Pyogenic granuloma-like Kaposi sarcoma: Case report and review of literature

Pyogenic granuloma-like Kaposi sarcoma (PGLKS) is a recent addition to histological variants of Kaposi sarcoma (KS). As the name implies, the appearance closely simulates benign pyogenic granuloma, both clinically and histologically, thus carrying a risk of misdiagnosis. PGLKS has been reported in both human immunodeficiency virus (HIV)-positive and negative patients. We report a case of PGLKS in a HIV-positive patient seen in our institution and performed a literature search on SNOMED, PubMed Central and Google Scholar databases for other reported PGLKS cases thus far. To the best of our knowledge, this is the first PGLKS case to be reported in Malaysia.


INTRODUCTION
Pyogenic granuloma-like Kaposi sarcoma (PGLKS) is a recent addition to histological variants of Kaposi sarcoma (KS). As the name implies, the appearance closely simulates benign pyogenic granuloma, both clinically and histologically, thus carrying a risk of misdiagnosis. PGLKS has been reported in both human immunodeficiency virus (HIV)-positive and negative patients. We report a case of PGLKS in a HIV-positive patient seen in our institution and performed a literature search on SNOMED, PubMed Central and Google Scholar databases for other reported PGLKS cases thus far. To the best of our knowledge, this is the first PGLKS case to be reported in Malaysia.

CASE PRESENTATION
A 28-year-old Malaysian man of Indian descent presented to the Dermatology Clinic with a two-week history of a swelling on his right 4 th toe, which bled on trauma. He was HIV-positive with CD4 count of 27 cells/mm 3 and was on highly active antiretroviral therapy (HAART) but had defaulted treatment. He was previously seen in the Dermatology Clinic for molluscum contagiosum on his face, but also defaulted follow-up.
Examination revealed a solitary lesion 2x2cm with firm, lobulated surface, clinically felt to be a pyogenic granuloma. The lesion was excised and sent for histopathological examination. Histopathological examination revealed a nodular, polypoid skin with surface hyperkeratosis and focal ulceration. The dermis was replaced by a cellular lesion composed of proliferated vascular channels, predominantly thin walled vessels, with slit-like anastomosing pattern and an epidermal collarette ( Figure 1). These were admixed with small arteriole-like vessels and occasional ectatic vessels. The vessels were lined by endothelial cells showing mild to moderate atypia. In other areas, the proliferation was almost solid and was composed of spindle cells with no obvious lumina ( Figure 2). Mitotic figures were easily identifiable (average of 6 in 10 high power fields). Extravasated red blood cells and haemosiderin pigments were abundant. Scattered eosinophilic globules were also seen ( Figure 3). The lesion appeared to extend to the deep excision margin. Immunohistochemistry showed positivity for CD34 and CD31 as well as for human herpesvirus 8 (HHV-8; Figure 4). The morphological features and immunohistochemical profile were consistent with pyogenic granuloma-like Kaposi sarcoma (PGLKS). Low-power view illustrating a polypoidal dermal vascular lesion with an epidermal collarette (black arrow). The dermal vascular proliferation showed a mixture of slit-like vascular channels with spindle cell proliferation and small capillary-sized and arteriole-like vessels, as well as ectatic vascular channels (black arrowheads). There was variation in the size and shape of the vascular channels containing red blood cells (black arrows), with proliferation of spindle cells with no obvious vascular lumina (black arrowheads). Red cell extravasation was prominent (yellow arrows). Mitoses were also seen (red arrows). There was mild to moderate nuclear atypia (black arrows), red blood cell extravasation and eosinophilic globules (black arrowheads), which are typical histopathological features of Kaposi sarcoma.   [16,17]. While a definitive conclusion about the cellular origin of KS remains elusive, theories of cell origin include reprogramming of HHV-8-infected blood vascular endothelial cells to lymphatic endothelial cells and uncommitted endothelial progenitor cell [16]. This partly explains the immunophenotype of the neoplastic cells within KS, which show overlapping lymphatic and vascular endothelial cell differentiation, thus precluding definitive conclusion drawn on the cell of origin.
The identification of HHV-8 in KS serves as the best supportive finding in the diagnosis of KS and its variants to date, thus discriminating it from other entities. In general, the main non-KS histopathological differential diagnosis in this setting may include other vascular lesions such as PG, bacillary angiomatosis, spindle cell haemangioma and angiosarcoma; the latter two may show predominant spindled morphology. Specifically referring to the case presented, the constellation of nodular/polypoid architecture with epidermal collarette, surface ulceration, vascular proliferation and a spindled component without overt atypia raised the main differential diagnoses of PG and PGLKS. Bacillary angiomatosis was considered less likely, given the lack of bacterial colonies and associated inflammation despite its association with HIV positivity. Spindle cell haemangioma usually contains vacuolated and epithelioid endothelial cells, and angiosarcoma typically shows overt endothelial atypia; both were absent in this case. When faced with a diagnostic dilemma, nuclear positivity for HHV-8 by immunohistochemistry (IHC) provides the most reliable supportive evidence for a diagnosis of PGLKS. We have also demonstrated HHV-8 positivity by immunohistochemistry in the current case. Given the wide availability and accessibility of the IHC, it quickly becomes preferable to the more tedious HHV8 DNA polymerase chain reaction (PCR) analysis. In addition, there is also a higher rate of false positivity linked to the latter [12].
In addition to HHV8 IHC, other vascular markers (CD31, CD34 and D2-40) have also been used to aid diagnosis to some extent. D2-40 is typically positive in KS cases, providing evidence of lymphatic differentiation of the neoplastic cells. D2-40 negativity however, does not exclude the diagnosis of PGLKS as previously illustrated by Cabibi D et al [6]. They hypothesized that the D2-40 negativity of the vessels may be due to the fact that this lesion represents a very early stage of the infection; in this setting, while more mature endothelial markers are focally and progressively lost, the lymphatic differentiation of endothelial CD34-positive precursors has not occurred yet. Alternatively, D2-40-negative PGLKS may represent HHV-8 infection of a pre-existing pyogenic granuloma [6].
In a HIV-positive individual, KS is an acquired immune deficiency syndrome (AIDS)-defining illness. AIDS-associated KS (AIDS-KS) commonly arises in the setting of low CD4 count and typically manifests as a disseminated disease. It is interesting to note that more than half of the PGLKS cases reported thus far are HIVnegative and show a predilection in males. Whether there are additional risk factors that predispose to the development of PGLKS in this cohort remain to be identified. In HIV-negative individuals and HIVpositive individuals on HAART, the behaviour of KS may be more indolent [17]. However, being a relatively new entity, the true behaviour of a localized PGLKS is yet to be determined.
It is important to think of PGLKS when encountering a PG-like lesion to avoid missing the diagnosis. The fact that PGLKS can be seen in both HIV-positive and HIV-negative patients and the fact it closely mimics typical PG clinically and on histopathological examination makes the diagnosis even more challenging. Particular attention to histomorphology, as well as utilizing ancillary studies where necessary will allow accurate identification of this uncommon lesion.

CONCLUSION
PGLKS is a rare entity with shared histological features of pyogenic granuloma and Kaposi sarcoma. Diagnosis relies on histopathological examination as these are often unsuspected clinically. Ancillary studies, especially HHV8 immunohistochemistry are useful in supporting its diagnosis.

Conflicts of interest
None.