FNA diagnosis of CD99 positive neuroblastoma: A diagnostic dilemma.

Tissue diagnosis of small round cell tumors relies heavily on immunohistochemical staining. Two of the small round cell tumors, namely neuroblastoma and primitive neuroectodermal tumor, have considerable morphologic overlap. Many studies suggest that CD99 positivity virtually excludes the diagnosis of neuroblastoma. We report a case of poorly differentiated neuroblastoma in which aberrant CD99 positivity led to diagnostic dilemma.

Introduction

Fine needle aspiration (FNAC) has limitations in classifying and typing small round cell tumors (SRCT) of childhood because of morphologic similarities and overlaps.[1-4] Adjunctive ancillary techniques like immunohistochemistry, cytogenetics, and molecular and ultrastructural studies are helpful in classifying these tumors.[2-4] Two of the small round cell tumors, namely primitive neuroectodermal tumors (PNET) and neuroblastoma (especially the poorly differentiated neuroblastomas), have considerable morphologic overlap. Although well-differentiated neuroblastoma shows prominent Homer-Wright rosettes and abundant neuropil as compared to PNET; this difference is not of value in separating poorly differentiated neuroblastomas. Immunohistochemistry using CD56, CD99 and NB84 are usually considered helpful in resolving this problem.[2] We report a case of poorly differentiated neuroblastoma where aberrant CD99 positivity on trucut biopsy posed a diagnostic dilemma.

Case Report

A four-year-old boy presented to the pediatric out-patient department (OPD) with acute onset pallor, fever and bone pains for four weeks with no significant findings on systemic examination. On investigation, the hemoglobin level was 7.9 gm% with a leucoerythroblastic picture on peripheral smear. Mild hepatosplenomegaly with large heterogeneous low-density retroperitoneal lymph nodes were observed on computed tomography scan (CT scan). Possibility of lymphoma was considered on radiology. Trephine iliac crest biopsy revealed infiltration by a malignant small round cell tumor. Flow cytometric examination of bone marrow showed the cells to be negative for CD45, and T and B cell markers, while CD 56 was positive.

CT guided FNAC was done from the retroperitoneal mass; both alcohol fixed and air dried smears were prepared and stained with Papanicolaou and May-Grünwald- Giemsa stains respectively. The FNA smears were highly cellular and showed predominantly singly dispersed small round to oval cells with high nuclear cytoplasmic ratio, hyper chromatic nuclei and inconspicuous small nucleoli. Few small aggregates and clusters of these cells with rudimentary rosette formation were seen. In few clusters, angulations and molding of nuclei was also observed. No eosinophilic fibrillar material (neuropil) could be seen in the background or within the rosettes. Cytomorphology was that of a poorly differentiated small round cell tumor [Figure 1a]. A trucut biopsy of the retro peritoneal mass was done which showed a tumor with similar morphology. Immunostain for CD99 (prediluted monoclonal mouse anti human antibody, Dako, Denmark) [Figure 1b] and neuron specific enolase (mouse anti NSE, Serotec, Oxford) done on the surgical biopsy specimen was strongly positive, whereas CD45 was negative. Based on these morphologic and immunohistochemical findings, a diagnosis of small round cell tumor consistent with a primitive neuroectodermal tumor (PNET) was suggested.

Figure 1

(a) Smears showing occasional rossetting (MGG, ×400); (b) CD99 positivity on biopsy (IHC, ×400)

Follow-up of patient revealed his urinary norepinephrine and vanillylmandelic acid levels (VMA, done by high performance liquid chromatography reference levels 1-8 mg/day)[5] to be raised. MIBG (Metaiodobenzylguanidine) scan revealed an MIBG avid tissue in the left suprarenal region with left abdominal lymphadenopathy with diffuse MIBG uptake noted in the entire skeleton.

The finding of small round cell tumor on bone marrow biopsy and FNA, elevated urinary levels of norepinephrine and VMA and MIBG avid tissue in the left suprarenal region led to the final diagnosis of neuroblastoma.

Discussion

Small round cell tumors have overlapping cytomorphologic features and are often undifferentiated or poorly differentiated; hence a definitive diagnosis on morphology alone is difficult.[3] Various ancillary studies including immunohistochemistry, ultra structural, cytogenetic and molecular techniques may provide additional information to aid in differential diagnosis.[1]

CD99 expression which is consistently present in the cells of Ewing's sarcoma/PNET may also be seen in over 90% of lymphoblastic lymphomas, 20 to 25% of primitive rhabdomyosarcomas, over 75% of poorly differentiated synovial sarcoma, approximately 50% of mesenchymal chondrosarcoma, the blastemal component of some cases of Wilms’ tumor and small cell osteosarcoma and desmoplastic round cell tumors.[6]

Many studies suggest that CD99 is negative in neuroblastoma and the presence of CD99 positivity virtually excludes the diagnosis of neuroblastoma.[6] Pohar – Marinsek[7] performed immunocytochemical staining with CD99 on 26 cases of neuroblastoma and found a weak reaction in eight cases. In another study, CD99 positivity has been reported in four out of nine FNA samples of neuroblastomas. In this study, 100% of the cells (strongly in two cases and weakly in two cases) were stained with CD99. But none of these cases showed positivity on surgical biopsy samples. One of the explanations for this disparity in immunostaining could be technical problems.[7-8] Conditions for immunostaining are more erratic in cytology as compared to histology and the specimen processing protocols have not been well established.[8]

In our case, we observed that a strong reaction with CD99 was present in more than 75% of the cells in the biopsy along with positive and negative controls.

CD56 is uniformly and strongly positive in neuroblastoma.[6] CD56 positivity, however, has been reported in other small round cell tumors including PNET. Reports of flow cytometry immunophenotyping showing strong expression of CD56 in Ewing tumor/PNET are available.[9] Our case also showed strong CD56 positivity on flow cytometry immunophenotyping on bone marrow aspirate.

With increased use of molecular techniques, it has been noted that tumors with proven molecular diagnosis can exhibit unusual patterns of immunohistochemical staining[2] Gautam et al.,[10] in their study of 37 cases on flow cytometric immunophenotyping and immunocytochemistry in the categorization of malignant small round cell tumors found positivity for markers in addition to what usually is expected in three cases on immunocytochemistry and in four cases on immunophenotyping. They attributed it to divergent differentiation in malignant small round cell tumors. However, in their study also, none of the cases of neuroblastoma were positive for CD99. As facilities like cytogenetics and molecular techniques are not available in all the laboratories, especially in the developing countries, a tissue diagnosis is often rendered based on cytomorphology and immunostaining that is comparatively cheaper and more widely available.

CD99 expression, which is not considered to be associated with neuroblastoma, was expressed in the above case. There was a strong clinical suspicion of neuroblastoma in this case as the patient was four years old and had presented with metastatic disease, so his urinary nor epinephrine and VMA levels were measured and were found to be raised. Urinary VMA and HVA (Homovanillic acid) or VMA and catecholamine levels are raised in 95 to 100% cases of neuroblastomas and can be used to differentiate Ewing's tumor/PNET from neuroblastoma.[11] MIBG is only rarely concentrated by non neural crest tumors and MIBG scintigraphy is useful in non-invasively establishing the diagnosis of neuroblastoma.[11]

CD99 positivity in a small round cell tumor in appropriate clinical setting should not exclude the possibility of a neuroblastoma. In small round cell tumors in children less than five years old, other non-invasive tests like measurement of urinary catecholamine levels and MIBG scan (if available) should be done instead of arriving at a diagnosis based on the results of immunohistochemistry alone.