The diagnostic performance of a simulated “short” gadoxetic acid-enhanced MRI protocol is similar to that of a conventional protocol for the detection of colorectal liver metastases

• Published in: European radiology 2021-04, Vol.31 (4), p.2451-2460
• Main Authors: Ghorra, Camille, Pommier, Romain, Piveteau, Arthur, Rubbia-Brandt, Laura, Vilgrain, Valérie, Terraz, Sylvain, Ronot, Maxime
• Format: Article
• Language: English
• Subjects: Aged; Colorectal cancer; Colorectal carcinoma; Colorectal Neoplasms - diagnostic imaging; Contrast Media; Diagnostic Radiology; Diagnostic systems; Gadolinium DTPA; Gastrointestinal; Histopathology; Humans; Imaging; Internal Medicine; Interventional Radiology; Lesions; Liver; Liver cancer; Liver Neoplasms - diagnostic imaging; Magnetic Resonance Imaging; Male; Medical diagnosis; Medical imaging; Medicine; Medicine & Public Health; Metastases; Metastasis; Middle Aged; Neuroradiology; Radiology; Retrospective Studies; Sensitivity and Specificity; Simulation; Ultrasound
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-020-07344-0

Objectives To compare the performance of standard and simulated short gadoxetic acid-enhanced MRI protocols for the detection of colorectal liver metastases (CRLM). Methods From 2008 to 2017, 67 patients (44 men (66%); mean age 65 ± 11 years old) who underwent gadoxetic acid-enhanced MRI during the initial work-up for colorectal cancer were included. Exams were independently reviewed by two readers blinded to clinical data in two reading sessions: (1) all acquired sequences (standard “long” protocol) and (2) only T2-weighted, diffusion-weighted, and hepatobiliary phase images (simulated “short” protocol). Readers characterized detected lesions using a 5-point scale (1—certainly benign to 5—certainly malignant). A lesion was considered a CRLM when the score was ≥ 3. The reference standard was histopathology or 12-month imaging follow-up. Chi-square, Student’s t , and McNemar tests were used for comparisons. Results A total of 486 lesions including 331 metastases (68%) were analyzed. The metastasis detection rate was 86.1% (95% CI 82–89.4)–86.7% (82.6–90) and 85.8% (81.6–89.2)–87% (82.9–90.2) with the short and long protocols, respectively ( p > 0.99). Among detected lesions, 92.1% (89.1–94.4)–94.8% (92.2–96.6) and 84.6% (80.8–87.7)–88.8% (85.4–91.5) were correctly classified with the short and long protocols, respectively ( p = 0.13 and p = 0.10). The results remained unchanged when lesions scored ≥ 4 were considered as CRLM. Conclusion The diagnostic performance of a simulated short gadoxetic enhanced-MR protocol including T2-weighted, diffusion-weighted, and hepatobiliary phase appears similar to that of a standard long protocol including dynamic phase images. Since this protocol shortens the duration of MR examination, it could facilitate the evaluation of patients with colorectal liver metastases. Key Points • The detection rate of colorectal metastases with a simulated, short, MRI protocol was similar to that of a standard protocol. • The performance of both protocols for the differentiation of metastases and benign lesions appears to be similar. • A short MR imaging protocol could facilitate the evaluation of patients with colorectal liver metastases.