Purpose: To compare the diagnostic performance of hepatobiliary phase (HBP) MRI sequences (HBP-MRI) versus complete-protocol MRI using gadoxetic acid, for secondary surveillance in patients with hepatocellular carcinoma (HCC) after curative-intent microwave ablation (MWA).Materials and methods: This IRB-approved and HIPAA-compliant retrospective cohort study comprised 46 patients [average age 66 years; 39 males and 7 females] treated with MWA for HCC at a tertiary care hospital from January 2017 to December 2019. Thirty-one patients had recurrence, and 15 additional cases without recurrence were randomly selected as a control group. All patients underwent complete-protocol gadoxetic acid-enhanced MRI every 3 months after MWA. After anonymization, three board-certified abdominal radiologists independently reviewed one pre-MWA complete MRI, and one post-MWA complete MRI. For all subsequent follow-up MRIs, the radiologists reviewed an abbreviated protocol consisting of only HBP and T2-weighted SSFSE sequences [223 abbreviated MRIs]. The identification of recurrence with the abbreviated protocol was compared with that using the reference standard complete-protocol MRIs. Recurrences, as identified in the full MRIs, were categorized as: local recurrence (local-R) – indicating active disease at the site of ablation as per the CT/MR LI-RADS v2018; intrahepatic distant recurrence (ID-R) - representing any new intrahepatic lesion ≥ 1 cm classified as LR-5 according to LI-RADS; and extrahepatic recurrence (E-R) – which included peritoneal nodules ≥ 1 cm or lymph nodes ≥ 2 cm in their short axis (defined by mRECIST guidelines).Results: Out of 46 patients, 31 (67.4%) had HCC recurrence. Within this group, 12 (38.7%) had a local-R, while 19 (61.3%) had non-local-R, which included ID-R in 17 (89.5%) patients and E-R in 2 (10.5%) patients. The remaining 15 (32.6%) patients showed no signs of recurrence. The Interreader agreement was notably high (κ = 0.67). The overall diagnostic performance of the abbreviated MRI protocol yielded a sensitivity (SEN) of 74% (95% confidence interval (CI) of 66.9%-79.3%), a specificity (SPE) of 92.2% (95%CI: 89.5%-94.2%), and a negative predictive value (NPV) of 91.3% (95%CI: 90.2%-93.4%). In detecting local-R, abbreviated MRI showed higher diagnostic accuracy with a SEN, SPE, and NPV of 89.6% (95%CI: 81.2%-93.7%), 87.7% (95%CI: 84.2%-89.5%), and 90.9% (95%CI: 85%-94.4%), respectively. For non-local-R sub-cohort, the SEN, SPE, and NPV were 70.6% (69.7%-71%), 91.1% (86.4%-94.4%), and 82.6% (82%-83.6%), respectively.Conclusion: Abbreviated MRI protocol, which includes hepatobiliary phase and T2-weighted SSFSE sequences, shows a significant negative predictive value and is particularly effective for identifying local recurrent HCC lesions. This suggests that it could be an acceptable secondary surveillance method for monitoring HCC patients following MWA.

Acknowledgements

No acknowledgement found.

References

No reference found.