To compare the role of DWI vs. gadoxetic-acid-disodium enhanced MRI in the detection of colorectal hepatic metastases.
Methods
Fifty-four patients with 115 hepatic metastases were included in this retrospective study, approved by the Ethical Board. All patients underwent intraoperative-ultrasound and surgical resection within two weeks after MRI. Images were grouped in 4 sets, which were analyzed by two radiologists in different sessions: unenhanced T1-T2w (set A), set A plus DWI (set B), set A plus gadoxetic-acid-disodium (set C), set A plus DWI plus gadoxetic-acid-disodium (set D). For each set, metastases presence/size/site was reported. Interobserver agreement and statistical significance were assessed by Cohen’s kappa and Mc-Nemar’s test, respectively.
Results
Readers’ agreement was always very good (k > 0.80). Mean sensitivity values were 84.3/92.1/95.6/97.3% for set A/B/C/D, respectively. Mean specificity, positive predicted, negative predicted, and accuracy values strongly and progressively increased in the various set too: from 62.5% (set A) to 85.0% as for specificity, from 92.8% to 97.3% as for positive predicted value, from 41.0% to 85.1% as for negative predicted value, and from 81.1% to 95.5% as for accuracy.
For each reader from set A to D, the number of false negatives progressively decreases.
Conclusions
For both readers, DWI improved all statistical parameters in the unenhanced examinations, as for nodules either smaller or greater than 1 cm, while in the EOB-enhanced examinations DWI prevalently increased specificity/negative predictive value.
