Background Ultrasound (US) is a promising diagnostic tool for the diagnosis of calcium pyrophosphate deposition disease (CPPD) but, to our knowledge, in the literature are not available data regarding the reliability between operators in the detection of calcium pyrophosphate (CPP) deposits.
Objectives The aim of this study is to evaluate the agreement between sonographers in the CPP deposits detection and to understand the main causes of discordance.
Methods Six sonographers participated to this study: 4 senior rheumatologists experienced in musculoskeletal ultrasound (MSUS), 1 radiologist expert in musculoskeletal imaging and 1 junior rheumatologist with 3 years' experience in MSUS. This study included 3 steps. In the first the participants defined the criteria for US detection of CPP deposits (this step was extensively described in another abstract). In the second phase the sonographers voted for the presence/absence of CPP deposits in a set of 120 images previously acquired (20 images by each participant) from patients affected or not by CPPD. During the third step the operators performed right knee US examination of eight patients and for each one gave a dichotomous score (presence/absence of CPP crystals). In the second phase the main objective was to test the validity of the criteria, while in the third step the purpose was to evaluate the effectiveness of criteria in a “real life” setting and to understand how the technical approach could affect the reliability.
During the second and the third phases, the dichotomous score was severally provided for menisci, hyaline cartilage, patellar tendon and synovial fluid. All the operators were blinded to patients' disease and to other operator's reports.
Results The overall agreement was good both for static images (k=0.6366) and dynamic scans (k=0.6190). The k value differs greatly between the structures assessed. It appears higher and homogenous for menisci (respectively 0.65 and 0.61 in static and dynamic images) while greater discordance was found at the level of patellar tendon (0,1675 for static images and 0.0505 for dynamic scans). Substantial difference was noted at the level of the hyaline cartilage, in fact for the static images the value of k was of 0.7728 while for the US examination it decreased to 0.1944
Conclusions US detection of CPP deposits appears easier at the level of fibrocartilage while the agreement decreased dramatically for the patellar tendon. This could be due to an absence of clear criteria for the identification of these deposits. The technical approach appeared very important in regards to CPP detection at the level of hyaline cartilage.
In conclusion, the overall agreement in US identification of CPP crystals appears good but could be improved with the revision of US criteria and the standardization of the technical approach as regards to the CPP identification at the level of the hyaline cartilage.
Disclosure of Interest None declared
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.