OBJECTIVE: To compare on-site and blinded off-site reading of baseline ultrasound (US) and contrast enhanced ultrasound (CEUS) for classification and characterisation of focal liver lesions. MATERIALS AND METHODS: 99 patients (57 women and 42 men, age range 18-89 years, mean age: 59 years) with 53 malignant and 46 benign liver lesions were studied with unenhanced US followed by contrast enhanced US after injection of 2.4 ml SonoVue® (Bracco, Milano, Italy). Image interpretation was performed on-site with clinical information available by consensus of two readers and off-site by two independent blinded readers at two different centers. Comparison of pre and post contrast scans and of the different readers was performed. Reference examinations were histology, intraoperative US, MRI or CT. RESULTS: Sensitivity for malignancy improved from 81/89/66% (on-site/off-site reader 1/2) before to 100/96/96% post contrast administration (p<0.05, except for reader 1). Specificity improved from 48/48/54% on baseline US to 89/80/76% on CEUS (p<0.05). Accuracy for specific lesion diagnosis was 62/59/50% pre and 90/77/72% post contrast (p<0.05). Classification and characterisation post contrast were mildly inferior for off-site reading. Agreement between on-site and off-site readers of unenhanced scans was fair (κ=0.29-0.39) while it was good for CEUS (κ=0.63-0.79). CONCLUSIONS: CEUS improves classification and characterisation of focal liver lesions and interobserver agreement compared to conventional US. Classification and characterisation post contrast were mildly but statistically significantly better for on-site than for off-site reading.
OBJECTIVE: To compare on-site and blinded off-site reading of baseline ultrasound (US) and contrast enhanced ultrasound (CEUS) for classification and characterisation of focal liver lesions. MATERIALS AND METHODS: 99 patients (57 women and 42 men, age range 18-89 years, mean age: 59 years) with 53 malignant and 46 benign liver lesions were studied with unenhanced US followed by contrast enhanced US after injection of 2.4 ml SonoVue® (Bracco, Milano, Italy). Image interpretation was performed on-site with clinical information available by consensus of two readers and off-site by two independent blinded readers at two different centers. Comparison of pre and post contrast scans and of the different readers was performed. Reference examinations were histology, intraoperative US, MRI or CT. RESULTS: Sensitivity for malignancy improved from 81/89/66% (on-site/off-site reader 1/2) before to 100/96/96% post contrast administration (p<0.05, except for reader 1). Specificity improved from 48/48/54% on baseline US to 89/80/76% on CEUS (p<0.05). Accuracy for specific lesion diagnosis was 62/59/50% pre and 90/77/72% post contrast (p<0.05). Classification and characterisation post contrast were mildly inferior for off-site reading. Agreement between on-site and off-site readers of unenhanced scans was fair (κ=0.29-0.39) while it was good for CEUS (κ=0.63-0.79). CONCLUSIONS: CEUS improves classification and characterisation of focal liver lesions and interobserver agreement compared to conventional US. Classification and characterisation post contrast were mildly but statistically significantly better for on-site than for off-site reading.
Authors: Alexander Hann; Lucas Bettac; Mark M Haenle; Tilmann Graeter; Andreas W Berger; Jens Dreyhaupt; Dieter Schmalstieg; Wolfram G Zoller; Jan Egger Journal: Sci Rep Date: 2017-10-06 Impact factor: 4.379