Wrist and Hand Ligament Injuries
Ligament injuries of the wrist and hand are common causes of pain, instability, and functional impairment, yet their diagnosis remains challenging. Imaging frequently reveals structural abnormalities, but their clinical significance is not always clear. This thesis therefore focuses not only on detecting abnormalities, but on identifying which findings are truly clinically meaningful and relevant for treatment. Chapter 2 investigates the prevalence of scapholunate interosseous ligament (SLIL) signal abnormalities on wrist MRI. Among 1,021 patients, SLIL signal changes were present in 31% of MRIs. Most patients belonged to the low clinical suspicion group, and prevalence increased with age. More than half had no documented prior wrist trauma. These findings demonstrate that SLIL signal abnormalities are common and should not automatically be interpreted as acute or clinically relevant pathology. Chapter 3 examines the relationship between extrinsic ligament injury and scapholunate diastasis in patients with MRI-confirmed scapholunate ligament injury. Among 101 patients, 40% had scapholunate diastasis greater than 2 mm. Injuries to both the volar and dorsal extrinsic ligaments were independently associated with diastasis. These findings suggest that clinically meaningful scapholunate instability may extend beyond the intrinsic scapholunate ligament and reflect a broader pattern of ligamentous insufficiency. Chapter 4 focuses on thumb ulnar collateral ligament (UCL) avulsion fractures. Among 114 patients, the avulsion fragment was, on average, similar in size to the UCL footprint. However, fragment size was not associated with surgery, whereas metacarpophalangeal joint instability was significantly associated with operative treatment. Thus, although radiographic morphology helps characterize the injury, clinical instability appears more important for treatment decision-making. Chapter 5 places these findings within the broader context of imaging for wrist ligament pathology. No single imaging modality fully resolves the diagnostic challenges. Radiography mainly demonstrates indirect signs, ultrasound is useful for superficial structures but operator dependent, CT provides excellent assessment of osseous anatomy but limited direct ligament visualization, and MRI allows direct visualization but has variable diagnostic performance. Artificial intelligence (AI) may provide additional value by improving standardization, reducing observer variability, supporting quantification, and facilitating more consistent and clinically meaningful interpretation. Chapter 6 further explores AI-based clinical prediction models. Such models may support individualized decision-making by integrating multimodal data and identifying complex patterns that may not be apparent through conventional interpretation alone. However, their clinical value depends on rigorous development, validation, transparent reporting, and demonstration of clinical impact. For wrist and hand ligament injuries, prediction models may ultimately help integrate imaging with factors such as age, trauma history, physical examination, and associated injury patterns. Overall, this thesis demonstrates that detecting a ligament abnormality is only the first step. Age, clinical history, associated injuries, instability, and examination findings determine whether an imaging abnormality is clinically meaningful. Future diagnostic approaches should therefore move beyond detection toward integrated, patient-specific interpretation, with advanced imaging and AI potentially supporting more accurate and treatment-oriented decision-making.