HARM revisited: Etiology of subarachnoid hyperintensities in brain FLAIR MRI

Background: The hyperintense acute reperfusion marker (HARM) describes a phenomenon with a hyperintense signal in the subarachnoid space in Fluid-Attenuated Inversion Recovery (FLAIR) magnetic resonance imaging (MRI) sequences, presumably based on blood–brain barrier breakdown in acute stroke with reperfusion. However, this imaging phenomenon was described in other medical conditions. Aim: Determination of the prevalence and associated clinical findings of this phenomenon in a large sample of patients with different neurological conditions. Methods: This is retrospective, single-center, observational study of 23,948 cerebral MRIs acquired in a Neurological University Clinic over 5 years. The prevalence of HARM, the underlying diagnosis, and damage pattern were examined by chart analysis; MRI was analyzed regarding the type of acute lesions, extent of microangiopathic lesions, and whether gadolinium-based contrast agent (GBCA) was given. Results: Among the MRI data, 84 images (0.35%) from 61 patients were HARM-positive without a subarachnoid signal abnormality in any other sequence. Etiologies were heterogeneous; 35 patients had a cerebrovascular disease (CVD; 19 patients received recanalization therapy), 12 patients had an inflammatory central nervous system (CNS) disease and 14 patients had epilepsy. GBCA was applied to 64% of the patients. Conclusion: HARM was a rare radiological finding in a range of different neurological pathologies, not limited to stroke, or to previous reperfusion therapy and was not dependent on previous GBCA administration. Our data suggest that the term is too narrow in terms of the concepts of the underlying pathology. We propose to use the term FLAIR Subarachnoid Hyperintensity (“FLASH”) phenomenon which might better reflect the observation that the radiological sign can be associated with a variety of central neurological conditions without a straightforward association with therapy.

Substrate Characterization and Outcome of Catheter Ablation of Ventricular Tachycardia in Patients With Nonischemic Cardiomyopathy and Isolated Epicardial Scar

Background: The substrate for ventricular tachycardia (VT) in left ventricular (LV) nonischemic cardiomyopathy may be epicardial. We assessed the prevalence, location, endocardial electrograms, and VT ablation outcomes in LV nonischemic cardiomyopathy with isolated epicardial substrate. Methods: Forty-seven of 531 (9%) patients with LV nonischemic cardiomyopathy and VT demonstrated normal endocardial (>1.5 mV)/abnormal epicardial bipolar low-voltage area (LVA, <1.0 mV and signal abnormality). Abnormal endocardial unipolar LVA (≤8.3 mV) and endocardial bipolar split electrograms and predictors of ablation success were assessed. Results: Epicardial bipolar LVA (27.3 cm 2 [interquartile range, 15.8–50.0]) localized to basal (40), mid (8), and apical (3) LV with basal inferolateral LV most common (28/47, 60%). Of 44 endocardial maps available, 40 (91%) had endocardial unipolar LVA (24.5 cm 2 [interquartile range, 9.4–68.5]) and 29 (67%) had characteristic normal amplitude endocardial split electrograms opposite the epicardial LVA. At mean of 34 months, the VT-free survival was 55% after one and 72% after multiple procedures. Greater endocardial unipolar LVA than epicardial bipolar LVA (hazard ratio, 10.66 [CI, 2.63–43.12], P =0.001) and number of inducible VTs (hazard ratio, 1.96 [CI, 1.27–3.00], P =0.002) were associated with VT recurrence. Conclusions: In patients with LV nonischemic cardiomyopathy and VT, the substrate may be confined to epicardial and commonly basal inferolateral. LV endocardial unipolar LVA and normal amplitude bipolar split electrograms identify epicardial LVA. Ablation targeting epicardial VT and substrate achieves good long-term VT-free survival. Greater endocardial unipolar than epicardial bipolar LVA and more inducible VTs predict VT recurrence.

NIMG-39. RADIOMIC ANALYSIS FOR NON-INVASIVE IN VIVO PROGNOSTIC STRATIFICATION OF DE NOVO GLIOBLASTOMA PATIENTS: A MULTI-INSTITUTIONAL EVALUATION FOR GENERALIZABILITY IN THE RESPOND CONSORTIUM

PURPOSE Multi-parametric MRI based radiomic signatures have highlighted the promise of artificial intelligence (AI) in neuro-oncology. However, inter-institution heterogeneity hinders generalization to data from unseen clinical institutions. To this end, we formulated the ReSPOND (Radiomics Signatures for PrecisiON Diagnostics) consortium for glioblastoma. Here, we seek non-invasive generalizable radiomic signatures from routine clinically-acquired MRI for prognostic stratification of glioblastoma patients. METHODS We identified a retrospective cohort of 606 patients with near/gross total tumor resection ( &gt;90%), from 13 geographically-diverse institutions. All pre-operative structural MRI scans (T1,T1-Gd,T2,T2-FLAIR) were aligned to a common anatomical atlas. An automatic algorithm segmented the whole tumors (WTs) into 3 sub-compartments, i.e., enhancing (ET), necrotic core (NC), and peritumoral T2-FLAIR signal abnormality (ED). The combination of ET+NC defines the tumor core (TC). Quantitative radiomic features were extracted to generate our AI model to stratify patients into short- (&lt; 14mts) and long-survivors ( &gt;14mts). The model trained on 276 patients from a single institution was independently validated on 330 unseen patients from 12 left-out institutions, using the area-under-the-receiver-operating-characteristic-curve (AUC). RESULTS Each feature individually offered certain (limited but reproducible) value for identifying short-survivors: 1) TC closer to lateral ventricles (AUC=0.62); 2) larger ET/brain (AUC=0.61); 3) larger TC/brain (AUC=0.59); 4) larger WT/brain (AUC=0.55); 5) larger ET/WT (AUC=0.59); 6) smaller ED/WT (AUC=0.57); 7) larger ventricle deformations (AUC=0.6). Integrating all features and age, through a multivariate AI model, resulted in higher accuracy (AUC=0.7; 95% C.I.,0.64-0.77). CONCLUSION Prognostic stratification using basic radiomic features is highly reproducible across diverse institutions and patient populations. Multivariate integration yields relatively more accurate and generalizable radiomic signatures, across institutions. Our results offer promise for generalizable non-invasive in vivo signatures of survival prediction in patients with glioblastoma. Extracted features from clinically-acquired imaging, renders these signatures easier for clinical translation. Large-scale evaluation could contribute to improving patient management and treatment planning. *Indicates equal authorship.

NIMG-55. AUGMENTED INTELLIGENCE IS SUPERIOR TO ARTIFICIAL INTELLIGENCE! HUMAN-COMPUTER SYNERGY FOR GENERATING HIGH QUALITY GLIOBLASTOMA SUB-REGION SEGMENTATIONS

PURPOSE Artificial intelligence (AI) is poised to improve diagnostic methods in neuro-oncologic imaging and contribute to patient management by analyzing pre-operative MRI scans. AI results are better interpreted by compartmentalizing glioblastoma into distinct sub-regions, i.e., necrotic core, enhancing tumor, peritumoral T2/FLAIR signal abnormality (ED). Manual delineation of these sub-regions by expert neuroradiologists is impractical, requiring hours for intricate cases. Computer-aided segmentation (CAS) can mitigate this issue but is limited in the quality of the produced segmentations. We hypothesize that CAS followed by expert refinements is more practical/time-efficient. METHODS CAS was used on a total of 359 glioblastoma patients with four MRI sequences (T1, T1Gd, T2, T2-FLAIR) from each patient. All segmentations were sent to expert neuroradiologist annotators for manual refinements. Once refined, our team including two senior attending neuroradiologists with ≥13 years of experience each, reviewed and either approved or returned the segmentations to individual annotators for further refinements. Total time required to refine and review the finalized segmentations was measured. RESULTS Following one round of refinements by expert annotators, 244/359 (68%) segmentations were approved by our team while 115/359 (32%) segmentations contained a variety of errors that required a second round of refinements. The most common observed errors were 1) missed ED in the anterior/inferior temporal lobes and corpus callosum (37/115 cases, 32%) and 2) erroneous segmentation of normal choroid plexus and blood vessels (14/115 cases, 12%). The expert annotators required 120 hours to refine all 359 segmentations, and our team required 26 additional hours to review them, resulting in 24 minutes/segmentation following CAS. CONCLUSION Our findings support the value of a well-communicated annotation protocol to coordinate CAS and expert annotators. With CAS, our team and expert annotators rapidly finalized segmentations for 359 glioblastoma patients, demonstrating the value of a synergistic approach to creating high quality tumor sub-region segmentations.

FLGS-02. What is the role of intraoperative MRI when performing a high-grade glioma resection using 5-ALA? A single center series reporting extent of resection of contrast-enhancing tumor and FLAIR signal abnormality with clinical outcomes

OBJECTIVE Use of intraoperative MRI (iMRI) and 5-ALA together is poorly investigated although both appear individually superior to standard resection alone in meta-analyses. METHODS We identified 54 patients who underwent 5-ALA guided resection of an intracranial high-grade glioma. Intraoperative ultrasound and frameless stereotactic neuronavigation with tractography were used in all cases. iMRI was selectively used in 33 cases. MRI scans were used to measure extent of resection (EOR) of T1-weighted contrast enhancing (T1WCE) tumor and FLAIR signal abnormality. Clinical data were collected prospectively and survival data retrospectively. RESULTS The mean EOR of T1WCE tumor was 93.7% (range 38 – 100%) with mean residual volume of 2.2 cc. The mean EOR of FLAIR signal abnormality was 59% (range 15–100%) with mean residual volume of 35.7 cc. Compared to using 5-ALA alone, using iMRI+5-ALA did not reduce the volume of residual T1WCE tumor (Mann-Whitney U, p = 0.557) and no more patients achieved complete resection (14/21 vs. 24/33, Chi-Square, p = 0.634). The volume of residual FLAIR signal abnormality was no different (Mann-Whitney U, p= 0.081) but more patients in the 5-ALA+iMRI group achieved resection of &gt;53.21% of the FLAIR signal abnormality, which is known to be a predictor of improved prognosis (20/31 vs 5/21, Chi-Square, p = 0.004*). There was no difference in new neurological deficits at 30 days (Fisher’s exact test, p=0.549). Median OS was 14.3 months and median PFS was 5.2 months. There were no significant differences in OS or PFS between patients operated with iMRI in addition to 5-ALA. CONCLUSIONS 5-ALA and iMRI both assist in safely maximizing EOR when combined with adjuncts such as awake craniotomy and cortical/subcortical mapping. The role and interaction of these two technologies needs further study to understand whether they are additive and what their influence on clinical outcomes is.

Encephalopathy With Alternating Hemispheric MRI Abnormalities

A 21-year-old woman with a long-standing history of migraine sought care at her local provider for a 1-week history of confusion and mixing up her words. She then had a witnessed seizure, with dyscognitive features and secondary generalization. On hospitalization, electroencephalography demonstrated left temporal theta slowing and sharp waves. Magnetic resonance imaging showed patchy T2-signal abnormality, nonenhancing, in the left temporal region (only a report was available). Thyroid peroxidase antibodies were increased at 271 IU/mL. A diagnosis of an autoimmune encephalopathy was made, and the patient was treated with phenytoin, levetiracetam, and high-dose corticosteroids, followed by a slow oral prednisone taper. The patient improved cognitively but had considerable emotional lability and an increase in headache frequency and severity and, thus, sought a second opinion. Blood was drawn for genetic testing. The patient died in her sleep a short time later, most likely in the context of sudden unexplained death in epilepsy. Her genetic testing results became available 1 month later, which showed findings consistent with MELAS syndrome: heteroplasmic sequence variation m.3243A>G (tRNA Leu) and homoplasmic rare variant m.2294A>G (16S rRNA). Encephalopathy or encephalitis of subacute onset with fluctuating course is not unique to autoimmune encephalitis. Common acquired metabolic disorders must be considered and excluded in all cases, such as deficiencies of vitamin B12 and folate, hypothyroidism, sepsis, and central nervous system–active medications.

Weakness and Dysarthria After Radiosurgery

A 60-year-old woman with a history of Sjögren syndrome had an episode of painful left eye vision loss. Brain magnetic resonance imaging showed an arteriovenous malformation adjacent to the left ventricular atrium. Although this was considered an asymptomatic lesion, the patient underwent stereotactic radiosurgery to reduce the risk of future growth and hemorrhage. Within days of the surgery, speech disturbance and weakness of the right arm and leg developed. Examination indicated a subcortical language deficit and an upper motor pattern of paresis of right-sided limbs. Considered in the differential diagnosis were new hemorrhage from the arteriovenous malformation, ischemic stroke, radiation-induced necrosis, abscess, and demyelinating disease. Magnetic resonance imaging of the head with and without gadolinium contrast was used to evaluate for these possibilities. Brain magnetic resonance imaging after onset of speech and motor symptoms demonstrated new areas of confluent T2 signal abnormality in the brainstem and deep white matter of the left hemisphere, with some accompanying enhancement but without evidence of hemorrhage or acute stroke. Slow conduction was noted in the left optic nerve on visual evoked potentials. The patient was diagnosed with optic neuritis, with subsequent evolution to multiple sclerosis in the setting of radiosurgery. The patient received intravenous methylprednisolone with full recovery of language function and partial recovery of the hemiparesis, which improved with rehabilitation such that she could walk without a gait aid. Because the patient had more than 1 episode, she was treated with interferon beta-1a for further prevention of multiple sclerosis relapse. Chronic neurotoxicity leading to subcortical dementia occurs in approximately 25% of patients undergoing whole-brain radiotherapy. Histologically, demyelination, as well as necrosis, can be a prominent feature. Among patients with multiple sclerosis or a clinically isolated syndrome, the risk of demyelinating events appears to increase after brain radiotherapy, within the field of treatment.

Progressive Myelopathy After Spine Surgery

A 69-year-old man with a progressive myelopathy for 2 years was referred for evaluation of suspected transverse myelitis. His medical history included discectomies, a severe episode of herpes simplex virus type 1 meningoencephalitis, and development of insidiously progressive numbness and weakness of his hands. Cervical spine magnetic resonance imaging showed 2 small, dural-based, gadolinium-enhancing lesions. Biopsy of these lesions showed only normal neural tissue. Subsequently, the dura was stripped away surgically from the lower cervical region, in an effort to remove these lesions. During the next year, a sensory level developed at about the level of the nipples (T4), along with a squeezing sensation on his trunk below. Imbalance and bilateral lower extremity weakness and numbness then developed. Magnetic resonance imaging showed a longitudinally extensive cord signal abnormality. The cause of the patient’s initial subjective hand numbness and weakness was indeterminate. The onset of severely progressive symptoms after surgical removal of those lesions and the reported stripping of dura made it likely that the progressive cord edema was due to chronic adhesive arachnoiditis. His prior meningoencephalitis was a potential additional risk factor for arachnoiditis. Computed tomography myelography showed a markedly abnormal spinal canal with scalloping of the cord contour, with delayed flow of contrast above C6-C7, consistent with arachnoid adhesions causing obstruction of normal cerebrospinal fluid flow. The patient was diagnosed with chronic adhesive arachnoiditis. A C4-C7 laminectomy and surgical lysis of the cord meningeal adhesions was performed, with subsequent intensive neurorehabilitation. Follow-up spinal cord magnetic resonance imaging 6 months after surgery showed improvement of the T2-signal abnormality but persistent myelomalacia and spinal cord atrophy. Adhesive arachnoiditis is an uncommon cause of progressive myelopathy resulting from an insult to the arachnoid meningeal layer, followed by inflammation and fibrosis. This process renders the arachnoid abnormally thick and adherent to the pia and dura mater. Abnormal adhesion of nerve roots or spinal cord to the dura produces neurologic impairment. Typical symptoms include back pain, paresthesias, lower limb weakness, and sensory loss. It is diagnosed clinically with supportive magnetic resonance imaging and computed tomography myelography findings.

Hirayama disease: the importance of flexion imaging

Hirayama disease is a rare cervical myelopathy characterised by asymmetrical upper limb weakness and muscle atrophy in the forearm and hand. MRI of the cervical spine plays an essential role in diagnosis, however, the characteristic findings are often only seen when the patient is imaged with the neck in flexion. We present a case of a 15-year-old male who presented with left forearm and hand weakness with muscle wasting. An MRI of the cervical spine with the neck in a neutral position demonstrated atrophy of the spinal cord with intrinsic signal abnormality between C5 and C7. Further imaging with the patient’s neck in flexion demonstrated the hallmark features of Hirayama disease. There was anterior displacement of the thecal sac and spinal cord, and an enlarged, crescent-shaped dorsal epidural space which enhanced following i.v. gadolinium administration. The atrophic segment of cord contacted the posterior vertebral bodies when the neck was in full flexion. This case highlights the importance of imaging patients suspected of having this entity with the neck in full flexion in order to make a diagnosis.