Artificial Intelligence-Assisted Advanced Analysis of Knee Imaging and Outcome Prediction

Official title

Artificial Intelligence-Assisted Advanced Analysis of Knee Imaging and Outcome Prediction for Ultrasound-Guided Injections

Overview

This study aims to develop and validate an artificial intelligence (AI)-assisted platform for musculoskeletal knee ultrasonography and to establish an interpretable prediction model for clinical outcomes following ultrasound-guided injection therapies in patients with degenerative knee disorders. The project seeks to improve the standardization, reproducibility, and clinical utility of knee ultrasound by reducing operator dependency and providing quantitative image analysis and outcome prediction. The study will be conducted in three phases. First, an AI foundation model for knee ultrasonography will be developed using standardized image acquisition protocols to enable automated localization, segmentation, and quantitative assessment of major anatomical structures, including tendons, ligaments, cartilage, fat pads, and peripheral nerves. Second, supervised machine learning models will be trained to classify normal and pathological ultrasound findings, including common degenerative and inflammatory abnormalities affecting the knee. Third, retrospective and prospective clinical data from approximately 150 patients receiving ultrasound-guided injection therapies will be integrated to develop and validate a predictive model for treatment outcomes using imaging biomarkers and clinical variables. Treatment response will be evaluated using validated patient-reported outcome measures, and explainable AI methods will be applied to improve model interpretability. The anticipated outcome of this study is the development of a comprehensive AI-assisted knee ultrasound platform that supports standardized image interpretation, quantitative assessment of musculoskeletal pathology, and personalized prediction of treatment response to ultrasound-guided injection therapies in degenerative knee disorders.

Detailed description

High-resolution musculoskeletal ultrasonography has become a first-line imaging modality because it enables real-time visualization and dynamic assessment with high accessibility and low cost. Nevertheless, ultrasound remains highly operator-dependent, resulting in variability in image acquisition and interpretation, which limits standardization and widespread implementation, particularly for complex joints such as the knee. Building on our established expertise in computational ultrasound and deep-learning-assisted dynamic shoulder analysis, including patented artificial intelligence (AI)-derived quantitative biomarkers, this three-year project aims to develop an AI platform for advanced knee ultrasound analysis and to construct a predictive model for clinical outcomes following ultrasound-guided injections in degenerative knee disorders.

In the first year, we will establish a normative AI foundation model for knee ultrasonography by developing automated localization and multi-structure segmentation of major anatomical components, including tendons, ligaments, cartilage, fat pads, and peripheral nerves. Standardized acquisition protocols will be implemented to ensure consistent image quality. A Faster Region-Based Convolutional Neural Network (Faster R-CNN) framework incorporating ResNet50, a Feature Pyramid Network, and a Region Proposal Network will be used to detect key bony landmarks, followed by a multi-structure segmentation engine and quantitative feature extraction modules (e.g., thickness, surface regularity, and tissue heterogeneity). Segmentation performance will be evaluated using Intersection-over-Union and Dice coefficients, while measurement reliability will be assessed using intraclass correlation coefficients, standard error of measurement, minimal detectable change, and Bland-Altman analyses.

In the second year, the platform will be expanded to differentiate pathological patterns in knee tendons, ligaments, cartilage, and fat pads. Expert clinicians will label each segmented structure as normal or abnormal and further annotate clinically relevant subtypes, such as tendinopathy, calcification, partial or full-thickness tears, synovial hypertrophy or effusion, cartilage wear or exposure, and meniscal degeneration or tear. Supervised learning models will be trained for classification and evaluated using accuracy, precision, recall (sensitivity), and F1-score.

In the third year, we will develop an outcome prediction model for ultrasound-guided injections by integrating retrospective and prospective real-world data from approximately 150 patients receiving common injection therapies, including intra-articular hyaluronic acid, dextrose prolotherapy or platelet-rich plasma, and peripheral nerve-targeted interventions. Treatment success will be defined using validated patient-reported outcome measures, including the Knee Injury and Osteoarthritis Outcome Score and the Patient Acceptable Symptom State, incorporating minimal clinically important difference thresholds. Feature selection methods and cross-validation will be applied to mitigate overfitting. Model performance will be assessed using area under the receiver operating characteristic curve, sensitivity, specificity, accuracy, F1-score, and calibration metrics, with Shapley Additive exPlanations employed to enhance interpretability. External validation will be performed if additional datasets become available.

This project is expected to deliver the first systematic AI-based normative atlas for knee ultrasonography and an interpretable outcome prediction framework, improving diagnostic consistency, reducing operator dependency, and enabling personalized, evidence-informed injection strategies.

Interventions

• Diagnostic test artificial intelligence-assisted advanced analysis
  The study will be conducted in three phases. First, an AI foundation model for knee ultrasonography will be developed using standardized image acquisition protocols to enable automated localization, segmentation, and quantitative assessment of major anatomical structures. Second, supervised machine learning models will be trained to classify normal and pathological ultrasound findings, including common degenerative and inflammatory abnormalities affecting the knee. Third, retrospective and prosp

Primary outcome measures

• AI Segmentation Performance for Normal Knee Structures [Time frame: Baseline (at ultrasound examination)]
• Diagnostic Accuracy of AI-Based Classification of Knee Pathologies [Time frame: Baseline (at ultrasound examination)]
• Accuracy of AI Prediction for Treatment Success [Time frame: 3 months after ultrasound-guided injection]

Secondary outcome measures (5)

• Knee Pain Intensity [Time frame: Baseline, 1 month, and 3 months]
• Knee Function [Time frame: Baseline, 1 month, and 3 months]
• Knee Injury and Osteoarthritis Outcome Score (KOOS) [Time frame: Baseline, 1 month, and 3 months]
• Patient Acceptable Symptom State (PASS) [Time frame: 3 months after treatment]
• Reliability of Ultrasound Measurements [Time frame: Baseline]

Eligibility criteria

Objective 1: Development of an AI-Based Normative Model for the Healthy Knee

Inclusion criteria

• Clinical diagnosis of healthy adult without major systemic disease
• Age ≥18 years
• Able to understand and follow study instructions
• Ambulatory without walking aids
• No pain in either knee for at least 6 months before enrollment

Exclusion criteria

• Previous knee surgery
• Rupture of one or more cruciate ligaments
• Knee injection within the preceding 6 months
• Major trauma involving the knee or periarticular region
• Rheumatic or autoimmune disease

Objective 2: Development of an AI-Based Model for the Identification of Pathological Knee Structures

Inclusion criteria

• Clinical diagnosis of radiographic knee osteoarthritis
• Age ≥18 years
• Knee pain in at least one knee during the preceding year
• Medical records confirming knee pain, soreness, or stiffness within 1 month before enrollment
• Radiographic evidence of knee osteoarthritis, defined by at least one of the following:
• Kellgren-Lawrence grade ≥2 on anteroposterior radiographs
• Kellgren-Lawrence grade ≥2 on skyline (patellofemoral) radiographs
• Superior or inferior patellar osteophytes or posterior tibial osteophytes on lateral radiographs

Exclusion criteria

• Systemic rheumatic disease (e.g., rheumatoid arthritis or ankylosing spondylitis)
• Malignancy
• Previous major knee trauma (including fracture)
• Previous knee surgery
• Intra-articular corticosteroid injection within the preceding 3 months

Objective 3: Development of an AI-Assisted Predictive Model for Injection Treatment Outcomes

Inclusion criteria

• Clinical diagnosis of radiographic knee osteoarthritis requiring ultrasound-guided injection therapy
• Age ≥18 years
• Knee pain in at least one knee during the preceding year
• Medical records confirming knee pain, soreness, or stiffness within 1 month before enrollment
• Radiographic evidence of knee osteoarthritis, defined by at least one of the following:
• Kellgren-Lawrence grade ≥2 on anteroposterior radiographs
• Kellgren-Lawrence grade ≥2 on skyline (patellofemoral) radiographs
• Superior or inferior patellar osteophytes or posterior tibial osteophytes on lateral radiographs
• Willingness to undergo ultrasound-guided injection therapy and complete scheduled follow-up assessments

Exclusion criteria

• Systemic rheumatic disease (e.g., rheumatoid arthritis or ankylosing spondylitis)
• Malignancy
• Previous major knee trauma (including fracture)
• Previous knee surgery
• Intra-articular corticosteroid injection within the preceding 3 months

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: Yes

Study design

Observational model

Cohort

Study locations

Taiwan · 1 center

• National Taiwan University Hospital Beihu Branch — Taipei

Identifiers

NCT: NCT07721116 · 202603036RINE

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗