Artificial intelligence software for diagnosing intracranial arterial occlusion in patients with acute ischemic stroke.

PURPOSE: To evaluate the diagnostic performance of AI software in diagnosing intracranial arterial occlusions in the proximal anterior circulation at CT angiography (CTA) and to compare it to manual reading performed in clinical practice. METHODS: Patients with acute ischemic stroke underwent CTA to detect arterial occlusion in the proximal anterior circulation. Retrospective review of CTA scans by two neuroradiologists served as reference standard. Sensitivity and specificity of AI software (StrokeViewer) were compared to those of manual reading using the McNemar test. The proportions of correctly detected occlusions in the distal internal carotid artery and/or M1 segment of the middle cerebral artery (large vessel occlusion [LVO]) and in the M2 segment of the middle cerebral artery (medium vessel occlusion [MeVO]) were calculated. RESULTS: Of the 474 patients, 75 (15.8%) had an arterial occlusion in the proximal anterior circulation according to the reference standard. Sensitivity of StrokeViewer software was not significantly different compared to that of manual reading (77.3% vs. 78.7%, P = 1.000). Specificity of StrokeViewer software was significantly lower than that of manual reading (88.5% vs. 100%, P < 0.001). StrokeViewer software correctly identified 40 of 42 LVOs (95.2%) and 18 of 33 MeVOs (54.5%). StrokeViewer software detected 8 of 16 (50%) intracranial arterial occlusions which were missed by manual reading. CONCLUSION: The current AI software detected intracranial arterial occlusion with moderate sensitivity and fairly high specificity. The AI software may detect additional occlusions which are missed by manual reading. As such, the use of AI software may be of value in clinical stroke care.

Experimental Multicenter and Multivendor Evaluation of the Performance of PET Radiomic Features Using 3-Dimensionally Printed Phantom Inserts.

The sensitivity of radiomic features to several confounding factors, such as reconstruction settings, makes clinical use challenging. To investigate the impact of harmonized image reconstructions on feature consistency, a multicenter phantom study was performed using 3-dimensionally printed phantom inserts reflecting realistic tumor shapes and heterogeneity uptakes. Methods: Tumors extracted from real PET/CT scans of patients with non-small cell lung cancer served as model for three 3-dimensionally printed inserts. Different heterogeneity pattern were realized by printing separate compartments that could be filled with different activity solutions. The inserts were placed in the National Electrical Manufacturers Association image-quality phantom and scanned various times. First, a list-mode scan was acquired and 5 statistically equal replicates were reconstructed. Second, the phantom was scanned 4 times on the same scanner. Third, the phantom was scanned on 6 PET/CT systems. All images were reconstructed using EANM Research Ltd. (EARL)-compliant and locally clinically preferred reconstructions. EARL-compliant reconstructions were performed without (EARL1) or with (EARL2) point-spread function. Images were analyzed with and without resampling to 2-mm cubic voxels. Images were discretized with a fixed bin width (FBW) of 0.25 and a fixed bin number (FBN) of 64. The intraclass correlation coefficient (ICC) of each scan setup was calculated and compared across reconstruction settings. An ICC above 0.75 was regarded as high. Results: The percentage of features yielding a high ICC was largest for the statistically equal replicates (70%-91% for FBN; 90%-96% for FBW discretization). For scans acquired on the same system, the percentage decreased, but most features still resulted in a high ICC (FBN, 52%-63%; FBW, 75%-85%). The percentage of features yielding a high ICC decreased more in the multicenter setting. In this case, the percentage of features yielding a high ICC was larger for images reconstructed with EARL-compliant reconstructions: for example, 40% for EARL1 and 60% for EARL2 versus 21% for the clinically preferred setting for FBW discretization. When discretized with FBW and resampled to isotropic voxels, this benefit was more pronounced. Conclusion: EARL-compliant reconstructions harmonize a wide range of radiomic features. FBW discretization and a sampling to isotropic voxels enhances the benefits of EARL-compliant reconstructions.

Interobserver reproducibility of diffusion-weighted MRI in monitoring tumor response to neoadjuvant therapy in esophageal cancer.

OBJECTIVE: To investigate the reproducibility of diffusion-weighted magnetic resonance imaging (DW-MRI) in assessing tumor response early in the course of neoadjuvant chemoradiotherapy in patients with operable esophageal cancer. METHODS: Eleven male patients (mean age 54.8 years) with newly diagnosed esophageal cancer underwent DW-MRI before and 10 days after start of chemoradiotherapy. Reproducibility of apparent diffusion coefficient (ADC) measurements by manual (freehand) and semi-automated volumetric methods was assessed. RESULTS: Interobserver reproducibility for the assessment of mean tumor ADC by the manual measurement method was good, with an ICC of 0.69 (95% CI, 0.36 to 0.85; P = 0.001). Interobserver reproducibility for the assessment of mean tumor ADC by the semi-automated volumetric measurement method was very good, with an ICC of 0.96 (95% CI, 0.91 to 0.98; P<0.001). CONCLUSION: Semi-automated volumetric ADC measurements have higher reproducibility than manual ADC measurements in assessing tumor response to chemoradiotherapy in patients with esophageal adenocarcinoma.

Cardiovascular autonomic regulation in preterm infants: the effect of atropine.

To study cardiovascular autonomic control, we assessed the effect of atropine on heart rate (HR) and blood pressure (BP) variability in 12 preterm infants (range 26-32 wk) before intubation for respiratory insufficiency. Spectral power analysis of R-R interval and systolic BP (SBP) series were estimated in a low-frequency (LF; 0.04-0.15 Hz) and high-frequency (HF; 0.4-1.5 Hz) band and evaluated for a 10-min period before and a 10-min period after atropine sulfate (0.01 mg/kg). Baroreceptor reflex (BR) functioning was estimated using transfer function analysis at LF (coherence, gain, and phase). Atropine resulted in a significant 12% increase in steady-state HR (p < 0.01) and unchanged SBP. For R-R interval series, the total spectral power decreased 6-fold (p < 0.01), which was predominantly due to a reduction in the LF band (16-fold; p < 0.01). In contrast, we observed a significant increase (25%; p < 0.05) in total spectral power of SBP series partly as a result of an increase in HF power. The LF power of SBP series was not altered. The median LF transfer gain (BR sensitivity) between SBP and R-R interval decreased from 4.2 to 1.4 ms/mm Hg (p < 0.01) after atropine. The LF phase relationship (BP leads R-R interval fluctuations by approximately 4 s) was not changed after atropine. In conclusion, even in preterm infants in distress, atropine modulates HR and BP variability, suggesting that BR-mediated parasympathetic control of heart rate is of significance for cardiovascular control at that age.

Beat-to-beat detection of fetal heart rate: Doppler ultrasound cardiotocography compared to direct ECG cardiotocography in time and frequency domain.

In order to obtain power spectral information on the fetal heart rate in stages of pregnancy earlier than labor an algorithm has been developed to calculate the fetal heart rate on a beat-to-beat basis from Doppler ultrasound cardiotocographic signals. The algorithm was evaluated by comparing the calculated fetal heart rate with the heart rate determined from direct ECG signals measured with a scalp electrode. Heart rates were compared both in time and frequency domain. In the time domain the results achieved by both methods correlate well (correlation coefficient = 0.977 (p < 0.001)), in the frequency domain the results correlate even better (correlation coefficient = 0.991 (p < 0.001)). Based on these findings, it can be concluded that the developed algorithm provides a valuable tool for obtaining power spectral information on the fetal heart rate in stages of pregnancy earlier than labor.

Noninvasive assessment of blood pressure variability in preterm infants.

The feasibility of measuring blood pressure (BP) variability by a noninvasive beat-to-beat finger arterial BP device (Finapres) was assessed in preterm infants. By application of the finger cuff around the infant's wrist, time and frequency domain (spectral power) analysis of noninvasive beat-to-beat BP signals were compared with intra-arterial measurements. A fast Fourier Transform was used to compute the spectral power density from 128-s periods. The low-frequency band (LF; 0.04-0.15 Hz) is partly associated with baroreflex activity. The high-frequency band (HF; 0.4-1.5 Hz) is associated with respiratory activity. In eight subjects above 1000 g, reliable signals could be obtained. We observed a high correlation between noninvasive and intra-arterial beat-to-beat systolic BP values (mean r value +/- SD, 0.87 +/- 0.11), with a gain close to 1 (mean gain +/- SD, 1.0 +/- 0.4 mm Hg/mm Hg). Finapres estimated beat-to-beat systolic BP changes more accurately than diastolic values. We found a very high amount of linear coupling, expressed as coherence function, between the power spectra of noninvasive and intra-arterial systolic BP measurements. For systolic BP, the (pooled) group mean +/- SEM coherence values were 0.93 +/- 0.00 and 0.91 +/- 0.01 for LF and HF fluctuations, respectively (NS). The wrist method of Finapres in neonates has limited value in estimating absolute BP but is useful in a clinical research situation, where identification of beat-to-beat changes in systolic BP is more important. Finapres provides a noninvasive tool for investigating autonomic cardiovascular regulation (baroreflex sensitivity, spectral analysis of BP fluctuations) in neonates.

Cardiovascular fluctuations and transfer function analysis in stable preterm infants.

To examine the baroreceptor reflex function, a beat-to-beat analysis between systolic blood pressure (SBP) and R-R interval fluctuations was studied in 10 stable appropriate-for-gestational age preterm infants (range, 27.2-33.7 wk) in the first postnatal week during quiet sleep. Spectral power analysis, using fast Fourier transform, and transfer functions (gain and phase difference) between SBP and R-R fluctuations were estimated in a low-frequency band (LF, 0.03-0.2 Hz) and high-frequency band (HF defined as the frequency band between the 10th and 90th centiles of the individual respiratory frequency). The LF/HF ratio reflects the sympathovagal balance. The mean frequency (+/-SD) of LF peaks was centered at 0.07 +/- 0.02 Hz. The mean frequency (+/-SD) of the individual HF band was 0.82 +/- 0.21 Hz. The LF/HF ratio in the R-R interval series [median, 29; interquartile range (IQR), 16-40] was higher than in the SBP series (median, 8; IQR, 4-14). The gain between R-R interval and SBP fluctuations (median, 4.2 ms/mm Hg; IQR, 2.4-5.0) in the LF band was higher than in the HF band (median, 1.7 ms/mm Hg; IQR, 1.4-3.0). SBP fluctuations lead R-R interval fluctuations in the LF band with a median phase difference of +96 degrees (IQR, 67-132). At LF the fluctuations in SBP precede changes in R-R interval with a time delay of 3.8 s. These observations indicate a dominant role of the sympathetic system in stable preterm infants in comparison with published adult values. Cross-spectral analysis allows a test for tracking the development of the sympathetic system in neonates.