abarabasch@ukaachen.de

To investigate whether an abbreviated liver-MRI protocol (AP), consisting of axial T2w-TSE, DWI and pre-contrast T1w-GRE in- and opposed phase (IP/OP) is appropriate to detect and classify liver metastases.

Ongoing reader study on so far 71 consecutive patients with solid tumors who underwent liver-MRI to search for metastatic disease. All patients underwent a standardized full protocol (FP) hepatic MRI at 1.5T with axial and coronal T2w-TSE ± fat-saturation (fs), axial T1w-GRE (IP/OP), axial and coronal DWI (b=0, 50, 800), multi-phase T1w-dynamic GRE before and after i.v. gadoxetate disodium, and axial T1w-GRE + fs in the hepatobiliary phase. Body radiologists with between 9 and 12 years of experience first reviewed only the abbreviated protocol, and made their diagnosis regarding the presence or absence of focal liver lesions, as well as to the likelihood of malignancy on a 5-point scale. Results of the full MRI-protocol as well as imaging follow-up or histopathology of liver lesion was used as ground truth.

Acquisition time for the AP was 10.3 min., vs 39.8 min. for the FP. Based on the interpretation of the AP images, reader 1 identified all 56 patients with focal liver lesions that were also identified by reading images of the FP; reader 2 identified 54 of the 55 patients with focal liver lesions identified by the respective FP readings. Accordingly, the sensitivity to identify patients with focal liver lesions based on the AP- vs. the FP-readings was 100% (for reader 1) and 98.2% (for reader 2). Regarding characterization of focal liver lesions (positive predictive value), reader 1 had one additional false-positive diagnosis with AP compared to FP, for a PPV of 92% vs.. 94% (35/38 vs. 34/36); 95%-CI: 79 %-98 % vs 81% -99%. Reader 2 had 3 additional false-positive diagnoses with AP vs. FP, for a PPV of 88% vs. 97% (30/34 vs. 33/34); CI: 73%-97% vs. 85%-100%). Average time to read the AP was 43sec. for reader 1 and 72sec. for reader 2.

An MRI acquisition time of 10.3 min. and a reading time between 43 and 72 sec. is sufficient to identify patients with malignant focal liver lesions; it offers identical sensitivity and similar specificity and PPV as does a full, dynamic contrast enhanced liver MRI protocol that included hepatobiliary phase imaging.

Abbreviated liver MRI seems useful to allow fast MRI-screening for metastatic spread to the liver in cancer patients.