Interobserver agreement, sensitivity and specificity for the diagnosis of horses with proximopalmar metacarpal lameness using radiography and scintigraphy – A retrospective study

Abstract

Lameness originating from the proximal metacarpal region is known as one of the most common causes of lameness in horses. Differential diagnosis is based on the exclusion of other possible causes. Diagnostic imaging modalities, such as radiography, ultrasonography and scintigraphy, are often used to ensure an accurate diagnosis. Radiography is used routinely, in clinical practice, often as the first modality to assess the integrity of the bone. However, radiography has been reported as an insensitive modality to assess changes in proximal metacarpal region such as desmitis and/or insertion desmopathy of the proximal suspensory ligament, which can lead cause lameness. In addition, radiographs of this area can be difficult to interpret. Increased opacify of the proximal lateral aspect of the third metatarsal bone can be seen in some clinically normal horses, often bilaterally. This may reflect increased loading of the lateral aspect of the bone and stressrelated remodeling. Nuclear scintigraphy can be extremely helpful in identifying stress-related bone injury typified by moderate or intense focal increased radiopharmaceutical uptake (IRU). However, a small proportion of horses with primary proximal suspensory desmitis show a similar pattern and intensity of IRU. Other horses with proximal suspensory desmitis have much less intense IRU compared with horses with a primary stress-related bone injury and, in many horses with primary suspensory ligament injury, radiopharmaceutical uptake is normal. Interestingly, these reports did not take the experience level of the interpreter into account. The current study therefore aimed to document the influence of the interpreter's experience on sensitivity and specificity of radiography and scintigraphy as diagnostic modalities for diagnosis proximal palmar metacarpal lameness. We hypothesised that sensitivity and specificity of radiography and scintigraphy for the diagnosis of proximal palmar metacarpal lameness increases with the experience of the interpreter. For this purpose, the medical records of horses which underwent a radiographic as well as scintigraphic examination of the proximal metacarpal region were reviewed. 32 limbs of 23 horses met the inclusion criteria of the study. Thereafter the horses were divided into two groups. The control group included the horses which were sound or had lameness located elsewhere than the proximal metacarpal region. The lame group included horses, which were diagnosed with lameness located at proximal metacarpal region using clinical examination, local anesthesia and other diagnostic modalities such as computed tomography, ultrasound and magnetic resonance imaging. The radiographs and the scintigraphs of each horse were examined blindly by three different interpreters with different levels of experience (certified radiologist, resident of ECVDI and intern). The results showed that the interpreter with high experience had higher positive and negative predicting value in both modalities compared to the less experienced interpreter. Radiography showed a sensitivity of 60.6% and specificity of 65.1% with a positive predictive value of 50.9% for all interpreters combined. In contrast, scintigraphy showed higher sensitivity (90.1%) and specificity (95%) as well as positive predictive value (90.9%) compared to radiography. The interobserver agreement was tested using Cohen kappa test and revealed bad agreement (0.23) between the interpreters for radiography, whereas it was better for scintigraphy (0.62). The agreement between the more expert interpreters for scintigraphy was significantly better (0.93) than other variants. In conclusion, the value of radiography and scintigraphy as diagnostic tools for palmar plantar metacarpal region is strongly dependent on the experience level of the interpreter.