Standardize and compare contrast-enhanced ultrasonographic digital images obtained with different technologies: how to overcome the subjectivity

Histomorphometry of the ligaments using a generic-purpose image processing software, a new strategy for semi-automatized measurements

Gold chloride technique can be combined with Adobe Photoshop® software to yield a quantitative assessment of the different areas in heterogeneous structures as are ligament. A semi-automatized method based on the sum of two- and three-dimensional morphological criteria upon colorimetric criteria allows the identification and measurement of the area occupied by a structure of interest. It also allows the quantification of color intensity to differentiate structures with similar staining avidity, like vessels and nerves. This computer-assisted, semiquantitative procedure for computerized morphometry is relatively simple to perform. The accuracy, efficiency, and reproducibility of this method based on a commercially available imaging program were considered adequate when tested on the anterior cruciate ligament of the cat. Image normalization by trained observers using a commercially available software package designed for photography, applied to a sample randomly chosen, has provided the means of making reproducible measurements of heterogeneous structures.

Resectable pancreatic adenocarcinoma: is the enhancement pattern at contrast-enhanced ultrasonography a pre-operative prognostic factor?

The aim of our study was to determine whether the enhancement pattern of pancreatic adenocarcinoma at contrast-enhanced ultrasonography (CEUS) is related to patient prognosis after resection. CEUS of 42 resected adenocarcinomas were retrospectively reviewed. Tumors were divided into two groups: group A=poorly vascularized (presence of avascular areas) or group B=well vascularized (absence of avascular areas). All lesions were resected and underwent pathological examination assessing tumor differentiation as: undifferentiated (poorly differentiated) or differentiated (moderately and well differentiated). Mean vascular density (MVD) was also evaluated. CEUS enhancement and pathology were correlated (Spearman's test). Survival was analyzed with the Kaplan-Meier method. Multivariate analysis was performed with the Cox regression model. There were 30 differentiated and 12 undifferentiated adenocarcinomas at pathology. At CEUS, 10 lesions were poorly vascularized, whereas 32 lesions were well vascularized. Positive correlation was observed between CEUS groups and tumoral differentiation (rs=0.51; p=0.001) and between CEUS and MVD (rs=0.74; p<0.0001). Median survival in patients with group A vascularization at CEUS was significantly lower than in group B (p=0.015). Cox proportional hazard model revealed the presence of poorly vascularized tumor at CEUS (p=0.0001) as a predictor of higher mortality. In conclusion, CEUS enables accurate depiction of the vascularization of adenocarcinoma, with positive correlation to histology grade and MVD.