Logo image
Back
379 - Implementation and Quality Assurance Monitoring of Imaging for the Multicenter PIKASO Trial to Prevent of Post-traumatic Osteoarthritis after ACL Reconstruction: Protocols and Experience from the First 100+ Patients
Abstract   Peer reviewed

379 - Implementation and Quality Assurance Monitoring of Imaging for the Multicenter PIKASO Trial to Prevent of Post-traumatic Osteoarthritis after ACL Reconstruction: Protocols and Experience from the First 100+ Patients

Xiaojuan Li, Mei Li, Richard Lartey, Stefan Zbyn, Sercan Tosun, Patrick Y. Yeh, Jennifer Jian, Zhiyuan Zhang, Amirtha V. Owens, Moriel Nessaiver, …
Osteoarthritis and cartilage, Vol.34(Supplement), pp.S279-S279
04/2026
DOI: 10.1016/j.joca.2026.01.384

View Online

Abstract

Purpose (the aim of the study): Design, implement, and monitor the quality of semi-quantitative and quantitative MRI (qMRI) measures as primary and secondary outcomes for the Preventing Injured Knees from OsteoArthritis: Severity Outcomes (PIKASO) trial. Methods: PIKASO, a multicenter, blinded, parallel, two-arm randomized controlled trial (RCT), evaluates the efficacy of 12-month dosing of oral metformin vs. placebo in reducing post-traumatic osteoarthritis (PTOA) development after anterior cruciate ligament reconstruction (ACLR). Primary outcomes are MRI-derived structural cartilage change (supplemented MRI Osteoarthritis Knee Score [sMOAKS]) at 24 months post-ACLR and patient-reported pain (Knee injury and Osteoarthritis Outcome Score [KOOS] pain subscale) average between 12 and 24 months post-ACLR. Secondary outcomes include qMRI (cartilage T1ρ, T2) changes at 12 and 24 months post-ACLR. Knee MRIs at 3T are obtained at baseline (pre-ACLR), and 12-months and 24-months post-ACLR. Knee frame radiographs (Synaflexer™, Clario) are collected at baseline and 24-months post-ACLR. The MRI protocol includes 2D and 3D TSE/FSE imaging for morphology, 3D DESS/MENSA imaging for morphometry and cartilage segmentation, 3D T1ρ/T2 qMRI for cartilage composition, and, for inflammation, fluid-attenuated inversion recovery (FLAIR) and optional dynamic contrast enhanced (DCE) (Table 1). The Arthritis Foundation Osteoarthritis Imaging Center (OIC) qualified site’s MR systems and knee coils by American College of Radiology (ACR) guidelines, followed by study protocol validation using phantoms and volunteers. The trial’s phantom quality assurance (QA) protocol includes monthly morphology scans to assess system and coil performance, along with quarterly qMRI measures (starting with 4-weekly calibration scans) using a novel musculoskeletal relaxometry phantom (Calimetrix) equipped with a built-in MR compatible thermometer for temperature correction of qMRI values. De-identified patient imaging data are transferred to the OIC within 24 hours of acquisition for QA, including 1) automated verification of proper MR parameters in DICOM headers; 2) radiologist image review for adequacy for sMOAKS grading; and 3) qMRI-QA with automated cartilage/meniscus segmentation and T1ρ/T2 quantification. Three-tier feedback from OIC is provided to sites within 72 hours based on adequacy for sMOAKS scoring (primary endpoint) and protocol adherence: PASS (ideal images), ADEQUATE (some issues, not affecting sMOAKS scoring, no re-scan), or FAIL (patient rescan required due to QA issues affecting sMOAKS scoring). Radiograph QA was similarly judged based on imaging quality and medial inter-marginal distance (IMD) measurement <3mm. Results: A total of 11 MR systems at 9 sites were qualified and enrolled for the trial (Table 2). Harmonized imaging protocols across 2 vendors and 5 software versions were developed and distributed electronically along with comprehensive SOPs and training. As of Oct 2025, all scanners have undergone 12-16 months of phantom QA, except for one unit brought online in Sept 2025. Only 7 of the 164 scheduled phantom QA scans across all magnets were missed (96% compliance rate); only 1 magnet had > 1-month lapse of QA. The overall longitudinal phantom repeatability root-mean-square coefficients of variation (RMS-CVs) across all sites were 2.8% (T1ρ) and 3.1% (T2), ranging 0.9%-5.0% for each site (Table 2). Monthly phantom QA identified two knee coils that developed problems; both resolved promptly, preventing patient data degradation/loss. Volunteer scan–rescan RMS-CVs across all sites were 3.1% (T1ρ) and 3.5% (T2), ranging 1.0%-5.7% for each site (Table 2). As of Oct 2025, a total of 112 patients have completed baseline MRI with 75% “PASS” and 25% “ADEQUATE” with no re-scans required; 109 patients (97%) completed baseline radiograph with 89% ideal images and no failed QA scores. Conclusions: A harmonized imaging protocol incorporating morphologic and qMRI has been successfully implemented across multiple vendors and sites in the PIKASO RCT. A rigorous QA workflow combining automated DICOM verification, radiologist review, and quantitative analysis produced high quality and inter-site consistent data. Excellent phantom and human scan–rescan repeatability confirmed reliability of qMRI measures.

Details

Metrics

1 Record Views
Logo image