Clinical Application of Automatic Assessment of Scoliosis Cobb Angle Based on Deep Learning.

INTRODUCTION: A recently developed deep-learning-based automatic evaluation model provides reliable and efficient Cobb angle measurements for scoliosis diagnosis. However, few studies have explored its clinical application, and external validation is lacking. Therefore, this study aimed to explore the value of automated assessment models in clinical practice by comparing deep-learning models with manual measurement methods. METHODS: The 481 spine radiographs from an open-source dataset were divided into training and validation sets, and 119 spine radiographs from a private dataset were used as the test set. The mean Cobb angle values assessed by three physicians in the hospital's PACS system served as the reference standard. The results of Seg4Reg, VFLDN, and manual measurement were statistically analyzed. The intra-class correlation coefficients (ICC) and the Pearson correlation coefficient (PCC) were used to compare their reliability and correlation. The Bland-Altman method was used to compare their agreement. The Kappa statistic was used to compare the consistency of Cobb angles at different severity levels. RESULTS: The mean Cobb angle values measured were 35.89° ± 9.33° with Seg4Reg, 31.54° ± 9.78° with VFLDN, and 32.23° ± 9.28° with manual measurement. The ICCs for the reliability of Seg4Reg and VFLDN were 0.809 and 0.974, respectively. The PCC and MAD between Seg4Reg and manual measurements were 0.731 (p<0.001) and 6.51°, while those between VFLDN and manual measurements were 0.952 (p<0.001) and 2.36°. The Kappa statistic indicated VFLDN (k= 0.686, p< 0.001) was superior to Seg4Reg and manual measurements for Cobb angle severity classification. CONCLUSION: The deep-learning-based automatic scoliosis Cobb angle assessment model is feasible in clinical practice. Specifically, the keypoint-based VFLDN is more valuable in actual clinical work with higher accuracy, transparency, and interpretability.

VLTENet: A Deep-Learning-Based Vertebra Localization and Tilt Estimation Network for Automatic Cobb Angle Estimation.

Scoliosis diagnosis and assessment rely upon Cobb angle estimation from X-ray images of the spine. Recently, automated scoliosis assessment has been greatly improved using deep learning methods. However, in such methods, the Cobb angle is usually predicted based on regression models that don't account for information of the spine structure. Alternatively, the Cobb angle can be estimated indirectly through landmark-detection and vertebra-segmentation, but this approach is still highly sensitive to small detection and segmentation errors. This paper proposes a novel deep-learning architecture, called the vertebra localization and tilt estimation network (VLTENet). This network boosts the Cobb angle estimation accuracy through employing vertebra localization and tilt estimation as network prediction goals. In particular, the VLTENet model innovatively combines a deep high-resolution network (HRNet) and a fully-convolutional U-Net architecture for capturing long-range contextual information, the overall structure, and local details in spinal X-ray images. A feature fusion channel attention (FFCA) module is also proposed to selectively emphasize more informative features and suppress less informative ones. In addition, a joint spine loss function (JS-Loss) is designed to account for the spine shape and other spatial constraints, so that the network focuses more on spine-related regions and ignore irrelevant background regions. Finally, we propose a new Cobb angle estimation method conforms with the clinical Cobb angle calculation guidelines, and produces accurate estimates for different types of scoliosis. Extensive experiments on the publically-available AASCE challenge dataset and on an in-house dataset demonstrated the superiority of our method for the task of automatic assessment of scoliosis.

A model of transplanting neonatal cardiomycytes into the wall of rat inferior vena cava / 中国胸心血管外科临床杂志

@#Objective     To establish a model of transplanting neonatal cardiomycytes into the wall of rat inferior vena cava. Methods     Neonatal cardiomyocytes (n=6, 5×106cells each, A group) or medium (n=6, B group) only were transplanted into the wall of inferior vena cava in female Fisher rats. At 21 days after transplantation, the contraction of transplanted cardiomyocytes was assessed and the inferior vena cava was processed for histology. Results     Distinct rhythmic beating of the vena cava at the site of cell transplantation before and after the aorties were clamped (at a rate 141± 47 rpm and 88± 44 rpm which was dramaticly lower than aortic beating, with a statistical difference at P value of 0.03). Cardiomyocyte was seen in 6 rats who had neonatal cardiomyocyte transplantation, but not in 6 rats receiving media. Hematoxylin and eosin staining showed viable cardiomyocytes in the wall of the vena cava in 6 rats treated with neonatal cardiomyocytes, but not in 6 rats receiving media. Conclusion     This study shows that neonatal cardiomyocytes can survive, mature and spontaneously and rhythmically contract after they are transplanted in the wall of inferior vena cava.

Microvascular damage after coronary artery bypass surgery: assessment using dobutamine stress myocardial contrast echocardiography.

BACKGROUND: Although dobutamine stress myocardial contrast echocardiography (DSMCE) has been widely used for the prediction of myocardial functional recovery, dynamic changes that occur at the microcirculatory level during stress have been studied limitedly. The objective of the present study was to use low-dose DSMCE to assess microvascular damage and predict myocardial functional recovery in coronary artery disease (CAD) patients receiving coronary artery bypass grafting. METHODS: Forty-six CAD patients were subjected to low-dose DSMCE, as well as echocardiography and coronary computed tomography angiography before revascularization, 1 year after coronary artery bypass grafting. Dynamic changes occurring at the microcirculatory level during stress were analyzed for the ability to predict functional recovery. Quantitative assessment of functional recovery was determined using myocardial blood flow (MBF) via receiver operating characteristic curve analyses. RESULTS: Patients who failed to recover had fewer changes in MBF (ΔMBF) at rest and with stress compared with the segments showing functional recovery. Semiquantitative changes (enhanced or reduced) of the myocardial perfusion score (ΔMPS) and quantitative changes in ΔMBF of stress myocardial contrast echocardiography enhanced the specificity of resting MPS and the sensitivity of wall motion scores (P < 0.05) for the prediction of functional recovery. CONCLUSIONS: Specific stress ΔMBF more accurately reflected the extent of microvascular damage compared with wall motion scores and resting MPS. ΔMBF and ΔMPS under stress myocardial contrast echocardiography provided higher accuracy than wall motion scores and resting MPS in predicting functional recovery in CAD patients after revascularization.

Effect of graft patency on the prediction of myocardial viability by dobutamine stress and myocardial contrast echocardiography before coronary artery bypass surgery.

BACKGROUND: Myocardial functional recovery after revascularization is considered the "gold standard" for myocardial viability (MV) assessment. However, the patency of the revascularized coronary artery affects myocardial functional recovery in patients subjected to coronary artery bypass grafting (CABG). The influence of graft patency on viability results has not been widely studied. PURPOSE: We evaluated the effect of graft patency on the prediction of MV after CABG by myocardial contrast echocardiography (MCE) and low-dose dobutamine stress echocardiography (LD-DSE). METHODS: Fifty-three subjects with chronic ischemic heart disease scheduled for CABG were divided randomly into groups A (n = 26) and B (n = 27). They underwent MCE and LD-DSE preoperatively. Patients were followed up 12 months after CABG. Group B patients underwent multislice computed tomography angiography to assess CABG patency, and patients with obstructed grafts were excluded. Group A patients were not subjected to multislice CT angiography. The accuracy of MCE and LD-DSE for assessing MV between the two groups was compared. RESULTS: The accuracy and positive predictive values of MCE and LD-DSE for predicting MV were higher in group B than in group A (p < 0.05). CONCLUSIONS: Preoperative LD-DSE and MCE ability to predict MV depends on the patency of CABG.