Emerging research suggests that an accelerated sub-five-minute knee MRI protocol may offer robust detection of structural abnormalities including tears of the anterior cruciate ligament (ACL), posterior cruciate ligament (PCL) and medial meniscus.
For the retrospective study, recently published in the American Journal of Roentgenology, researchers evaluated the use of an accelerated 3T MRI of the knee — which combines parallel imaging (PI), simultaneous multi-slice acceleration and deep learning (DL) powered super-resolution — for 124 adult patients (mean age of 46) who underwent arthroscopic surgery.
The study authors found that the sub-five-minute knee MRI protocol provided 100 percent sensitivity, 99 percent specificity and 99 percent accuracy for the detection of anterior cruciate ligament (ACL) tears.
For the detection of posterior cruciate ligament (PCL) tears, the abbreviated knee MRI protocol offered 100 percent sensitivity, specificity, and accuracy, according to the researchers. They also noted 90 percent sensitivity, 95 percent specificity and 94 percent accuracy for medial meniscus tears.
“(Deep learning) image reconstruction methods enable highly accelerated clinical PI-SMS knee MRI with better image quality than conventional methods and substantially shorter scan times, adding value through maintained diagnostic accuracy,” wrote lead study author Jan Vosshenrich, M.D., who is affiliated with the Departments of Radiology at the Grossman School of Medicine at New York University and University Hospital Basel in Basel, Switzerland, and colleagues.
Three Key Takeaways
1. High diagnostic performance for ligament tears. The abbreviated MRI protocol demonstrated 100 percent sensitivity and specificity for detecting PCL tears, and 100 percent sensitivity with 99 percent specificity for ACL tears, indicating excellent diagnostic accuracy for major ligament injuries.
2. Effective detection of meniscal and cartilage abnormalities.
The protocol achieved 90 percent sensitivity and 95 percent specificity for medial meniscus tears, and 85 percent sensitivity and 88 percent specificity for cartilage defects, showing strong potential for evaluating a range of structural abnormalities.
3. Advanced imaging efficiency with deep learning. The integration of deep learning-powered super-resolution enhanced image quality, allowing for substantially shorter scan times (under five minutes) without compromising diagnostic accuracy, which could improve workflow and patient throughput.
For cartilage defects, the study authors noted an 85 percent sensitivity, 88 percent specificity and an 88 percent accuracy for the accelerated knee MRI protocol.
“The DL-based super-resolution augmentation of images contributed to excellent cartilage lesion detection in our study. It in essence quadruples the spatial resolution by doubling the matrix size along both in-plane axes, thereby enhancing image detail and visualization of small anatomic structures,” emphasized Vosshenrich and colleagues.
(Editor’s note: For related content, see “Study Examines Impact of Deep Learning on Fast MRI Protocols for Knee Pain,” “Study Assesses Potential of Seven-Minute AI-Enhanced 3T MRI of the Shoulder” and “Image IQ Quiz: 30-Year-Old Patient with Knee Pain.”)
Beyond the inherent limitations of a retrospective single-center study, the authors conceded possible patient selection bias with the study’s emphasis on patients who had 3T MRI and arthroscopic surgery. They cautioned against broad extrapolation of the study findings to patients who had different MRI field strengths and the use of other deep learning algorithms for image reconstruction. The researchers also conceded that use of a simplified grading system may have enhanced diagnostic performance of the accelerated MRI protocol for cartilage defects.
