Making Sense of ‘Porto‐Sinusoidal Vascular Disorder’: What Does It Mean for the Pathologist and the Patient?

Abstract

Histological features that occur in the non-cirrhotic liver of patients with portal hypertension (PH), as in idiopathic non-cirrhotic portal hypertension (INCPH), include portal and periportal vascular alterations and architectural changes, most notably nodular regenerative hyperplasia (NRH) [1, 2]. In recent years, it has been noted that the same constellation of features can been seen in liver biopsies from non-cirrhotic patients without PH [3]. Recognition of this patient group appeared to create a ‘diagnostic gap’ which the vascular liver disease interest group (VALDIG) sought to address this by proposing an overarching concept of porto-sinusoidal vascular disease (PSVD) [3]. They suggested that the diagnosis of this ‘novel entity’ could be made using a simple algorithm which is based on the presence of specific or non-specific clinical signs of PH and specific or non-specific histological features of PSVD. The schema requires exclusion of cirrhosis by an adequate liver biopsy; PSVD can then be diagnosed by the presence of either one specific clinical sign of portal hypertension (varices, PH bleeding, porto-collaterals on imaging) or one ‘specific’ histological lesion (obliterative portal venopathy [OPV] or portal vein stenosis, NRH, incomplete septal fibrosis/cirrhosis). Furthermore, they suggested that the diagnosis could be made with the combination of one non-specific sign of portal hypertension (ascites, platelets < 150 000 mm3, spleen ≥ 13 cm) and one non-specific histological lesion (portal tract abnormalities including herniated portal vein branches and hypervascularised portal tracts, architectural disturbance such as irregular distribution of tracts and central veins, non-zonal sinusoidal dilatation and mild perisinusoidal fibrosis). It is to the great credit of the VALDIG consortium that it has tackled this important and challenging topic. However, we and others [4, 5] have concerns that trying to group all components of the spectrum together as ‘PSVD’ runs the risk of portraying PSVD as a distinct disease entity. Indeed, the term PSVD has been variably adopted by the scientific community but variably interpreted. From our PubMed search, since the 2019 VALDIG paper, there have been 43 clinical research publications or case reports that have used this term, although interestingly only 65% specifically applied the VALDIG criteria and pathologists were only involved in 74% of the studies. The VALDIG authors [3] acknowledged that there were potential disadvantages with their approach, in particular that it may be too simplistic and decrease accuracy of diagnosis. There are several areas of weakness in their algorithm. First, it explicitly excludes cases in which there are hepatic vein abnormalities despite the well-known overlap clinically and histopathologically of such abnormalities with the various conditions that form PSVD. Furthermore, despite the name, PSVD does not include sinusoidal obstruction syndrome on the basis that it is a microvascular disease. Second, the basis for attributing features that comprise either specific or non-specific histological lesions is somewhat arbitrary and not based on any pathogenetic considerations. An earlier paper by Aggarwal et al. [6] proposed that several of the changes described as non-specific by VALDIG were components of OPV itself. Third, including incomplete septal fibrosis/cirrhosis as a specific histological lesion is difficult to understand, given that this is a change which can occur in a myriad of chronic liver diseases in which there has been fibrosis regression or remodelling. The inclusion of mild perisinusoidal fibrosis is also challenging as this is also seen in a variety of conditions, most notably steatotic liver disease, which is not excluded from PSVD as a concurrent condition. We would suggest that if it is used in this context it might be better to refer to mild ‘diffuse or non-zonal’ perisinusoidal fibrosis associated with non-zonal sinusoidal dilatation. Fourth, Premkumar and Anand [5] have argued that geographical and disease population considerations need to be taken into account for clinical validation before adopting the PSVD concept. Given these issues, and that liver biopsy is essential for the diagnosis of PSVD, we suggest that it is timely to redefine the histological features of PSVD, at least as it applies to diagnostic histopathology in routine practice. We recognise, however, that there may be merit in maintaining the VALDIG algorithm for much needed future prospective long-term follow-up studies to ensure comparability across different cohorts and studies. We suggest that PSVD should consistently refer to porto-sinusoidal vascular ‘disorder’ rather than ‘disease’ as the latter term describes an abnormal condition affecting the structure or function of part of the body, usually associated with specific signs and symptoms. Adoption of ‘disorder’ in PSVD terminology by various groups has increased in recent years including the VALDIG group [7, 8]. We also suggest that the list of histological components of PSVD should be streamlined to (i) exclude incomplete septal fibrosis and mild perisinusoidal fibrosis, (ii) adopt the terminology of Guido et al. [2] for example, portal vein stenosis rather than OPV and (iii) remove the distinction between specific and non-specific features (Table 1). When dealing with biopsies from patients with possible PSVD, the presence or absence of all five features should be described in the report. Portal vein stenosis Portal and periportal vascular abnormalities Nodular regenerative hyperplasia Other architectural abnormalities: Irregular distribution of portal tracts and hepatic veins Non-zonal sinusoidal dilatation and/or capillarisation (CD34 positivity) with or without perisinusoidal fibrosis A significant problem is that some of the histological features are subtle. One study [9] demonstrated significant inter-observer variation for lobular and portal-based features with Kappa values ranging from only 0.12 to 0.32. This has raised concerns that we may under-recognise PSVD [9]. However, our experience with referral practice is that conversely that the subtle nature of the changes (e.g. portal vein herniation) can lead to over-reporting. There is an urgent need to define thresholds for documenting the presence of all five histological changes which could be facilitated by the generation of a widely available atlas of key features with clear definitions to improve consistency. How then should the surgical pathologist approach liver biopsies in which there is a possibility of PSVD? There are several settings in which this may occur. First, the biopsy may have been done because of clinical evidence of PH and to establish whether or not there is cirrhosis. The second is in the investigation of patients clinically considered to have probable INCPH. In such cases, where histological changes of PSVD are present, these should be documented. Our recommendation is that in cases where there is unequivocal portal vein stenosis or NRH in particular, they should be reported as showing features that fall within the spectrum of PSVD, with a recommendation that further clinical studies should be undertaken to confirm or exclude PH. If only one or more of the other lesions are present, the report should indicate that these features can be seen in the spectrum of PSVD, but in the absence of definite PH, their clinical significance is currently uncertain. Third, the biopsy may have been undertaken during the investigation of asymptomatic abnormal liver function tests, particularly cholestatic changes. This is a not uncommon background for PSVD without PH [10]. Again the presence or absence of the various histological vascular lesions should be clearly documented and may be a trigger for additional investigations given that PSVD without PH may be associated with pro-thrombotic conditions or other systemic conditions such as immunological disorders [11]. One of the main issues with the use of the term PSVD in patients without definite PH is the uncertainty about its clinical relevance. Data are sparse to clarify the longer term significance in such cases. Three recent studies [11-13] have suggested that there is either slow or no progression of PSVD in the absence of portal hypertension. In a recent paper in Liver International, Zhang et al. [12] showed none of their 131 adult patients with PSVD, but without definite PH was found to have developed PH or associated complications after a median of almost 4-year follow-up. Di Giorgio et al. [13] reported similar findings in a smaller paediatric cohort. Nonetheless, abnormal liver function tests (aminotransferases or gamma-glutamyl transpeptidase) persist in a proportion of patients [11, 12]. In the absence of more substantial long-term studies, and notwithstanding the difficulties with the VALDIG algorithm, we concur with Kmeid et al. [4] that there is benefit in separating cases of those with INCPH from PSVD without PH to facilitate such outcome studies, but this will require refinement and standardisation of the features used to make the latter diagnosis. In summary, pathologists should recognise that PSVD is not a single disease entity. Instead, it is a composite of several abnormal histological vascular and architectural features. Specifying the observed histological features (such as NRH, portal vein stenosis and others) is more informative and accurate than designating PSVD as a definitive diagnosis. All authors have approved the final version of the manuscript and have given their consent for publication. The authors declare no conflicts of interest. The authors have nothing to report.

Topics & Concepts