Bug spray worse than the bite?

In a retrospective analysis of 87 cases of Caucasian childhood non-Hodgkin malignant lymphoma (NHML) from Lyon, France, all the case were diffuse lymphomas, but 47 were diagnosed as monomorphic small non-cleaved NHML, pathologically indistinguishable from Burkitt's lymphoma (BL). BL could then be the most frequent childhood lymphoma in France. This homogeneous series allows better definition of the characteristics of BL within NHML. Age distribution is similar to that of endemic BL, with a sex ratio of 3.7/1. Abdominal masses are initially present in 68% of the cases, whereas jaw is involved in only 4%. The disease is characterized by its overwhelming evolution in the absence of therapy. However, complete remission (CR) is usually obtained after the first chem-therapy regimen. Most relapses occur at 3-8 months. Death could be related to cerebrospinal fluid (CSF) involvement, local recurrence or secondary marrow involvement. Ninety per cent of the patients alive with no evidence of disease (NED) 8 months after CR can be considered as definitely cured. Our study on Caucasian children with NHML indicates that, from histological and clinical criteria, nearly half the cases are very similar to African BL. Even though EBV rarely associated with our cases, BL could be a worldwide lymphoma. CONFUSION AND CONTROVERSY persist with regard to the eponym "Burkitt's lymphoma" (BL) which was used initially to refer to a clinical and epidemiological entity in African children (Burkitt, 1958, 1962). Burkitt's initial report described a tumour "involving the jaw in African children" and "dependent on climatic factors" and suggested "the implication of an infectious agent in its aetiology" (Burkitt, 1958, 1962). In 1964, Epstein et al. first described, in a cultured Burkitt-cell line, the herpes-type virus now known as the Epstein-Barr virus (EBV). In tropical Africa and in New Guinea, 96-97% of cases are associated with EBV African endemic BL (EBL) is now recognized as a clear epidemiological entity. The question whether this lymphoma is unique and restricted to endemic regions of the world, or whether it can also be found elsewhere, arose with 3 reports in The answer should have been provided by the plethora of reports from many countries, published between 1966 and 1970 (see: Philip et al., 1980b; Lenoir et al., unpublished). However , there was a tendency to contrast EBL with BL arising outside Africa (non-endemic BL), the latter being sporadic , mainly abdominal in presentation, and usually not EBV-associated (Dorfman,

Age distribution is similar to that of endemic BL, with a sex ratio of 3.7/1.
Abdominal masses are initially present in 68% of the cases, whereas jaw is involved in only 4%.
The disease is characterized by its overwhelming evolution in the absence of therapy. However, complete remission (CR) is usually obtained after the first chemtherapy regimen. Most relapses occur at 3-8 months. Death could be related to cerebrospinal fluid (CSF) involvement, local recurrence or secondary marrow involvement. Ninety per cent of the patients alive with no evidence of disease (NED) 8 months after CR can be considered as definitely cured.
Our study on Caucasian children with NHML indicates that, from histological and clinical criteria, nearly half the cases are very similar to African BL. Even though EBV rarely associated with our cases, BL could be a worldwide lymphoma.
CONFUSION AND CONTROVERSY persist with regard to the eponym "Burkitt's lymphoma" (BL) which was used initially to refer to a clinical and epidemiological entity in African children (Burkitt, 1958(Burkitt, , 1962). Burkitt's initial report described a tumour "involving the jaw in African children" and "dependent on climatic factors" and suggested "the implication of an infectious agent in its aetiology" (Burkitt, 1958(Burkitt, , 1962. In 1964, Epstein et al. first described, in a cultured Burkittcell line, the herpes-type virus now known as the Epstein-Barr virus (EBV). In tropical Africa and in New Guinea, 96-97% of cases are associated with EBV (see: Klein, 1975;IARC, 1981) and African endemic BL (EBL) is now recognized as a clear epidemiological entity.
The question whether this lymphoma is unique and restricted to endemic regions of the world, or whether it can also be found elsewhere, arose with 3 reports in -1966(Dorfman, 1965O'Conor et al., 1965;Wright, 1966). The answer should have been provided by the plethora of reports from many countries, published between 1966 and 1970 (see: Philip et al., 1980b;Lenoir et al., unpublished). However, there was a tendency to contrast EBL with BL arising outside Africa (non-endemic BL), the latter being sporadic, mainly abdominal in presentation, and usually not EBV-associated (Dorfman, 1965;Cohen et al., 1969;Hirshaudt et al., 1973;Ablashi et al., 1974;Levine & Cho 1974;Banks et al., 1975).
The confusion and controversy existed because most of the reports did not take into consideration the clear conclusions formalized in 1969 by a World Health Organization expert committee . The committee carried out a detailed and objective study of cases gathered from Africa and other countries, and concluded that BL is a clear, pathological entity which is found all over the world and has features distinct from those of other undifferentiated lymphomas and leukaemias. It recommended that the tumour be referred to as a "malignant lymphoma undifferentiated Burkitt type".
The purpose of our study was to conduct a retrospective analysis of 87 cases of NHML in children. Two of us (RGM and PAB) were asked to identify cases to which the WHO definition of BL could be applied; and we report here 47 cases of BL found in Caucasian children from Lyon, France, between 1965 and1979. This homogeneous series makes it possible to define epidemiological and clinical features of BL in France. We hope that these data will stimulate new pathological studies of previously reported series, since BL appears to be a clear clinico-pathological entity within childhood NHML.

MATERIAL AND METHODS
Between 1965 and 1979, 146 patients under 16 years of age were treated for NHML at the Centre Leon Berard,Lyon (75 patients) and at other chemotherapy units in Lyon (71 patients). Since 36 cases with incomplete clinical records were excluded, 110 were reviewed (Philip et al., 1980b). Non-Caucasian cases and patients for whom initial sections were not available were also left out; thus our analysis was finally carried out on 87 Caucasian patients. Sections were reexamined for 76 patients (average of 3 stained sections from each patient) and new sections were available for 11 cases (at least 7 sections per patient). Sections routinely stained with haematoxylin and eosin and embedded in paraffin were thus available for all the cases. New sections were stained with haematoxylin and eosin, periodic-acid-Schiff, May-Griinwald-Giemsa, Wilder silver reticulin impregnation and Giemsa by the Lennert technique.
Histological data.-Histological material from all 87 cases before treatment was reviewed independently by 2 of us, without knowledge of the clinical data. R.G.M. reviewed the cases using the Kiel classification (Gerard-Marchant et al., 1974). Details of this classification in childhood NHML are given in Table I. P.A.B. reviewed the cases using a modification of the Rappaport classification (Bryon et al., 1981).  Single tumour (extranodal) with regional node involvement.
Two or more nodal areas on the same side of the diaphragm.
Two single (extranodal) tumours with or without regional node involvement on the same side of the diaphragm Two or more nodal areas above and below the diaphragm.
All the primary intrathoracic tumours.
All extensive primary intra-abdominal disease.
All paraspinal or epidural tumours. Details of this classification for children are given in Table II. Cases over which there was disagreement were reexamined, when complete clinical and follow-up information was made available to the pathologists. When both reviewers agreed on the diagnosis of BL, the case was included in the final series.
Clinical data.-The cases considered to be BL were first studied on the basis of a complete pretreatment evaluation, i.e., complete blood count, measurements of serum urea, uric acid, electrolytes, serum lactic dehydrogenase, bilirubin, alkaline phosphatase and transaminase, marrow aspirate (1-4 per patient. average 2), urine analysis, i.v. pyelogram, chest roentgenograms, examination of cerebrospinal fluid for malignant cells and protein, and careful clinical examination.
Staging was done using Murphy (1977) and Ziegler (1977) classifications (Table III). In cases of localized BL (Murphy, Stages I and II; Ziegler, Stages A and AR) lymphography has been performed since 1975 to avoid errors of staging.
Before 1973, patients were treated with radiotherapy at the initial site and with a mono-or bi-chemotherapeutic regimen (cyclophosphamide and vincristine) for 2 years (33 patients). After 1973, patients were treated with polychemotherapy (Gout-Lemerle et al., 1976), with a combination of cyclophosphamide, vincristine, prednisone and adriamycin (COPAD) (14 patients). Patients were judged to be in complete remission (CR) if there was no clinically palpable tumour and if all previously abnormal parameters had returned to normal. Survival curves, using the 2 staging classifications, were calculated by the method of Kaplan and Meier (1958).

Histopathology
The 2 pathologists examining histological material without knowledge of the clinical data agreed over 71/87 cases (81-6%); they agreed initially that 43 of these 71 were BL.
Results of the second review (in the light of clinical data and of the other reviewer's opinion) are summarized in Table IV. Agreement was reached to discard 4 cases that were not considered was final disagreement, they were considered as BL by R.G.M., and as lymphoblastic non-convoluted, histiocytic noncleaved, mixed histiocytic, and unclassified by P.A.B. The data can be summarized as follows: (1) After discussion, 47 cases were diagnosed as BL. The 2 pathologists agreed that, in Giemsa-stained sections, BL is a diffuse lymphoma composed of monomorphic lymphoid cells of uniform maturity with a round nucleus and 3-4 nucleoli; they have a moderate amount of cytoplasm, which is well defined, deeply basophilic and usually contains clear vacuoles.
(2) One of us (P.A.B.), using a histomorphometric technique (Bryon et al., 1981), identified 2 types of cell in the BL group: 30 cases with a typical, monomorphic, small non-cleaved cell, and 17 cases with a slightly larger cell. No differences were found between the 2 groups with regard to age, staging, initial presentation or survival (Philip et al., 1980b).

Clinical data
The results obtained from the 47 cases are summarized in Figs. 1-4, and Tables V and VI.
Sex and age distribution There were 37 males and 10 females among the patients (ratio 3.7: 1).
The age distribution of the patients is given in Fig. 1, indicating a peak incidence between the ages of 6 and 9 (18/47 patients). The distribution between 21 and 16 years shows an average for the whole group of 8-5 years. Table V, by far the most frequent site at presentation (in 68% of cases, 32 patients) was the abdomen. Lymph-node involvement was seen initially in 9 patients (19.1%) (7 with head   and neck localization) and other sites in 6 patients; viz. s.c.

Staging
Two major classifications made it possible to distinguish 2 groups of patients. Localized disease (Murphy Stages I and II, Ziegler Stages A and AR) was seen initially in 14 patients (29.8%). This group of patients clearly has a good prognosis in both staging systems (c.f. Figs Fig. 2. series). BL accounted for 56.6% of our 83 cases. The fact that this value is higher than in other series reported in the literature (Garwicz et al., 1974: 19%- . 4). When the 13 patients who are treated. Lymphoblastic or non-abdomot go into remission are excluded, inal lymphomas in children frequently 3urvival of the remainder reaches present as leukaemias and are usually h. Achievement of CR is thus a received by other hospitals in the area, r factor for a good prognosis. whereas cases of BL, which is most o kinds of relapse can occur during frequently a solid abdominal tumour, are nent (Table VI). CSF involvement is usually sent to the Centre Leon Berard. early complication, especially in It is quite surprising to read the recent ats with extensive disease, and ocreport by Wollner et al. (1980) of 35 d in 6/29 cases at 1-4 months of abdominal lymphomas in children, indition. Later, there may be local cating that none of them was a BL. It is rence (9/29) and secondary marrow our opinion that Berard's report that BL vement (8/29). Other sites of relapse accounts for one third of childhood NHML rare (1 in lymph nodes and 1 in the outside Africa (Cossmann & Berard, 1980) ).
is more likely to be epidemiologically e survival curve was very different correct than are our data. We (Lenoir that of other NHML in children et al., unpublished) have found 635 cases 4). Most fatal relapses occurred of BL reported in the literature that 3 8 months of evolution even though should be considered to be non-endemic of the patients who relapsed before BL. The present report represents 7 % of nths died later (1 after 20 months, the cases published in the literature. It is er 4 years). Twenty patients were our opinion that most of the abdominal with no evidence of disease (NED) lymphomas reported as "undifferentiated niths after achieving CR, and 18 are lymphomas" by the classification of Rapclive and can be considered cured. paport, could be in fact true BL; our data relapsed with marrow involvement are not exceptional, and BL must be .d 32 months after diagnosis (Table distinguished pathologically  These data also show that Burkitt-type lymphoma is a clear clinical entity. The age-incidence curve, which indicates a peak before the age of 10 years, is similar to that of the recent Olweny report from Uganda (Olweny et al., 1980) and to other reports from different parts of the world (Lenoir et al., unpub.). The finding in the present series that BL is mainly (68%) an abdominal lymphoma confirms previous reports from outside Africa (Dorfman, 1965;Hoogstraten, 1967;Levine et al., 1982). Other initial sites of presentation were also found, especially lymph-node involvement, but jaw tumours were found in only 2 cases, whereas 72% of patients described in Uganda showed a jaw swelling (Olweny et al., 1980). There is thus a clear clinical difference between African and non-African BL. Nevertheless, BL should not be a synonym for "jaw tumour", as numerous authors have assumed, who identified BL only when a jaw tumour was present (Willemin et al., 1966;Cuevas et al., 1976;Joncas & Rioux, 1977). This assumption represented a major bias in the diagnoses of the 625 cases cited above (Lenoir et al., unpub).
From our data, the characteristics of BL outside Africa are summarized in Table VII. We have found that patients alive NED 8 months after CR can be considered to have a 90 % chance of complete cure. This short evolution suggests that non-endemic BL could be used as a model for therapeutic research. In our opinion, such research should be carried out primarily on the 3 principal sites of complication (CSF, local recurrence and marrow). One of the most interesting avenues of research in BL is the use of intensive chemotherapy followed by mar-row autografts (Ziegler, 1977;Appelbaum et al., 1978). Recent observations by our group (unpublished) confirm the high efficiency of this regimen even in patients in a second CR.
BL is thus a clinico-pathological entity in Caucasian children. It is also an immunological entity; when 10 of our 47 cases were studied by immunological methods, all tumour cells were found to be B cells with surface immunoglobulins. This finding is confirmed by all the data in the literature (Mann et al., 1976). BL is also a cytogenetic entity: in 4 of our cases, the tumour cells were studied by banding techniques, and 3 translocations of (8;14) and one variant translocation, t (8;22), were found (Bertrand et al., 1981). The 8; 14 translocation, initially reported in African cases as a 14q + (Manolov & Manolova, 1972), has been found in most non-endemic BLs (Zech et al., 1976). Two variant translocations, t (2 ;8) and t (8 ;22), were then described in both endemic and non-endemic BL (Bernheim et al., 1981). These translocations are now considered to be a characteristic feature of BL in all parts of the world. Finally, whereas 96% of endemic cases are EBV-associated, an association rate of only 10-20% was found in this series (Philip et al., 1980a;Lenoir, unpub.).
It must be emphasized that the present study is strictly pediatric. Levine's studies in the US (Levine et al., 1982) clearly show that BL is not only a pediatric lymphoma, and unpublished data by members of our group confirm this conclusion.
The only major differences apparent between non-endemic and endemic BL are the high incidence, the high rate of EBV association and the high jaw-tumour TABLE VII.-Summary of characteristics of non-endemic Burkitt-type lymphoma 1. In Western countries, BL represents at least one third of childhood NHML. 2. The primary site of the tumour is usually abdominal or pelvic. 3. The disease is characterized by acute evolution in the absence of therapy. 4. The immediate risk, especially for extensive disease, is CFS involvement. 5. The second risk is local recurrence or secondary marrow involvement.
6. The evolution of the disease is short, 90% of relapses being < 8 months from diagnosis. 7. Children alive NED after 12 months are probably cured. frequency of the latter. The relationship between incidence and EBV association should be considered to favour the role of EBV in the pathogenesis of African BL. Our report of 100% EBV association in Arab Algerian children from a nonendemic area may indicate, however, that socio-economic level is more important than incidence or geography as an explanation of the EBV association (Philip et al., 1 980a).
BL is first and foremost a histopathologically distinct lymphoma. A retrospective analysis should now be made of large series of pathological specimens in order to establish the actual incidence of BL among childhood NHML. The term "Burkitt-type lymphoma" should be used in all classifications.