Case report / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
7. Conclusion Our initial experience with the use of this computerized screening tool to detect neurocognitive deficits in brain tumor patients was mostly positive. Testing time is very short with approximately 25 min, which lead to generally good acceptance and compliance. Also, the pattern of tested functions shows a reasonable overlap with neurocognitive domains that are often found to be compromised in neuro-oncological patients. In this first analysis, we could show that the majority of patients with malignant gliomas, metastases and also with meningiomas show significant deficits in various neurocognitive domains. These deficits are mostly unchanged or even improved after surgery. It has also been shown that neurocognitive deficits are not necessarily linked to brain tumor location but seem to more of a global phenomenon. We conclude that the use of this screening test can facilitate the assessment of overall performance and neurocognitive functions preoperatively and special need for rehabilitation and psycho-social care in the follow-up period. Conflict of interest The authors declare that they have no conflict of interest. Ethical approval All procedures performed in this study were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For this type of study formal consent is not required. Acknowledgement We would like to thank our medical students for their help in conducting the tests. References [1] Archibald YM, Lunn D, Ruttan LA, et al. Cognitive functioning in long-term survivors of high-grade glioma. J Neurosurg 1994;80:247–53. [2] Armstrong CL. The predictive value of longitudinal neuropsychologic assessment in the early detection of brain tumor recurrence. Cancer 2003;97 (3):649–56. [3] Brown PD, Buckner JC, O’Fallon JR, et al. Effects of radiotherapy on cognitive function in patients with low-grade glioma measured by the folstein minimental state examination. J Clin Oncol 2003;21:2519–24. [4] Correa DD. Cognitive functions in brain tumor patients. Hematol Oncol Clin North Am 2006;20:1363–76. [5] Correa DD, Maron L, Harder H, et al. Cognitive functions in primary central nervous system lymphoma: literature review and assessment guidelines. Ann Oncol 2007;18:1145–51. [6] Dijkstra M, van ND, Stalpers LJ, et al. Late neurocognitive sequelae in patients with WHO grade I meningioma. J Neurol Neurosurg Psychiatry 2009;80:910–5. [7] Douw L, Klein M, Fagel SS, et al. Cognitive and radiological effects of radiotherapy in patients with low-grade glioma: long-term follow-up. Lancet Neurol 2009;8:810–8. [8] Fliessbach K, Hoppe C, Schlegel U, et al. NeuroCogFX–a computer-based neuropsychological assessment battery for the follow-up examination of neurological patients. Fortschr Neurol Psychiatr 2006;74:643–50. [9] Fliessbach K, Rogowski S, Hoppe C, et al. Computer-based assessment of cognitive functions in brain tumor patients. J Neurooncol 2010;100:427–37. [10] Gehring K, Aaronson NK, Taphoorn MJ, et al. Interventions for cognitive deficits in patients with a brain tumor: an update. Expert Rev Anticancer Ther 2010;10:1779–95.
[11] Gehring K, Roukema JA, Sitskoorn MM. Review of recent studies on interventions for cognitive deficits in patients with cancer. Expert Rev Anticancer Ther 2012;12:255–69. [12] Gehring K, Sitskoorn MM, Aaronson NK, et al. Interventions for cognitive deficits in adults with brain tumours. Lancet Neurol 2008;7:548–60. [13] Herman MA, Tremont-Lukats I, Meyers CA, et al. Neurocognitive and functional assessment of patients with brain metastases: a pilot study. Am J Clin Oncol 2003;26:273–9. [14] Hoppe C, Fliessbach K, Schlegel U, et al. NeuroCog FX: computerized screening of cognitive functions in patients with epilepsy. Epilepsy Behav 2009;16:298–310. [15] Kashiwazaki D, Takaiwa A, Nagai S, et al. Reversal of cognitive dysfunction by total removal of a large lateral ventricle meningioma: a case report with neuropsychological assessments. Case Rep Neurol 2014;6:44–9. [16] Kayl AE, Meyers CA. Does brain tumor histology influence cognitive function? Neuro Oncol 2003;5:255–60. [17] Kerrigan S. Mental incapacity in patients undergoing neuro-oncologic treatment: a cross-sectional study. Neurology 2014;83(6):537–41. [18] Klein M. The prognostic value of cognitive functioning in the survival of patients with high-grade glioma. Neurology 2003;61(12):1796–8. [19] Klein M, Duffau H, De Witt Hamer PC. Cognition and resective surgery for diffuse infiltrative glioma: an overview. J Neurooncol 2012;108:309–18. [20] Klein M, Taphoorn MJ, Heimans JJ, et al. Neurobehavioral status and healthrelated quality of life in newly diagnosed high-grade glioma patients. J Clin Oncol 2001;19:4037–47. [21] Koizumi H, Ideguchi M, Iwanaga H, et al. Cognitive dysfunction might be improved in association with recovered neuronal viability after intracranial meningioma resection. Brain Res 2014;1574:50–9. [22] Meyers CA, Brown PD. Role and relevance of neurocognitive assessment in clinical trials of patients with CNS tumors. J Clin Oncol 2006;24:1305–9. [23] Meyers CA, Hess KR. Multifaceted end points in brain tumor clinical trials: cognitive deterioration precedes MRI progression. Neuro Oncol 2003;5:89–95. [24] Meyers CA, Hess KR, Yung WK, et al. Cognitive function as a predictor of survival in patients with recurrent malignant glioma. J Clin Oncol 2000;18:646–50. [25] Meyers CA, Wefel JS. The use of the mini-mental state examination to assess cognitive functioning in cancer trials: no ifs, ands, buts, or sensitivity. J Clin Oncol 2003;21:3557–8. [26] Raysi DS, Mariano M, Mazza M, et al. Cognitive deficits in patients with low and high grade gliomas. J Neurosurg Sci 2013;57:259–66. [27] Satoer D, Vork J, Visch-Brink E, et al. Cognitive functioning early after surgery of gliomas in eloquent areas. J Neurosurg 2012;117:831–8. [28] Scheibel RS, Meyers CA, Levin VA. Cognitive dysfunction following surgery for intracerebral glioma: influence of histopathology, lesion location, and treatment. J Neurooncol 1996;30:61–9. [29] Talacchi A, d’Avella D, Denaro L, et al. Cognitive outcome as part and parcel of clinical outcome in brain tumor surgery. J Neurooncol 2012;108:327–32. [30] Talacchi A, Santini B, Savazzi S, et al. Cognitive effects of tumour and surgical treatment in glioma patients. J Neurooncol 2011;103:541–9. [31] Taphoorn MJ, Heimans JJ, Snoek FJ, et al. Quality of life and neuropsychological functions in long-term low-grade glioma survivors. Int J Radiat Oncol Biol Phys 1994;29:1201–2. [32] Taphoorn MJ, Klein M. Cognitive deficits in adult patients with brain tumours. Lancet Neurol 2004;3:159–68. [33] Taylor BV, Buckner JC, Cascino TL, et al. Effects of radiation and chemotherapy on cognitive function in patients with high-grade glioma. J Clin Oncol 1998;16:2195–201. [34] Teixidor P, Gatignol P, Leroy M, et al. Assessment of verbal working memory before and after surgery for low-grade glioma. J Neurooncol 2007;81:305–13. [35] Tucha O, Smely C, Preier M, et al. Preoperative and postoperative cognitive functioning in patients with frontal meningiomas. J Neurosurg 2003;98:21–31. [36] van ND, Klein M, Stalpers LJ, et al. Differential effect of surgery and radiotherapy on neurocognitive functioning and health-related quality of life in WHO grade I meningioma patients. J Neurooncol 2007;84:271–8. [37] Waagemans ML, van ND, Dijkstra M, et al. Long-term impact of cognitive deficits and epilepsy on quality of life in patients with low-grade meningiomas. Neurosurgery 2011;69:72–8. [38] Weitzner MA, Meyers CA. Quality of life and neurobehavioural functioning in patients with malignant gliomas. Baillieres Clin Neurol 1996;5:425–39. [39] Weitzner MA, Meyers CA. Cognitive functioning and quality of life in malignant glioma patients: a review of the literature. Psychooncology 1997;6:169–77. [40] Witgert ME, Meyers CA. Neurocognitive and quality of life measures in patients with metastatic brain disease. Neurosurg Clin N Am 2011;22:79–85. vii.
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test Markus Hoffermann a,⇑, Lukas Bruckmann a, Kariem Mahdy Ali a, Karla Zaar a, Alexander Avian b, Gord von Campe a a b
Department of Neurosurgery, Medical University of Graz, Graz, Austria Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Graz, Austria
a r t i c l e
i n f o
Article history: Received 18 August 2016 Accepted 15 October 2016 Available online xxxx Keywords: Neurocognitive function Neuropsychological assessment Brain tumor Brain surgery Brain tumor surgery NeuroCogFX
a b s t r a c t Neurocognitive assessment becomes increasingly important in neuro-oncology. The presence and degree of neurocognitive deficits in patients with brain tumors appear to be important not only as outcome measures but also in treatment planning and as possible prognostic markers for tumor-progression. Common screening methods for neurocognitive deficits are often insufficient in uncovering subtle changes or harbor the risk of being observer-dependent and time-consuming. We present data of brain tumor patients screened by a computer-based neurocognitive assessment tool before and after surgery. 196 patients with tumor resections were tested at our institution using the NeuroCog FxÒ software 2 days before and 3–4 months after surgery. Additionally to the test results, patient-related information, such as age, sex, handedness, level of education, pre- and postoperative neurological status, KPS, location and histopathological diagnosis were recorded. These prospectively collected results were correlated in the here presented retrospective study. The majority of patients with malignant gliomas, metastases and meningiomas showed significant deficits in various neurocognitive domains, most of them improved or did not decline in their postoperative neurocognitive performances. Interestingly, there was no significant correlation of neurocognitive deficits and brain tumor location. In future, standardized neuropsychological assessment should become an essential part of the management and care of patients with brain tumors to provide a more personalized and tailored treatment. Further studies will improve the understanding of the influence of various treatment modalities on neuro-cognition. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Brain tumors are heterogeneous diseases ranging from benign neoplasias, e.g. Meningioma WHO grade I, to very aggressive tumors like Glioblastoma WHO grade IV. Epileptic seizures, motor or sensory deficits, headaches, cranial nerve deficits or neuropsychological changes are possible initial symptoms. A neurosurgical intervention is usually the first crucial step in the treatment of brain tumor patients, at least to establish a histopathological diagnosis or to achieve tumor removal if it is safely possible. Tumor resection procedures are entailed by the risk to impinge on various functions of the brain, depending on the localization of the treated lesion. In addition to a surgical procedure, most brain tumor patients need adjuvant treatment like radio- and/or chemotherapy. The fact that possible side-effects of brain tumor specific therapies ⇑ Corresponding author at: Dept. of Neurosurgery, Medical University of Graz, Auenbruggerplatz 29, 8036 Graz, Austria. Fax: +43 316138513368. E-mail address: [email protected] (M. Hoffermann).
don’t only affect simply measureable neurological functions like motor function but also more complex neurocognitive functions seems obvious and has become an important part of brain tumor research [4,6,16,20,22,31,32,37–40]. It is now well accepted that brain tumors and related treatments can impair cognitive functions. Not only neurocognitive deficits themselves, but also their possible predictive and prognostic value [18,24] and possible treatment strategies [10–12,22] have gained growing interest in the past. Especially in malignant brain tumors, a decline in neurocognitive functions may even be a measurable sign for tumor progression [2,23]. Although deficits in neuropsychological functions of brain tumor patients have been investigated in a number of studies, the latter predominantly provide cross-sectional data from already treated patients and there is very little focus on the influence of surgery on these deficits [27,29]. We therefore decided to perform prospective longitudinal assessments of individual neurocognitive functions in our setting of patient care. In this retrospective study we share our initial experiences with the use of a computer-based neurocognitive screening tool. The
http://dx.doi.org/10.1016/j.jocn.2016.10.030 0967-5868/Ó 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
2
Case report / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
test was implemented at our institution in 2009 and meanwhile it has been shown by Kerrigan et al. that screening for neurocognitive deficits in brain tumor patients is not only of scientific interest but might also prevent subjective misperception of the patients’ mental capacity and thereby of their ability to provide informed consent [17]. Neuropsychological assessment is usually a time-consuming procedure that has to be carried out by specialized neuropsychologists and provides detailed information on various neuropsychological functions, e.g. language processing, memory, calculation or spatial orientation. Of course, not all of these functions may be of immediate clinical interest in the majority of brain tumor patients. This may contribute to the fact that testing for neuropsychological deficits mostly is not part of routine patient examination during brain tumor treatment. But, as shown before, these deficits seem to be of prognostic and predictive value and their revelation may contribute to an optimal treatment design [2,17,18]. Considering these facts, the department of Neurosurgery Graz has implemented a screening test battery for neurocognitive deficits in the routine pre-operative check-up and the first post-operative follow-up of patients with brain tumors requiring craniotomy procedures. Before the implementation of a neuropsychological screening tool, one has to make sure that the desired test is short (approx. 30 min.), repeatable, shows good psychometric properties (validity, reliability, and population norms) and is sensitive to changes in cognitive function, highly standardized and relatively simple to administer. It should also be completable by most patients, even those with considerable cognitive deficits, to avoid selection bias [22]. These features recommended by Meyers et al. in 2006 can be extended as proposed by Correa et al. in a review of 2007, which provided the basis for the development of NeuroCogFxÒ: [9] Assessment of cognitive domains sensitive to tumor and treatment effects, standardized training procedures and certification for individuals involved in test administration and availability in different languages [5]. In contrast to the Mini Mental State Examination [13,22,25], a dementia screening tool that is also widely used in brain tumor trials [3,33], we found that the neurocognitive screening software NeuroCogFXÒ meets the majority of these requirements to a high extent [8,9]. This software has originally been designed for neurocognitive screening in a variety of neurological disorders, especially epilepsy, at the Department of Epileptology, University of Bonn, Germany, and it has been standardized by testing 242 healthy individuals. Therefore, age standardization, critical differences to judge individual performances and an extensive assignment of subtests to neuropsychological domains are provided. Sensitivity and specificity for individual diagnosis of fractional neuropsychological dysfunction range from medium to high, but to a certain extent, training effects and only medium retest-reliability have to be taken into consideration for the interpretation of test results [8]. Recently, Fliessbach et al. have published data on practicability, retest reliability, practice effects, critical differences and validity for neurocognitive assessment with NeuroCogFxÒ in brain tumor patients enrolled in the German Glioma Network. In this study, the software was used parallel to a battery of established neuropsychological tests. Practicability was found to be good, retest reliability was medium-sized for most subtests in the control group (retest reliability r1 2 = 0.5–0.7) and brain tumor group (r1 2 = 0.6–0.8), but low in the 2-back test and simple reaction time in brain tumor patients (r1 2 = 0.18 and 0.33, respectively). Significant practice effects were seen in all subtests except the 2-back test and simple reaction time. These effects were not found when testing healthy individuals for a third or fourth time. The study revealed highly significant and strong correlations between NeuroCogFXÒ subtests and corresponding established tests (Pearson correlation r = 0.43–0.80). Regarding validity of the test, the authors were able to show in a factor analysis, that
the software subtests represent 5 important cognitive domains, most of which are typically altered in brain tumor patients: (I) psychomotor speed, (II) attention/executive functions and visual working memory, (III) verbal memory and word fluency, (IV) verbal short-term memory and (V) figural memory. However, the factors attention/executive functions and visual working memory may be underrepresented [9]. Mean test duration is 25 min and the software design allows standardized, repeatable and simple application, in principle even by trained non-academic staff. With eight different subtests, the domains verbal short time memory, working memory, reaction time, selective attention, susceptibility to interference or cognitive flexibility, verbal learning and recognition, and phonemic-literal verbal fluency are assessed. We provide only a short summary of the tested neurocognitive domains in Table 1, as detailed descriptions of the test properties have been published previously [8,9,14].
2. Materials and methods During routine preoperative checkup 1–2 days before undergoing craniotomy procedures and at their first postoperative followup exams after three to four months, a total of 196 patients (116 female, 80 male) with surgically treatable intracranial tumors underwent neurocognitive screening assessment using the NeuroCogFXÒ software. Patient’s mean age was 56.6 ± 13.9 years, ranging from 18.6 to 80.3 years (f: 56.6 ± 14.4, m: 56.7 ± 13.2). All tests were performed from 2009 to 2012. Patients treated for pituitary adenomas were not included. All patients were informed that their test results will have no impact on their treatment plan. Testing was conducted by the two main authors (M.H. and L.B.) or occasionally by medical students under their supervision. Clinical data regarding handedness, focal neurologic deficits, seizures and Karnofsky Performance Status were stored in a prospective database. The histopathological diagnoses were categorized into 9 subgroups (Meningioma, High Grade Glioma [= HGG = Glioma WHO III and IV] and Low Grade Glioma [=LGG = Glioma WHO I and II], Metastasis, Vestibular Schwannoma, Hemangioblastoma, Lymphoma and Epidermoid Cyst). In order to provide comparable tumor location data, categorization into 15 subgroups was performed (Supratentorial: frontal, rolandic, parieto-occipital and temporal in left and right hemisphere; frontal midline-dominant lesions; Infratentorial: cerebello-pontine angle or hemispheres, midline cerebellar lesions and brainstem lesions). The NeuroCogFXÒ subtest results and total scores with percentile ranks were automatically saved by the software in text form. These were transferred to IBM SPSS Statistics (Release 20.0.0. 2011. Chicago (IL), USA: SPSS Inc., an IBM Company). For interpretation of individual test results, the graduation system proposed by the test developers was used. Individual scores were given percentile ranks according to previously published normative data. Test results were categorized as follows: percentile rank (PR) = 0: very poor; PR < 3: poor; PR < 16: marginal; PR 16–84: normal; PR > 84: very good. The same classification was used for interpretation of subtests, overall score and overall test quality. The ‘‘overall score” of NeuroCogFXÒ is given in standard values, with a mean value of 100 and standard deviation of 10. Categories were merged for group comparisons (frontal vs. other hemispheric, left vs. right hemisphere, left temporal vs. left frontal) and analysis of changes in pre- and postoperative test as ‘‘impaired” (marginal, poor or very poor, PR < 16) and ‘‘not impaired” (normal and very good, PR P 16). Either Chi-square Test or Fisher’s exact Test were used for group comparisons. McNemar Test was used for analysis of changes between
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
3
M. Hoffermann et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx Table 1 Description of NeuroCogFXÒ subtests (modified from Hoppe C et al., 2009; Fliessbach K et al., 2010). Subtest
Function
Task
Measures
Validation test
Digit Span
Verbal short-term memory
Score: number of correct responses
Two Back
Working memory
Digit span forward (WMS-R) d2 (total score)
Simple Reaction
Alertness
Successive visual presentation of single digits (1/s) from a digit sequence with increasing length (3–9); two trials for each span; immediately recall the digit sequence by typing Continuous presentation of single digits (1/s); react (press spacebar) as fast as possible if present digit equals the second to the last digit React as fast as possible when a blue circle occurs (press spacebar)
Go/No Go
Selective attention
React (Go) as fast as possible if a blue circle occurs (press spacebar) but ignore yellow circles (No Go)
Reaction time: median of reaction times for correct reactions
Inverted Go/No Go
Selective attention and cognitive flexibility Verbal learning and recognition
React (Go) as fast as possible if a yellow circle occurs (press spacebar) but ignore blue circles (No Go)
Reaction time: median of reaction times for correct reactions
Three trials of same word list learning (12 words), subsequent yes/no recognition tests (test items: distractor items 1:2; spacebar press indicates ‘‘yes”) plus delayed recognition test Three trials of same figure list learning (7 checkerboard patterns with four delighted fields in a 3 3 matrix); subsequent yes/no recognition tests (spacebar press indicates ‘‘yes”) plus delayed recognition test Producing orally words with random first letter (L, P or S) within 1 min as fast as possible
Score: number of correctly recognized test items minus number of incorrect classifications/2 Score: number of correctly recognized test items minus number of incorrect classifications/2
AVLT (learning)
Score: number of correct words
Semantic word fluency
Verbal Memory
Figural Memory
Figural learning and recognition
Phonematic Fluency
Phonematic literal word fluency
Score: number of correct reactions minus number of incorrect reactions Reaction time: median of reaction times
TAP mean reaction times TAP mean reaction times TAP mean reaction times
ROCF (memory quotient)
Table 2 Performance of all patients in the eight different NeuroCogFXÒ subtests (n and %). Pre
Digit span 2-Back test Reaction test Go/no Go Inverse Go/no Go Verbal memory Figural memory Verbal fluency
Post
Impaired
Not impaired
No change
Improvement
Decline
55 (29.1%) 52 (35.9%) 67 (35.6%) 76 (40.9%) 60 (32.6%) 81 (45.0%) 60 (33.3%) 90 (49.2%)
134 (70.9%) 93 (64.1%) 121 (64.4%) 110 (59.1%) 124 (67.4%) 99 (55.0%) 120 (66.7%) 93 (50.8%)
139 (73.5%) 88 (60.7%) 136 (72.3%) 129 (69.4%) 123 (66.8%) 120 (66.7%) 120 (66.7%) 117 (63.9%)
31 (16.4%) 33 (22.8%) 23 (12.2%) 40 (21.5%) 31 (16.8%) 36 (20.0%) 47 (26.1%) 30 (16.4%)
19 (10.1%) 24 (16.6%) 29 (15.4%) 17 (9.1%) 30 (16.3%) 24 (13.3%) 13 (7.2%) 36 (19.7%)
pre- and postoperative results. No multiple testing p-value adjustment was performed since this was an exploratory analysis in a very heterogeneous population. Continuous data are presented as mean, ± SD and range. Categorical data are given as counts and percentages. A p-value of <0.05 was considered statistically significant. The conduction of this study was approved by the local ethics committee of the Medical University of Graz. 3. Results Most of the 196 patients were right-handed (n = 177, 90.3%), 11 were left-handed (5.6%) and 8 patients reported to have no dominant hand (4.1%). The predominant histopathological diagnosis was Meningioma, followed by HGG. Details on histopathological diagnosis are given in Table 3. The most common tumor location was left frontal. See Table 4 for detailed information on tumor locations.
The following educational level according to the International Standard Classification of Education was observed in the cohort: 19.9% reached level 2 (second stage of basic education), 62.2% level 3b (upper secondary education), 8.2% level 3a (upper secondary education, which provides direct access to tertiary education) and 7.7% level 5a (tertiary education). Four patients stated they had no education. Median KPS was 100 (min 60, max 100) preoperatively and 100 (min 30, max 100) postoperatively. Mean test duration was 24.7 ± 4.4 min preoperatively and 22.9 ± 3.6 postoperatively. 4. Test results of all patients The NeuroCogFXÒ overall score could not be calculated in 39 cases, when not all subtests were conducted correctly. Most comprehension problems occurred with the 2-back test. This test could not be completed correctly in 15.8% of preoperative tests and 17.2% of postoperative tests. In the preoperative setting, 6.5% of patients
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
4
Case report / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
Table 3 Histopathological diagnoses and Overall NeuroCogFXÒ Score (range) before and after surgery (AThe NeuroCogFXÒ overall score could not be calculated in 39 cases, when not all subtests were conducted correctly). Overall scoreA (n = 157)
Diagnosis
No. of patients (n = 196)
%
Meningioma High Grade Glioma Low Grade Glioma Metastasis Vestibular schwannoma Others (Lymphoma, Hemangioblastoma, Epidermoid)
76 55 23 27 12 3
38.8 28.1 11.7 13.8 6.1 1.5
Pre 93 91 97 91 96 96
(88–99) (84–97) (93–99) (80–102) (89–98) (95–97)
Post
Sign. change (p-value)
98 93 96 90 97 92
0.001 0.736 0.813 0.925 0.166
(90–102) (86–99) (90–100) (86–96) (94–103) (89–95)
Table 4 Tumor locations and Overall NeuroCogFXÒ Score (with standard deviation) before and after surgery. (AThe NeuroCogFXÒ overall score could not be calculated in 39 cases, when not all subtests were conducted correctly. BCPA = cerebellopontine angle). Diagnosis
No. of patients (n = 196)
%
Left frontal Left rolandic Left parieto-occipital Left temporal Right frontal Right rolandic Right parieto-occipital Right temporal Midline frontal Cerebellum left side Cerebellum right side Cerebellum midline CPA left B CPA right B Brainstem
44 14 17 18 25 8 14 11 10 6 7 6 10 5 1
22.4 7.1 8.7 9.2 12.8 4.1 7.1 5.6 5.1 3.1 3.6 3.1 5.1 2.6 0.5
also had trouble with comprehension of the figural memory test and 4% with the verbal memory test. During the follow-up tests, 11 patients (5.5%) were found to be in too poor clinical condition or to have neurologic deficits that would compromise test results significantly, so these were not tested again. The figural memory test could not be conducted correctly in 2.5% of postoperative patients and the verbal memory test in 2%. Two Patients (1%) refused re-testing. In preoperative tests, the most poor or very poor subtest results were seen in the verbal memory and the verbal fluency test. Significant improvements at follow-up performance were seen in the Go/no Go test (not impaired: preoperative 59% vs. follow up 72%; p = 0 .003) and the figural memory test (not impaired: preoperative 67% vs. follow up 86%; p < 0.001). No significant improvement or worsening was seen in the other subtests. For details, see Table 2. 5. Test results and histology 5.1. Subtest results in high grade gliomas Preoperative deficits (marginal, poor and very poor performance) were most commonly (61.4%) seen in the verbal memory test but also in 56.3% in the Go/no Go, 50.0% in the verbal fluency, 49.0% in the reaction, 40.6% in the 2-back, 36.4% in the figural memory, 34.8% in the inverse Go/no Go, and 32.7% in the digit span tests. A significant improvement was observed in the Go/no Go test (15 out of 27 impaired improved, five out of 21 not impaired declined, p = 0.041). A tendency toward an improvement was seen in the figural memory test (twelve out of 16 impaired improved, four out of 28 not impaired declined, p = 0.077). 5.2. Subtest results in low grade gliomas Preoperative deficits (marginal, poor and very poor performance) were most commonly (52.6%) seen in the 2-back test and
Overall scoreA (n = 157) Pre
Post
Sign. change (p-value)
92 95 92 90 90 93 94 92 95 92 91 97 99 98 99
91 (85–100) 97 (90–103) 94 (92–100) 90 (86–98) 94 (90–102) 97 (88–98) 97 (93–107) 95 (91–100) 99 (97–113) 93 (89–103) 94 (87–97) 90 (86–118) 101 (96–115) 97 (97–106) 101
0.703 0.685
(84–96) (88–105) (88–95) (88–97) (78–100) (89–95) (92–14) (89–96) (94–105) (84–98) (83–104) (97–102) (90–111) (94–98)
0.404 0.037
0.142
also in 43.5% in the digit span, 39.1 % in the verbal memory and figural memory, 30.4% in the verbal fluency, 26.1% in the Go/no Go and reaction tests and 21.7% in the inverse Go/no Go test. A tendency toward a decline was observed in the verbal fluency test (one out of seven impaired improved, seven out of 16 not impaired declined, p = 0.070) and a tendency toward an improvement was seen in the digit span test (five out of ten impaired improved, none out of 13 not impaired declined, p = 0.063). 5.3. Subtest results in meningiomas Preoperative deficits (marginal, poor and very poor performance) were most commonly (46.7%) seen in the verbal fluency test and also in 41.9% in the verbal memory, 38.7% in the Go/no Go, 36.0% in the inverse Go/no Go and reaction time, 31.1% in the figural memory and in the 2-back, and 26.3% in the digit span test. A significant improvement was observed in the figural memory test (20 out of 23 impaired improved, five out of 51 not impaired declined, p = 0.004). A tendency toward an improvement was seen in the Go/no Go (15 out of 29 impaired improved, six out of 46 not impaired declined, p = 0.078), 2-back test (13 out of 19 impaired improved, five out of 42 not impaired declined, p = 0.096) and verbal memory tests (16 out of 31 impaired improved, seven out of 43 not impaired declined, p = 0.093). 5.4. Subtest results in metastases Preoperative deficits (marginal, poor and very poor performance) were most commonly (62.5%) seen in the verbal fluency test and also in 47.4% in the 2-back, 41.7% in the verbal memory and in the figural memory, 36.0% in the Go/no Go and inverse Go/no Go, 30.8% in the digit span and the reaction tests. A significant improvement was observed in the figural memory test (nine out of ten impaired improved, one out of 14 not impaired declined, p = 0.021).
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
M. Hoffermann et al. / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
5.5. Test results and tumor location Comparisons between subtest results (marginal, very poor and poor vs. normal and very good) in several different tumor localizations groups (frontal vs. other hemispheric, left vs. right hemisphere, left temporal vs. left frontal) did mostly not depict significant differences except for the preoperative Go/no Go test, in which patients with left temporal tumors showed more marginal, poor or very poor results (p = 0.022) and the postoperative figural memory test, in which patients with tumors in the right hemisphere showed more marginal, poor or very poor results (p = 0.011). 6. Discussion In our series, patients with malignant brain tumors (HGG and Metastases) scored significantly lower in the preoperative screening than patients with more favorable diagnoses. In the follow-up tests, there was no significant rise in impaired overall results. Talacchi et al. have shown that postoperative cognitive decline in glioma patients is likely to be linked to significant MR imaging alterations, e.g. brain edema, and that HGG patients are more likely to experience overall improvement of cognitive functions after surgery [30]. The glioma patients in our study did not seem to improve to a great extent, unlike patients with Meningiomas. This observation is most likely related to the infiltrating and by itself destroying growth pattern of gliomas as compared to pure mass effect by extra-axial lesions or the often fast-recovering tumor edema that can be observed in Meningiomas. Our results do not allow a statistically significant correlation between tumor histology and impaired cognitive domains, but alertness, selective attention and verbal memory seemed to be more affected in HGG patients than in most other groups, especially LGG. Interestingly, the LGG patients performed worse in verbal short-term and working memory tests. It has also been shown by Raysi Dehcordi et al. that the pattern of pre- and postoperative deficits of LGG patients differs from those of HGG patients, who showed marked impairment in working memory, fluency function and processing speed both pre- and postoperatively [26]. In our study, we were able to show an overall post-operative improvement or at least no significant decline in most cognitive domains in all histology groups, although preoperative deficits were evident in a high number of patients with HGG, Metastasis and Meningioma. Together with the fact that our follow-up tests were performed at three months, before radiation or chemotherapy effects are to be expected, we conclude that surgery by itself is unlikely to provoke neurocognitive disturbances in most brain tumor patients. Similar findings have previously been reported by some of the few groups who have investigated pre- and postoperative deficits in patients with LGG, HGG and Meningioma [1,15,21,26,34]. Of course, since the study cohort consists of patients who were eligible for tumor surgery, it can be assumed that there is high selection bias. The natural course of the tumor diseases was not part of this investigation, but one would expect further decline in cognitive function without surgery or other tumor specific treatment. Although Meningioma is a common type of benign brain tumors, little is known about cognitive function in patients suffering from this condition. In contrast to what clinicians might presume, van Nieuwenhuizen et al. have shown in a study on Meningioma patients, that patients with this disease show significant impairment of neurocognitive function [36]. Tucha et al. conducted tests of cognitive function preoperatively and at a 4–9-
5
month postoperative follow-up on patients who underwent frontal meningioma resection. Except in the case of working memory, comparisons between preoperative and postoperative assessments of cognition revealed. No significant differences in memory, visuoconstructive abilities, or executive function were found, but a postoperative improvement of attention function was observed along with slight improvement in working memory. The authors concluded that frontal Meningioma resection does not negatively affect cognition [35]. Our study supports these findings, since the meningioma group scored significantly better in the postoperative overall results. Furthermore, we also found the domain of working memory significantly improved in these patients, but independent of tumor location. We were not able to depict significant differences in the impairment of certain cognitive domains when comparing various tumor locations independent of histology. There was only a trend towards a higher degree of improved postoperative results in right hemisphere lesions when compared to left hemisphere tumors. It can be assumed that neurocognitive changes in brain tumor patients are more of a global phenomenon, supporting network theories rather than a location-oriented point of view. Scheibel et al. have shown that patients with left-hemispheric tumors (glioblastomas and other histological entities) showed worse performance in verbal tasks and patients with right-sided lesions scored lower in facial recognition, for example. These laterality effects were similar to those found in focal neurological disorders, such as cerebrovascular disease [28]. In contrast to these findings, our own results are supported by Klein et al., who have shown for diffuse infiltrating gliomas that brain tumor-related cognitive deficits are not restricted to any single domain, and that even in cases where deficits were related to tumor localization, patients tended to exhibit global deficits [19]. Rather than reflecting a local effect, these and our own findings indicate that whole brain cooperation is essential for intact cognitive performance. Douw et al. also conducted a much longer follow-up study of survivors among the patients in the aforementioned study by Klein et al., at a mean of 12 years after initial diagnosis. Attentional dysfunction was found to have deteriorated significantly in patients who received radiotherapy. This deterioration was independent of tumor lateralization, fraction dose, size, age, extent of resection and antiepileptic drug use. A progressive decline in attentional function was observed even among patients who received fraction doses that were considered to be safe (62 Gy) [7]. The results of these studies suggest that radiotherapy has a high impact on the cognitive status on a long-term basis. A possible drawback of our study is that the effects of radiotherapy, chemotherapy and anti-epileptic drugs are not reflected in our final analysis. These factors have been part of our initial analyses but failed to show any significant effect because of low power. This data is not shown in order to avoid over-interpretation of possible findings. Another limitation can be seen in the follow-up timeframe. Since this is a retrospective study, the intention was to identify a high number of comparable cases. Therefore, only the test results before surgery and at the first follow-up exam were used because in the majority of cases, the test was not conducted more than once post-operatively. In our opinion, repetitive testing is still advisable especially in glioma patients, since a decline in function can be indicative of tumor progression. It is also important to be aware of possible training effects in the re-test scenario, which have not been systematically assessed in this study population. The lack of additional Quality of life, depression and anxiety assessment is another drawback of this study, since these factors can influence neurocognitive performance per se. Also, it has to be taken into account that the software is not designed to detect deficits in executive functions.
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030
6
Case report / Journal of Clinical Neuroscience xxx (2016) xxx–xxx
7. Conclusion Our initial experience with the use of this computerized screening tool to detect neurocognitive deficits in brain tumor patients was mostly positive. Testing time is very short with approximately 25 min, which lead to generally good acceptance and compliance. Also, the pattern of tested functions shows a reasonable overlap with neurocognitive domains that are often found to be compromised in neuro-oncological patients. In this first analysis, we could show that the majority of patients with malignant gliomas, metastases and also with meningiomas show significant deficits in various neurocognitive domains. These deficits are mostly unchanged or even improved after surgery. It has also been shown that neurocognitive deficits are not necessarily linked to brain tumor location but seem to more of a global phenomenon. We conclude that the use of this screening test can facilitate the assessment of overall performance and neurocognitive functions preoperatively and special need for rehabilitation and psycho-social care in the follow-up period. Conflict of interest The authors declare that they have no conflict of interest. Ethical approval All procedures performed in this study were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For this type of study formal consent is not required. Acknowledgement We would like to thank our medical students for their help in conducting the tests. References [1] Archibald YM, Lunn D, Ruttan LA, et al. Cognitive functioning in long-term survivors of high-grade glioma. J Neurosurg 1994;80:247–53. [2] Armstrong CL. The predictive value of longitudinal neuropsychologic assessment in the early detection of brain tumor recurrence. Cancer 2003;97 (3):649–56. [3] Brown PD, Buckner JC, O’Fallon JR, et al. Effects of radiotherapy on cognitive function in patients with low-grade glioma measured by the folstein minimental state examination. J Clin Oncol 2003;21:2519–24. [4] Correa DD. Cognitive functions in brain tumor patients. Hematol Oncol Clin North Am 2006;20:1363–76. [5] Correa DD, Maron L, Harder H, et al. Cognitive functions in primary central nervous system lymphoma: literature review and assessment guidelines. Ann Oncol 2007;18:1145–51. [6] Dijkstra M, van ND, Stalpers LJ, et al. Late neurocognitive sequelae in patients with WHO grade I meningioma. J Neurol Neurosurg Psychiatry 2009;80:910–5. [7] Douw L, Klein M, Fagel SS, et al. Cognitive and radiological effects of radiotherapy in patients with low-grade glioma: long-term follow-up. Lancet Neurol 2009;8:810–8. [8] Fliessbach K, Hoppe C, Schlegel U, et al. NeuroCogFX–a computer-based neuropsychological assessment battery for the follow-up examination of neurological patients. Fortschr Neurol Psychiatr 2006;74:643–50. [9] Fliessbach K, Rogowski S, Hoppe C, et al. Computer-based assessment of cognitive functions in brain tumor patients. J Neurooncol 2010;100:427–37. [10] Gehring K, Aaronson NK, Taphoorn MJ, et al. Interventions for cognitive deficits in patients with a brain tumor: an update. Expert Rev Anticancer Ther 2010;10:1779–95.
[11] Gehring K, Roukema JA, Sitskoorn MM. Review of recent studies on interventions for cognitive deficits in patients with cancer. Expert Rev Anticancer Ther 2012;12:255–69. [12] Gehring K, Sitskoorn MM, Aaronson NK, et al. Interventions for cognitive deficits in adults with brain tumours. Lancet Neurol 2008;7:548–60. [13] Herman MA, Tremont-Lukats I, Meyers CA, et al. Neurocognitive and functional assessment of patients with brain metastases: a pilot study. Am J Clin Oncol 2003;26:273–9. [14] Hoppe C, Fliessbach K, Schlegel U, et al. NeuroCog FX: computerized screening of cognitive functions in patients with epilepsy. Epilepsy Behav 2009;16:298–310. [15] Kashiwazaki D, Takaiwa A, Nagai S, et al. Reversal of cognitive dysfunction by total removal of a large lateral ventricle meningioma: a case report with neuropsychological assessments. Case Rep Neurol 2014;6:44–9. [16] Kayl AE, Meyers CA. Does brain tumor histology influence cognitive function? Neuro Oncol 2003;5:255–60. [17] Kerrigan S. Mental incapacity in patients undergoing neuro-oncologic treatment: a cross-sectional study. Neurology 2014;83(6):537–41. [18] Klein M. The prognostic value of cognitive functioning in the survival of patients with high-grade glioma. Neurology 2003;61(12):1796–8. [19] Klein M, Duffau H, De Witt Hamer PC. Cognition and resective surgery for diffuse infiltrative glioma: an overview. J Neurooncol 2012;108:309–18. [20] Klein M, Taphoorn MJ, Heimans JJ, et al. Neurobehavioral status and healthrelated quality of life in newly diagnosed high-grade glioma patients. J Clin Oncol 2001;19:4037–47. [21] Koizumi H, Ideguchi M, Iwanaga H, et al. Cognitive dysfunction might be improved in association with recovered neuronal viability after intracranial meningioma resection. Brain Res 2014;1574:50–9. [22] Meyers CA, Brown PD. Role and relevance of neurocognitive assessment in clinical trials of patients with CNS tumors. J Clin Oncol 2006;24:1305–9. [23] Meyers CA, Hess KR. Multifaceted end points in brain tumor clinical trials: cognitive deterioration precedes MRI progression. Neuro Oncol 2003;5:89–95. [24] Meyers CA, Hess KR, Yung WK, et al. Cognitive function as a predictor of survival in patients with recurrent malignant glioma. J Clin Oncol 2000;18:646–50. [25] Meyers CA, Wefel JS. The use of the mini-mental state examination to assess cognitive functioning in cancer trials: no ifs, ands, buts, or sensitivity. J Clin Oncol 2003;21:3557–8. [26] Raysi DS, Mariano M, Mazza M, et al. Cognitive deficits in patients with low and high grade gliomas. J Neurosurg Sci 2013;57:259–66. [27] Satoer D, Vork J, Visch-Brink E, et al. Cognitive functioning early after surgery of gliomas in eloquent areas. J Neurosurg 2012;117:831–8. [28] Scheibel RS, Meyers CA, Levin VA. Cognitive dysfunction following surgery for intracerebral glioma: influence of histopathology, lesion location, and treatment. J Neurooncol 1996;30:61–9. [29] Talacchi A, d’Avella D, Denaro L, et al. Cognitive outcome as part and parcel of clinical outcome in brain tumor surgery. J Neurooncol 2012;108:327–32. [30] Talacchi A, Santini B, Savazzi S, et al. Cognitive effects of tumour and surgical treatment in glioma patients. J Neurooncol 2011;103:541–9. [31] Taphoorn MJ, Heimans JJ, Snoek FJ, et al. Quality of life and neuropsychological functions in long-term low-grade glioma survivors. Int J Radiat Oncol Biol Phys 1994;29:1201–2. [32] Taphoorn MJ, Klein M. Cognitive deficits in adult patients with brain tumours. Lancet Neurol 2004;3:159–68. [33] Taylor BV, Buckner JC, Cascino TL, et al. Effects of radiation and chemotherapy on cognitive function in patients with high-grade glioma. J Clin Oncol 1998;16:2195–201. [34] Teixidor P, Gatignol P, Leroy M, et al. Assessment of verbal working memory before and after surgery for low-grade glioma. J Neurooncol 2007;81:305–13. [35] Tucha O, Smely C, Preier M, et al. Preoperative and postoperative cognitive functioning in patients with frontal meningiomas. J Neurosurg 2003;98:21–31. [36] van ND, Klein M, Stalpers LJ, et al. Differential effect of surgery and radiotherapy on neurocognitive functioning and health-related quality of life in WHO grade I meningioma patients. J Neurooncol 2007;84:271–8. [37] Waagemans ML, van ND, Dijkstra M, et al. Long-term impact of cognitive deficits and epilepsy on quality of life in patients with low-grade meningiomas. Neurosurgery 2011;69:72–8. [38] Weitzner MA, Meyers CA. Quality of life and neurobehavioural functioning in patients with malignant gliomas. Baillieres Clin Neurol 1996;5:425–39. [39] Weitzner MA, Meyers CA. Cognitive functioning and quality of life in malignant glioma patients: a review of the literature. Psychooncology 1997;6:169–77. [40] Witgert ME, Meyers CA. Neurocognitive and quality of life measures in patients with metastatic brain disease. Neurosurg Clin N Am 2011;22:79–85. vii.
Please cite this article in press as: Hoffermann M et al. Pre- and postoperative neurocognitive deficits in brain tumor patients assessed by a computer based screening test. J Clin Neurosci (2016), http://dx.doi.org/10.1016/j.jocn.2016.10.030