The present case is the second documented case of Kaposiform hemangioendothelioma (KHE) in the right tonsil of a 2-year-old child, who referred to us with a clinical history of episodes of tonsillitis with ipsilateral neck swelling, since birth.
During one of the clinical episodes, the patient had acute dysphagia and dysponea, wherein he was radiologically diagnosed with a peritonsillar abscess and therefore, underwent right-sided tonsillectomy, elsewhere. On review of histopathology slides, the differential diagnoses included a juvenile hemangioma, Kaposi's sarcoma, myofibromatosis and hemangiopericytoma. Presence of irregular, infiltrating lobules of spindle cells with a "kaposiform" pattern, forming slit-like, crescentic capillaries with platelet thrombi, eosinophilic bodies and prominent areas of lymphangiomatosis were helpful in differentiating it from a juvenile hemangioma [1
]. However, fortunately, the present case was not associated with KMP. Although, the platelet count was towards lower side of range, the patient did not present with features of life threatening thrombocytopenia and or anaemia. Even though a Kaposiform hemangioendothelioma is known to occur with KMP, this association has been noted in 42% cases, in a study by Lyons et al [2
], wherein the authors documented 8 cases in the head and neck region, including 4 cases associated with KMP, while 3 cases unassociated with KMP. None of the cases in that study was noted in the tonsil region. KMP is more commonly seen in cases occurring in abdominal than somatic sites. Lately, Gruman et al [5
] have also documented 10 cases of Kaposiform hemangioendothelioma, unassociated with KMP, including 3 cases in head and neck region, but none in the tonsil. Despite a Kaposiform pattern of tumor cells, including eosinophilic bodies, a Kaposi's sarcoma was ruled out in view of presence of several dilated lymphatic channels, focal capillary formation, lack of nuclear atypia and mitosis within tumor cells, along with HHV8 negativity. This reinforces lack of a common pathway for a Kaposiform hemangioendothelioma and a Kaposi's sarcoma. In spite of SMA positivity, aforementioned histological features and diffuse CD34 immunoreactivity and focal CD31 positivity within tumor cells, ruled out a myofibromatosis that has been documented at this site and in similar aged patients [7
]. Infact, the present case was initially reported as myofibromatosis at another laboratory. Variable SMA positivity within tumor cells, presumably in the pericytes, has been documented in a KHE [2
]. This reinforces application of an optimal panel of IHC markers with the already described histomorphological 'clues' for a KHE. Additional IHC markers like isoform 1, GLUT-1, a glucose transporter protein and Lewis Y antigen (LeY) have been found useful in differentiating KHE from a juvenile hemangioma (JH), as these are not expressed in KHE, in contrast to a juvenile hemangioma [2
]. Ki-67 was noted in few tumor nuclei as similarly described by Lyon et al [2
], who noted a contrasting prominent staining in cases of JH.
Presence of several co-existing dilated lymphatic vessels was a significant 'clue' in diagnosis of a KHE. It has been documented that approximately two-thirds of KHE, when carefully studied, exhibit lymphatic abnormalities comprising thin-walled vessels that surround vascular tumor nodules and often extend outward. One of the reasons that have been hypothesized for this association is that the development of KHE begins with a lymphatic malformation onto which a vascular component is engrafted. Another hypothesis is that KHE initially produces lymphatic endothelial growth factors (for example VEGF-C), that leads to proliferation of adjacent lymphatics, as noted in other tumors [2
]. Site-wise, tonsil, as noted in the present case, seems to be a "fertile soil" for the development of this unusual tumor, with vascular and lymphatic components. Lately, D-240 has been identified as a useful marker for highlighting lymphatic endothelial cells [10
]. However, in view of present unavailability of this marker in our laboratory, it was not included in the IHC panel. Nonetheless, histopathological features were unequivocal for presence of substantial lymphatic component, wherein the lymphatic channels were negative for CD34 and CD31, in contrast to the lobules of spindle cells [2
]. Aforementioned histological features and lack of KMP in the present case were overlapping with a tufted hemangioma [11
]. A similar co-existence of lymphangiectasia with vascular tumor nodules is seen in a tufted angioma. KHE and tufted angioma are probably same part of the spectrum. Cases of an acquired tufted angioma have been described with KMP, as well as cases of KHE have been described without KMP [5
]. The platelet count in the present case was towards lower side of the range, but no symptoms of coagulopathy were noted, excluding a KMP.
Interestingly, on postoperative imaging in the present case, coexisting lymphangioma was also identified. This was a discrete lesion in the parapharyngeal region, excluding the possibility of the extension from the main lesion. This could possibly have been additional reason for transient ipsilateral neck swelling, since birth, reflective of episodic secondary inflammation.
Therapeutically, KHE, in isolation, is a candidate for complete surgical excision. Increasing size, risk of coagulopathy are indicators for therapeutic interventions in such cases. Medical treatment is included in cases associated with KMP [13
]. KMP was lacking in the present case. Cases of KHE, unassociated with KMP have been followed-up without treatment and have shown no disease and even tumor regression in a few such cases [5
]. Surgical excision in this case was performed elsewhere, presumably without clear resection margins, as a result of preoperative clinicoradiological impression of an inflammatory lesion. In view of postoperative imaging results that showed cystic lesion, indicative of coexisting lymphangiomas, the patient was offered sclerotherapy at our hospital. He has been recommended for 4 cycles of sclerotherapy on a 2 monthly basis.
In conclusion, KHE is an uncommon tumor with a distinct clinicopathologic features, including IHC profile and differs from a Kaposi's sarcoma and its other histological mimics. Careful attention towards its histopathological features, including its association with lymphatic component, coupled with IHC, is helpful in its identification, including at rare sites like tonsil in the present case. A coexisting lymphangiomas was a unique feature that led to incorporation of sclerotherapy in the present case. Surgical excision with follow-up is the treatment mainstay in most cases.