MyMedR

Displaying all 2 publications

Abstract:

Sort:

Fulltext Anatomically realistic aortic dissection simulator as a potential training tool for point-of-care ultrasound

Rahmah M, Al-Ashwal RH, Salim MIM, Lam YT, Hau YW

J Ultrason, 2024 Feb;24(94):1-9.
PMID: 38343785 DOI: 10.15557/jou.2024.0002

AIM: Simulators for aortic dissection diagnosis are limited by complex anatomy influencing the accuracy of point-of-care ultrasound for diagnosing aortic dissection. Therefore, this study aimed to create a healthy ascending aorta and class DeBakey, type II aortic dissection simulator as a potential point-of-care ultrasound training model.
MATERIAL AND METHODS: 3D mould simulators were created based on computed tomography images of one healthy and one DeBakey type II aortic dissection patient. In the next step, two polyvinyl alcohol-based and two silicone-based simulators were synthesised.
RESULTS: The results of the scanning electron microscope assessment showed an aortic dissection simulator's surface with disorganised surface texture and higher root mean square (RMS or Rq) value than the healthy model of polyvinyl alcohol (RqAD = 20.28 > RqAAo = 10.26) and silicone (RqAD = 33.8 > RqAAo = 23.07). The ultrasound assessment of diameter aortic dissection showed higher than the healthy ascending aorta in polyvinyl alcohol (dAD = 28.2 mm > dAAo = 20.2 mm) and Si (dAD = 31.0 mm > dAAo = 22.4 mm), while the wall thickness of aortic dissection showed thinner than the healthy aorta in polyvinyl alcohol, which is comparable with the actual aorta measurement. The intimal flap of aortic dissection was able to replicate and showed a false lumen in the ultrasound images. The flap was measured quantitatively, indicating that the intimal flap was hyperechoic.
CONCLUSIONS: The simulators were able to replicate the surface morphology and echogenicity of the intimal flap, which is a linear hyperechoic area representing the separation of the aorta wall.
Fulltext Performance analysis of seven Convolutional Neural Networks (CNNs) with transfer learning for Invasive Ductal Carcinoma (IDC) grading in breast histopathological images

Voon W, Hum YC, Tee YK, Yap WS, Salim MIM, Tan TS, et al.

Sci Rep, 2022 Nov 10;12(1):19200.
PMID: 36357456 DOI: 10.1038/s41598-022-21848-3

Computer-aided Invasive Ductal Carcinoma (IDC) grading classification systems based on deep learning have shown that deep learning may achieve reliable accuracy in IDC grade classification using histopathology images. However, there is a dearth of comprehensive performance comparisons of Convolutional Neural Network (CNN) designs on IDC in the literature. As such, we would like to conduct a comparison analysis of the performance of seven selected CNN models: EfficientNetB0, EfficientNetV2B0, EfficientNetV2B0-21k, ResNetV1-50, ResNetV2-50, MobileNetV1, and MobileNetV2 with transfer learning. To implement each pre-trained CNN architecture, we deployed the corresponded feature vector available from the TensorFlowHub, integrating it with dropout and dense layers to form a complete CNN model. Our findings indicated that the EfficientNetV2B0-21k (0.72B Floating-Point Operations and 7.1 M parameters) outperformed other CNN models in the IDC grading task. Nevertheless, we discovered that practically all selected CNN models perform well in the IDC grading task, with an average balanced accuracy of 0.936 ± 0.0189 on the cross-validation set and 0.9308 ± 0.0211on the test set.

Filters

Please provide feedback to Administrator ([email protected])

External Links