Pigmented villonodular synovitis (PVNS) of the knee joint: magnetic resonance imaging (MRI) using standard and dynamic paramagnetic contrast media. Report of 52 cases surgically and histologically controlled

Abstract

Purpose: Pigmented villonodular synovitis (PVNS) is a rare proliferative disorder of the synovial membrane, exhibiting benign behaviour from a biological point of view. This kind of synovial hyperplasia leads to the formation of villi and nodules characterized by deposit of intracellular haemosiderin. It primarily involves young adults, the peak age being between the second and fourth decade of life. It may appear either in a diffuse or a localized (nodular) form. The joint most affected is the knee and diffuse PVNS is the most common form. Diagnostic imaging techniques, particularly MRI, allow lesion identification, suggesting a diagnosis. However, such diagnosis can be confirmed only on histology as the final diagnosis of PVNS, and therefore the possibility of differential diagnosis with other haemorrhagic and chronic hyperplastic synovites, is based on the detection of intracellular haemosiderin components. The aim of this study is to evaluate the usefulness of MRI, which might be completed with the intravenous injection of contrast medium, in the characterization of such pathological picture.

Materials and methods: From January 1999 to December 2002, we evaluated 52 patients presenting knee swelling, pain and functional impairment. Only 19 patients had a history of trauma. All patients underwent MRI using a dedicated 0.2 T unit or a whole-body' 1.5 T unit. In 30 cases the baseline examination was completed with intravenous injection of contrast medium, followed by dynamic 3D-SPGR sequences at 45, 90, 135 and 225 seconds from the initial injection. These dynamic sequences were then processed by means of early and late subtractions, evaluating the regions of interest (ROI) positioned in the areas with higher post-contrast enhancement. Thirty-eight patients had been previously submitted to Ultrasonography (US), whereas twenty-five patients to Computed Tomography (TC). Later, all patients underwent surgery. Only two patients required an arthrotomy. We then retrospectively evaluated the imaging findings obtained, comparing them with the histological data. Patients affected by autoimmune and systemic inflammatory disorders were excluded from this study.

Results: The suspected diagnosis of PVNS was confirmed in 44/52 patients examined. CT examination allowed to detect the presence of a synovial proliferation with densitometric values ranging from 55 to 75 Hounsfield Units (HU) in all cases. In 11 cases, US examination revealed the presence of nodular hyperechoic structures surrounded by anechoic areas, with no differentiation between diffuse and nodular forms. Baseline MRI images showed no differential features among the various histological forms detected. In fact, the nodular structures demonstrated intermediate-to-low intensity signal in all sequences performed. Contrast enhanced MRI showed the presence of areas of inhomogeneous signal due to the increased intensity signal of hypervascular areas. The analysis of vascular dynamics demonstrated a characteristic exponential intensity/time curve both in diffuse and localized forms.

Discussion: The definition of pigmented villonodular synovitis was first employed by Jaffé in 1941 to describe the benign proliferative inflammatory nature of such pathology, characterized by a thickened and hyperplastic synovia organized into villi and nodules, leading to deposition of intracellular haemosiderin pigments. Presently, Authors prefer to include in hemorrhagic synovites all chronic and haemorragic synovial disorders, regardless of the aetiopathogenesis (rheumatoid arthritis, arthropathy secondary to haemorrhagic diathesis, chronic articular traumatism, haemangioma, synovial sarcoma). PVNS involves young adults, with no gender preference; it affects the knee joint in 66-80% of cases, with no typical symptomatology. The absolute absence of any characteristic feature makes a correct differential diagnosis difficult. So far, the only possibility to diagnose PVNS is based on the histological examination: presence of intracellular and subsynovial haemosiderin pigments, predominance of nodular structures as compared to villi, presence of macrophage multinucleate cells, production of collagen, mitotic cellular elements. Therefore, the possibility of characterizing PVNS using MRI is based on detection of a higher number of nodules as compared to villi, as the presence of haemosiderin is always characterized by low signal intensity on T2-weighted images, both intra- and extracellularly. New information on MRI semeiotics has come from the use of post-contrast enhanced dynamic sequences which are able to provide semi-quantitative data on CM velocity distribution within the hyperplastic areas. However, in our experience, dynamic-enhanced MRI did not provide any differential feature between PVNS and the other chronic hemorrhagic forms. Any inflammatory pathology leads to an increased capillary permeability with progressive deposit of CM in the area of interest. In all cases examined, the maximum deposit of contrast medium was observed in the extracellular phase, with a delayed wash-out.

Conclusions: PVNS of the knee presents a difficult differential diagnosis. In many cases, only MRI is able to identify the presence of haemosiderin precipitates within the nodules characterizing the lesion. The use of standard and dynamic contrast media seems unable to provide additional diagnostic information. Thus, the diagnosis still pertains to histology.