Comparison of machine learning-based CT fractional flow reserve with cardiac MR perfusion mapping for ischemia diagnosis in stable coronary artery disease

OBJECTIVES

To compare the diagnostic performance of machine learning (ML)-based computed tomography-derived fractional flow reserve (CT-FFR) and cardiac magnetic resonance (MR) perfusion mapping for functional assessment of coronary stenosis.

METHODS

Between October 2020 and March 2022, consecutive participants with stable coronary artery disease (CAD) were prospectively enrolled and underwent coronary CTA, cardiac MR, and invasive fractional flow reserve (FFR) within 2 weeks. Cardiac MR perfusion analysis was quantified by stress myocardial blood flow (MBF) and myocardial perfusion reserve (MPR). Hemodynamically significant stenosis was defined as FFR ≤ 0.8 or > 90% stenosis on invasive coronary angiography (ICA). The diagnostic performance of CT-FFR, MBF, and MPR was compared, using invasive FFR as a reference.

RESULTS

The study protocol was completed in 110 participants (mean age, 62 years ± 8; 73 men), and hemodynamically significant stenosis was detected in 36 (33%). Among the quantitative perfusion indices, MPR had the largest area under receiver operating characteristic curve (AUC) (0.90) for identifying hemodynamically significant stenosis, which is in comparison with ML-based CT-FFR on the vessel level (AUC 0.89, p = 0.71), with comparable sensitivity (89% vs 79%, p = 0.20), specificity (87% vs 84%, p = 0.48), and accuracy (88% vs 83%, p = 0.24). However, MPR outperformed ML-based CT-FFR on the patient level (AUC 0.96 vs 0.86, p = 0.03), with improved specificity (95% vs 82%, p = 0.01) and accuracy (95% vs 81%, p < 0.01).

CONCLUSION

ML-based CT-FFR and quantitative cardiac MR showed comparable diagnostic performance in detecting vessel-specific hemodynamically significant stenosis, whereas quantitative perfusion mapping had a favorable performance in per-patient analysis.

CLINICAL RELEVANCE STATEMENT

ML-based CT-FFR and MPR derived from cardiac MR performed well in diagnosing vessel-specific hemodynamically significant stenosis, both of which showed no statistical discrepancy with each other.

KEY POINTS

• Both machine learning (ML)-based computed tomography-derived fractional flow reserve (CT-FFR) and quantitative perfusion cardiac MR performed well in the detection of hemodynamically significant stenosis. • Compared with stress myocardial blood flow (MBF) from quantitative perfusion cardiac MR, myocardial perfusion reserve (MPR) provided higher diagnostic performance for detecting hemodynamically significant coronary artery stenosis. • ML-based CT-FFR and MPR from quantitative cardiac MR perfusion yielded similar diagnostic performance in assessing vessel-specific hemodynamically significant stenosis, whereas MPR had a favorable performance in per-patient analysis.

Glymphatic system in the thalamus, secondary degeneration area was severely impaired at 2nd week after transient occlusion of the middle cerebral artery in rats

Background and objectives

The glymphatic system is a recently discovered cerebrospinal fluid transport system and little is known about its dynamic changes after stroke. This study aimed to dynamically observe the structural and functional changes of the impaired glymphatic system in the thalamus after ischemic stroke by pathology and MRI.

Materials and methods

Ischemic stroke was induced by the middle cerebral artery occlusion (MCAO) model. A total of 20 Sprague-Dawley rats were randomly assigned into four groups: sham, MCAO 1 week, MCAO 2 week, and MCAO 2 month. All rats successively underwent neurological examination, dynamic contrast-enhanced MRI (DCE-MRI), and immunofluorescence staining. Immunofluorescence staining of glial fibrillary acidic protein (GFAP), aquaporin-4 (AQP4), ionized calcium-binding adaptor molecule 1 (Iba1), and beta-amyloid precursor protein (APP) were done in thalamus ventroposterior nucleus.

Results

The astrocyte and microglial activation and the APP deposition in the MCAO 2 week group were the highest (P < 0.05 for all). The AQP4 polarization rates of the MCAO 2 week and 2 month groups were the lowest (P < 0.05 for all). Although there was no correlation between histological changes and MRI metrics in all four groups (P > 0.05 for all), the tendency of the APP deposition was nearly consistent with the one of the contrast agent retention in DCE-MRI.

Conclusion

The glymphatic system in the thalamus was severely impaired at 2nd week after MCAO, and may be revealed by DCE-MRI. This study may provide a relevant theoretical basis for making a thorough inquiry of the mechanism of brain injury after stroke and clinical treatment of ischemic stroke and help readers appreciate the importance of DCE-MRI.

Establishment and validation of a radiological-radiomics model for predicting high-grade patterns of lung adenocarcinoma less than or equal to 3 cm

Objective

We aimed to develop a Radiological-Radiomics (R-R) based model for predicting the high-grade pattern (HGP) of lung adenocarcinoma and evaluate its predictive performance.

Methods

The clinical, pathological, and imaging data of 374 patients pathologically confirmed with lung adenocarcinoma (374 lesions in total) were retrospectively analyzed. The 374 lesions were assigned to HGP (n = 81) and non-high-grade pattern (n-HGP, n = 293) groups depending on the presence or absence of high-grade components in pathological findings. The least absolute shrinkage and selection operator (LASSO) method was utilized to screen features on the United Imaging artificial intelligence scientific research platform, and logistic regression models for predicting HGP were constructed, namely, Radiological model, Radiomics model, and R-R model. Also, receiver operating curve (ROC) curves were plotted on the platform, generating corresponding area under the curve (AUC), sensitivity, specificity, and accuracy. Using the platform, nomograms for R-R models were also provided, and calibration curves and decision curves were drawn to evaluate the performance and clinical utility of the model. The statistical differences in the performance of the models were compared by the DeLong test.

Results

The R-R model for HGP prediction achieved an AUC value of 0.923 (95% CI: 0.891-0.948), a sensitivity of 87.0%, a specificity of 83.4%, and an accuracy of 84.2% in the training set. In the validation set, this model exhibited an AUC value of 0.920 (95% CI: 0.887-0.945), a sensitivity of 87.5%, a specificity of 83.3%, and an accuracy of 84.2%. The DeLong test demonstrated optimal performance of the R-R model among the three models, and decision curves validated the clinical utility of the R-R model.

Conclusion

In this study, we developed a fusion model using radiomic features combined with radiological features to predict the high-grade pattern of lung adenocarcinoma, and this model shows excellent diagnostic performance. The R-R model can provide certain guidance for clinical diagnosis and surgical treatment plans, contributing to improving the prognosis of patients.

Correlation of computed tomography quantitative parameters with tumor invasion and Ki-67 expression in early lung adenocarcinoma

The purpose of the study is to investigate the correlation of computed tomography (CT) quantitative parameters with tumor invasion and Ki-67 expression in early lung adenocarcinoma.The study involved 141 lesions in 141 patients with early lung adenocarcinoma. According to the degree of tumor invasion, the lesions were assigned into (adenocarcinoma in situ + minimally invasive adenocarcinoma) group and invasive adenocarcinoma (IAC) group. Artificial intelligence-assisted diagnostic software was used to automatically outline the lesions and extract corresponding quantitative parameters on CT images. Statistical analysis was performed to explore the correlation of these parameters with tumor invasion and Ki-67 expression.The results of logistic regression analysis showed that the short diameter of the lesion and the average CT value were independent predictors of IAC. Receiver operating characteristic curve analysis identified the average CT value as an independent predictor of IAC with the best performance, with the area under the receiver operating characteristic curve of 0.893 (P < .001), and the threshold of -450 HU. Besides, the predicted probability of logistic regression analysis model was detected to have the area under the curve of 0.931 (P < .001). The results of Spearman correlation analysis showed that the expression level of Ki-67 had the highest correlation with the average CT value of the lesion (r = 0.403, P < .001).The short diameter of the lesion and the average CT value are independent predictors of IAC, and the average CT value is significantly positively correlated with the expression of tumor Ki-67.

Acutely Inhibiting AQP4 With TGN-020 Improves Functional Outcome by Attenuating Edema and Peri-Infarct Astrogliosis After Cerebral Ischemia

Background

Ischemic stroke is one of the leading causes of human death and disability. Brain edema and peri-infarct astrocyte reactivity are crucial pathological changes, both involving aquaporin-4 (AQP4). Studies revealed that acute inhibition of AQP4 after stroke diminishes brain edema, however, its effect on peri-infarct astrocyte reactivity and the subacute outcome is unclear. And if diffusion-weighted imaging (DWI) could reflect the AQP4 expression patterns is uncertain.

Methods

Rats were subjected to middle cerebral artery occlusion (MCAO) and allocated randomly to TGN 020-treated and control groups. One day after stroke, brain swelling and lesion volumes of the rats were checked using T2-weighted imaging (T2-WI). Fourteen days after stroke, the rats successively underwent neurological examination, T2-WI and DWI with standard b-values and ultra-high b-values, apparent diffusion coefficient (ADC) was calculated correspondingly. Finally, the rats’ brains were acquired and used for glial fibrillary acidic protein (GFAP) and AQP4 immunoreactive analysis.

Results

At 1 day after stroke, the TGN-020-treated animals exhibited reduced brain swelling and lesion volumes compared with those in the control group. At 14 days after stroke, the TGN-020-treated animals showed fewer neurological function deficits and smaller lesion volumes. In the peri-infarct region, the control group showed evident astrogliosis and AQP4 depolarization, which were reduced significantly in the TGN-020 group. In addition, the ultra-high b-values of ADC (ADCuh) in the peri-infarct region of the TGN-020 group was higher than that of the control group. Furthermore, correlation analysis revealed that peri-infarct AQP4 polarization correlated negatively with astrogliosis extent, and ADCuh correlated positively with AQP4 polarization.

Conclusion

We found that acutely inhibiting AQP4 using TGN-020 promoted neurological recovery by diminishing brain edema at the early stage and attenuating peri-infarct astrogliosis and AQP4 depolarization at the subacute stage after stroke. Moreover, ADCuh could reflect the AQP4 polarization.

Prediction of future imagery of lung nodule as growth modeling with follow-up computed tomography scans using deep learning: a retrospective cohort study

Background

Risk prediction models of lung nodules have been built to alleviate the heavy interpretative burden on clinicians. However, the malignancy scores output by those models can be difficult to interpret in a clinically meaningful manner. In contrast, the modeling of lung nodule growth may be more readily useful. This study developed a CT-based visual forecasting system that can visualize and quantify a nodule in three dimensions (3D) in any future time point using follow-up CT scans.

Methods

We retrospectively included 246 patients with 313 lung nodules with at least 1 follow-up CT scan. For the manually segmented nodules, we calculated geometric properties including CT value, diameter, volume, and mass, as well as growth properties including volume doubling time (VDT), and consolidation-to-tumor ratio (CTR) at follow-ups. These nodules were divided into growth and non-growth groups by thresholding their VDTs. We then developed a convolutional neural network (CNN) to model the imagery change of the nodules from baseline CT image (combined with the nodule mask) to follow-up CT image with a particular time interval. The model was evaluated on the geometric and radiological properties using either logistic regression or receiver operating characteristic (ROC) curve.

Results

The lung nodules consisted of 115 ground glass nodules (GGN) and 198 solid nodules and were followed up for an average of 354 days with 2 to 11 scans. The 2 groups differed significantly in most properties. The prediction of our forecasting system was highly correlated with the ground truth with small relative errors regarding the four geometric properties. The prediction-derived VDTs had an area under the curve (AUC) of 0.857 and 0.843 in differentiating growth and non-growth nodules for GGN and solid nodules, respectively. The prediction-derived CTRs had an AUC of 0.892 in classifying high- and low-risk nodules.

Conclusions

This proof-of-concept study demonstrated that the deep learning-based model can accurately forecast the imagery of a nodule in a given future for both GGNs and solid nodules and is worthy of further investigation. With a larger dataset and more validation, such a system has the potential to become a prognostication tool for assessing lung nodules.

An In Vivo Data-Based Computational Study on Sitting-Induced Hemodynamic Changes in the External Iliac Artery

Although sedentary behavior (characterized by prolonged sitting without otherwise being active in daily life) is widely regarded as a risk factor for peripheral artery disease (PAD), underlying biomechanical mechanisms remain insufficiently understood. In this study, geometrical models of ten external iliac arteries were reconstructed based on angiographic data acquired from five healthy young subjects resting in supine and sitting (mimicked by side lying with bent legs) positions, respectively, which were further combined with measured blood flow velocity waveforms in the common iliac arteries (with each body posture being maintained for 30 min) to build computational models for simulating intra-arterial hemodynamics. Morphological analyses showed that the external iliac arteries suffered from evident bending deformation upon the switch of body posture from supine to sitting. Measured blood flow velocity waveforms in the sitting position exhibited a marked decrease in mean flow velocity while increase in retrograde flow ratio compared with those in the supine position. Hemodynamic computations further revealed that sitting significantly altered blood flow patterns in the external iliac arteries, leading to a marked enlargement of atheroprone wall regions exposed to low and oscillatory wall shear stress (WSS), and enhanced multidirectional disturbance of WSS that may further impair endothelial function. In summary, our study demonstrates that prolonged sitting induces atheropromoting hemodynamic changes in the external iliac artery due to the combined effects of vascular bending deformation and changes in flow velocity waveform, which may provide important insights for understanding the involvement of biomechanical factors in sedentary behavior-related PAD.

Dependence of radiomic features on pixel size affects the diagnostic performance of radiomic signature for the invasiveness of pulmonary ground-glass nodule

OBJECTIVE

To investigate the effect of reducing pixel size on the consistency of radiomic features and the diagnostic performance of the downstream radiomic signatures for the invasiveness for pulmonary ground-glass nodules (GGNs) on CTs.

METHODS

We retrospectively collected the clinical data of 182 patients with GGNs on high resolution CT (HRCT). The CT images of different pixel sizes (0.8mm, 0.4mm, 0.18 mm) were obtained by reconstructing the single HRCT scan using three combinations of field of view and matrix size. For each pixel size setting, radiomic features were extracted for all GGNs and radiomic signatures for the invasiveness of GGNs were built through two modeling pipelines for comparison.

RESULTS

The study finally extracted 788 radiomic features. 87% radiomic features demonstrated inter pixel size variation. By either modeling pipeline, the radiomic signature under small pixel size performed significantly better than those under middle or large pixel sizes in predicting the invasiveness of GGNs (p's value <0.05 by Delong test). With the independent modeling pipeline, the three pixel size bounded radiomic signatures shared almost no common features.

CONCLUSIONS

Reducing pixel size could cause inconsistency in most radiomic features and improve the diagnostic performance of the downstream radiomic signatures. Particularly, super HRCTs with small pixel size resulted in more accurate radiomic signatures for the invasiveness of GGNs.

ADVANCES IN KNOWLEDGE

The dependence of radiomic features on pixel size will affect the performance of the downstream radiomic signatures. The future radiomic studies should consider this effect of pixel size.

Impaired glymphatic system in secondary degeneration areas after ischemic stroke in rats

BACKGROUND

Pathomechanism of secondary degeneration in remote regions after ischemic stroke has not been totally clarified. Contrast-enhanced MRI with injecting Gd-DTPA in cisterna magna (CM) is regarded as an efficient method to measure glymphatic system function in brain. Our research aimed at evaluating glymphatic system changes in secondary degeneration areas by contrast-enhanced MRI.

METHODS

Ischemic stroke was induced by left middle cerebral artery occlusion (MCAO) model. A total of 12 Sprague-Dawley rats were randomly divided into three groups: control group with sham operations (n=4), the group of acute phase (1 day after MCAO) (n=4), and the group of subacute phase (7 days after MCAO) (n=4). Contrast-enhanced MRI was performed in 1days or 7days after operations respectively. All rats received an intrathecal injection of Gd-DTPA (2μl/min, totally 20μl) and high-resolution 3D T1-weighted MRI for 6 h. The time course of the signal-to-noise ratio (SNR) in substantia Nigra (SN) and ventral thalamic nucleus (VTN) was evaluated between two hemispheres in all rats.

RESULTS

In control group without ischemia, time-to-peak of SNR in SN was earlier than that in VTN. There were no differences of SNR between two hemispheres after intrathecal Gd-DTPA administration. In the group of acute phase, MRI revealed similar time course and time-to-peak of SNR between ipsilateral and contralateral VTN, while a tendency of higher SNR in ipsilateral SN than contralateral SN at 4h, 5h, 6h after Gd-DTPA injection. And time-to-peak of SNR was similar in bilateral SN. In the group of subacute phase, time-to-peak of SNR was similar in bilateral VTN, while longer in ipsilateral SN compared with contralateral side. In addition, SNR in T1WI in ipsilateral was significantly higher than SNR in contralateral SN and VTN at 5h (VTN, P= 0.003; SN, P=0.004) and 6h (VTN, P=0.015; SN, P=0.006) after Gd-DTPA injection.

CONCLUSION

Glymphatic system was impaired in ipsilateral SN and VTN after ischemic stroke, which may contribute to neural degeneration.

Non-periodicity of blood flow and its influence on wall shear stress in the carotid artery bifurcation: An in vivo measurement-based computational study

Although arterial blood flow is physiologically non-periodic under resting conditions, periodic flow assumption has been widely adopted in most hemodynamic studies. So far, it remains unclear how the non-periodicity of blood flow would influence local hemodynamic parameters, especially wall shear stress (WSS) that associates closely with endothelial function and vascular disease. In this study, numerical simulations of blood flows in sixteen normal carotid artery bifurcations were performed under measured non-periodic and averaged periodic flow conditions, respectively, with the obtained results being compared in terms of five typical WSS metrics (i.e., mean WSS (MWSS), time-averaged WSS (TAWSS), oscillatory shear index (OSI), transverse WSS (transWSS), and average temporal gradient of WSS (WSSTG)) in the atheroprone low-WSS regions. It was found that simplifying the physiologically non-periodic flow condition into a periodic one did not significantly alter the major features of WSS distribution, but resulted in underestimations of some WSS metrics. Specifically, the degree of underestimation was largest (27.2% ± 8.3%) for WSSTG, smallest (0.5% ± 0.4%) for MWSS, while moderate (5.1% ± 3.2% ~ 9.2% ± 4.1%) for other WSS metrics. Statistical analyses revealed that the cycle-to-cycle variability of flow velocity waveform (var-V) and the planarity of internal carotid artery correlated strongly with the periodic flow assumption-induced underestimations of WSS metrics. These findings suggest that taking the non-periodic characteristic of blood flow into consideration could be important for studying hemodynamics in arteries with a large var-V or specific morphological characteristics, especially when WSSTG is the main hemodynamic parameter of concern.

Combining systemic and stereotactic MEMRI to detect the correlation between gliosis and neuronal connective pathway at the chronic stage after stroke

BACKGROUND

The early dysfunction and subsequent recovery after stroke, characterized by the destruction and remodeling of connective pathways between cortex and subcortical regions, is associated with neuroinflammation. As major components of the inflammatory process, reactive astrocytes have double-edged effects on pathological progression. The temporal patterns of astrocyte and neuronal pathway activity can be revealed by systemic and stereotactic manganese-enhanced magnetic resonance imaging (MEMRI), respectively. In the present study, we aimed to detect an association between astrocyte activity and recovery of neuronal connective pathways by combining systemic with stereotactic MEMRI.

METHODS

Fifty adult rats, divided into two groups, underwent a 60-min occlusion of the middle cerebral artery. The groups were given either a systemic administration or stereotactic injection of MnCl2 at 1, 3, 7, and 14 days after stroke and underwent MRI 4 and 2 days later, respectively. Immunofluorescence (IF) of group 1 was conducted to corroborate the results. Repetitive behavioral testing was also performed with all rats at 1, 3, 7, and 14 days to obtain a functional score.

RESULTS

Ring- or crescent-shaped enhancements formed in the striatal peri-infarct regions (STR) at 11 and 18 days. This was concurrent with the activity of glial fibrillary acidic protein (GFAP)-positive astrocytes, which mainly localized at the peri-infarct region and significantly increased in number at 11 and 18 days after stroke. Microglia/macrophages, detected by IF, mainly localized in the lesion core, rather than in the region of enhancement. The ipsilateral substantia nigra (SN) revealed Mn-related signal enhancement reduction and subsequent signs of the recovery process at 3 to 5 days and 9 to 16 days, respectively. Behavioral testing showed that sensorimotor functions were initially disturbed, but subsequently recovered at 7 and 14 days.

CONCLUSIONS

We found a positive temporal correlation between astrogliosis and the recovery of neuronal connective pathways at the chronic stage by using the in vivo method of MEMRI. Our results highlighted the potential contribution of astrocytes to the neuronal recovery of these connective pathways.