Safety and technical efficacy of early minimally invasive endoscopy-guided surgery for intracerebral haemorrhage: the Dutch Intracerebral haemorrhage Surgery Trial pilot study.

BACKGROUND: Previous randomised controlled trials could not demonstrate that surgical evacuation of intracerebral haemorrhage (ICH) improves functional outcome. Increasing evidence suggests that minimally invasive surgery may be beneficial, in particular when performed early after symptom onset. The aim of this study was to investigate safety and technical efficacy of early minimally invasive endoscopy-guided surgery in patients with spontaneous supratentorial ICH. METHODS: The Dutch Intracerebral Haemorrhage Surgery Trial pilot study was a prospective intervention study with blinded outcome assessment in three neurosurgical centres in the Netherlands. We included adult patients with spontaneous supratentorial ICH ≥10mL and National Institute of Health Stroke Scale (NIHSS) score ≥2 for minimally invasive endoscopy-guided surgery within 8 h after symptom onset in addition to medical management. Primary safety outcome was death or increase in NIHSS ≥4 points at 24 h. Secondary safety outcomes were procedure-related serious adverse events (SAEs) within 7 days and death within 30 days. Primary technical efficacy outcome was ICH volume reduction (%) at 24 h. RESULTS: We included 40 patients (median age 61 years; IQR 51-67; 28 men). Median baseline NIHSS was 19.5 (IQR 13.3-22.0) and median ICH volume 47.7mL (IQR 29.4-72.0). Six patients had a primary safety outcome, of whom two already deteriorated before surgery and one died within 24 h. Sixteen other SAEs were reported within 7 days in 11 patients (of whom two patients that already had a primary safety outcome), none device related. In total, four (10%) patients died within 30 days. Median ICH volume reduction at 24 h was 78% (IQR 50-89) and median postoperative ICH volume 10.5mL (IQR 5.1-23.8). CONCLUSIONS: Minimally invasive endoscopy-guided surgery within 8 h after symptom onset for supratentorial ICH appears to be safe and can effectively reduce ICH volume. Randomised controlled trials are needed to determine whether this intervention also improves functional outcome. TRIAL REGISTRATION: Clinicaltrials.gov : NCT03608423, August 1st, 2018.

Subclassification of the Koos grade 2 vestibular schwannoma into 2a and 2b for individualized patient care: A validity and reliability study.

OBJECTIVE: Vestibular schwannoma (VS) growth of ≥2 mm during serial MRI observation, irrespective of size, is the benchmark for treatment initiation in almost all centers. Although the probability of less optimal outcomes significantly increases in VS closer to the brainstem, early intervention does not improve long-term quality of life. Moving beyond the recommendation of definitive treatment for all VS after detected growth, we subclassified Koos 2 tumors based on extrameatal extension and relation to the brainstem. The aim of the current study was to evaluate the Koos 2 subclassification's validity and the inter-and intra-rater reliability of the entire Koos classification. METHODS: Six experts, including neurosurgeons, otorhinolaryngologists and radiologists from two tertiary referral centers, classified 43 VS MRI scans. Validity of the Koos 2 subclassification was evaluated by the percentage agreement against the multidisciplinary skull base tumor board management advice. Inter- and intra-rater reliability were calculated using the intraclass correlation coefficient (ICC). RESULTS: Validity was almost perfect in Koos 2a VSs with a 100% agreement and 87.5% agreement for Koos 2b. Inter-rater reliability for all Koos grades was significantly excellent (ICC 0.91; 95%CI 0.866 to 0.944, p= <0.001). Five raters had an excellent intra-rater reliability (ICC > 0.90; p= <0.01) and one rater had a good intra-rater reliability (ICC 0.88; 95% CI 0.742 to 0.949). CONCLUSIONS: Although multiple factors influence decision-making, the classification of Koos 2a and 2b with excellent inter- and intra-rater reliability, can aid in recommending treatment initiation, moving beyond detected tumor growth, aiming to optimize patient centered care.

Genotype and Phenotype Analyses of a Novel WFS1 Variant (c.2512C>T p.(Pro838Ser)) Associated with DFNA6/14/38.

The aim of this study is to contribute to a better description of the genotypic and phenotypic spectrum of DFNA6/14/38 and aid in counseling future patients identified with this variant. Therefore, we describe the genotype and phenotype in a large Dutch-German family (W21-1472) with autosomal dominant non-syndromic, low-frequency sensorineural hearing loss (LFSNHL). Exome sequencing and targeted analysis of a hearing impairment gene panel were used to genetically screen the proband. Co-segregation of the identified variant with hearing loss was assessed by Sanger sequencing. The phenotypic evaluation consisted of anamnesis, clinical questionnaires, physical examination and examination of audiovestibular function. A novel likely pathogenic WFS1 variant (NM_006005.3:c.2512C>T p.(Pro838Ser)) was identified in the proband and found to co-segregate with LFSNHL, characteristic of DFNA6/14/38, in this family. The self-reported age of onset of hearing loss (HL) ranged from congenital to 50 years of age. In the young subjects, HL was demonstrated in early childhood. At all ages, an LFSNHL (0.25-2 kHz) of about 50-60 decibel hearing level (dB HL) was observed. HL in the higher frequencies showed inter-individual variability. The dizziness handicap inventory (DHI) was completed by eight affected subjects and indicated a moderate handicap in two of them (aged 77 and 70). Vestibular examinations (n = 4) showed abnormalities, particularly in otolith function. In conclusion, we identified a novel WFS1 variant that co-segregates with DFNA6/14/38 in this family. We found indications of mild vestibular dysfunction, although it is uncertain whether this is related to the identified WFS1 variant or is an incidental finding. We would like to emphasize that conventional neonatal hearing screening programs are not sensitive to HL in DFNA6/14/38 patients, because high-frequency hearing thresholds are initially preserved. Therefore, we suggest screening newborns in DFNA6/14/38 families with more frequency-specific methods.

Abdominopelvic CT Image Quality: Evaluation of Thin (0.5-mm) Slices Using Deep Learning Reconstruction.

BACKGROUND. Because thick-section images (typically 3-5 mm) have low image noise, radiologists typically use them to perform clinical interpretation, although they may additionally refer to thin-section images (typically 0.5-0.625 mm) for problem solving. Deep learning reconstruction (DLR) can yield thin-section images with low noise. OBJECTIVE. The purpose of this study is to compare abdominopelvic CT image quality between thin-section DLR images and thin- and thick-section hybrid iterative reconstruction (HIR) images. METHODS. This retrospective study included 50 patients (31 men and 19 women; median age, 64 years) who underwent abdominopelvic CT between June 15, 2020, and July 29, 2020. Images were reconstructed at 0.5-mm section using DLR and at 0.5-mm and 3.0-mm sections using HIR. Five radiologists independently performed pairwise comparisons (0.5-mm DLR and either 0.5-mm or 3.0-mm HIR) and recorded the preferred image for subjective image quality measures (scale, -2 to 2). The pooled scores of readers were compared with a score of 0 (denoting no preference). Image noise was quantified using the SD of ROIs on regions of homogeneous liver. RESULTS. For comparison of 0.5-mm DLR images and 0.5-mm HIR images, the median pooled score was 2 (indicating a definite preference for DLR) for noise and overall image quality and 1 (denoting a slight preference for DLR) for sharpness and natural appearance. For comparison of 0.5-mm DLR and 3.0-mm HIR, the median pooled score was 1 for the four previously mentioned measures. These assessments were all significantly different (p < .001) from 0. For artifacts, the median pooled score for both comparisons was 0, which was not significant for comparison with 3.0-mm HIR (p = .03) but was significant for comparison with 0.5-mm HIR (p < .001) due to imbalance in scores of 1 (n = 28) and -1 (slight preference for HIR, n = 1). Noise for 0.5-mm DLR was lower by mean differences of 12.8 HU compared with 0.5-mm HIR and 4.4 HU compared with 3.0-mm HIR (both p < .001). CONCLUSION. Thin-section DLR improves subjective image quality and reduces image noise compared with currently used thin- and thick-section HIR, without causing additional artifacts. CLINICAL IMPACT. Although further diagnostic performance studies are warranted, the findings suggest the possibility of replacing current use of both thin- and thick-section HIR with the use of thin-section DLR only during clinical interpretations.

High-Accuracy Nodal Staging of Head and Neck Cancer With USPIO-Enhanced MRI: A New Reading Algorithm Based on Node-to-Node Matched Histopathology.

OBJECTIVES: Ultrasmall superparamagnetic iron oxide (USPIO)-enhanced magnetic resonance imaging (MRI) is a potential diagnostic tool for lymph node assessment in patients with head and neck cancer. Validation by radiologic-pathologic correlation is essential before the method is evaluated in clinical studies. In this study, MRI signal intensity patterns of lymph nodes are correlated to their histopathology to develop a new USPIO-enhanced MRI reading algorithm that can be used for nodal assessment in head and neck cancer patients. MATERIALS AND METHODS: Ten head and neck cancer patients underwent in vivo USPIO-enhanced MRI before neck dissection. An ex vivo MRI of the neck dissection specimen was performed for precise coregistration of in vivo MRI with histopathology. Normal clinical histopathological workup was extended with meticulous matching of all lymph nodes regarded as potentially metastatic based on their in vivo MRI signal intensity pattern. On the basis of histopathology of resected nodes, in vivo MRI signal characteristics were defined separating benign from malignant lymph nodes. RESULTS: Fifteen of 34 node-to-node correlated lymph nodes with remaining signal intensity on T2*-weighted MRI were histopathologically metastatic and 19 were benign. Radiological analysis revealed that metastatic lymph nodes showed equal or higher MRI signal intensity when compared with lipid tissue on T2*-weighted MGRE sequence (15/16 lymph nodes; 94%), whereas healthy lymph nodes showed lower (17/19 lymph nodes; 89%) or complete attenuation of signal intensity (273/279; 98%) when compared with lipid tissue on T2*-weighted MGRE. Histopathology of all resected specimens identified 392 lymph nodes. Six lymph nodes with (micro)metastases were missed with in vivo MRI. Whether these 6 lymph nodes were correlated to a nonmalignant lymph node on in vivo MRI or could not be detected at all is unclear. CONCLUSIONS: We developed a new reading algorithm to differentiate benign from malignant lymph nodes in head and neck cancer patients on the basis of their appearance on high-resolution T2*-weighted USPIO-enhanced MRI. Next steps involve validation of our reading algorithm to further improve the accuracy of neck lymph node staging with USPIO-enhanced MRI in prospective clinical studies with larger number of patients.

Assessing radiation-induced carotid vasculopathy using ultrasound after unilateral irradiation: a cross-sectional study.

BACKGROUND: Increased head and neck cancer (HNC) survival requires attention to long-term treatment sequelae. Irradiated HNC survivors have a higher ischemic stroke risk. However, the pathophysiology of radiation-induced vasculopathy is unclear. Arterial stiffness could be a biomarker. This study examined alterations in intima-media thickness (IMT) and stiffness-related parameters, shear wave (SWV) and pulse wave velocity (PWV), in irradiated compared to control carotids in unilateral irradiated patients. METHODS: Twenty-six patients, median 40.5 years, 5-15 years after unilateral irradiation for head and neck neoplasms underwent a bilateral carotid ultrasound using an Aixplorer system with SL18-5 and SL10-2 probes. IMT, SWV, and PWV were assessed in the proximal, mid, and distal common (CCA) and internal carotid artery (ICA). Plaques were characterized with magnetic resonance imaging. Measurements were compared between irradiated and control sides, and radiation dose effects were explored. RESULTS: CCA-IMT was higher in irradiated than control carotids (0.54 [0.50-0.61] vs. 0.50 [0.44-0.54] mm, p = 0.001). For stiffness, only anterior mid-CCA and posterior ICA SWV were significantly higher in the irradiated side. A radiation dose-effect was only (weakly) apparent for PWV (R2: end-systolic = 0.067, begin-systolic = 0.155). Ultrasound measurements had good-excellent intra- and interobserver reproducibility. Plaques had similar characteristics but were more diffuse in the irradiated side. CONCLUSIONS: Increased CCA-IMT and SWV in some segments were seen in irradiated carotids. These alterations, even in young patients, mark the need for surveillance of radiation-induced vasculopathy. TRIAL REGISTRATION: clinicaltrials.gov ( https://clinicaltrials.gov/ct2/show/NCT04257968 ).

Long-term cognitive, psychosocial, and neurovascular complications of unilateral head and neck irradiation in young to middle-aged adults.

BACKGROUND: With a growing, younger population of head and neck cancer survivors, attention to long-term side-effects of prior, often radiotherapeutic, treatment is warranted. Therefore, we studied the long-term cognitive effects in young adult patients irradiated for head and neck neoplasms (HNN). METHODS: Young to middle-aged adults with HNN (aged 18-40 years) and treated with unilateral neck irradiation ≥ 5 years before inclusion underwent cardiovascular risk and neuropsychological assessments and answered validated questionnaires regarding subjective cognitive complaints, fatigue, depression, quality of life, and cancer-specific distress. Additionally, magnetic resonance imaging (MRI) of the brain was performed to assess white matter hyperintensities (WMH), infarctions, and atrophy. RESULTS: Twenty-nine patients (aged 24-61, 13 men) median 9.2 [7.3-12.9] years post-treatment were included. HNN patients performed worse in episodic memory (Z-score = -1.16 [-1.58-0.34], p < 0.001) and reported more fatigue symptoms (Z-score = 1.75 [1.21-2.00], p < 0.001) compared to normative data. Furthermore, patients had a high level of fear of tumor recurrence (13 patients [44.8%]) and a heightened speech handicap index (13 patients [44.8%]). Only a small number of neurovascular lesions were found (3 infarctions in 2 patients and 0.11 [0.00-0.40] mL WMH), unrelated to the irradiated side. Cognitive impairment was not associated with WMH, brain atrophy, fatigue, or subjective speech problems. CONCLUSIONS: HNN patients showed impairments in episodic memory and an increased level of fatigue ≥ 5 years after radiotherapy compared to normative data. Cognitive impairments could not be explained by WMH or brain atrophy on brain MRI or psychological factors. TRIAL REGISTRATION: Clinicaltrials.gov ( https://clinicaltrials.gov/ct2/show/NCT04257968 ).

Exploring the missing heritability in subjects with hearing loss, enlarged vestibular aqueducts, and a single or no pathogenic SLC26A4 variant.

Pathogenic variants in SLC26A4 have been associated with autosomal recessive hearing loss (arHL) and a unilateral or bilateral enlarged vestibular aqueduct (EVA). SLC26A4 is the second most frequently mutated gene in arHL. Despite the strong genotype-phenotype correlation, a significant part of cases remains genetically unresolved. In this study, we investigated a cohort of 28 Dutch index cases diagnosed with HL in combination with an EVA but without (M0) or with a single (M1) pathogenic variant in SLC26A4. To explore the missing heritability, we first determined the presence of the previously described EVA-associated haplotype (Caucasian EVA (CEVA)), characterized by 12 single nucleotide variants located upstream of SLC26A4. We found this haplotype and a delimited V1-CEVA haplotype to be significantly enriched in our M1 patient cohort (10/16 cases). The CEVA haplotype was also present in two M0 cases (2/12). Short- and long-read whole genome sequencing and optical genome mapping could not prioritize any of the variants present within the CEVA haplotype as the likely pathogenic defect. Short-read whole-genome sequencing of the six M1 cases without this haplotype and the two M0/CEVA cases only revealed previously overlooked or misinterpreted splice-altering SLC26A4 variants in two cases, who are now genetically explained. No deep-intronic or structural variants were identified in any of the M1 subjects. With this study, we have provided important insights that will pave the way for elucidating the missing heritability in M0 and M1 SLC26A4 cases. For pinpointing the pathogenic effect of the CEVA haplotype, additional analyses are required addressing defect(s) at the RNA, protein, or epigenetic level.

Diagnosing necrotizing external otitis on CT and MRI: assessment of pattern of extension.

BACKGROUND AND PURPOSE: Necrotizing external otitis (NEO) is a serious complication of external otitis. NEO can be classified according to-anterior, medial, posterior, intracranial, and contralateral-extension patterns. Currently there is no consensus on the optimal imaging modality for the identification of disease extension. This study compares NEO extension patterns on MR and CT to evaluate diagnostic comparability. METHODS: Patients who received a CT and MR within a 3-month interval were retrospectively examined. Involvement of subsites and subsequent spreading patterns were assessed on both modalities by a radiologist in training and by a senior head and neck radiologist. The prevalence of extension patterns on CT and MR were calculated and compared. RESULTS: All 21 included NEO cases showed an anterior extension pattern on CT and MR. Contrary to MR, medial extension was not recognized on CT in two out of six patients, and intracranial extension in five out of eight patients. The posterior extension pattern was not recognized on MR. Overall, single anterior extension pattern (62%) is more prevalent than multiple extension patterns (38%). CONCLUSION: All anterior NEO extension pattern were identified on CT as well as MR. However, the medial and intracranial spreading patterns as seen on MR could only be identified on CT in a small number of patients. The posterior spreading pattern can be overlooked on MR. Thus, CT and MR are complimentary for the initial diagnosis and work-up of NEO as to correctly delineate disease extent through the skull base.

Excessive toxicity of cabozantinib in a phase II study in patients with recurrent and/or metastatic salivary gland cancer.

AIM: Because the tyrosine kinases c-MET and vascular endothelial growth factor receptors (VEGFR) are often overexpressed in salivary gland cancer (SGC), this study evaluated the efficacy and safety of cabozantinib in patients with recurrent/metastatic (R/M) SGC. PATIENTS AND METHODS: A single-centre phase II study was conducted. Patients with immunohistochemical c-MET-positive R/M SGC were included in three cohorts: adenoid cystic carcinoma (ACC); salivary duct carcinoma (SDC) and other miscellaneous SGCs. No prior systemic treatments were required. Patients started cabozantinib 60 mg once daily. The primary outcome was the objective response rate (ORR). Secondary outcomes included survival, safety and quality of life. Per Simon-two-stage design, depending on efficacy, a maximum of 43 patients would be included. RESULTS: In total, 25 patients were included until premature closure owing to severe toxicity. Six patients (24%) had grade 3-5 wound complications, occurring at a median of 7.1 months on cabozantinib treatment (range 2.1-12.6). Remarkably, four of these six patients developed this complication in the area prior exposed to high-dose radiotherapy. Other grade ≥3 adverse events in >1 patient were hypertension (20%), diarrhoea (8%) and dehydration (8%). Twenty-one patients were evaluable for response; 1/15 ACC (ORR: 7%); 1/4 SDC and 0/2 patients with other miscellaneous SGC responded. Median progression-free survival was 9.4 months (95% confidence interval [CI] 7.4-11.4 months), 7.2 months (95%CI 0.0-15.1) and 6.9 months (95%CI 0.0-15.1), respectively. CONCLUSION: This study showed too many severe cabozantinib-associated wound complications in patients with SGC, especially in prior irradiated areas. Therefore, the study closed prematurely. The efficacy in the limited number of evaluable patients was low to moderate. TRIAL REGISTRATION: This trial was registered on ClinicalTrials.gov: NCT03729297.

Effective low-dose sirolimus regimen for kaposiform haemangioendothelioma with Kasabach-Merritt phenomenon in young infants.

AIMS: Management of kaposiform haemangioendotheliomas (KHE) with Kasabach-Merritt phenomenon is challenging in young infants who are subjected to developmental pharmacokinetic changes. Sirolimus, sometimes combined with corticosteroids, can be used as an effective treatment of KHE. Simultaneously, toxicities such as interstitial pneumonitis related to the use of sirolimus may be fatal. As infants have a very low CYP3-enzyme expression at birth, which rises during ageing, we hypothesize that a reduced metabolization of sirolimus might lead to high sirolimus serum levels and low dose may be sufficient without the side effects. METHODS: A case series of 5 infants with kaposiform haemangioendothelioma with Kasabach-Merritt phenomenon was analysed retrospectively. All infants were treated with sirolimus 0.2 mg/m2 every 24 or 48 hours according to their age. Prednisone was added to the therapy for additional effect in 4 patients. RESULTS: In all patients, low dose of sirolimus led to therapeutic sirolimus levels (4-6 ng/mL). All infants (aged 4 days-7 months) had a complete haematological response, without serious adverse events. In all patients, the Kasabach-Merritt phenomenon resolved, the coagulation profile normalized and tumour size reduction was seen. CONCLUSION: Low-dose sirolimus treatment is safe for infants with kaposiform haemangioendothelioma and Kasabach-Merritt phenomenon. It is essential to realize that during the first months of life, metabolism is still developing and enzymes necessary to metabolise drugs like sirolimus still have to mature. To avoid toxic levels, the sirolimus dosage should be based on age and the associated pharmacological developments.

Merkel Cell Carcinoma: New Trends.

Merkel cell carcinoma (MCC) is a rare neuroendocrine tumor of the skin mainly seen in the elderly. Its incidence is rising due to ageing of the population, increased sun exposure, and the use of immunosuppressive medication. Additionally, with the availability of specific immunohistochemical markers, MCC is easier to recognize. Typically, these tumors are rapidly progressive and behave aggressively, emphasizing the need for early detection and prompt diagnostic work-up and start of treatment. In this review, the tumor biology and immunology, current diagnostic and treatment modalities, as well as new and combined therapies for MCC, are discussed. MCC is a very immunogenic tumor which offers good prospects for immunotherapy. Given its rarity, the aggressiveness, and the frail patient population it concerns, MCC should be managed in close collaboration with an experienced multidisciplinary team.

Prognostic and Predictive Value of Integrated Qualitative and Quantitative Magnetic Resonance Imaging Analysis in Glioblastoma.

Glioblastoma (GBM) is the most malignant primary brain tumor for which no curative treatment options exist. Non-invasive qualitative (Visually Accessible Rembrandt Images (VASARI)) and quantitative (radiomics) imaging features to predict prognosis and clinically relevant markers for GBM patients are needed to guide clinicians. A retrospective analysis of GBM patients in two neuro-oncology centers was conducted. The multimodal Cox-regression model to predict overall survival (OS) was developed using clinical features with VASARI and radiomics features in isocitrate dehydrogenase (IDH)-wild type GBM. Predictive models for IDH-mutation, 06-methylguanine-DNA-methyltransferase (MGMT)-methylation and epidermal growth factor receptor (EGFR) amplification using imaging features were developed using machine learning. The performance of the prognostic model improved upon addition of clinical, VASARI and radiomics features, for which the combined model performed best. This could be reproduced after external validation (C-index 0.711 95% CI 0.64-0.78) and used to stratify Kaplan-Meijer curves in two survival groups (p-value < 0.001). The predictive models performed significantly in the external validation for EGFR amplification (area-under-the-curve (AUC) 0.707, 95% CI 0.582-8.25) and MGMT-methylation (AUC 0.667, 95% CI 0.522-0.82) but not for IDH-mutation (AUC 0.695, 95% CI 0.436-0.927). The integrated clinical and imaging prognostic model was shown to be robust and of potential clinical relevance. The prediction of molecular markers showed promising results in the training set but could not be validated after external validation in a clinically relevant manner. Overall, these results show the potential of combining clinical features with imaging features for prognostic and predictive models in GBM, but further optimization and larger prospective studies are warranted.

Myotonic dystrophy and recurrent pleomorphic adenomas: Case report and association hypothesis.

We report a case of a patient with concurrent myotonic dystrophy and recurrent pleomorphic adenoma and hypothesize the association between both diseases. A 58-year-old man with classic myotonic dystrophy type 1 was diagnosed with pleomorphic adenoma. Appropriate treatment was commenced. Massive recurrences occurred within 15, 28 and 22 months respectively, after repeated surgical removal. Three case reports on similar occurrences of synchronous myotonic dystrophy and pleomorphic adenoma are discussed and an association between both disease entities is hypothesized. A conceivable association between myotonic dystrophy and pleomorphic adenoma is hypothesized by upregulation of the Wnt/Beta-catenin signaling pathway, initiated by a decreased expression of microRNA, pleomorphic adenoma gene 1 induced Beta-catenin accumulations and alterations in tumor suppressor genes and oncogenes due to RNA processing defects induced by the expanded repeat in the DMPK gene.

Novel Diagnostic Approaches for Assessment of the Clinically Negative Neck in Head and Neck Cancer Patients.

In head and neck cancer, the presence of nodal disease is a strong determinant of prognosis and treatment. Despite the use of modern multimodality diagnostic imaging, the prevalence of occult nodal metastases is relatively high. This is why in clinically node negative head and neck cancer the lymphatics are treated "electively" to eradicate subclinical tumor deposits. As a consequence, many true node negative patients undergo surgery or irradiation of the neck and suffer from the associated and unnecessary early and long-term morbidity. Safely tailoring head and neck cancer treatment to individual patients requires a more accurate pre-treatment assessment of nodal status. In this review, we discuss the potential of several innovative diagnostic approaches to guide customized management of the clinically negative neck in head and neck cancer patients.

Cerebral Artery and Vein Segmentation in Four-dimensional CT Angiography Using Convolutional Neural Networks.

PURPOSE: To implement and test a deep learning approach for the segmentation of the arterial and venous cerebral vasculature with four-dimensional (4D) CT angiography. MATERIALS AND METHODS: Patients who had undergone 4D CT angiography for the suspicion of acute ischemic stroke were retrospectively identified. A total of 390 patients evaluated in 2014 (n = 113) or 2018 (n = 277) were included in this study, with each patient having undergone one 4D CT angiographic scan. One hundred patients from 2014 were randomly selected, and the arteries and veins on their CT scans were manually annotated by five experienced observers. The weighted temporal average and weighted temporal variance from 4D CT angiography were used as input for a three-dimensional Dense-U-Net. The network was trained with the fully annotated cerebral vessel artery-vein maps from 60 patients. Forty patients were used for quantitative evaluation. The relative absolute volume difference and the Dice similarity coefficient are reported. The neural network segmentations from 277 patients who underwent scanning in 2018 were qualitatively evaluated by an experienced neuroradiologist using a five-point scale. RESULTS: The average time for processing arterial and venous cerebral vasculature with the network was less than 90 seconds. The mean Dice similarity coefficient in the test set was 0.80 ± 0.04 (standard deviation) for the arteries and 0.88 ± 0.03 for the veins. The mean relative absolute volume difference was 7.3% ± 5.7 for the arteries and 8.5% ± 4.8 for the veins. Most of the segmentations (n = 273, 99.3%) were rated as very good to perfect. CONCLUSION: The proposed convolutional neural network enables accurate artery and vein segmentation with 4D CT angiography with a processing time of less than 90 seconds.© RSNA, 2020.