Clinical Neuroradiology最新文献

Purpose: Intracranial hemorrhage (ICH) is a life-threatening condition requiring rapid diagnostic and therapeutic action. This study evaluates whether Artificial intelligence (AI) can provide high-quality ICH diagnostics and turnaround times suitable for routine radiological practice.

Methods: A convolutional neural network (CNN) was trained and validated to detect ICHs on DICOM images of cranial CT (CCT) scans, utilizing about 674,000 individually labeled slices. The CNN was then incorporated into a commercial AI engine and seamlessly integrated into three pilot centers in Germany. A real-world test-dataset was extracted and manually annotated by two experienced experts. The performance of the AI algorithm against the two raters was assessed and compared to the inter-rater agreement. The overall time ranging from data acquisition to the delivery of the AI results was analyzed.

Results: Out of 6284 CCT examinations acquired in three different centers, 947 (15%) had ICH. Breakdowns of hemorrhage types included 8% intraparenchymal, 3% intraventricular, 6% subarachnoidal, 7% subdural, < 1% epidural hematomas. Comparing the AI's performance on a subset of 255 patients with two expert raters, it achieved a sensitivity of 0.90, a specificity of 0.96, an accuracy of 0.96. The corresponding inter-rater agreement was 0.84, 0.98, and 0.96. The overall median processing times for the three centers were 9, 11, and 12 min, respectively.

Conclusion: We showed that an AI algorithm for the automatic detection of ICHs can be seamlessly integrated into clinical workflows with minimal turnaround time. The accuracy was on par with radiology experts, making the system suitable for routine clinical use.

Purpose: The modified Rankin Scale (mRS) at 90 days is the primary outcome in most acute stroke studies, but the long follow-up period has disadvantages. The National Institutes of Health Stroke Scale (NIHSS) at 24 h shows a strong, but imperfect, association with 90-day mRS. This study examines the association between 24-hour NIHSS and 90-day mRS and reasons for discrepancies.

Methods: Data are from the ESCAPE-NA1 thrombectomy patients. To address the non-normality distribution of the NIHSS and include deceased patients, a 7-point ordinal score was generated by grouping 24-hour NIHSS. The association of ordinal 24-hour NIHSS and 90-day mRS was assessed with adjusted ordinal logistic regression. Differences in baseline and treatment/post-treatment variables were compared between patients with discordant and concordant outcomes.

Results: One-thousand-seventy-six patients with available 24-hour NIHSS and 90-day mRS were included (median 24-hour NIHSS 6[IQR: 2-14], median 90-day mRS 2[IQR: 1-4]). Ordinal 24-hour NIHSS was associated with 90-day mRS (adjusted cOR 2.53 [95%CI 2.33-2.74]). Forty-eight (4.5%) patients had discordant outcomes. Of those, 19(1.8%) had 24-hour NIHSS < 6 and 90-day mRS5-6; all of which had ≥ 1 severe adverse event, most commonly pneumonia (6[31.6%]) or recurrent stroke (4[21.1%]). Twenty-nine patients (2.7%) had 24-hour NIHSS > 14 and 90-day mRS 0-2. In these patients, baseline NIHSS and ASPECTS was lower, and collateral status was worse.

Conclusion: An ordinal NIHSS score that includes death at 24 h shows a strong association with 90-day mRS, suggesting that it could be used as an alternative outcome. Patients with discrepant outcomes differed from the remaining patients regarding their baseline NIHSS, ASPECTS, collateral status, and post-stroke complications.

Purpose: Direct aspiration first pass technique (ADAPT) has been the standard for aspiration thrombectomy (AT) in anterior circulation large vessel occlusion (AC-LVO) with modifications of the technique based on devices used and difficulties encountered. We introduce the ASCEND technique (Aspiration with Steam-shaped Catheter, Excluding additioNal Devices), and hypothesize that it improves catheter navigation, leading to time and cost savings in MT.

Methods: This is a single institute, retrospective, pre-post intervention study, including consecutive patients with AC-LVO who underwent AT as first-pass technique. Patients were divided into two groups based on the first-pass technique attempted (ASCEND vs conventional ADAPT). Baseline characteristics, primary outcomes (first pass time, total procedure time, total additional devices and device cost) and secondary outcomes (recanalization, complications) were compared between groups. Multiple linear regression models were built for primary outcomes to look for the effect of steam-shaping when covariates are present to reflect real-world setting. Multi-national survey performed to introduce the technique and feedback obtained.

Results: ASCEND (n = 39) and cADAPT (n = 40) groups were similar in baseline clinical characteristics. Anterior genu and ophthalmic segment were crossed in 94.9%, clot reached in 84.6%, and entire MT procedure completed in 59.0% of patients of the ASCEND group without use of additional materials. Groups were similar in performance and safety indicators. ASCEND technique was superior to conventional ADAPT with less first-pass time (8.9 vs 14.7 min), total procedure time (20.2 vs 35.4 min), additional devices used (0 vs 3) and cost involved (2083 vs 5830 £) per procedure (all P < 0.05). Multiple linear regression models maintained improved primary outcomes with steam-shaping (all P < 0.05). Neurointerventionalists who tried ASCEND (n = 11) affirmed that it was safe and likely to save time and cost involved.

Conclusion: ASCEND technique, involving a simple step of steam-shaping the aspiration catheter during MT can provide huge benefits in time and cost savings, without compromise of performance or safety.

Purpose: Deep learning (DL) methods for detecting large vessel occlusion (LVO) in acute ischemic stroke (AIS) show promise, but the effect of computed tomography angiography (CTA) image quality on DL performance is unclear. Our study investigates the impact of improved image quality from Photon Counting Computed Tomography (PCCT) on LVO detection in AIS using a DL-based software prototype developed by a commercial vendor, which incorporates a novel deep learning architecture.

Materials and methods: 443 cases that underwent stroke diagnostics with CTA were included. Positive cases featured vascular occlusions in the Internal Carotid Artery (ICA), M1, and M2 segments of the Middle Cerebral Artery (MCA). Negative cases showed no vessel occlusion on CTA. The performance of the DL-based LVO detection software prototype was assessed using Syngo.via version VB80.

Results: Our study included 267 non-occlusion cases and 176 cases. Among them, 150 cases were scanned via PCCT (no occlusion = 100, ICA and M1 = 41, M2 = 9), while 293 cases were scanned using conventional CT (no occlusion = 167, ICA and M1 = 89, M2 = 37). Independent of scanner type, the algorithm showed sensitivity and specificity of 70.5 and 98.9% for the detection of all occlusions. DL algorithm showed improved performance after excluding M2 occlusions (sensitivity 86.2%). After stratification by scanner type, the algorithm showed significantly a trend towards better performance (p = 0.013) on PCCT CTA images for the detection of all occlusions (sensitivity 84.0%, specificity 99%) compared to CTA images from conventional CT scanner (sensitivity 65.1%, specificity 98.8%). The detection of M2 occlusions was also better on PCCT CTA images (sensitivity 55.6%) compared to conventional scanner CTA images (sensitivity 18.9%), but the sample size for M2 occlusions was limited, and further research is needed to confirm these findings.

Conclusion: Our study suggests that PCCT CTA images may offer improved detection of large vessel occlusion, particularly for M2 occlusions. However further research is needed to confirm these findings. One of the limitations of our study is the inability to exclude the presence of a perfusion deficit, despite ruling out vascular occlusion, due to the lack of CT perfusion (CTP) imaging data. Future research may investigate CNNs by leveraging both CTA and CTP images from PCCT for improved performance.

Background and importance: Flow diverters (FDs) provide curative endovascular treatment for wide-necked sidewall aneurysms. The efficacy of FDs for bifurcation or branching sidewall aneurysms is probably limited. We used anatomical flow diversion (AFD) for intractable large cerebral aneurysms. We report our experiences with AFD.

Methods: The concept of AFD is the transformation from the bifurcation or branching sidewall type to the nonbranching sidewall type. Linearization of the parent artery by stenting, intentional branch occlusion, and aneurysmal coil embolization were performed. Furthermore, bypass surgery is performed for patients intolerant to branch occlusions. We evaluated the clinical outcomes of intractable aneurysms treated with AFD.

Results: AFD was performed in seven unruptured large aneurysms. Aneurysmal locations were the top of the basilar artery (BA), BA-superior cerebellar artery (SCA), internal carotid artery (IC)-posterior communicating artery (PcomA), and IC terminal. The mean dome diameter was 17.0 ± 4.6 mm. Six patients underwent bypass surgery. The occluded branches were the PCA + SCA, PcomA, and anterior cerebral artery (ACA) A1. An FD was used in three patients and a neck bridge stent in four patients. No intraprocedural complications occurred. Two postprocedural ischemic complications occurred in one patient. Six (86%) patients demonstrated a modified Rankin Scale (mRS) 0 at the 3-month follow-up, and one with an ischemic complication showed an mRS 5. Complete occlusion of all aneurysms was maintained with a median follow-up duration of 60 months.

Conclusion: AFD is useful for intractable large cerebral aneurysms with high curability, although safety verification is required.

Purpose: Since perfusion imaging may be unavailable in smaller hospitals, alternative imaging selection methods for acute ischemic stroke can improve outcomes and optimize resources. This study assessed the safety and effectiveness of using imaging criteria other than DEFUSE 3 and DAWN for thrombectomy beyond 6 h from symptom onset in patients stroke in the anterior circulation.

Methods: This is a retrospective, single-center analysis of consecutive patients with large vessel occlusion in the anterior circulation undergoing thrombectomy. Patients were categorized into two groups based on the collateral status (moderate collaterals and good collaterals).

Results: Among 198 patients, 106 (54%) met the inclusion criteria and were analyzed. Good collateral status was observed in 78 (74%) patients. Patients with good collaterals showed significantly lower mRS scores at discharge and at 90 days compared to their counterparts with moderate collateral status (4 (3-4) vs. 4 (4-5); p = 0.001 and 2 (0-4) vs. 6 (3-6); p < 0.001, respectively). More patients with good collateral status achieved favorable outcomes at 90 days compared to those with moderate status (48 (61.5%) vs. 5 (17.9%); p < 0.001). Good collaterals were an independent predictor of good clinical outcomes at 90 days (OR = 1.31, 95% CI: 1.13-1.53, p < 0.001).

Conclusion: Selecting patients for endovascular treatment of acute ischemic stroke using non-contrast CT and CT angiography shows 90-day outcomes similar to the DAWN and DEFUSE-3 trials. Using collateral status on CT angiography can predict favorable outcomes after mechanical thrombectomy in resource-limited settings where perfusion imaging is unavailable.

Purpose: Glioma is a complex cancer comprising various subtypes and mutations, which may have different metabolic characteristics that can potentially be investigated and identified using perfusion imaging. Therefore, the aim of this work was to use radiomics and machine learning analysis of arterial spin labeling MRI data to automatically differentiate glioma subtypes and mutations.

Methods: A total of 495 Arterial Spin Labeling (ASL) perfusion imaging datasets from the UCSF Glioma database were used in this study. These datasets were segmented to delineate the tumor volume and classified according to tumor grade, pathological diagnosis, and IDH status. Perfusion image data was obtained from a 3T MRI scanner using pseudo-continuous ASL. High level texture features were extracted for each ASL dataset using PyRadiomics after tumor volume segmentation and then analyzed using a machine learning framework consisting of ReliefF feature ranking and logistic model tree classification algorithms.

Results: The results of the evaluation revealed balanced accuracies for the three endpoints ranging from 55.76% (SD = 4.28, 95% CI: 53.90-57.65) for the tumor grade using 25.4 ± 37.21 features, 62.53% (SD = 2.86, 95% CI: 61.27-63.78) for the mutation status with 23.3 ± 29.17 picked features, and 80.97% (SD = 1.83, 95% CI: 80.17-81.78) for the pathological diagnosis which used 47.3 ± 32.72 selected features.

Conclusions: Radiomics and machine learning analysis of ASL perfusion data in glioma patients hold potential for aiding in the diagnosis and treatment of glioma, mainly for discerning glioblastoma from astrocytoma, while performance for tumor grading and mutation status appears limited.