Regenerative Hepatic Pseudotumor: A New Pseudotumor of the Liver

SUMMARY

Cases of new pseudotumor of the liver were collected from multiple medical centers. Four resection and 4 biopsy specimens were collected, including 4 women and 4 men with an average age of 48±15 years, range 28 to 73. The lesions were visible on imaging, but were either ill-defined or had indeterminate features for characterization. They ranged in size from 2 to 9 cm and were multiple in five cases. The resection specimens showed lesions that had vague borders but were visible in juxtaposition to the normal liver on gross examination. Histologically, the lesions also had ill-defined borders and were composed of benign reactive liver parenchyma. Central vein thrombi were seen in 5 cases and portal vein thrombi in 2 cases. These vascular changes were associated reactive parenchymal changes including sinusoidal dilation, patchy bile ductular proliferation, and portal vein abnormalities. All lesions lacked the histological findings of hepatic adenomas, focal nodular hyperplasia, or other known tumors and psuedotumors of the liver. In summary, this study provides a detailed description of a new pseudotumor of the liver: a reactive, hyperplastic mass like lesion that forms in association with localized vascular thrombi, for which we propose the term mass regenerative hepatic pseudotumor (RHP). This lesion can closely mimic other benign or malignant hepatic tumors on imaging and histology.

1. Introduction

Pseudotumors of the liver are non-neoplastic, mass forming lesions that can mimic a neoplasm in imaging studies, on gross examination, and sometimes on histology. Focal nodular hyperplasia and inflammatory pseudotumors are the most common pseudotumors, but there are a number of less common entities. Recently, a new type of pseudotumor was encountered, visible by imaging and on gross examination, but composed entirely of benign reactive parenchyma. The lesion was recognized in biopsy specimens and in resection specimens that were directed at liver lesions identified by imaging but that were negative for tumors and for known pseudotumors on histological examination, and instead show a distinctive pattern of benign reactive changes in response to vascular flow abnormalities. This non-neoplastic, mass forming lesion was briefly described in a few liver pathology text books [1, 2], but no comprehensive study has been performed. In order to clarify the clinical and histological findings, and in order to more fully establish histological criteria, cases of regenerative hepatic pseudotumors (RHP) were collected from multiple institutions.

2. Materials and Methods

After IRB approval, cases were collected from the consult and in-house files of three institutions. Because the pseudotumor did not have a formal name, a search of pathology files by diagnosis was not possible and natural language searches were not productive. Thus, cases were collected when they became known to the authors through either routine clinical sign-out, consult sign-out, or review of cases in QA conferences. Case number 3 was used to illustrate this entity in the prior pathology text books [1, 2]. Available clinical records and imaging reports were reviewed.

All cases were centrally reviewed. Routine stains performed at the time of sign out were examined. Cases were searched for vascular thrombi and their location noted. Sinusoidal dilation was scored as on a scale of 0–3: 0 = absent, 1= mild (limited to zone 3 sinusoids), 2 = moderate (involving zones 2 and 3 sinusoids) and 3 = marked (extending into zone 1). Bile ductular proliferation was also scored on a scale of 0–3; 0 = absent or minimal, 1 = mild (present in two or fewer portal tracts), 2 = moderate (present in more than two portal tracts or striking ductular proliferation in any portal tract), or 3 = marked (striking ductular proliferation in most portal tracts). Cases were also evaluated for portal tract approximation, defined as areas where portal tracts are abnormally close together, implying areas of parenchymal atrophy. Additional findings were recorded including lobular arteries, portal vein abnormalities (herniation, atrophy, dilatation), elastosis, and thickened hepatic plates on reticulin stain, the latter implying active regeneration.

3. Results

We identified 8 cases in 4 men and 4 women, with an average age at presentation of 48±15 years, range 28 to 73 years (Table 1). Collected cases included biopsies (N=4) and resections (N=4). In addition, one of the biopsies (case No. 8) had a follow up liver transplant for an underlying genetic disorder (Table 1).

Table 1.

Clinical and imaging findings in regenerative hepatic pseudotumors

Case	type	Age	gender	Clinical presentation	Size	Location (lobe)	Imaging findings including follow-up
1	Biopsy	63	M	Not available	2	Right	CT: lesion in right hepatic lobe US: no lesion visible Follow-up CT: no lesion visible CT with contrast: vague 2 cm lesion, biopsied
2	Resection	36	F	Not available	9.1, exophytic mass Background liver with numerous smaller lesions	Left	MRI with contrast: multiple lesions (10, 9, 6, 3, 2) with some intrinsic T1 signal and some peripherally increased T2 signal. Post contrast there was avid contrast enhancement on arterial phase and fade on portal venous phase; hypo intense on the hepatobiliary phase with Eovist. Radiology report favored inflammatory hepatic adenoma Follow-up: lesions get smaller over time when OCP are stopped
3	Resection	49	F	Imaging as part of follow up for stage II T3N0 rectal adenocarcinoma	2.9, 1.3	right	MRI without contrast: two lesions CTA with and without contrast (2 months later): slight growth in size CT without contrast (3 months later): stable in size.
4	Resection	29	F	Not available	?	Right lobe, segment 5	CT with contrast: 4 cm lesion that is centrally enhancing and peripherally hypodense on early phase; isodense on delayed series. DDX; FNH vs adenoma
5	Biopsy	73	M	Not available Hx of multiple myeloma	6	left	PET CT: focal increased tracer activity Non-contrast CT: not visible MRI (3 months later): 4.4 × 6 cm geographical region that is T1 hypointense and T2 isointense, with delayed enhancement
6	Biopsy	43	M	Follow up of stage T3 N1B colon carcinoma	1.2 Two other small lesions	Right lobe Segment 6	Hypointense on ultrasound. Hypodense lesion on CT. Hypointense on T1W, Isointense on T2W, Hypointense on hepatobiliary phase with Eovist. Diffuse fatty change in the liver parenchyma. Follow up CT and MRI showed reduction in size for all 3 lesions. No increased FDG with PET.
7	Biopsy	56	F	Follow-up for squamous cell carcinoma	4.3 Second smaller lesion	Right, segment 5	Irregular, low density on contrast enhanced CT Hyperintense on T2, mild restricted diffusion. Hypointense on arterial phase, isointense on portal venous phase and mildly hyperintense during delayed phase. No increased FDG uptake on PET. Hypointense on hepatobiliary phase with Eovist. Follow-up: stable at 5 years
8	Bx and F/U OLT	28	F	Chronic portal vein thrombosis, complement C3 disorder	Numerous, up to 2 cm	Throughout liver	T2 hyperintense, mildly hyperintense on DWI Triple phase CT shows arterial phase enhancing, hyperdense on portal venous phase and shows washout in delayed phase raising concerns for HCC

3.1. Clinical findings

Information on clinical presentation was limited, but most patients were asymptomatic and the lesions were discovered during evaluation for other abdominal disease concerns. For example, four of the patients had histories of malignancies and the lesions were discovered during routine surveillance for recurrent disease. Overall, five patients had potential hypercoagulable states, including the 4 persons with malignancies and 1 person with chronic portal vein thrombosis from a complement disorder (Table 1).

3.2. Imaging findings

In all cases, the lesions were visible on imaging, but were either ill-defined or had indeterminate features for characterization. For example, in case 1, the lesion was initially visible on non-contrast CT, but not on a follow-up CT or on an ultrasound. When the lesion was still present on a subsequent contrast enhanced CT, it underwent resection. The lesions ranged in size from 2 to 9 cm and were multiple in 5 cases. The lesions were located in the right lobe (N=5), left lobe (N=2), or throughout the liver (N=1).

Imaging did not show a uniform pattern of findings (Table 1), but in several cases suggested a differential diagnosis of adenoma or focal nodular hyperplasia. Eovist was performed in three cases and was hypointense in all cases.

3.3. Gross findings

On gross examination of the 4 resected specimens, the lesions consisted of vaguely nodular yellow-tan-red areas. No central scars were seen. None of the lesions had capsules and they blended into the liver at their edges, but were visible in juxtaposition to the normal liver on gross examination. There was no hemorrhage or necrosis.

3.4. Histological findings

The histological findings are summarized in Table 2. Figure 1 illustrates the histological findings in a biopsy specimen, Figures 2 and 3 illustrate the findings in resection specimens, while figures 4 and 5 show a biopsy specimen with follow-up resection.

Table 2.

Summary of histological findings in regenerative hepatic pseudotumors

Case	thrombi	Sinusoidal dilation (scale 0–3)	Bile ductular proliferation (scale 0–3)	Portal tract approximation	Lobular arteries	other
1	Central	2	1	mild	no	Associated with focal subcapsular scar Focal elastosis changes Regenerative changes on reticulin with mild plate thickening
2	Portal	2	2	mild	yes	Prominent arteries in PT
3	Portal	3	1	mild	no	Portal vein dilatation Focal portal vein herniation Focal portal tract elastosis
4	Central	2	1	no	yes
5	Central	0	1	mild	no	Portal veins atrophic with some missing
6	None	3	0	no	yes	Focal portal vein herniation Regenerative changes on reticulin with mild plate thickening
7	central	3	0	no	yes	Focal zone 3 hepatocyte drop out Regenerative changes on reticulin with mild plate thickening
8	Central on bx	3	2	equivocal	no	Prominent arteries in portal tracts Regenerative changes on reticulin with mild plate thickening

Figure 1.

Case 1, biopsy. Panel A. The lesion shows patchy but striking sinusoidal dilatation and congestion. Panel B. At low power magnification, there were areas of subtle increase in the number of portal tracts. In this image taken with a 20X lens, 4 closely approximated portal tracts are evident (arrows). Panel C. A thrombosed central vein. Panel D. A VVG shows the same thrombosed vein (same field as prior image). Panel E. The central vein is inconspicuous and shows elastotic change. Panel F. A VVG stain suggests this could be a remotely thrombosed central vein.

Figure 2.

Case 2, resection. Panel A. The lesion (upper image) blends almost imperceptibly into the background liver. A thin line has been added to separate the lesion from the background liver. Panel B. At low power magnification, there is striking but patchy sinusoidal dilatation. Panel C. Intralobular arteries were present. Panel D. Multiple vessels in the portal tracts are thrombosed. In some portal tracts, the vessels were significantly distorted as well as occluded. Panel E. A nearly occluded portal vein at higher magnification. Panel F. Many portal tracts showed bile ductular proliferation. 1G. Prominent arteries were present in many portal tracts. 1H. The background liver also showed nodular regenerative hyperplasia.

Figure 3.

Case 3, resection. Panel A. At low power, the lesion mimics an inflammatory hepatic adenoma. Panel B. At higher power magnification, however, the lesion has true portal tracts, instead of the faux portal tracts commonly present in inflammatory hepatic adenomas. Panel C. A remotely thrombosed portal vein is seen. Panel D. Many portal tracts have prominent, hyperplastic arteries.

Figure 4.

Case 8, needle biopsy. Panel A. At low power, there is marked sinusoidal dilatation. Panel B. A thrombosed central vein is present. Panel C. The portal tracts showed patchy bile ductular proliferation. Panel D. One of the portal tracts shows a portal vein with an unusual morphology.

Figure 5.

Follow-up resection on case 8. Panel A. At low power, the lesion (upper part of panel) blends imperceptibly with the background liver (lower part of image). Panel B. The lesion shows patchy, marked sinusoidal dilatation. Panel C. The portal tracts varied in their appearance throughout the lesion and were irregularly distributed. In this large field taken with a 2X lens, no portal tract were evident. Panel D. In this image, the portal vein is small, but the portal tract otherwise appears normal. Panel E. This portal tract has largely been resorbed, leaving only a remnant. Panel F. Other lesions in the background liver included three regenerative nodules with a solid appearance. Panel G. One pseudotumor lesion had at its periphery multiple foci of telangiectatic vessels. Panel F. A higher-magnification image of the telangiectatic vessels.

The lesions were visible at low power, had ill-defined borders, and were composed of benign liver parenchyma with interspersed portal tracts. Portal vein thrombosis was seen in two cases (Panel 2E, 3C) and central vein thrombi in five cases (Figure 1C, 1D, 4B). The thrombi were subacute to chronic with fibrosis. The biopsies also showed reactive parenchymal changes including sinusoidal dilation (N=7/8) (Figure 1A, 2B, 3A, 4A, 5B) and bile ductular proliferation (Figure 4C). Lobular arteries were readily identifiable on H&E in 4 cases (Figure 2C).

Portal tracts were present throughout the lesions. The portal tracts, however, showed portal tract approximation (N=4 cases) (Figure 1B) and irregular spacing in the surgical resection specimens (Figure 5C). In one resection specimen, small remnants of portal tracts were evident (Figure 5E). The arteries within the portal tracts were sometimes prominent (2 cases) (Figure 2G, 3D). None of the cases had intra-lesional bands of fibrosis.

Other findings included focal subcapsular scarring with mesothelial proliferation (N=1), focal elastosis (N=2), and regenerative changes on reticulin stains (N=4). At the cytological level, the individual hepatocytes within the lesions were similar to hepatocytes outside the lesion. In some cases, the hepatocytes were subtly atrophic compared to the non-lesional hepatocytes, but otherwise there was no differences in morphology and no cytological atypia.

3.5. Non-lesional background liver

None of the cases had fibrosis in the background liver by imaging or by histology. In two of the resection specimens, the portal tracts in the background liver had abnormal arterialization. One case (case No. 2) showed nodular regenerative hyperplasia in the background liver (Figure 2H).

3.6. Other specimens

In case 3, a prior biopsy was performed (not available for review) and signed out as nonspecific changes. An initial biopsy on case 4 was diagnosed as a hepatic adenoma (not available for review), but the resection was negative for an adenoma, instead showing RHP. In case 8, a prior biopsy (not available for review), was signed out as suggesting vascular outflow disease. A follow-up biopsy (index biopsy) showed a RHP and a subsequent follow-up liver transplant (for congenital complement disorder) showed a non-cirrhotic liver with diffuse vascular changes in the background liver, as well as focal telangiectasias (Figure 5G, H) associated with a small scar consistent with a biopsy site, three regenerative nodules (Figure 5F) and one RHP.

4. Discussion

Recognized pseudotumors of the liver [3–7] can be composed primarily of hepatocytes, inflammatory cells, inflammation/fibrosis, necrosis, or extracellular material such as amyloid (Table 3). This study describes an additional pseudotumor, regenerative hepatic psuedotumor (RHP). RHP are associated with vascular flow abnormalities, but they do not have the histological findings of focal nodular hyperplasia. The lesion is potentially more common than is evident by this study, as many cases are likely signed out by pathologists descriptively, as nonspecific changes.

Table 3.

Known pseudotumors in non-cirrhotic livers

Pseudotumor and representative reference	Brief description
Regenerative hepatic psuedotumor	Sinusoidal dilatation/congestion with vascular thrombi, regenerative changes, abnormal portal veins and or arteries.
Focal nodular hyperplasia	Nodules of hepatocytes surrounded by bands of fibrosis containing mild ductular proliferation, often with central scars. Eccentric thick walled vessels in central scars.
Segmental atrophy and nodular elastosis	Collapsed hepatic parenchyma/elastosis
Inflammatory pseudotumor	Inflamed fibrous tissue
Pseudolymphoma	Reactive lymphoid follicles
Sarcoidoma	Confluent granulomas and fibrosis
Amyloidoma	Composed of amyloid
Hypoxic pseudolobular necrosis	Necrotic hepatocytes

At low power, hepatic RHP show portal tracts, ill-defined borders, and lack intralesional fibrosis. The core histological findings reflect the presumed etiology of vascular flow changes, with variable degrees of portal vein abnormalities (atrophy, absence, herniation), portal artery abnormalities (hyperplasia), sinusoidal dilatation, lobular arterialization, and portal vein or central vein thrombosis. Other histological findings are also likely a reaction to the vascular flow changes, including patchy bile ductular proliferation [8] and focal elastosis [3]. The hepatocytes can show mild atrophy, but there is no atypia. The data in this study do not clearly indicate whether the thrombi are the cause of the lesion or a secondary finding.

RHP were identified in two settings. The first setting is that of prothrombotic risk factors, suggesting the lesions resulted from either portal or central vein thrombi. The second setting was that of livers with a background of vascular abnormalities. The etiology of these vascular abnormalities was not evident, but the diffuse nature of the findings suggested the condition affected the entire liver. In one case (follow-up liver explant, Case 8), other proliferative lesions (regenerative nodules) were also evident.

The differential for RHP includes non-mass forming lesions such as nodular regenerative hyperplasia, vascular outflow disease, and nonspecific changes next to an unsampled mass lesion. Nodular regenerative hyperplasia and vascular outflow disease are both diffuse patterns of injury, in contrast to the localized RHP. Thus, a diagnosis of RHP requires a radiographically or grossly visible geographic lesion. Nonspecific changes next to an unsampled mass lesion is harder to exclude on biopsy specimens, and a biopsy diagnosis should indicate the importance of making sure the lesion seen on imaging was adequately sampled. The presence of vascular thrombi would favor a RHP.

In terms of mass forming lesions, the differential can include focal nodular hyperplasia, inflammatory hepatic adenomas, or segmental atrophy/nodular elastosis, depending on the histological findings. Fully developed focal nodular hyperplasias have well defined parenchymal nodularity caused by bands of fibrosis, findings lacking in RHP. It is true that early focal nodular hyperplasias can lack fully developed fibrous bands [9], but they typically still have some degree of intralesional fibrosis. Early focal nodular hyperplasias also lack the vascular thrombi of RHP and generally do not have significant sinusoidal dilatation. Both RHP and focal nodular hyperplasia [10, 11] can have intralesional arterialization of the lobules and large thick-walled arteries. In focal nodular hyperplasia, however, the thick walled arteries tend to be larger and located in bands of fibrosis, not in portal tracts. Immunostains for glutamine synthetase are helpful, as RHP lack the typical map-like staining pattern seen in focal nodular hyperplasia.

Inflammatory hepatic adenomas commonly have sinusoidal dilatation and faux portal tracts with bile duct like structures [12], but they lack true portal tracts, do not have vascular thrombi, and have stronger and more diffuse CRP/SAA staining than the background liver. The pseudotumor segmental atrophy/nodular elastosis is also strongly associated with vascular thrombi but differs in its morphology and imaging findings [3, 13, 14]. The early stages are characterized by extensive parenchymal collapse, bile ductular proliferation, and mixed inflammation, while later stages show a progression towards no inflammation or ductular proliferation, with the histology dominated by extensive elastosis [3]. Elastosis was present in two cases of RHP, but only as focal findings.

Several earlier case reports and small cases series have described benign liver lesions that do not fit into the current classification system of hepatic tumors or pseudotumors and noted that vascular flow abnormalities were the likely cause, leading to localized hyperplastic and reactive changes that were visible on imaging [15–18]. While the histological descriptions are not always sufficient to be sure, we suspect that at least some of these represent RHP. We did not identify any cases arising in cirrhotic livers, but cannot exclude the possibility that related lesions can occur in cirrhotic livers. By analogy, focal nodular hyperplasia occurs in non-cirrhotic livers, but morphologically similar focal nodular hyperplasia-like lesions have been reported in cirrhotic livers [19, 20].

The natural history of RHPs are unclear, but based on limited information, at least some of them can be stable over time, while others can diminish in size on follow-up imaging. There is no histological or imaging evidence that they transform into focal nodular hyperplasia.

In summary, we describe a new hepatic pseudotumor of the liver that results from localized vascular flow changes, showing a constellation of findings that includes a geographic lesion on imaging and or gross examination, sinusoidal dilatation, portal vein and hepatic artery abnormalities, portal and or central vein thrombi, and associated parenchymal regenerative and reactive changes. The lesion is distinct from focal nodular hyperplasia and from segmental atrophy/nodular elastosis, the two other pseudotumors of the liver that are caused by vascular flow changes. RHP can mimic both hepatic adenomas and focal nodular hyperplasia, but careful attention to the histological findings leads to the correct diagnosis.