Usefulness of deep learning-assisted identification of hyperdense MCA sign in acute ischemic stroke: comparison with readers' performance.

PURPOSE: To evaluate the usefulness of deep learning-assisted diagnosis for identifying hyperdense middle cerebral artery sign (HMCAS) on non-contrast computed tomography in comparison with the diagnostic performance of neuroradiologists. MATERIALS AND METHODS: We obtained 46 HMCAS-positive and 52 HMCAS-negative test samples extracted using 50-pixel-diameter circular regions of interest. Five neuroradiologists undertook an initial diagnostic performance test by describing the HMCAS-positive prediction rate in each sample. Their diagnostic performance was compared with that of a deep convolutional neural network (DCNN) model that had been trained using another dataset in our previous study. In the second test, readers could reference the prediction rate of the DCNN model in each sample. RESULTS: The diagnostic performance of the DCNN for HMCAS showed an accuracy of 81.6% and area under the receiver-operating characteristic curve (AUC) of 0.869, whereas the initial diagnostic performance of neuroradiologists showed an accuracy of 78.8% and AUC of 0.882. The second diagnostic test of neuroradiologists with reference to the results of the DCNN model showed an accuracy of 84.7% and AUC of 0.932. In all readers, AUC values were higher in the second test than the initial test. CONCLUSION: The ability of DCNN to identify HMCAS is comparable with the diagnostic performance of neuroradiologists.

Spatial coefficient of variation in pseudo-continuous arterial spin labeling cerebral blood flow images as a hemodynamic measure for cerebrovascular steno-occlusive disease: A comparative 15O positron emission tomography study.

Pseudo-continuous arterial spin labeling (pCASL) is a completely non-invasive method of cerebral perfusion measurement. However, cerebral blood flow (CBF) quantification is hampered by arterial transit artifacts characterized by bright vascular signals surrounded by decreased signals in tissue regions, which commonly appear in patients with reduced cerebral perfusion pressure. The spatial coefficient of variation (CoV) of pCASL CBF images has been proposed as an alternative region-of-interest (ROI)-based hemodynamic measure to predict prolonged arterial transit time (ATT). This retrospective study investigates the utility of spatial CoV by comparison with 15O positron emission tomography (PET). For patients with cerebrovascular steno-occlusive disease ( n = 17), spatial CoV was positively correlated with ATT independently measured by pulsed arterial spin labeling ( r = 0.597, p < 0.001), confirming its role as an ATT-like hemodynamic measure. Comparisons with 15O PET demonstrated that spatial CoV was positively correlated with vascular mean transit time ( r = 0.587, p < 0.001) and negatively correlated with both resting CBF ( r = -0.541, p = 0.001) and CBF response to hypercapnia ( r = -0.373, p = 0.030). ROI-based spatial CoV calculated from single time-point pCASL can potentially detect subtle perfusion abnormalities in clinical settings.

Ischemic White Matter Lesions Associated With Medullary Arteries: Classification of MRI Findings Based on the Anatomic Arterial Distributions.

OBJECTIVE: The purposes of this article are to describe the important role of the medullary arteries in the pathogenesis of cerebral vascular disease and to present a classification of MRI findings of ischemic white matter lesions for use in elucidating pathogenesis. CONCLUSION: From the viewpoint of the anatomy of the medullary arteries, the pattern of medullary artery-related ischemic changes and infarcts can be classified into four types: 1, ischemic leukoaraiosis; 2, infarcts involving individual medullary arteries; 3, watershed infarcts; and 4, territorial infarcts.

Impact of subject head motion on quantitative brain (15)O PET and its correction by image-based registration algorithm.

OBJECTIVE: Subject head motion during sequential (15)O positron emission tomography (PET) scans can result in artifacts in cerebral blood flow (CBF) and oxygen metabolism maps. However, to our knowledge, there are no systematic studies examining this issue. Herein, we investigated the effect of head motion on quantification of CBF and oxygen metabolism, and proposed an image-based motion correction method dedicated to (15)O PET study, correcting for transmission-emission mismatch and inter-scan mismatch of emission scans. METHODS: We analyzed (15)O PET data for patients with major arterial steno-occlusive disease (n = 130) to determine the occurrence frequency of head motion during (15)O PET examination. Image-based motion correction without and with realignment between transmission and emission scans, termed simple and 2-step method, respectively, was applied to the cases that showed severe inter-scan motion. RESULTS: Severe inter-scan motion (>3 mm translation or >5° rotation) was observed in 27 of 520 adjacent scan pairs (5.2 %). In these cases, unrealistic values of oxygen extraction fraction (OEF) or cerebrovascular reactivity (CVR) were observed without motion correction. Motion correction eliminated these artifacts. The volume-of-interest (VOI) analysis demonstrated that the motion correction changed the OEF on the middle cerebral artery territory by 17.3 % at maximum. The inter-scan motion also affected CBV, CMRO2 and CBF, which were improved by the motion correction. A difference of VOI values between the simple and 2-step method was also observed. CONCLUSIONS: These data suggest that image-based motion correction is useful for accurate measurement of CBF and oxygen metabolism by (15)O PET.

Malignant astrocytic tumors: clinical importance of apparent diffusion coefficient in prediction of grade and prognosis.

PURPOSE: To retrospectively assess the apparent diffusion coefficient (ADC) for prediction of malignancy and prognosis of malignant astrocytic tumors. MATERIALS AND METHODS: The institutional review board approved this study and did not require patient informed consent. Findings from 37 consecutive patients (21 men, 16 women; mean age, 43 years) with pathologically proved malignant astrocytic tumors that included 22 glioblastomas (GBMs) and 15 anaplastic astrocytomas (AAs) were retrospectively evaluated. The minimum ADC value of each tumor was preoperatively determined from several regions of interest defined in the tumor, preferably with avoidance of cystic or necrotic components, on ADC maps derived from isotropic diffusion-weighted images. Surgical intervention, followed by radiation therapy, was undertaken in all cases according to hospital protocol. Immunohistologically, Ki-67 labeling index (LI), indicating cell proliferation, was also determined. The patients were classified into two groups, progressive and stable, according to the 2-year observation after the initial treatment. Correlation analysis (Pearson product moment correlation), Student t test, Welch test, receiver operating characteristic analysis, and Kaplan-Meier method with log-rank test were used for statistical evaluation. RESULTS: There was a significant negative correlation between minimum ADC and Ki-67 LI (r = -0.562, P < .001). The mean minimum ADC (0.834 x 10(-3) mm2 x sec(-1)) of GBM was significantly lower than that (1.06 x 10(-3) mm2 x sec(-1)) of AA (P < .001, Student t test). The mean minimum ADC (0.80 x 10(-3) mm2 x sec(-1)) of the progressive group was significantly lower than that (1.037 x 10(-3) mm2 x sec(-1)) of the stable group (P < .001). The cutoff value of 0.90 x 10(-3) mm2 x sec(-1) for minimum ADC for differentiation of patients with a favorable prognosis from those with a poor prognosis provided the best combination of sensitivity (79%) and specificity (81%) (receiver operating characteristic analysis). The significant difference in the prognosis between two groups classified by using this cutoff value of minimum ADC was noted (P = .002, log-rank test). CONCLUSION: The minimum ADC of malignant astrocytomas can provide additional information about their clinical malignancy related to posttreatment prognosis.

Changes in brain activation patterns associated with learning of Korean words by Japanese: an fMRI study.

We used functional magnetic resonance imaging (fMRI) to explore the change in brain activation associated with the learning of Korean words written in Han-gul characters (K-words) by young Japanese at two stages. Subjects were 12 right-handed native Japanese without previous knowledge of Korean words and characters. On the first day they were taught the pronunciation and meaning of 20 K-words. Then, after the first fMRI session (on day 2), they were given a set of 20 cards with the words and corresponding photographs. They also received a tape and were instructed to memorize the 20 K-words by studying them every day until the day of the second fMRI session (day 16). During the fMRI sessions, 20 Japanese words written in kana syllabograms (J-words) and the 20 previously presented K-words, as well as 20 new K-words (Kn-words) were presented visually for silent reading. The first J-word reading, relative to the first K-word reading, showed activation in the left angular gyrus. K-word reading relative to J-word reading during both sessions showed activation in occipital regions. Within these activated areas, session by condition interaction was found only in the left angular gyrus. The interaction between session and condition resulted from the fact that the differences in blood oxygenation-level-dependent signals between K-words and J-words and between Kn-words and J-words were significantly greater in the first session than in the second session. From the results, we concluded that patterns of brain activation changed as the memory of the 20 K-words became fixed through daily practice and that reading of both Korean words and Japanese syllabograms engaged the left angular gyrus.

Neural basis of temporal context memory: a functional MRI study.

Temporal context information is crucial to understanding human episodic memory. Human lesion and neuroimaging data indicate that prefrontal regions are important for retrieving temporal context memory, although the exact nature of their involvement is still unclear. We employed functional magnetic resonance imaging (fMRI) to elucidate the neural basis of two kinds of temporal context memory: the temporal order of items between lists and within a list. On the day of the fMRI experiment, subjects memorized a list of 30 pictures in the morning and another list of 30 pictures in the afternoon. During the scanning session, the subjects performed three tasks. In a between-lists task, they were asked to judge the temporal order between two items that had been presented in different lists. In a within-list task, they were asked to judge the temporal order between two items that had been presented in a single list. We found bilateral prefrontal activities during these two temporal context memory tasks compared with a simple item-recognition task. Furthermore, in direct comparison between these two tasks, we found differential prefrontal activities. Thus, right prefrontal activity was associated with temporal order judgment of items between lists, whereas left prefrontal activity was related to temporal order judgment of items within a list. These results indicate that retrieval processes of two kinds of temporal context memory are supported by different, but overlapping, sets of cerebral regions. We speculate that this reflects different cognitive processes for retrieving temporal context memory between separate episodes and within a single episode.

Neural basis of the retrieval of people's names: evidence from brain-damaged patients and fMRI.

The aim of this study was to identify the neuroanatomical basis of the retrieval of people's names. Lesion data showed that patients with language-dominant temporal lobectomy had impairments in their ability to retrieve familiar and newly learned people's names, whereas patients with language-nondominant temporal lobectomy had difficulty retrieving newly learned people's names. Functional magnetic resonance imaging experiments revealed activations in the left temporal polar region during the retrieval of familiar and newly learned people's names, and in the right superior temporal and bilateral prefrontal cortices during the retrieval of newly learned information from face cues. These data provide new evidence that the left anterior temporal region is crucial for the retrieval of people's names irrespective of their familiarity and that the right superior temporal and bilateral prefrontal areas are crucial for the process of associating newly learned people's faces and names.