Archives of Medicine and Health Sciences

: 2018  |  Volume : 6  |  Issue : 1  |  Page : 64--69

Immunohistochemical characterization of small round blue cell tumors of childhood at Ile-Ife, Nigeria: A 10-Year retrospective study

Ifeoma Florence Ezejiofor1, Kayode Adelusola1, Muheez Alani Durosinmi2, Lorenzo Leoncini3, Willians Olufemi Odesanmi1, Maria Raffaella Ambrosio3, S Lazzi3, Rinde O. O. Olaofe1, Gloria Gbutorano3,  
1 Department of Morbid Anatomy and Forensic Medicine, Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, Osun State, Nigeria
2 Department of Haematology and Blood Transfusion, Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, Osun State, Nigeria
3 Department of Human Pathology and Oncology, University of Siena, Siena, Italy

Correspondence Address:
Dr. Ifeoma Florence Ezejiofor
Department of Morbid Anatomy and Forensic Medicine, Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife, Osun State


Background: Immunostains when used in correlation with clinical site of tumours and morphology permits accurate and specific diagnosis of these undifferentiated tumours. Materials and Methods: A ten-year retrospective research of the histopathological and immunohistochemical features of small round blue cell tumours (SRBCT) in OAUTHC was analyzed. Pathology reports of all SRBCT and their blocks were retrieved and recut slides reviewed to determine each tumour types. Acute myelocytic lymphoma/leukeamias and Glioblastoma multiformis were excluded from SRBCT of childhood. Eighty four (84) cases that fulfilled the inclusion criteria were analyzed using immunohistochemistry. Results: The age range of presentation was 0-15 years (Mean 5.98±3.964 year S.D.). Retinoblastoma and Wilms' tumours were the commonest histological sub-types with a percentage of 20.0% each followed by Burkitt lymphoma (17.6%), rhabdomyosarcoma (9.8%) which was the only soft tissue sarcoma found. The least represented was supra-tentorial CNS-PNET (1.2%). Eighty four blocks of SRBCT were subjected to different panels of immunohistochemistry. Of all these tumours 36 cases had a change in diagnosis: 23 cases had an initial, pre-immunohistochemical umbrella diagnosis ranging from NHL, SRBCT to no pathological diagnosis at all. Seven (7) cases with initial diagnosis ranging from Ewing's sarcoma (1 case), retinoblastoma (1 case) to rhabdomyosarcoma (5 cases) were finally confirmed with immunostains as Burkitt lymphoma (BL) while one case of BL was finally confirmed as rhabdomyosarcoma. The last five of the 36 cases were totally non-neoplastic lesions but had histology diagnosis ranging from NHL, spinal cell sarcoma to periosteal osteosarcoma. The immunostains done on these 5 cases revealed erythroid hyperplasia with dyserythropoiesis, cellular neurofibroma, fibrous dysplasia, reactive follicular hyperplasia and normal retinal tissue. Conclusion: Application of immunohistochemistry does indeed enhance the diagnostic accuracy of these undifferentiated tumors.

How to cite this article:
Ezejiofor IF, Adelusola K, Durosinmi MA, Leoncini L, Odesanmi WO, Ambrosio MR, Lazzi S, Olaofe RO, Gbutorano G. Immunohistochemical characterization of small round blue cell tumors of childhood at Ile-Ife, Nigeria: A 10-Year retrospective study.Arch Med Health Sci 2018;6:64-69

How to cite this URL:
Ezejiofor IF, Adelusola K, Durosinmi MA, Leoncini L, Odesanmi WO, Ambrosio MR, Lazzi S, Olaofe RO, Gbutorano G. Immunohistochemical characterization of small round blue cell tumors of childhood at Ile-Ife, Nigeria: A 10-Year retrospective study. Arch Med Health Sci [serial online] 2018 [cited 2023 Mar 23 ];6:64-69
Available from:

Full Text


Childhood cancer is fast becoming an important pediatric problem in several parts of Africa including Nigeria, and this is due to progressive decline in infectious and nutritional diseases; this change may be attributed to the increased awareness and use of immunization against childhood killer diseases.[1] It is estimated that more than 85% of childhood cancer cases occurred in resource-poor countries, and it is possible that this will increase to 90% in the next two decades as a result of an expected rapid increase in the young population.[2] In most parts of Africa and other developing countries, childhood malignancies vary between 5%–15% and 4.3%–12.5% of all malignant tumors.[3],[4] In Nigeria, malignant tumors of childhood in Jos, Kano, Ibadan, and Calabar constitute 10%, 10.9%, 12%, and 12.5%, respectively, of all malignant tumors.[5],[6],[7]

There are marked variations in the incidence of different types of childhood cancers between different populations in the world. Leukemias predominate in Caucasians and are linked to high socioeconomic status while Burkitt lymphoma (BL) predominates in most developing countries and is associated with Epstein–Barr virus and chronic malaria.[8] Small round blue cell tumors (SRBCTs) are characterized as sheets of cells with small, hyperchromatic round nuclei and scanty cytoplasm and represent a diverse group of tumors which pose a number of problems in diagnosis. Today, SRBCTs are diagnosed with the use of multiple parameters of which immunohistochemistry plays an important role.

 Materials and Methods

This is a retrospective study of 84 diagnosed cases of SRBCT in the Obafemi Awolowo University Teaching Hospitals Complex (OAUTHC), Ile-Ife, Nigeria, between the periods of January 2002 to February 2011. The pathology reports were retrieved, and information about the age, sex, and anatomical site and clinical and histological diagnosis was extracted. The paraffin blocks of all 84 available cases of SRBCT which met inclusion criteria diagnosed in the study period were retrieved. Fresh sections were prepared and stained with hematoxylin and eosin (H and E) and examined using multi-headed light microscope (LEICA).

Immunohistochemical studies were done by the indirect immunoperoxidase method on formalin-fixed paraffin-embedded (FFPE) sections. This was carried out in collaborations with Policlinico Santa Maria Le Scotte protocol under the supervision of Prof. Lorenzo Leoncini of the Department of Human Pathology and Oncology, Anatomical Pathology Section, University of Siena, Italy.

Four-micrometer (4 μm) thick sections of FFPE tissue blocks were made and mounted on a positively charged glass slides (ProbeOn Plus; Fisher Scientific, Pittsburgh, PA, USA). The tissue sections were deparaffinized by passing them through different grades of xylene and then rehydrated in decreasing ethanol concentrations. The antigens were appropriately retrieved by heat-induced epitope retrieval one and two (HIER1 and 2) with the use of BOND-III machine by Leica. Endogenous peroxidase was blocked using 3% hydrogen peroxide. The panel of antibodies used depended on the clinical site and morphology of the lesions. The tissues were incubated with different primary antibodies and diaminobenzene, a chromogen, was used as a final substrate. Immunostaining was performed on Bond-Max automated immunostainer (Leica Microsystem, Bannockburn, IL, USA) with different antibodies including epithelial membrane antigen (EMA), glycophorin (Dako), vimentin, desmin, myogenin, pan-cytokeratin (CK), CD99, synaptophysin, glial fibrillary acidic protein (GFAP), CD56, WT1, terminal deoxynucleotidyl transferase (Tdt), CD45, CD20, CD3, CD10, B-cell lymphoma (BCL)-2, Ki-67, S-100, and CD34 (Novocastra) with control in parallel.


A total of 132 cases of childhood malignant neoplasms were documented in the OAUTHC histopathology records over a 10-year period of which 109 (82.6%) cases were classified as SRBCT; acute myelocytic leukemias and glioblastoma were excluded from SRBCTs. Only 84 cases fulfilled the inclusion criteria for this study and were appropriately grouped using the WHO classification and immunohistochemistry. Of the 84 cases, 58 (68.6%) were males while 26 (31.6%) were females with male-to-female ratio of 2.1:1 [Figure 1]. The mean age was 5.98 ± 3.964 years standard deviation [Figure 1]. The age range at presentation was 0–15 years. The peak age incidence in both males and females was similar with highest tumor incidence found between the ages 0–5 years with frequencies of 34 (58.6%) and 12 (44.5%), respectively. This was followed by ages 6–10 years in both sexes with frequencies of 15 (25.9%) and 9 (33.3%) while the least represented tumor burden was found between ages 11–15 years with frequencies of 9 (15.5%) and 6 (22.2%), respectively. The male sex, however, appeared to have greater number of tumor burden in all age groups [Figure 1].{Figure 1}

The most common histological subtypes of SRBCT recorded in this study were retinoblastoma and Wilms' tumor with frequencies of 17 (20.0%) each, followed by BL with a frequency of 15 (17.6%). Rhabdomyosarcoma was the only soft-tissue sarcoma found with a frequency of 8 (9.8%). There was no case of Ewing's sarcoma found in this study, but 3 cases of peripheral primitive neuroectodermal tumor (p-PNET) with a frequency of 3.3% were seen. The osteosarcomas had frequency of 5 (6.1%) while diffuse large BCL (DLBCL), intermediate group lymphoma, ependymomas, and poorly differentiated carcinoma had frequencies of 2 (2.4%) each [Table 1].{Table 1}

The most common childhood brain tumor represented was medulloblastoma with a frequency of 4 (4.9%) followed by ependymoma with a frequency of 2 (2.4%). The least represented was supratentorial central nervous system (CNS)-PNET with only one case (1.2%) [Table 1].

Five cases of osteosarcoma (small cell variants) were found with a frequency of 5.9% while three cases of neuroblastoma were seen with a frequency of 3 (3.7%)

The distributions of the main tumor groups varied markedly with age and between sexes. Males had highest tumor percentages in all the different histology subtypes of SRBCTs with the exception of poorly differentiated carcinoma where both sexes had equal predilection and both cases occurred as nasopharyngeal mass [Table 1].



All the cases of retinoblastoma occurred in age bracket of 0–5 years with a frequency of 17 (37.0%) and had a strong male preponderance. Seventeen (94.4%) of the 18 cases previously diagnosed as retinoblastoma stained positive with synaptophysin and therefore retained their final immunohistochemically confirmed diagnosis of retinoblastoma. Only one case was negative for synaptophysin but stained positive for CD45, CD20, CD10, and negative for Tdt and BCL-2. It had marked surface membrane staining pattern in CD45, CD20, and CD10 confirming its final diagnosis as BL.

Wilms' tumor

Like retinoblastoma, also accounted for 20% of overall SRBCT cases studied. Most of its presentation occurred within ages 0–5 years with a frequency of 14 (30.4%), and few cases were found between 6 and 10 years with a frequency of 3 (12.5%) [Table 2]. Immunostaining for WT1 was detected in 17 cases of the 84 tumors studied. There were marked cytoplasmic and nuclear staining patterns in almost all the blastema tumor cells. The epithelial component had moderate-to-marked cytoplasmic stain for pan-CK while the stroma had moderate-to-marked stains for vimentin [Figure 2]a and [Figure 2]b.{Table 2}{Figure 2}


Immunostaining for myogenin was detected in 8 cases of 84 tumors studied and was subtyped as rhabdomyosarcoma. Seven of these cases were formally diagnosed as rhabdomyosarcoma while only one had initial diagnosis as BL and was an intra-abdominal mass. These tumors had moderate-to-marked nuclear staining pattern with myogenin and were finally confirmed as rhabdomyosarcoma.

Brain tumors

Four cases of 84 tumors studied were finally confirmed as medulloblastoma; two of the cases had no morphologic diagnosis but were confirmed with marked cytoplasmic synaptophysin-positive stain as well as location of the tumor. All four cases were cerebellar mass lesion. Two cases of ependymomas were confirmed using marked cytoplasmic positivity for GFAP with negative EMA. One had initial diagnosis of just PNET but finally reclassified as ependymoma. One case of supratentorial CNS-PNET showed marked cytoplasmic synaptophysin-positive stain but negative stain for CD99.

Peripheral primitive neuroectodermal tumor

Three cases of p-PNET were confirmed using marked membranous CD99 positivity, morphological diagnosis, and clinical site [Figure 3]a and [Figure 3]b.{Figure 3}


Accounted for 3.5% of overall tumors studied. Its peak age incidence was found between ages 0–5 and 11–15 years of life with frequencies of 1 (2.2%) and 2 (13.3%), respectively [Table 2]. It had strong male preponderance with male-to-female ratio of 2:1. Three cases of neuroblastoma were also confirmed using marked membranous CD56 stain, negative pan-CK, and CD99 membranous stain with site of tumor location [Figure 4]a,[Figure 4]b,[Figure 4]c,[Figure 4]d.{Figure 4}

Poorly differentiated carcinoma

Accounted for 2.3% of overall tumor studied. It was the least represented group of SRBCT seen and had equal sex predilection. It occurred equally in ages 6–10 and 11–15 years of age. Two cases of the 84 tumors studied were diagnosed as poorly differentiated carcinoma and were finally confirmed immunohistochemically as so with positive marked cytoplasmic staining with pan-CK [Figure 5]a and [Figure 5]b.{Figure 5}

Results of immunohistochemistry

Of the 84 cases of SRBCT that were studied using immunohistochemistry, all Wilms' tumor had WT1 values of +3. Seventy-five percent (75%) of medulloblastoma had synaptophysin, CD56, and GFAP values of 3+++ while 25% had 2++ and were all negative for EMA. Two cases of ependymomas had GFAP values of 2++ but negative stains for EMA.

Three cases of neuroblastoma had CD56 values of 3+++ but negative for pan-CK, CD99, and desmin. Three cases of p-PNETs had CD99 of 3+++ while supratentorial CNS-PNET was negative for CD99 but had synaptophysin value of 3+++. Retinoblastoma also had synaptophysin value of 3+++.

Rhabdomyosarcoma had myogenin value of 3+++ and desmin value of 2+++ while osteosarcoma had a vimentin value of 2++. Poorly differentiated carcinoma had pan-CK and EMA values of 3+++.

In lymphomas, BL, DLBCL, and intermediate group lymphoma (DLBCL/BL), all had CD45 and CD20 values of 3+++. They also had CD10 value of 3+++ while BCL-2 was negative in all cases of BL except in DLBCL and intermediate group which showed values of +1 each. However, Ki-67 had +3 value (100%) in BL but <95% in intermediate group and <90% in DLBCL.

Two cases of acute lymphoblastic lymphoma (ALL) had negative reaction in CD20 but Tdt value of +3 as well as CD3 value of +3 while one case was negative value for CD3 which had three positive values for both CD20 and Tdt.

On the whole, 36 cases had a change in diagnosis. Of these 36 cases, 23 cases had an initial, preimmunohistochemical umbrella diagnosis ranging from NHL, SRBCT, to no pathological diagnosis at all. Seven cases with initial diagnosis ranging from Ewing's sarcoma (1 case), retinoblastoma (1 case), to rhabdomyosarcoma (5 cases) were finally confirmed with immunostains as BL while one case of BL was finally confirmed as rhabdomyosarcoma. Five cases with complete missed diagnosis were shown in [Table 3].{Table 3}


Increasing burden of malignant tumor, including SRBCT, has necessitated the need for accurate diagnosis to determine the magnitude of the problem. Immunohistochemistry is, therefore, very important routine accessory tool in making diagnosis of SRBCT.[9] However, because of difficulty in obtaining antibodies, unavailability of power supply and its instability as well as financial constraints, this investigative tool is not established in most laboratories in Nigeria.

Several studies have implicated the usefulness and complementary role of immunohistochemistry as an adjunct to morphologic diagnosis of SRBCT.[10] Its input improves the accuracy of diagnosis and initiates appropriate therapeutic and management protocol to be applied. The work done by Naresh et al.[11] shows its application and usefulness in both resource-poor and resource-rich countries. Sebire et al.[12] who worked on immunohistochemical findings in embryonal small round cell tumors with molecular diagnostic confirmation emphasize that diagnosis of pediatric tumors relies heavily on immunostains. Leong and Wannakrairot [10] subjected 557 archival tissue blocks of poorly differentiated round cell tumors and spindle cell tumors to immunostains and had definitive diagnosis in 420 cases (75.4%). They also concluded that immunohistochemistry provided definitive diagnosis in 70% of round cell tumors and 92% of spindle cell tumors.[10] Thomas and Olu-Eddo [13] in Ibadan teaching hospital reported a positive predictive value of morphologic diagnosis of lymphoma in routine H and E sections as 60% with negative predictive value of 86.7%. The index study reported a positive predictive value of morphologic diagnosis of lymphoma and rhabdomyosarcoma, based on routine H and E-stained sections as 77.3% and 96.5%, respectively, and negative predictive values as 92.4% and 98.7%, respectively. Meanwhile, the positive predictive values based on of morphologic diagnosis of retinoblastoma, carcinoma, and nephroblastoma approached 100%. High diagnostic accuracies observed in retinoblastoma, Wilms' tumor, and carcinoma were probably due to specific clinical location of these tumors, morphology features, and epithelial differentiation in cases of Wilms' tumors and carcinomas. These epithelial tumors showed marked CK expressions. Lymphomas, which collectively stand as one of common childhood tumors, are often misdiagnosed and therefore will benefit exclusively from immunohistochemical stains. Thomas and Olu-Eddo [13] also reported that morphological diagnosis of lymphoma and rhabdomyosarcoma based on routine H and E-stained sections showed a sensitivity of 75% and 62.5% with specificity of 76.5% and 94.1%, respectively. The sensitivity values of lymphoma in this index case when subtyped into different histology groups based on H and E-stained sections dropped significantly to 0% for ALL, 46.7% for BL, and 50% for DLBCL indicating crucial need for immunohistochemical importance in definitive diagnosis of individual lymphoma tumors. Therefore, lymphomas and rhabdomyosarcoma are often misdiagnosed, and since most of these tumors constitute the commonest childhood tumor in Nigeria, immunohistochemistry must be applied as a routine technique in their diagnosis. The work done by Thomas and Olu-Eddo [13] also showed the importance of immunostains in diagnosis, where 6 out of 25 cases previously diagnosed as lymphoma had a significant change in diagnosis when immunohistochemistry was applied. The index case also showed change in diagnosis in 36 out of 84 cases when immunohistochemistry was applied.

Majority of studies revealed both in local and international journals and showed the suitability of archival tissue block specimen in our laboratory for immunohistochemistry. The index study reaffirms the suitability of archival tissue specimens in our laboratories for immunohistochemical staining as well.

Therefore, there is a need for applying routine immunostains in archival tissue blocks of small undifferentiated tumors for accurate diagnosis and proper classification. A significant number of researchers have supported the importance of immunohistochemistry as a complementary tool in routine morphological diagnosis of tumors, especially SRBCT of childhood.[9] It is, therefore, recommended that a panel of at least seven antibodies be applied in this resource deprived country like ours in making diagnosis of these tumors. These antibodies should include myogenin, CD45, CD99, synaptophysin, WT1, CK, and vimentin. However, these markers must be selected based on clinical site and morphological features. For retinoblastoma and Wilms' tumor that have specific clinical site, minimum of two markers should be recommended such as synaptophysin and WT1 and CK, respectively. However, for rhabdomyosarcoma, only myogenin is adequate in making the diagnosis, but for positive CD45 lymphomas, few other panels of immunostains such as BCL-2, Ki67, and CD10 should be added for further subclassification. We suggest further study on 5-year survival rate for these children when accurate diagnosis and proper treatment is given.

 Conclusion and Recommendations

Application of immunohistochemistry in this study has shown a significant improvement in the diagnostic accuracy of these undifferentiated tumors. We, therefore, recommend minimal immunopanel use in conjunction with clinical location and morphologic features as part of criteria to be considered in our poor resource setting to separate these undifferentiated SRBCTs. This will help to initiate appropriate treatment and improve health status of these children and reduce unnecessary wastage of chemotherapy given on cases of unconfirmed diagnosis.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Mandong MB, Angyo IA, Zoakah AI. Paediatric solid malignant tumours in Jos university teaching hospital: A hospital-based study. Niger J Med 2000;9:52.
2Yaris N, Mandiracioglu A, Büyükpamukcu M. Childhood cancer in developing countries. Pediatr Hematol Oncol 2004;21:237-53.
3Hadley LG, Rouma BS, Saad-Eldin Y. Challenge of pediatric oncology in Africa. Semin Pediatr Surg 2012;21:136-41.
4Jamal S, Mamoon N, Mushtaq S, Luqman M. Pattern of childhood malignancies: Study of 922 cases at armed forces institute of pathology (AFIP), Rawalpindi, Pakistan. Asian Pac J Cancer Prev 2006;7:420-2.
5Tanko NM, Echejoh GO, Manasseh NA, Mandong MB, Uba AF. Paediatric solid tumours in Nigerian children: A changing pattern? Afr J Paediatr Surg 2009;6:7-10.
6Ochicha O, Gwarzo AK, Gwarzo D. Pediatric malignancies in Kano, Northern Nigeria. World J Pediatr 2012;8:235-9.
7Akang EE. Tumors of childhood in Ibadan, Nigeria (1973-1990). Pediatr Pathol Lab Med 1996;16:791-800.
8Stiller CA. Epidemiology and genetics of childhood cancer. Oncogene 2004;23:6429-44.
9Dehner LP. On trial: A malignant small cell tumor in a child: Four wrongs do not make a right. Am J Clin Pathol 1998;109:662-8.
10Leong AS, Wannakrairot P. A retrospective analysis of immunostaining in the identification of poorly differentiated round cell and spindle cell tumours: Result, regents and costs. Pathology 1992;24:256-60.
11Naresh KN, Ibrahim HA, Lazzi S, Rince P, Onorati M, Ambrosio MR, et al. Diagnosis of Burkitt lymphoma using an algorithmic approach – applicable in both resource-poor and resource-rich countries. Br J Haematol 2011;154:770-6.
12Sebire NJ, Gibson S, Rampling D, Williams S, Malone M, Ramsay AD, et al. Immunohistochemical findings in embryonal small round cell tumors with molecular diagnostic confirmation. Appl Immunohistochem Mol Morphol 2005;13:1-5.
13Thomas JO, Olu-Eddo AA. Immunocytochemistry in the diagnosis of small blue cell tumours of childhood. West Afr J Med 2006;25:199-204.