The effectiveness of computed tomography (CT) in comparison with ultrasound scan (US) for the diagnosis of liver fibrosis: A Structured Literature Review

Abstract

Abstract Background The liver provides biochemical support for virtually every other organ in the body, and damage to the liver as a result of inflammation or infection, or overuse of alcohol or drugs can lead to the development of fibrosis. Diagnosis of liver fibrosis usually requires some form of medical imaging, often either ultrasound or computed tomography (CT). The “gold standard” of diagnosis remains acquisition of a biopsy sample for histological analysis, however, there are risks associated with this procedure. Aim This review seeks to assess the recent literature for diagnostic accuracy data for both CT and ultrasound in comparison with biopsy to establish which of these imaging techniques is the most sensitive and specific. Methods The MEDLINE database (via PubMed), the Cochrane database, the Public Library of Science, and Google Scholar were interrogated for primary research studies that compared CT or ultrasound examinations against biopsy for diagnosis or monitoring of known or suspected liver fibrosis. Inclusion and exclusion criteria were applied to maintain clear focus of the review question. Results A total of six studies were selected for inclusion in this review, all of which focussed on the diagnostic accuracy of their chosen imaging technique in comparison with biopsy. The studies were assessed for quality, and the reported sensitivity and specificity data was extracted and weighted to provide combined values for each of the imaging techniques. Depending upon the ultrasound technique (conventional, elastic, or Doppler), the sensitivity ranged between 67% and 100%, while specificity was between 60% and 97%. CT scanning resulted in greater sensitivity and specificity, both scoring above 90%.

Conclusion CT scans are shown to provide the greatest levels of sensitivity and specificity, followed closely by Doppler ultrasound of the hepatic vasculature. There was, however, considerable heterogeneity across the studies identified for inclusion, so the combined data should be interpreted with a degree of caution. Further research in this area will likely involve the application of artificial intelligence and deep learning engines to supplement the role of radiographers, which could potentially considerably increase the efficiency of medical imaging departments.