Objective: To improve the radiographic assessment of cartilage loss, as measured by joint space width (JSW) in patients with osteoarthritis (OA) of the knees required to detect the effect of structure modifying drugs in OA trials. This was achieved by determining which of 3 nonfluoroscopic radiographic views--standing extended, semiflexed, and schuss--produced the most accurate radioanatomic positioning of the joint and greater reproducibility in joint repositioning and JSW measurement.
Methods: Knees from 74 patients with OA of the knees who had medial tibiofemoral compartment JSW > or =2 mm in all views were studied. For all 3 radiographic views, accuracy in the radioanatomic positioning of the knee was determined for both joint rotation and flexion. Reproducibility in joint repositioning and JSW measurement were determined from the difference between repeat examinations taken within 2 h.
Results: About 86% of knees in the 3 views had accurate rotational position of the joint at each visit. Radioanatomically, knees in the semiflexed view were significantly more accurately positioned in regard to knee flexion (p<0.0005) than in the schuss view, which in turn was better (p<0.014) than in the extended knee view. Joint repositioning was significantly more reproducible in the semiflexed (p<0.0001) than in the extended knee, which was better (p<0.013) than in the schuss position. JSW measurement was significantly more reproducible in the semiflexed (p<0.014) than both schuss and extended knee positions, which were not significantly different from each other.
Conclusion: Protocols defining the nonfluoroscopic radiographic procedures for the semiflexed view provide the most accurate radioanatomic joint positioning, and the most reproducible joint repositioning and JSW measurement. Using this method significantly fewer knees would be required to detect significant JSW changes in a structure modifying drug trial compared to the schuss and the extended knee positions.