Introduction There are few epidemiological data on the longitudinal follow-up of nodular lesions in cirrhotic patients.

Methods The Royal Free Hospital database was searched for all reports of MRI scans of the liver in which the term “nodule” was used. 630 such scans were identified in 369 individual patients between 1 January 2006 and 1 January 2011. Patients were then excluded if there was <1-year follow-up (45), if hepatocellular carcinoma (HCC) was identified on the initial scan (135), if an alternative diagnosis was made (129) or if no significant arterialised lesion was reported despite previous suspicion on ultrasound scanning (31). Retrospective analysis was, therefore, performed on 29 cirrhotic patients with a diagnosis of regenerative, indeterminate or dysplastic arterialised nodules at baseline and >1 year total follow-up with MRI and alpha-fetoprotein (AFP) surveillance.

Results The median age at first scan was 55 years (range 36–73) and the most common aetiologies of cirrhosis were hepatitis C (55%), hepatitis B (10%) and alcohol (25%) with 10% other. The median follow-up period was 22 months (12–56) and the median number of scans 4 (2–10). At baseline nodules were described as indeterminate in 48%, regenerative in 45% and dysplastic in 7%. The prevalence of HCC after 1, 2, 3 and 4 years of follow-up was 5 (17%), 9 (31%), 11 (38%) and 11 (38%) respectively, with the highest incidence occurring in the first 2 years of follow-up (82% of cases). The median size of nodule at baseline in those who developed HCC was 10 mm (5–18) and it was 9 mm (5–26) in those who did not. There was no association between the description of the nodule and the likelihood of developing HCC (five arising from indeterminate and 6 from regenerative nodules) and the two nodules initially described as dysplastic did not transform to tumour. AFP results were not informative.

Conclusion 31% of lesions described as arterialised nodules when first scanned developed into HCC within 2 years and initial radiological features consistent with “regenerative” nodules are not reassuring. Patients with discrete arterialised nodules should be enrolled in enhanced surveillance programmes.

Competing interests None declared.