Unclear focal liver lesions in contrast-enhanced ultrasonography--lessons to be learned from the DEGUM multicenter study for the characterization of liver tumors

Abstract

Purpose: To discuss the difficulties of contrast-enhanced ultrasound (CEUS) in a large multi-center trial.

Materials and methods: CEUS was performed on 1349 liver lesions with an unclear diagnosis after native ultrasound using a standardized protocol (phase inversion; low MI < 0.4; Sonovue Bolus 1.2 - 4.8 ml). The early arterial, arterial, portal venous and late phase > 2 min. were documented. The diagnosis based on CEUS results was compared to the final diagnosis (histology: n = 1006; MRI: n = 269; CT: n = 269 - multiple examinations possible).

Results: Of the 1349 enclosed liver lesions, 20 could not be definitively diagnosed even using all diagnostic steps including histology (the others were proven to be benign n = 573 or malignant n = 756). Of the 1349 unclear liver lesions, 1257 could be differentiated with an accuracy of 90.3% using CEUS. The sensitivity, specificity, and positive and negative predictive value for malignant liver lesions was 95.8%, 83.1%, 88.2% and 93.7% respectively. 92 liver lesions (6.8%) could not be definitively diagnosed using CEUS. Most of them were benign (n = 67) on final diagnosis. The CEUS diagnosis was wrong for 39 lesions. However, only 8 lesions classified as benign by CEUS turned out to be malignant. In 3 cases HCC proven by histology was incorrectly diagnosed by CEUS as adenoma and 2 lesions incorrectly diagnosed by CEUS as FNH turned out to be an HCC and a metastasis. Two lesions diagnosed by CEUS as hemangiomas turned out to be an HCC and a metastasis. One lesion classified as benign by CEUS was ultimately diagnosed as a lymphoma.

Conclusion: Even in this multi-center trial, CEUS proved to be an excellent method for clarifying liver lesions remaining unclear after native ultrasound. The CEUS diagnosis of benign was only incorrect in a few cases.