Effective automatic detection of anterior cruciate ligament injury using convolutional neural network with two attention mechanism modules

Abstract

Background To develop a fully automated CNN detection system based on magnetic resonance imaging (MRI) for ACL injury, and to explore the feasibility of CNN for ACL injury detection on MRI images. Methods Including 313 patients aged 16 – 65 years old, the raw data are 368 pieces with injured ACL and 100 pieces with intact ACL. By adding flipping, rotation, scaling and other methods to expand the data, the final data set is 630 pieces including 355 pieces of injured ACL and 275 pieces of intact ACL. Using the proposed CNN model with two attention mechanism modules, data sets are trained and tested with fivefold cross-validation. Results The performance is evaluated using accuracy, precision, sensitivity, specificity and F1 score of our proposed CNN model, with results of 0.8063, 0.7741, 0.9268, 0.6509 and 0.8436. The average accuracy in the fivefold cross-validation is 0.8064. For our model, the average area under curves (AUC) for detecting injured ACL has results of 0.8886. Conclusion We propose an effective and automatic CNN model to detect ACL injury from MRI of human knees. This model can effectively help clinicians diagnose ACL injury, improving diagnostic efficiency and reducing misdiagnosis and missed diagnosis.

Refined Detection and Classification of Knee Ligament Injury Based on ResNet Convolutional Neural Networks

“…Specifically, in our five-fold cross-validation, our model exhibited an average accuracy of 93.217%, with an average F1 score of 92.006% and a ROC-AUC of 0.97690. These results are comparatively higher than those reported in the study by Liang et al [14], indicating a stronger performance in terms of both accuracy and reliability in detecting ACL injuries.…”

“…In conclusion, while the study by Liang et al [14] introduces an innovative approach by incorporating attention mechanisms into CNN for ACL-injury detection, the presented approach further advances the field by achieving higher diagnostic accuracy and reliability.…”

“…"Effective automatic detection of anterior cruciate ligament injury using convolutional neural network with two attention mechanism modules" [14] paper presents a novel approach in the application of convolutional neural networks (CNNs) for detecting anterior cruciate ligament (ACL) injuries from MRI scans. By integrating two attention mechanism modules, somewhat similar to the ones used in advanced architectures suited for natural language processing (NLP), like transformers into their CNN architecture, the authors aim to enhance the model's focus on relevant features within the MRI images, potentially increasing its diagnostic accuracy.…”

“…Key enhancements include: Convolutional neural networks (CNNs) have emerged as the primary choice for image analysis, demonstrating their strength, particularly within the medical-imaging field. They frequently assist radiologists in the detection and classification of a range of health conditions [11][12][13][14][15][16][17][18][19][20][21]. A comparative review of these cited works [11][12][13][14][15][16][17] that tackle the diagnosis of ACL condition using deep neural networks is undertaken in Section 4, Discussions, where we examine their methodologies and findings in relation to our own study.…”

The Role of Artificial Intelligence in Anterior Cruciate Ligament Injuries: Current Concepts and Future Perspectives

“…Since 2021, there has been an exponential increase in studies on custom architecture CNNs for the diagnosis of ACL injuries applied to MRI, and currently there are various DL models developed, such as VGG16, VGG19, U-Net, AdaBoost, XGBoost, Xception, MRPyrNet, Inception ResNet-v2, RadImageNet, and Inception-v3 DTL [35][36][37][38][39][40][41][42][43][44][45][46][47][48][49][50][51]. Awan et al introduced a method that utilizes a tailored 14-layer ResNet-14 configuration of a CNN, which processes data in six distinct directions.…”