Spontaneous regression of focal nodular hyperplasia

Unusual presentation of more common disease/injury

Spontaneous regression of focal nodular hyperplasia: a pathological report H Laumonier,1 F Leblanc,2 C Balabaud,3 P Bioulac-Sage4 1

Department of Radiology, Hôpital St André CHU Bordeaux, Bordeaux, France Department of Surgery, Hôpital St André CHU Bordeaux, Bordeaux, France 3 Hôpital St André CHU Bordeaux, Bordeaux, France 4 Department of Pathology, Hôpital Pellegrin CHU Bordeaux, Bordeaux, France 2

Correspondence to H Laumonier, [email protected]

Summary Focal nodular hyperplasia (FNH) is a benign tumour of the liver, occurring in 0.6–3% of the general population. Most lesions are diagnosed incidentally. With the increasing use and improvement of diagnostic imaging, FNH is being observed more often. It has been shown, using radiological approaches, that most FNH remain stable, or even regress, over a long follow-up period. In addition, it is extremely rare that FNH were discovered in elderly. However, to our knowledge, there is no pathological report illustrating the regression of FNH. We report here a case showing the pathological changes occurring during the regression phase of FNH, with dense fibrous tissue mixed with arteries replacing hepatocytes nodules.

BACKGROUND Focal nodular hyperplasia (FNH) is a benign tumour of the liver, occurring in 0.6–3% of the general population.1 It consists of circumscribed nodules of the hyperplastic parenchyma surrounded by more or less continuous fibrous bands radiating from a stellate, fibrotic scar containing dystrophic vessels, mainly thick-walled arteries. Most lesions are diagnosed incidentally. Surgery is not recommended, except in few cases when the tumour compresses neighbouring organs, liver vessels or bile ducts. With the increasing use and improvement of diagnostic imaging, FNH is being observed more often. In the past, FNH has been one of the most challenging hepatic lesions for radiologists; however, recent technical advances in imaging have improved diagnostic accuracy.2 3 In rare cases, when imaging and liver biopsy have failed to confirm the diagnosis, final diagnosis is achieved only at the time of surgery.4 However, recently the pathological diagnosis of FNH has been greatly facilitated by the use of glutamine synthetase immunostaining.5 6 Follow-up studies have been performed only in small series of patients. It has been shown using radiological approaches that most FNH remain stable or even regress over a long follow-up period.7–9 In addition, it is extremely rare that FNH were discovered in elderly. However, to our knowledge there is no pathological report illustrating the regression of FNH.

CASE PRESENTATION A 60-year-old woman was admitted to a local hospital for an acute abdominal pain on 18 February 2009. On CT scan, a large subcapsular haematoma was discovered. There was no haemorrhagic shock. She was admitted to our digestive intensive care unit. This overweight patient (body mass index 29) was treated for hypertension for a short period of time when she BMJ Case Reports 2010; doi:10.1136/bcr.02.2010.2704

was 16 years old and then since 1994. She was also treated for high cholesterol level (past 10 years) and hypertriglyceridaemia (past 4 years). There was a family history of heart diseases (her mother, the father and brother of the mother; her brother had died at a young age of heart infarcts or arterial hypertension). She smoked 1 pack of cigarettes/day (for 42 years) and took contraceptive pills for 20 years. She was menopaused at the age of 45 with a substitutive hormonal treatment until 2000. In 2005 she had an MRI showing at least three FNH (figure 1), one located in the right lobe (58 mm), and two (40 and 30 mm) located in the left lobe, as well as liver steatosis. The three nodules had characteristic MRI pattern of FNH with strong arterial enhancement and typical central scar in particular. On the last MRI (2009) there was another tumour with strong arguments for a hepatocellular carcinoma (HCC) (data not shown) with an intrahepatic rupture (figure 2) and subcapsular haematoma. The three FNH were still present (figure 2) with a significant shrinkage (segment VII: 41 mm vs 58 mm diameter in 2005, segment II: 21 mm vs 40 mm, segment II (data not shown): 16 mm vs 30 mm). Furthermore in 2009, MRI features of the segment VII lesion had changed and were no more typical of FNH. This lesion was in particular heterogenous in T2W with an atypical hypointense central scar and did not exhibit the characteristic strong arterial enhancement after gadolinium-contrast injection. In the portal venous and delayed gadoliniumenhanced sequences, the lesion showed progressive enhancement suggestive of fibrous component. The patient underwent a right hepatectomy and a segment II tumorectomy on 26 February 2009. The follow-up was uneventful except for a parietal abcess.

1 of 8

Figure 1 2005 MRI showing typical FNHs. (A) Phase opposed T1 image showing nodules in segments VII (58 mm) and II (40 mm), appearing hyperintense in a steatotic liver. (B) Fat suppressed T2W scan showing both hyperintense lesions, with a high signal intensity linear central area corresponding with the scar (arrows). (C) Gadolinium-enhanced fat-suppressed T1W sequence (arterial phase) showing strong enhancement of the left-lobe lesion (arrow) (D) showing strong enhancement of the segment VII lesion with hypointense central scar (arrow). (E,F) Gadolinium-enhanced fat-suppressed T1W image (delayed phase) showing persistent hyperintensity of the three lesions due to low signal intensity of the steatotic liver parenchyma and sustained enhancement of the central scar.

Figure 2 2009 MRI. (A) Fat-suppressed T2W image shows a large subcapsular haematoma and shrinkage of the segment VII FNH (41 mm) which appears heterogenous with atypical hypointense central scar (arrow). (B) Gadolinium-enhanced fat-suppressed T1W sequence (arterial phase):lack of arterial enhancement of the segment VII nodule and shrinkage of the segment II lesion (21 mm) which still exhibits strong contrast uptake. (C) Gadolinium-enhanced fat-suppressed T1W image (delayed phase) showing sustained enhancement in both lesions but without a clearly defined central scar.

2 of 8

BMJ Case Reports 2010; doi:10.1136/bcr.02.2010.2704

Figure 3 (A,B) Fresh resected specimen. (A) Cut section of the haematoma and part of the hepatocellular carcinoma (HCC); (B) cut section of the shrunked FNH; the haematoma is also visible (right).


3 of 8

Figure 4 The FNH is made of two parts: the fibrotic pink part in the middle limited by the dotted line and the peripheral part corresponding to a more classical aspect of FNH. The fibrotic parts contain mainly arteries (arrows).

Figure 5

4 of 8

The fibrotic zone (upper part) is devoid of glutamine synthetase (GS) staining, compared to the rest of the FNH.


Figure 6

There is no cytokeratin 7 (CK7) staining in the fibrotic area (upper part) compared to its periphery.

Figure 7

In the fibrotic area, there are many arteries with thick wall and narrow lumen. SMA, smooth muscle antibody.


5 of 8

Figure 8

No ductular reaction was observed in the area devoid of hepatocytes and with numerous arteries.

Figure 9 Cut section of segment three FNH (seen in the upper right corner). Glutamine synthetase (GS) well delineates the nodule. Compared to GS localisation in the non-tumoral liver, restricted to a few hepatocytes rows around hepatic veins, in FNH the staining is more diffuse, respecting the fibrous area.

6 of 8


Pathological results The resected liver portion had a weight of 2000 g. In the right lobe, the large subcapsular hematoma (15 × 10 cm) was in contact with a subcapsular multinodular necrotico haemorrhagic poorly limited tumour (figure 3). This tumour was a grade 3–4 Edmonson HCC (data not shown). Nearby the HCC, there was another whitish/brown, hard tumour (4 × 3 cm) (figure 3). At the microscopic level, part of the tumour was a recognisable FNH, with hepatocytes nodules surrounded by fibrotic bands, inflammatory infiltrate and mild ductular reaction (figures 4 and 5), and with a typical GS immunostaining (figure 6). The other part of the tumour was more fibrotic (figures 4 and 5), with numerous arteries exhibiting extremely thick walls and a very narrow lumen (figure 7). In this area devoid of hepatocytes, glutamine synthetase staining was negative (figure 6), and there was no ductular reaction (figure 8). In segment II, there was a typical FNH (figure 3) with large steatotic areas (data not shown) and the characteristic GS staining (figure 9). The non-tumoral liver was steatosic (80%) (data not shown).

DIFFERENTIAL DIAGNOSIS In this case, no differential diagnosis was discussed. In the absence of a previous diagnosis of FNH, the other diagnosis to be discussed on the MRI were liver cell adenoma and metastasis.

OUTCOME AND FOLLOW-UP In May 2010, the patient was in good health.

DISCUSSION

▲ ▲


Learning points ▲

Although FNH regression has been described in a couple of case reports,10–14 there are only few data describing this phenomenon at the pathological level.7–9 In one study,9 the volumes of FNH were found stable in six cases, decreased in 10 cases and increased in two cases, using CT and MRI observation after a mean follow-up period of 37.6 months. In another report where sonographic follow-up was carried out,7 it was reported that size reduction of FNH occurred in 8 of 16 patients after a mean followup period of 33 months. In the last published series of patients,8 with the longest follow-up period (a mean of 42 months), most lesions (70.6%) were stable in size; only 1 lesion (2.9%) increased, and the others (26.5%) decreased or even disappeared over time. FNH more frequently occurs in women between the ages of 20 and 50.4 Because of the female predominance and the young age at onset, a role of female hormones has been suggested, but definite evidence has not been provided yet. It has been considered previously that oral contraceptive use and pregnancy might be associated with the enlargement of FNH; however, one study indicated that use of contraceptive pills did not increase the incidence of FNH.15 Kuo et al8 investigated the possible factors associated with the evolution in FNH. Although there was no difference between FNH with or without regression, the significant factors associated with disappearance of FNH were older age (p < 0.05), nodule diameter