Huizheng Wang¹, Weiyin Vivian Liu², Xingyao Yu³, Wen Chen⁴, Ling Sang⁴, and Kejun Wang^4
1Department of Radiology, Taihe Hospital, Jinzhou Medical University Union Training Base, Shiyan,Hubei, China, ²GE Healthcare, Beijing, China, ³Biomedical Engneering College,Hubei University of Medicine, Shiyan,Hubei, China, ⁴Department of Radiology, Taihe Hospital, Hubei University of Medicine, Shiyan,Hubei, China

Image follow-up for intervention effectiveness of osteoarthritis is necessary to avoid or slow down the development of joint space stenosis, bone sclerosis and osteophyte formation. Compared to ZTE-MRI with the feature of nominal zero echo time, MAGIC, synthetic MRI, offers not only anatomical structure images but also T1, T2, and PD maps. We found a significant correlation between joint space value and meniscus volume in reflection of antagonism effect of joint space and surrounding tissues. Proton density weighted synthetic MR images (PD-MAGIC) offered as good as ZTE-MR structure images and might supplement quantitative information clinically to improve  follow-up diagnostic performance.

Lin Mu¹, Ying Qiu², Yun Pei³, Yi Zhu⁴, and Ke Jiang^4
1Radiology, The First Hospital of Jilin University, Changchun, China, ²The First Hospital of Jilin University, Changchun, China, ³College of Electronic Science and Engineering, Jilin University, Changchun, China, ⁴Philips Healthcare, Beijing, China

The use of three dimensional (3D) volumetric acquisition in clinical settings has been limited due to long scan time. A deep learning-based reconstruction algorithm allows shortening of scan time and provide comparable overall image quality when compared with standard sequences. Adaptive-CS-Net, a deep neural network previously introduced at the 2019 fast MRI challenge, was expanded and presented here as a Compressed-SENSE Artificial Intelligence (CS-AI) reconstruction. The purpose of the study is to determine the feasibility of 3D PDWI accelerated with CS-AI for evaluating the knee image quality and compared with SENSE and standard Compressed-SENSE.

Junru Zhong¹, Yongcheng Yao¹, Sheheryar Khan², Fan Xiao¹, Dόnal G. Cahill¹, James F. Griffith¹, and Weitian Chen^1
1Department of Imaging and Interventional Radiology, The Chinese University of Hong Kong, Sha Tin, NT, Hong Kong, ²School of Professional Education & Executive Development, The Hong Kong Polytechnic University, Kowloon, Hong Kong

We present a deep learning-based knee osteoarthritis grading system that automatically provides a binary classification for cartilage degeneration. The system was trained on MRI data sets applying MOAKS grading from the Osteoarthritis Initiative (OAI). The proposed method achieved an accuracy of 0.75 to 0.83, despite being conducted on highly imbalanced data sets. Significant improvement in accuracy is expected with more balanced data sets.

Jiajia Li¹, Shufang Wei¹, Xianchang Zhang², Jing An³, Esther Raithel⁴, and Yinghui Ge^1
1Fuwai Central China Cardiovascular Hospital, Zhengzhou, China, ²MR Collaboration, Siemens Healthineers Ltd, Beijing, China, ³Siemens Shenzhen Magnetic Resonance Ltd, Shenzhen, China, ⁴Siemens Healthcare GmbH, Erlangen, Germany

Hemophilic arthropathy (HA) is a serious complication of haemophilia, characterized by bone and cartilage damage due to recurrent joint bleeding. Existing evidence showed HA has a fast disease progression. MRI-based evaluation plays an important role in the detection, categorization and staging of soft tissue and osteochondral changes in HA, which is important for assessing and monitoring treatment outcomes. However, currently semi-quantitative scoring system based on conventional MRI images cannot give an objective assessment of knee cartilage damage. To this end, this study for the first time used an automatic segmentation algorithm to quantitatively evaluate the morphological change of knee cartilage in HA.

Elena Voronkova^1,2, Petr Menshchikov^3,4, Ilya Melnikov¹, Andrei Manzhurtsev^1,4, Maxim Ublinskii¹, Denis Vorobyev¹, Dmitriy Kupriyanov³, and Tolib Akhadov^1
1Clinical and Research Institute of Emergency Pediatric Surgery and Trauma, Moscow, Russian Federation, ²National Research Nuclear University MEPhI, Moscow, Russian Federation, ³Philips Healthcare, Moscow, Russian Federation, ⁴Institute of Biochemical Physics, Russian Academy of Sciences, Moscow, Russian Federation

This paper has demonstrated the benefits of segmentation into the deep, intermediate, and superficial layers in analysis of T2 maps of patellar cartilage with chondromalacia. In the deep and intermediate layers, the Т2 values were significantly different in all the patients groups, while the T2 values in the superficial layer did not depend on the disease grade. Compared with the whole cartilage assessment, this approach increased the sensitivity and specificity of chondromalacia staging by 17% and therefore can significantly increase the clinical efficiency of T2 mapping.

Renchi Ma¹, Huiyu Qiao², Xinrong Qiao¹, Huijun Chen², Xihai Zhao², and Zhuozhao Zheng^1
1Department of Radiology, Beijing Tsinghua Changgung Hospital, School of Medicine, Tsinghua University, Beijing 102218, China, Beijing, China, ²The Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing 100084, China, Beijing, China

T1, T2, and T2* mapping provide valuable information for evaluating joint cartilage. However, few studies simultaneously quantified T1, T2, and T2* map to evaluate knee cartilage. In this study, a three-dimensional (3D) simultaneous quantitative T1-T2-T2* mapping (SQUMA) sequence is utilized in knee cartilage. Excellent correlation and good consistency in measuring T1, T2 and T2* values were found between 3D SQUMA sequence and referenced sequences in phantom studies. The T1, T2 and T2* values of knee cartilage were 949.90±82.59 ms, 36.85±2.25 ms and 27.89±5.74 ms, respectively, with good repeatability. 3D SQUMA sequence is feasible for in vivo evaluation of knee cartilage.

Keyan Yu^1,2, Chuanyang Zheng³, Jiaping Hu¹, Lijie Zhong¹, Xiaodong Zhang¹, and Qi Dou^3
1Department of Medical Imaging, The Third Affiliated Hospital of Southern Medical University, Guangzhou, China, ²Department of Radiology, Peking University Shenzhen Hospital, Shenzhen, China, ³Department of Computer Science & Engineering, The Chinese University of Hong Kong, Hong Kong, China

Infrapatellar fat pad (IPFP) is an important risk factor for the incident of radiographic knee osteoarthritis (iROA)^{1, 2}. However, the potential of being an independent biomarker to predict iROA is untapped. Deep learning (DL) is a set of algorithms that enable computers to discover complicated patterns in large data sets3. In this study, we train a DL model to predict iROA with auto-segmented IPFP, comparing it to the DL model set up with corresponding whole knee MR images (MRI). The results reveal that IPFP alteration can predict iROA independently comparably to the whole knee MRI at one year before iROA.

Xiaoxuan Wang¹, Xiaowen Ma¹, Xiaocheng Wei², and Haiyan Li^1
1Department of Magnetic Resonance Imaging, Hong Hui Hospital of Xi’an Jiaotong University, Xi’an, China, ²GE Healthcare, Beijing, China

In this study, we aim to investigate whether DTI is superior to conventional MRI in evaluating acute ACL injury. It was concluded that the mean FA values is lower and the mean ADC values is higher in the ACL injury group than control group. The sensitivity, specificity, accuracy and Kappa value of DTI were higher than those of conventional MRI.

Xiaolian Su¹, Guangyu Tang¹, and Pu-Yeh Wu^2
1Department of Radiology, Shanghai tenth People’s Hospital, Tongji University School of Medicine, Shanghai, China, shanghai, China, ²GE Healthcare, Beijing, China, Beijing, China

Recent advances in UTE-MRI techniques development allow for quantitative measurements of tissues with very short T2. However, in vivo experiments are needed to extend its application in clinical practice. In this in vivo study, we found that both UTE-MTR and UTE-adiabatic-T1rho values indicated a moderate correlation with MOAKS Grading. Meanwhile, both of them had moderate diagnostic efficacy for the mild cartilage degeneration (MOAKS = 1). These findings suggested that quantitative UTE-MRI techniques have a great potential as imaging biomarkers for early OA diagnosis in clinic.

Takayuki Sakai¹, Jihun Kwon², Masami Yoneyama², Daichi Murayama¹, Quin Lu³, Arjun Desai⁴, Akshay S Chaudhari⁵, Tosiaki Miyati⁶, Shigehiro Ochi¹, and Atsuya Watanabe^7,8
1Radiology, Eastern Chiba Medical Center, Chiba, Japan, ²Philips Japan, Tokyo, Japan, ³Philips Healthcare NA, San Francisco, CA, United States, ⁴Electrical Engineering, Stanford University, Stanford, CA, United States, ⁵Radiology, Stanford University, Stanford, CA, United States, ⁶Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa, Japan, ⁷General Medical Services, Chiba University Graduate School of Medicine, Chiba, Japan, ⁸Orthopaedic Surgery, Eastern Chiba Medical Center, Chiba, Japan

Multi-Interleaved X-prepared TSE with inTUitive RElaxometry (MIXTURE) is a recently developed 3D magnetization-prepared turbo spin-echo sequence. MIXTURE provides 3D high-resolution isotropic knee morphometry and T1ρ relaxometry in a clinically reasonable scan time. In this study, we integrated MIXTURE with the into a Deep-learning, Open-Source Medical Analysis (DOSMA) framework for deep learning-powered musculoskeletal MR image and evaluated its clinical usefulness in patients with knee joint pain. We show that the combination of MIXTURE and DOSMA enables a comprehensive whole knee MR exam including 3D isotropic submillimeter imaging, T2- and T1ρ-mapping, and standardized image analysis. The clinical insight MIXTURE-DOSMA brings is demonstrated. 

Peng Luo¹, Wentao Hu², Guanwu Li¹, and Yongming Dai^2
1Department of Radiology, Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, China, ²United Imaging Healthcare, Shanghai, China

Evaluation of early-stage knee osteoarthritis (OA) has long been of wide concern, while MRI has proved its value in this field by the richness of contrasts and non-invasive nature. However, MRI detection for composition change in OA still calls for deeper researches. ADC and T₂ measured by conventional MRI are only an average of intra-voxel components and assumed to be independent. However, by performing a hybrid multidimensional MR imaging (HM-MRI) method, an implied correlation between ADC and T₂ in articular cartilage has emerged. Moreover, V_normal might offer extra information in detecting cartilage degeneration at early stage of OA than conventional ADC/T₂ parameters.

Yanjing Zhang^1,2, Li Zhang^3,4, Wanzhen Yao^1,2, Siyu Dai^1,2,5,6, Dingbo Shu^1,2, Jianping Ding^2,5,7, and Mengxiao Liu^8
1Department of Radiology, The Affiliated Hospital of Hangzhou Normal University, Hangzhou, China, ²Institute of Sport Medicine, Hangzhou Normal University, Hangzhou, China, ³Department of Radiology, The Children’s Hospital, Zhejiang University School of Medicine, Hangzhou, China, ⁴National Clinical Research Center for Child Health, Hangzhou, China, ⁵School of Clinical Medicine, Hangzhou Normal University, Hangzhou, China, ⁶Faculty of Medicine, The Chinese University of Hong Kong, HongKong, China, ⁷1 Department of Radiology, The Affiliated Hospital of Hangzhou Normal University, Hangzhou, China, ⁸MR scientific Marketing, Diagnostic Imaging, Siemens Healthineers Ltd, Shanghai, China

This research investigated the difference in T2* value of knee cartilage between amateur runners and non-exercisers, and examined  the trends in the T2* value of knee cartilage in amateur runners before and after a half marathon. The results showed that the T2* value in lateral femorotibial joint of amateur runners was lower than non-exercisers. And the T2* value increased significantly after half-marathon running, and returned to the baseline level after 3 days. This indicates that long-term running may be beneficial to joint health; the effect of a half-marathon exercise on cartilage is reversible.

Xu Xu¹, Chunchao Xia², Wen Zeng², Wanlin Peng², Yongming Dai³, Ke Xue³, Zhenlin Li², Zhenlin Li², and Zhenlin Li^2
1Department of Radiology, west China hospital of Sichuan University, Chengdu, China, ²west China hospital of Sichuan University, Chengdu, China, ³MR Collaboration, Central Research Institute, United Imaging Healthcare, Shanghai, China

In knee MRI, patients with pain or limited motion can have difficult to endure a long-time imaging position. At the same time, the relatively high spatial resolution and good image quality were required in such musculoskeletal MRI. Therefore, shortening scan time and improving the image quality is the aim of current knee MRI.Our study compared conventional parallel imaging (PI) and artificial intelligence assisted compressed sensing (ACS) in terms of imaging quality and diagnostic performance in knee MRI, and achieved a 5-minute comprehensive examination of the knee, without compromising either image quality or visualization of anatomical and pathological structures.

Li Jujia¹, Jian Zhao¹, Xuesong Zhang¹, Ping Zhang¹, Congcong Ren¹, Yan Zeng¹, Ranxu Zhang¹, Ming Wang¹, and Xiaoyue Zhou^2
1The Third Hospital of Hebei Medical University, Shijiazhuang, China, ²Siemens Healthineers, Ltd., Shanghai, 201318, China., Shanghai, China

T2 mapping can evaluate the dynamic changes of the meniscal microstructure. The study was designed to evaluate the T2 values of the menisci of amateur marathon volunteers at three different timings. Compared with the T2 values of menisci one week before running, the values 12 h after running increased significantly (P <0.05), and the values 2 months after running did not change significantly (P > 0.05). The result suggests that marathons do not necessarily cause irreversible knee damage. 

Yongcheng Yao¹, Dόnal G. Cahill¹, James F. Griffith¹, and Weitian Chen^1
1The Chinese University of Hong Kong, New Territories, Hong Kong

Thickness mapping techniques are important in the  management of osteoarthritis. Since most thickness mapping approaches require two surfaces as inputs, surface segmentation is important for articular cartilage thickness mapping. In this work, a systematic approach is proposed for automatic articular cartilage surface segmentation from a volume segmentation mask, which includes cartilage surface reconstruction, body surface construction, surface normal estimation, directional voxels searching, restricted surface dilation, and surface close operation. The proposed method can be used for reliable cartilage thickness mapping.