Fusion of photon-counting CT and MRI improves assessment of the knee: a pilot study

The complementary information gained by combining photon counting CT (PCCT) and MRI may enhance the evaluation of the osteochondral unit. Our aim is to evaluate whether fusing PCCT and MRI improves diagnostic clarity and segmentation consistency for osteochondral lesions compared with MRI alone.

24 participants (15 with knee OA and 9 healthy controls) underwent PCCT and 3T MRI between October 2024 and May 2025. Images were co-registered and fused using a mutual information-based algorithm. Two musculoskeletal radiologists rated diagnostic clarity on 5-point Likert scales. Inter-reader agreement was quantified with the intraclass correlation coefficient (ICC) and Cohen kappa. Another musculoskeletal radiologist and a radiology trainee performed manual segmentations of bones, cartilage, meniscus, and osteophytes on PCCT-MRI fusion or MRI alone. Inter- and intra-rater segmentation consistency was evaluated using Dice similarity coefficient and surface Dice (sDice). Paired comparisons were analyzed with the Wilcoxon signed-rank test.

Fusion imaging yielded significantly higher clarity scores for cartilage thickness (3.91 vs 4.95), articular surface (3.93 vs 4.63), and osteophyte boundaries (4.00 vs 4.83; all p < .05). Inter-reader agreement improved (ICC, from 0.43 to 0.88; Cohen kappa, from 0.42 to 0.89). Quantitative analysis showed greater inter- and intra-rater consistency for cartilage (Dice, 0.70 vs 0.81; sDice, 0.66 vs 0.78) and bone (Dice, 0.93 vs 0.95; sDice, 0.61 vs 0.86; all p < .01).

PCCT-MRI fusion enhances diagnostic clarity and inter-reader agreement. Fusion also increases segmentation consistency for cartilage and bone, supporting its potential clinical value for comprehensive osteochondral evaluation.