Predicting muscle invasion in bladder cancer by deep learning analysis of MRI: comparison with vesical imaging-reporting and data system.

OBJECTIVES: To compare the diagnostic performance of a novel deep learning (DL) method based on T2-weighted imaging with the vesical imaging-reporting and data system (VI-RADS) in predicting muscle invasion in bladder cancer (MIBC). METHODS: A total of 215 tumours (129 for training and 31 for internal validation, centre 1; 55 for external validation, centre 2) were included. MIBC was confirmed by pathological examination. VI-RADS scores were provided by two groups of radiologists (readers 1 and readers 2) independently. A deep convolutional neural network was constructed in the training set, and validation was conducted on the internal and external validation sets. ROC analysis was performed to evaluate the performance for MIBC diagnosis. RESULTS: The AUCs of the DL model, readers 1, and readers 2 were as follows: in the internal validation set, 0.963, 0.843, and 0.852, respectively; in the external validation set, 0.861, 0.808, and 0.876, respectively. The accuracy of the DL model in the tumours scored VI-RADS 2 or 3 was higher than that of radiologists in the external validation set: for readers 1, 0.886 vs. 0.600, p = 0.006; for readers 2, 0.879 vs. 0.636, p = 0.021. The average processing time (38 s and 43 s in two validation sets) of the DL method was much shorter than the readers, with a reduction of over 100 s in both validation sets. CONCLUSIONS: Compared to radiologists using VI-RADS, the DL method had a better diagnostic performance, shorter processing time, and robust generalisability, indicating good potential for diagnosing MIBC. KEY POINTS: • The DL model shows robust performance for MIBC diagnosis in both internal and external validation. • The diagnostic performance of the DL model in the tumours scored VI-RADS 2 or 3 is better than that obtained by radiologists using VI-RADS. • The DL method shows potential in the preoperative assessment of MIBC.

Multiple hypointense vessels on susceptibility-weighted imaging predict early neurological deterioration in acute ischaemic stroke patients with severe intracranial large artery stenosis or occlusion receiving intravenous thrombolysis.

BACKGROUND AND PURPOSE: Early neurological deterioration (END) is a common feature in patients with acute ischaemic stroke (AIS) receiving thrombolysis. This study aimed to investigate whether the presence of multiple hypointense vessels (MHVs) on susceptibility-weighted imaging (SWI) could predict END in patients with the anterior circulation AIS treated with recombinant tissue plasminogen activator (r-tPA). METHODS: This was a retrospective study focusing on AIS patients suffering from symptomatic stenosis or occlusion of the middle cerebral artery or internal carotid artery with r-tPA treatment. We collected clinical variables and initial haematological and neuroimaging findings. MHVs were measured on SWI performed after intravenous thrombosis and were defined as the presence of a greater number of veins or veins of a larger diameter with greater signal loss on SWI than those of the contralesional haemisphere. The degree of hyperintensity of MHVs was classified into four grades: none, subtle, moderate and extensive. END was defined as an increase in the National Institutes of Health Stroke Scale score by 2 points during the first 48 hours after the onset of symptoms. Multivariate logistic regressions were conducted to investigate the predictors of END. RESULTS: The study included 61 patients (51 males and 10 females) with a mean age of 62.4±12.6 years. Thirty-five (57.4%) patients presented with MHVs: 8 (13.1%) were graded as subtle MHVs, while 23 (37.7%) and 4 (6.6%) were graded as moderate or extensive MHVs, respectively. Twenty patients (32.8%) presented with END. Logistic regression analysis showed that compared with patients without MHVs, moderate MHVs (adjusted OR 5.446, 95% CI 1.360 to 21.800; p=0.017) and extensive MHVs (adjusted OR 15.240, 95% CI 1.200 to 193.544; p=0.036) were significantly associated with END. CONCLUSIONS: MHVs might be a useful predictor of END in AIS patients with symptomatic large artery stenosis or occlusion after r-tPA treatment.

Asymmetrical cortical vessel sign predicts prognosis after acute ischemic stroke.

INTRODUCTION: To assess whether the asymmetrical cortical vessel sign (ACVS) on susceptibility-weighted imaging (SWI) could predict 90-day poor outcomes in anterior circulation acute ischemic stroke (AIS) patients treated with recombinant tissue plasminogen activator (r-tPA). METHODS: Clinical data of consecutive patients with anterior circulation AIS treated with r-tPA were retrospectively analyzed. Clinical variables included age, sex, vascular risk factors, NIHSS score, onset to treatment time, and initial hematologic and neuroimaging findings. Follow-up was performed 90 days after onset. Poor outcome was defined as a modified Rankin scale (mRS) ≥3 at 90 days. RESULTS: A total of 145 patients were included, 35 (24.1%) patients presented with ACVS (≥Grade 1) on SWI. Fifty-three (36.6%) patients had a poor outcome at 90 days. ACVS (≥Grade 1) occurred in 21 (39.6%) patients with poor outcome compared with 14 (15.2%) patients with favorable outcome (p = .001). Univariate analysis indicated that age, NIHSS score on admission, previous stroke, hemorrhagic transformation, severe intracranial large artery stenosis or occlusion (SILASO), and ACVS were associated with 90-day poor outcome (p < .05). Since SILASO and ACVS were highly correlated and ACVS had different grades, we used three logistic regression models. Results from the three models showed that ACVS was associated with 90-day poor outcome. CONCLUSIONS: In r-tPA-treated patients with anterior circulation AIS, ACVS might be a helpful neuroimaging predictor for poor outcome at 90 days.

High Neutrophil-to-Lymphocyte Ratio Predicts Hemorrhagic Transformation in Acute Ischemic Stroke Patients Treated with Intravenous Thrombolysis.

BACKGROUND: The relationship between the neutrophil-to-lymphocyte ratio (NLR) and hemorrhagic transformation (HT) in acute ischemic stroke (AIS) treated with intravenous thrombolysis (IVT) remains unclear. This study assessed whether high NLR is associated with HT in this population. METHODS: Data were prospectively collected for continuous patients with AIS treated with IVT and retrospectively analyzed. Clinical variables included age, sex, vascular risk factors, National Institutes of Health Stroke Scale (NIHSS) score, onset-to-treatment time, and initial hematologic and neuroimaging findings. HT was confirmed by imaging performed within 3 days after IVT. Symptomatic HT (sHT) was defined as NIHSS score increased by 4 points compared with that on admission according to previously published criteria. The NLR value was based on the blood examination before IVT, and high NLR was defined as ≥75th percentile. RESULTS: The study included 285 patients (201 (70.5%) males, the mean age was 62.3 years (range 29-89)). Seventy-two (25.3%) patients presented with HT, including three (1.1%) with sHT. The median NLR was 2.700 (1.820-4.255, interquartile range). Seventy-one (24.9%) patients had a high NLR (≥4.255) on admission. Univariate analysis indicated that patients with HT had higher NIHSS scores (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (P < 0.001), systolic blood pressure (SBP), platelet counts, lymphocyte counts, and NLR (. CONCLUSIONS: High NLR could be a useful marker for predicting HT in AIS patients after IVT.

Clinical characteristics and magnetic resonance imaging findings in nine patients with nonalcoholic Wernicke's encephalopathy: a retrospective study.

PURPOSE: Wernicke's encephalopathy (WE) is a severe neurological disorder caused by thiamine deficiency. The most common cause of WE is alcoholism. However, there is a significant paucity of information in the existing literature relating to nonalcoholic WE. In this study, we investigated the clinical characteristics and neuroimaging findings of nine patients with nonalcoholic WE. PATIENTS AND METHODS: We retrospectively collated clinical data from nine patients who had been diagnosed with WE in accordance with established criteria including age, gender, risk factors and clinical manifestations. We also collated initial hematological and neuroimaging findings. RESULTS: The mean age of the nine patients was 54.0±17.1 years; four of these patients (44.4%) were male. All nine patients had a history of fasting (range, 5-47 days) prior to WE. Four of the nine patients (44.4%) exhibited the classical triad, and eight (88.9%) showed alterations in mental status. Magnetic resonance imaging (MRI) scans showed that all nine patients had symmetric lesions of the medial thalamus. MRI also revealed other WE-related lesions in mammillary bodies (22.2%), the periaqueductal region (55.6%), the tectal plate of the midbrain (77.8%), cranial nerve nuclei (77.8%) and in the symmetric subcortical white matter (11.1%). CONCLUSION: Our analysis showed that fasting is a common cause of WE in nonalcoholic patients and that MRI is a useful tool for the diagnosis of WE. The most common MRI findings were symmetrical lesions of the medial thalamus lesions, followed by the tectal plate of the midbrain and cranial nerve nuclei.

Effect of heart rate on the diagnostic accuracy of 256-slice computed tomography angiography in the detection of coronary artery stenosis: ROC curve analysis.

The aim of the present study was to investigate the effect of heart rate (HR) on the diagnostic accuracy of 256-slice computed tomography angiography (CTA) in the detection of coronary artery stenosis. Coronary imaging was performed using a Philips 256-slice spiral CT, and receiver operating characteristic (ROC) curve analysis was conducted to evaluate the diagnostic value of 256-slice CTA in coronary artery stenosis. The HR of the research subjects in the study was within a certain range (39-107 bpm). One hundred patients suspected of coronary heart disease underwent 256-slice CTA examination. The cases were divided into three groups: Low HR (HR <75 bpm), moderate HR (75≤ HR <90 bpm) and high HR (HR ≥90 bpm). For the three groups, two observers independently assessed the image quality for all coronary segments on a four-point ordinal scale. An image quality of grades 1-3 was considered diagnostic, while grade 4 was non-diagnostic. A total of 97.76% of the images were diagnostic in the low-HR group, 96.86% in the moderate-HR group and 95.80% in the high-HR group. According to the ROC curve analysis, the specificity of CTA in diagnosing coronary artery stenosis was 98.40, 96.00 and 97.60% in the low-, moderate- and high-HR groups, respectively. In conclusion, 256-slice coronary CTA can be used to clearly show the main segments of the coronary artery and to effectively diagnose coronary artery stenosis. Within the range of HRs investigated, HR was found to have no significant effect on the diagnostic accuracy of 256-slice coronary CTA for coronary artery stenosis.