Journal of Clinical Neuroscience 19 (2012) 411–414 Contents lists available at SciVerse ScienceDirect Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn Clinical Study Normal or non-diagnostic neuroimaging studies prior to the detection of malignant primary brain tumors Paul B. Thaler a, Jian Yi Li b, Yakov Isakov c, Karen S. Black d, Michael Schulder e, Alexis Demopoulos f,⇑ a Department of Plastic Surgery, NYU Langone Medical Center, New York, NY, USA Department of Pathology and Laboratory Medicine, North Shore-Long Island Jewish Health System, Hofstra North Shore-LIJ School of Medicine, Lake Success, NY, USA c Department of Neurology, North Shore–Long Island Jewish Health System, Manhasset, NY, USA d Department of Radiology, Hofstra North Shore-LIJ School of Medicine, Manhasset, NY, USA e Department of Neurosurgery, Hofstra North Shore-LIJ School of Medicine, Manhasset, NY, USA f Tisch Cancer Institute and Department of Neurosurgery, Mount Sinai School of Medicine, One Gustave L Levy Place, Box 1079, New York, NY 10029-6574, USA b a r t i c l e i n f o Article history: Received 18 August 2011 Accepted 2 September 2011 Keywords: MRI Oncology Primary brain tumor a b s t r a c t We aimed to describe a single institution experience of neuroimaging failure to demonstrate malignant primary brain tumors. We retrospectively reviewed case histories for all newly diagnosed adult patients with malignant primary brain tumors treated at a single institution between 1 July 2006 and 30 June 2008. We specifically looked at patients in whom neuroimaging was normal or non-diagnostic at initial presentation. Among 193 patients with malignant primary brain tumors, there were 102 with World Health Organization (WHO) grade IV gliomas (glioblastoma multiforme, GBM), 54 with anaplastic gliomas, 18 with low grade gliomas, and 19 with primary central nervous system lymphomas (PCNSL). Initial imaging was normal in nine patients and abnormal but non-diagnostic in an additional eight patients with primary brain cancer. Normal or non-diagnostic neuroimaging was not uncommon among patients with GBM. Dramatic, rapid tumor growth is possible. Close interval clinical and radiographic follow-up can be important especially in the management of elderly patients presenting with seizures and nondiagnostic neuroimaging studies. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction With increasing access to advanced neuroimaging techniques in Emergency Departments (ED), contrast-enhanced CT scans and MRI are often performed to investigate new neurologic complaints within days or even hours of onset. Confusion arises when such studies are normal or reveal non-specific abnormalities despite overt symptoms or signs. Patients are often perplexed when later studies reveal abnormalities typical for aggressive, malignant neoplasms. 2. Materials and methods As part of a specialty neuro-oncology service, the neurosurgeon (M.S.) and the neuro-oncologist (A.D.) retrospectively identified 17 patients with a malignant primary brain tumor with non-diagnostic neuroimaging (CT scan or MRI) from a total cohort of consecutive, newly diagnosed patients with a malignant brain tumor treated between 1 July 2006 and 30 June 2008. Data on clinical characteristics were gathered by a retrospective review of patients’ ⇑ Corresponding author. Tel.: +1 212 241 6756; fax: +1 646 537 8549. E-mail address: alexis.demopoulos@mssm.edu (A. Demopoulos). 0967-5868/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jocn.2011.09.002 charts. A neuroradiologist (K.B.) re-reviewed the first imaging study for evidence of malignancy. P53 immunostaining data were captured, if available. All tumor diagnoses were verified histologically by a neuropathologist [J.Y.L.]. 2.1. Imaging All patients underwent 1.5-Tesla MRI. Slice acquisition was limited to 5 mm cuts. Sequences performed included diffusionweighted imaging, non-contrast axial and sagittal T1-weighted MRI, three-dimensional post-contrast T1-weighted MRI, and axial pre-contrast T2-weighted MRI (including fluid-attenuated inversion recovery [FLAIR] MRI). 3. Results Of the 193 newly diagnosed, consecutive patients with brain tumors diagnosed between 1 July 2006 and 30 June 2008, 102 patients had World Health Organization (WHO) grade IV gliomas (glioblastoma multiforme, GBM), 18 had low grade gliomas (LGG), 54 had anaplastic gliomas (AG), and 19 had primary central nervous system lymphomas (PCNSL). No patients with LGG or AG had a history of normal neuroimaging. Normal neuroimaging was 412 P.B. Thaler et al. / Journal of Clinical Neuroscience 19 (2012) 411–414 found in eight of 102 patients with GBM (8%) and one of 19 patients with PCNSL (5%). Initially three patients had CT scans and six had MRI, including the patient with PCNSL. Normal neuroimaging was reported in an additional two patients with GBM, but the films were unavailable for independent confirmation. These patients were included in the total number of patients, but were excluded from patients with normal neuroimaging. Clinical characteristics of patients with normal neuroimaging preceding diagnosis are presented in Table 1. Of the nine patients with malignant brain tumor with normal neuroimaging confirmed by independent review, three presented with seizures, one with weakness, two with cognitive disturbance, and three with falls. The patient with PCNSL presented with confusion. The median age was 77 years (range: 38–89 years). The median interval between normal and abnormal imaging was 5.5 months (range: 2–76.6 months). Most patients (six of nine) underwent repeat neuroimaging within six months. Immunostaining for p53 was positive in three patients, negative in three patients and unavailable in two (Table 1). 3.1. Non-diagnostic abnormal imaging All patients with non-specific abnormalities underwent MRI at presentation. Non-specific T2-weighted abnormalities were seen in seven of 102 patients with GBM and one of 54 with AG. Initial imaging showed multifocal abnormalities in four patients. All patients presented with seizures. The median age was 64 years (range: 47–77 years). Repeat imaging demonstrated enhancing disease a median of six weeks later and within three months in all patients (range: 0.7–2.9 months). P53 immunostaining was negative in four, positive in one and unavailable in three patients. P53 immunopositivity is a marker for secondary transformation of lower grade gliomas to higher grade gliomas. De novo or ‘‘primary’’ glioblastomas are associated with negative p53 immunostaining. Although our data are incomplete, p53 immunonegativity was present in seven of the 11 specimens available for analysis. Two of the three patients with the longest interval between normal and abnormal scans had p53-positive tumors, suggesting secondary transformation without prior radiographic correlate. The p53 immunostain results from our small series correlate well with the published data where about 25% of primary GBM have TP53 mutations while TP53 mutations are significantly more frequent (>65%) in secondary GBM).1,2 3.2. Illustrative patients Fig. 1 includes initial and follow-up images of six representative patients, including three whose first scans were unremarkable. Patient 1, an 80-year-old woman, presented to the ED with a generalized tonic–clonic seizure. The initial neuroimaging included T2-fluid attenuated inversion recovery (FLAIR) and T1-weighted post-contrast images, which were unremarkable and the patient was discharged home. She returned with recurrent spells 3.3 months later. Repeat imaging demonstrated right temporal lobe hyperintensity on T2-weighted MRI, with enhancement on T1-weighted MRI (Fig. 1). Biopsy demonstrated GBM. Patient 3, a 77-year-old man, presented to the ED with hemiparesis and was found to have a large, robustly enhancing mass lesion involving the right hemisphere. The patient reported a two-year history of generalized weakness – diagnosed as adult onset myotonic dystrophy – a prior MRI was obtained and re-reviewed. This MRI was performed 20.2 months earlier and was normal (Fig. 1). Patient 8, a 90-year-old right-handed man with a history of stroke, was brought to the ED by his family for an episode of left-sided weakness and aphasia. Cranial neuroimaging was unremarkable (Fig. 1). Bilateral carotid stenosis was found and he Table 1 Clinical and radiographic characteristics of patients with normal or non-diagnostic neuroimaging studies prior to the detection of malignant primary brain tumors Patient no. Age (years) Sex Presenting complaint First scan type Findings Next scan type Reason for subsequent imaging Time between scans (months) Pathology p53 status Normal neuroimaging preceding diagnosis 1 80 F Seizure, GTC 2 66 F Falls 3 77 M Weakness 4 38 F Seizure, GTC 5 44 F Falls 6 56 F Seizures 7 80 M Falls 8 90 M Memory loss MRI CT MRI MRI CT MRI CT MRI Negative Negative Negative Negative Negative Negative Negative Negative MRI MRI MRI MRI MRI MRI MRI MRI Seizure Headache Hemiparesis Seizure Headache Seizure Confusion Memory loss 3.3 5.9 20.2 5.1 2.0 76.6 32.5 4.5 GBM GBM GBM GBM GBM GBM GBM GBM + Non-specific abnormality preceding diagnosis 9 77 F Seizure, focal MRI MRI Seizure 1.5 GBM NA 10 53 F Seizure, focal MRI MRI Hemiparesis 0.8 AG NA 11 64 M Seizure, GTC MRI MRI Seizure 1.1 GBM 12 47 M Seizure, GTC MRI MRI Lethargy 2.9 GBM 13 64 M Seizure, GTC MRI MRI Seizure 2.0 GBM 14 73 F Seizure, focal MRI MRI Seizure 2.2 GBM + 15 59 M Seizure, GTC MRI MRI Seizure 1.6 GBM NA 16 71 F Seizure, focal MRI T2 changes T2 changes T2 changes T2 changes T2 changes * T2 changes * T2 changes * T2 changes * MRI Seizure 0.7 GBM Confusion 3.6 (MRI), 1.6 (CT) PCNSL Primary central nervous system lymphoma with normal neuroimaging preceding diagnosis 17 81 M Confusion MRI and CT Negative MRI + NA + NA NA = negative, + = positive, AG = anaplastic glioma, CT = computed tomography, GBM = glioblastoma multiforme, MRI = magnetic resonance imaging, NA = not available, PCNSL = primary central nervous system lymphoma. Indicates that T2-weighted changes were multifocal. * P.B. Thaler et al. / Journal of Clinical Neuroscience 19 (2012) 411–414 413 Fig. 1. Initial imaging studies (two left-hand columns) and repeat imaging studies (two right-hand columns) of representative patients. Patient 1, who presented with a generalized tonic–clonic seizure, had unremarkable initial neuroimaging: (far left) axial T2-fluid attenuated inversion recovery (FLAIR) and (mid left) axial T1-weighted postcontrast images. Repeat imaging demonstrated right temporal lobe hyperintensity on T2-weighted MRI, with enhancement on T1-weighted enhanced MRI. Patient 3, a 77year-old man who presented with hemiparesis, had (far left FLAIR, mid left T1-weighted) prior axial MRI that appeared normal. (Mid right FLAIR, far right T1-weighted enhanced) follow-up axial MRI revealed a large, robustly enhancing mass lesion involving the right hemisphere. Patient 8, a 90-year-old right-handed man with a history of stroke, presented with left-sided weakness and aphasia. (Far left FLAIR, mid left T1-weighted) axial cranial neuroimaging was unremarkable. (Mid right FLAIR, far right T1weighted enhanced) repeat axial neuroimaging revealed a mass lesion. Patient 11, a 64-year-old woman who presented with confusion and seizures, had (far left) coronal FLAIR, and (mid left) axial T1-weighted images showing multiple nonenhancing abnormalities involving the left temporal lobe. (Mid right) Coronal FLAIR and (far right) axial T1-weighted enhanced repeat imaging revealed multiple enhancing regions within the T2-weighted hyperintensity. Patient 12, a 47-year old man, presented with lethargy, after ‘‘normal’’ neuroimaging at another hospital performed after a complex partial seizure: review of the (far left FLAIR, mid left T1-weighted enhanced) axial scans revealed a T2-weighted hyperintensity in the medial right frontal lobe. Our repeat MRI (axial, mid right FLAIR, far right T1-weighted enhanced) demonstrated progression to a large left frontal tumor. Patient 17, an 81-year-old man, presented with confusion and agitation, axial T2-weighted (far left) and contrast-enhanced T1-weighted (mid left) MRI was unremarkable. Two months later, the (mid right CT scan) image was also unremarkable. A further two months later, he developed lethargy and coma, and (far right) T1weighted enhanced axial MRI showed a large frontal lesion. 414 P.B. Thaler et al. / Journal of Clinical Neuroscience 19 (2012) 411–414 underwent carotid endartectomy (not shown). His left-sided weakness improved, but his aphasia and confusion worsened. Repeat neuroimaging performed on 4.5 months later revealed a mass lesion (Fig. 1). Biopsy demonstrated GBM. Patient 11, a 64-year-old woman, presented to the ED with confusion and seizures, and was found to have multiple non-enhancing abnormalities involving the left temporal lobe (Fig. 1). An encephalitis work-up and body imaging were unrevealing (not shown). She improved on anti-seizure medication and was discharged home with a plan for repeat neuroimaging. Repeat imaging, performed two months later, revealed multiple enhancing regions within the T2-weighted hyperintensity (Fig. 1). Biopsy demonstrated GBM. Patient 12, a 47-year-old man, presented to our ED with complaints of lethargy. Three months earlier, he had suffered a complex partial seizure and underwent normal neuroimaging at another hospital. Outside films were obtained and re-reviewed. A region of T2-weighted hyperintensity was appreciated in the medial right frontal lobe (Fig. 1). This abnormality had not been remarked upon at that institution. Our neuroimaging demonstrated progression to a large left frontal tumor, found to be GBM. Patient 17, an 81-year-old man with a past medical history of alcoholism and alcohol withdrawal syndrome, was brought by his wife to the ED with confusion and agitation. Neuroimaging was unremarkable (Fig. 1) and he was eventually discharged home. He returned 2 months later with progressive confusion and agitation. His CT scan was unremarkable (Fig. 1). Two months later, he developed lethargy and coma, requiring intubation and sedation, and was brought by ambulance to the ED, where repeat imaging demonstrated a large frontal lesion (Fig. 1). 4. Discussion This retrospective study carries significant risk for recollection bias. To minimize such risks, we included all consecutive malignant primary brain tumor patients diagnosed by us at a single institution within a specified two year period. In addition, patients who reported normal neuroimaging studies that could not be verified were included in the total patient number, but excluded as neuroimaging failures. However, a much larger study than our 193 patients is needed to accurately predict early neuroimaging failures in patients with malignant brain tumors. We included three patients who underwent normal neuroimaging several years before a diagnostic study. We do not know whether symptoms at that time – seizures, weakness, and falls – were related to the later development of a primary brain tumor. Whether disease remained symptomatically occult for several years or developed several years later is unknown. However, excluding those patients leaves five of 102 patients with GBM (5%) with normal neuroimaging. Our experience suggests that a single high quality neuroimaging study performed in the emergent setting may be insufficient to detect all malignant primary brain tumors. Repeating such studies two to three months later may be helpful, especially in otherwise healthy, elderly patients suffering unprovoked seizures. Non-specific neuroimaging abnormalities may change rapidly to include findings typical for patients with malignant gliomas. Among the entire cohort of patients with malignant glioma with non-enhancing MRI findings, there were 18 patients with LGG, one with AG, and seven with GBM (eight of 26 or 30.8%). Since standard of care includes early neoadjuvant treatment, re-imaging (within three months) led to earlier therapeutic interventions. Incidental findings are well known among patients undergoing advanced imaging.3,4 The rate of finding such abnormalities is estimated at 2.7% or once in every 37 scans.5 Some investigators propose screening asymptomatic populations with neuroimaging for early malignant primary brain tumor detection.6 In three large cohort series, no primary malignant high grade glioma was discovered by screening MRI of asymptomatic individuals, although three low grade gliomas were found in 6277 adults, while 679 patients had other, non-malignant abnormalities.3,4,6 More recently, a meta-analysis of several studies, including a total of 19,559 patients, found neoplasms in 0.7%. Although description of the prevalence of each specific tumor type was impaired by either a lack of subtyping or non-specific classifications, there were 135 neoplastic findings, including eight low grade gliomas, 72 meningiomas, and 14 unspecified neoplasms.5 Patients with malignant brain tumors often ask whether MRI performed years earlier would have led to earlier diagnosis. Our cohort included three patients with normal neuroimaging years before diagnosis of GBM (Fig. 1). Screening MRI would not have disclosed tumors in these three patients. Our findings suggest routine population screening using neuroimaging will fail to identify patients who later develop WHO grade III or IV gliomas. Older patients who have seizures or focal deficits should undergo follow-up imaging within several months of presentation. In addition, patients who have had normal CT scans or MRI even a short time before presentation with new neurological symptoms should have new studies done. 5. Conclusions Patients who have normal brain CT scans or MRI after an accident, or because of new neurological symptoms, may prove to have high grade gliomas upon repeat imaging done months or years after the initial studies. If non-specific abnormalities are seen on T2-weighted MRI alone, follow-up imaging should be done within, at most, three months. References 1. Ohgaki H, Dessen P, Jourde B, et al. Genetic pathways to glioblastoma: a population-based study. Cancer Res 2004;64:6892–9. 2. Batchelor TT, Betensky RA, Esposito JM, et al. Age-dependent prognostic effects of genetic alterations in glioblastoma. Clin Cancer Res 2004;10:228–33. 3. Katzman GL, Dagher AP, Patronas NJ. Incidental findings on brain magnetic resonance imaging from 1000 asymptomatic volunteers. JAMA 1999;282:36–9. 4. Vernooij MW, Ikram MA, Tanghe HL, et al. Incidental findings on brain MRI in the general population. N Engl J Med 2007;357:1821–8. 5. Morris Z, Whiteley WN, Longstreth Jr WT, et al. Incidental findings on brain magnetic resonance imaging: systematic review and meta-analysis. BMJ 2009;339:b3016. 6. Landy HJ, Lee TT, Potter P, et al. Early MRI findings in high grade glioma. J Neurooncol 2000;47:65–72.