Patterns, Movement and Clinical Diagnosis of Abdominal Adhesions

Patterns in normal abdominal movement captured with medical imaging can be recognised by a trained radiologist but the process is time consuming. Abdominal adhesions present a diagnostic problem in which the radiologist is asked to detect abnormal movement that may be indicative of pathology. This paper postulates that the use of image analysis can augment the diagnostic abilities of the radiologist in respect of adhesions. Proof of concept experiments were conducted in-silico to explore the effectiveness of the technique. The results indicate that trained participants are accurate in their assessment of abnormalities when supplied with additional information from image analysis techniques. However without the additional information, participants made incorrect diagnoses on many occasions. ROC methods were used to quantify the outcomes of the in-silico experiment.