Prognostic Factors in Childhood Acute Lymphoblastic Leukaemia J. Armstrong, M. Pomeroy, J. J. Fenneily, *S. Cahalane
Department of Oncology, the Children's Hospital, Temple Street, Dublin. Summary 58 patients under the age of ~ years with acute lymphoblastic leukaemia were managed from 1971 to 1985. We analysed their overall survival from diagnosis to assess the prognostic significance of clinical and laboratory features present at diagnosis. Factors which were statistically significant included white blood count, platelet count, palpable lymph node enlargement, mediastinal widening on chest x-ray and palpable splenic enlargement. The purpose of this study was to identify that subset of patients which might have benefitted from more intensive treatment. Introduction Prognostic stratification is important in leukaemia and other malignancies. It identifies those patients likely to achieve a good response to standard therapy and'. indicates the probability of cure. In addition it identifies those patients at diagnosis who are at high risk of relapse and where treatment intensification, including investigational regimes may be justified. Such patients may be for allogenic bone marrow transplant. Autologous bone marrow transplant is in the experimental phase and conceivably could have a role in the management of certain subsets of patients. Patients and methods -- We studiedpatients whore~eivedtheir treatment at the Children's Hospital, Temple Street, Dublin and were treated by one of the authors*. All patients except one, were caucasians and were all below the age of 14. There were 35 males and 23 females. In the early years of the study at the time of diagnosis of most of the patients immune typing was not available. This was subsequently carded out by Dr. M. Greaves of the Imperial Cancer Research Fund in the United Kingdom. Immune typing of 27 patients revealed that 23 were of the common ALL type, 3 were of the T cell type and one of the null type.
Diagnosis was conf'Lrmed by peripheral blood examination, bone marrow examination and by cytochemical staining, specifically using the Periodic Acid Schiff reaction. When they became available chromosomal studies were carried out. All patients were managed in a single room in a manner reported previously1. In recent years a laminar air flow unit has been utilized, fifty-four patients were enrolled in the United Kingdom Acute Lymphoblastic Leukaemia (UKALL) trials of the time2 and were managed according to protocol. Three patients were treated in 1971 with standard induction by vincristine and steroids, re-induction using duanorubicin, cytosine, arabinoside and 6-thioguanine and maintenance with mercaptopurine and methotrexate. One patient with T cell leukaemia and mediastinal widening on chest x-ray was treated with the Berlin Frankfurt Munster Regime3. Survival rates were calculated using life table methods and differences between groups were tested using the logran K test*. Results 54 of 58 patients (93%) achieved remission. Between 1971 and 1973 four patients died during induction. One patient was admitted to hospital with meningeal leukaemia and a white cell count of 499 x 109/litre, a platelet count of 7 x 109/litre and a haemoglobin of 4.9 gms. This patient died 1 day after admission. A second patient was admitted to hospital in fulminant renal failure due to uric acid nephropathy, had a white cell count of 156 x 109/litre, and died three days after hospital admission. One child developed a .staphlococcal septicaemia and died six days after diagnosis. One child who received initial induction therapy at another hospital, was admitted to our hospital with a haemophilus influenzae pneumonia which failed to respond to intensive antibiotic treatment and died 2 weeks after diagnosis. Cumulative survival at five years was 53.38% (S.E. 0.07). Figure 1 shows 40
survival versus time. At five years the curve began to plateau. We also evaluated patients for 10 years overall survival. This was 467% (S.E. 0.08). Forty-seven patients were aged between 2 and 10 years. This group had a 63 % 5 year survival. 9 patients were less than 2 years at diagnosis, and 2 of these patients are currently alive and disease free, five years and nine months, and four years after diagnosis respectively. Two patients died shortly after diagnosis, one discussed above with wCC 499x 109/litre and meningeal leukaemia, and another patient with haemophilus pneumonia. Five patients achieved remission but subsequently relapsed and died at a mean of thirty months after diagnosis. One of these patients had clinically palpable splenomegaly and one patient had a white cell count of 57 x 109/litre and a platelet count of 10 x 109/litre. Hence three of these seven patients had adverse prognostic factors. Two. patients were over 10 years old at diagnosis and both died within two years. Analysis of survival according to sex showed an initially superior survival for girls which later approximated that of the boys and there was no statistically difference (p=0.48). White cell count at diagnosis was found to be significant. The lowest tertile of the group (wcc below 5.0 x 109/litre) had a survival of 80% at 5 years. The middle tertile (wcc 5.1-13.0 x 109/litre) had a 5 year survival of 61% (p--0.001) patients who had a wcc less than 20 x 109/litre had a survival of 66%; those with a wcc of above this had a rate of 12% (p=0.001). The lowest terfile of platelet counts at presentation (below 20 x 109/litre) had a 5 year survival of 26%. - The middle tertile (20.1-93.0 x 109/litre) had a rate of 62% and the highest tertile had a rate of 72% (p=0.014). Haemoglobin levels at presentation were of no prognostic significance. Patients with palpable nodal involvement had a 74% mortality at 5 years. Those without palpable nodes or with
VoL ,59 No. 2
Childhood acute lymphoblastic leukaemia 41 109/litre, and died two years and eight months from diagnosis. Table I summarises the prognostic factors at diagnosis of acute lymphoblastic leukaemia found to be statistically significant.
100
80
Discussion
40 t.) 1.1.1
IX.
20
0
i
I
2
I
i
4
I
I
6
I
i
8
I
I
10
TIME (YEARS) Fig. 1 - Overall survival in acute active lymphoblastic leukaemia (A.L.L.) S.E. =0.08. nodes thought not to be significant had a 5 year mortality of 37% (p=0.017). Those with mediastinal lymph node enlargement on chest x-ray had a statistically significant poorer survival (p=0.018). Those with palpable splenic involvement had a 5 year mortality of 71% versus 34% in those without splenomegaly (p=0.02). Hepatic involvement did not confer a worse prognosis. 16 of 58 patients presented with bone pain or joint related symptoms. The presence of these symptoms or of abnormal skeletal roentgenograms did not confer a statistically significant change in prognosis. Ten patients presented with weight loss had a 79% five year mortality versus 43% for those without (p=0.162). Immune pheno type was known in 27 patients of whom 23 common were ALL. All 3 patients of the T cell type were boys, with mediastinal widening on chest x-ray at diagnosis. One patient aged 9 years, had a white cell count of 156 x 109/litre, gross lymphadenopathy and hepatosplenomegaly and died three clays after diagnosis. A second patient who
had a white cell count of 137 x 109/litre, gross lymphadenopathy and splenomegaly, achieved remission but had a bone marrow and meningeal relapse and died one year and four months after diagnosis. One boy had a normal white cell count 7.2 x 109/lilre but had marked lymphadenopathy and splenomegaly at diagnosis. He is alive and apparently disease free at close of study two years and nine months after diagnosis. Only 1 patient was documented to be of NULL cell type. He was 1 year and five months at diagnosis and had a white cell count of 43.8 x 109./litre. He is alive at study closure without relapse (survival to date three years and three months). Marrow lymphocyte karyotype analysis showed hyperdiploidy in one patient, who is alive and disease free six years after diagnosis. One "pseudodiploid" patient diagnosed at five years and seven months had a normal white cell count at diagnosis. He died one year and five months from diagnosis. One patient had Down's Syndrome. He was age three years at diagnosis, had a white cell count of 30 x
We identified several statistically significant prognostic factors, high white cell count, low platelet count, clinically detectable lymph node enlargement, and mediastinal lymph node enlargement on chest x-ray (Table I). We did not establish prognostic significance for sex, clinically detectable hepatomegaly, weight loss, skeletal pain or radiological abnormalities or haemoglobin levels. Age is generally found in the literature to be an important prognostic factor 5,6,7,8.9and though most patients in this study (45/ 58) were aged between 2 and 10 years those less than 2 years or over ten years had an increased incidence of relapse. Overall survival figures at 5 and 10 years mirrored closely the figures of Carr et al (Irish Journal of Medical Science, January 1988, 2-4). Palpable lymphadenopathy conferred a bad prognosis in our patients and this was statistically significant. Other studies have shown decreased survival in those with bulky adenopathy but generally this has not been statistically significant 1~ The presence of bulky visceral disease has been recognised as signifying a poor prognosis. In our study palpable splenomegaly was a poor prognostic factor, as noted in other studies. Palpable hepatomegaly was not found to be a statistically significant factor though recent studies have shown such indices of liver involvement as serum glutamate pyruvate transaminase (SGPT) andlactic dehydrogenase (LDH) to be significant indicators of prognoTABLE I Statistically significant prognostic prognostic factors at diagnosing acute lymphoblastic leukaemia. Prognostic factors High white cell count at diagnosis 20 x 109/litre Low Platelet Count 20 x 109/litre Palpable lymph node enlargement Mediastrinal widening on chest x-ray Palpable splenomegaly
P Value 0.001 0.014 0.017 0.018 0.02
i.J,M.S. February, 1990
42 Armstrong et al. sis t2d3. White cell count elevated at diagnosis has been accepted for many years as indicating poor prognosis. We showed a graded survival according to white cell count even at the lower levels. Those with counts below normal had the best prognosis. The analysis of data in most previous studies has concentrated on the progressive decrease in survival in proportion to a white cell count elevated above normal. The majority of patients in the study group, however, were accrued during the 1970's. In the 1980's treatment programmes such as Berline Frankfurt Munster Regime and UKALL have improved results for patients with high white cell counts over 50 x 109/litre though there may be increased treatment toxicity. Immune type is a significant prognostic factor. The "common" ALL type indicates the best prognosis t4. The majority of patients in most series are in this category. T cell leukaemia occurs more commonly in boys, particularly in the older age group, in association with a high white cell count and abnormal chest x-ray, and therefore includes a number of independent poor prognostic variables. The null type of ALL is recognised as having an intermediate prognosis. Reports on the prognostic significance of patient sex have been conflicting. Some reports indicate a generally better outcome for females 15,tr.tT,ls.t9,whereas other studies have not shown improved survival for females3,2~ In some of these studies poor outlook for boys can in part be attributed to T cell leukaemia and the occurrence of testicular relapse. 16 of 58 patients presented with bone pain or joint related symptoms. Other studies have shown a similar frequency at presentation z2,2~. These symptoms are not related to an adverse prognosis. However, they may be the only initial manifestation and this may lead to a delay in diagnosisu. If the patient is misdiagnosed and treated with Prednisone for a prolonged period it may worsen the prognosis 22. Haemoglobin levels were not found to be of prognostic significance. A haemoglobin level of above 10 g has been shown to confer a poor prognosis n. It has also been shown that severe anaemia at presentation (haemoglobin
below 5 g) confers a poor prognosis 25. Other factors of recognized prognostic significance which were not assessed in this study, include negro race and lower socioeconomic group. Whereas several studies27,2s,29 have shown chromosomal abnormalities to be prognostically significant, only three patients were documented in this study to have such abnormalities. Hyperdiploidy is associated with a more favourable prognosis than normal karyotype and the one patient with hyperdiploidy responded well to treatment. Both "pseudodiploidy" and hypodiploidy are indicators of a poor prognosis. The one hypodiploid patient in this study died one year and five months from diagnosis. Recent improvements in cytogenetic techniques have allowed the detection of subtle chromosomsal abnormalities in 60 to 90% of children w i t h A L L 3~ While the prognostic significance of such abnormalities as translocations was not addressed in this study, it has been noted32 that there is poor prognosis in such translocations as T (8, 14), T (4:11) or T (9:22). A number of proto-oncogenes are located at sites of recurring chrosomal rearrangements. There is a close correlation between, translocation 8:14 and B-cell ALL (FABL3) and ph-positive acute lymphoblastic leukaemia is an unfavourable prognostic factor-u. It is likely that further progress in this area will improve our understanding of the pathogenesis of acute lymphoblastic Leukaemia. The aim of this study was to identify the features in our patients which confer a poor prognosis. Resulting from such studies patients are now allocated to more intensive therapy, programmes B.F.M.; or U.K.A.L.L. with or without consolidation. This has been shown to improve European survival32,33,34,35,36.In the BFM studies they have deyised a simple mathematical model to assign their patients to prognostic groups which are given appropriate therapy37. Thus by intensification of chemotherapy regimes the outlook for "poor prognosis" patients can be improved. It is hoped that this strategy, with the possible future use of new techniques, such as allogenic bone marrow transplant will continue the progress made in the management of this disease.
References
1. Fennelly, J. J., O'Donoghue, S, Kavanagh, T. et al. Management of childhood acute leukaemia. Journal of the Irish Medical Association. 1974: 67, 589-594. 2. Petn, J., Eden, O. B., Liffeyman, J. et al. Improvement in prognosis for children with acute lymphoblastic leukaemia. Lancet 1986: 408-411. 3. Berlin, Springer-Verlag, Lampert, F., Heinz, G., Langerman, H-J. et al. Acute lymphoblastic leukaemia: Current status of therapy in children. Recent results in cancer research. Vol. 93, 1984. 4. Peto, R., Pike, M. C., Armitage, P. et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. British J. Cancer. 1977. 5. Crist, W., Pullen, J., Boyett, J. et al. Acute lymphoid leukaemia in adolescents: Clinical and biologic features predict a poor prognosis - a pediatric oncology groups study. Journal Clinical Oncology 1988: 6, 34-43. 6. Blacklock, H., Matthews, J., Buchanan, J. et al. Improving survival poor acute lymphoblastic leukaemia in adolescents and adults. Cancer 1981: 48, 1931-1935. 7. Amadori, S., Meloni, G., Baccarani, M. et al. Long term survival in adolescent and adult acute lymphoblastic leukaemia. Cancer 1983: 52, 30-34. 8. Jacobs, A. D., Gale, R. P. Recent advances in the hology and treatment of acute lymphoblastic leukaemia in adults. N., England Medical Journal 1984: 311, 1219-1231. 9. Crist, W., Pullen, J., Boyett, J. et at. Clinical and biologic features predict a poor prognosis in acute lymphoid leukaemia in infants. A pediatric oncology group study, Blood 1986: 67, 135-140. 10. Simonem J. J., Verzosa, M. S., Rudy, A. J. et al. Initial features and prognosis in 383 children with acute lymphocytic leukaemia. Cancer 1975: 36, 2099-2108. 11. Miller, R., D., Leikin, S., Albo, V. et al. Prognostic factors and therapy in acute lymphoblastic leukaemia of childhood. CCG141. Cancer 1983: 51, 1041-1049. 12. Raotanen, J., Simens, M. A. Elevated serum transaminase activity at diagnosis is associated with rapidly progression disease in children with acute lymphoblastic leuk-aema. Cancer 1988: 61,754-757. 13. Piu, C-H., Dodge, R. K., DaM, G. V. et al. Serum lactic dehydrogenase level has prognostic value in childhood acute lymphoblastic leukaemia. Blood, Vo166, No. 4, October '85, pp. 778-782. 14, Fliedner, V. Von. Contribution of immunological markers to the diagnosis and prognosis of human leukaemia. Eur. J. Cancer Clin Oncol. 1984: Vol. 21, No. 3,347-353. 15. Sather, H., Miller, D., Besbit, M. etal. Differences in prognosis for boys and girls with acute lymphoblastic leukaemia. The Lancet 1981: 1,739-43. 16. Mutts, M. J. Baumer. Sex and prognosis in childhood acute leukaemia. Lancet 1978: ii, 128-129. 17. Motts, J. Baumer, Gentle, T. Sex and prognosis in childhood acute lymphocytic leukaemia. Lancet 1978: ii, 673.
Vol.159 No.2
18. Robinson, L, Sather, H., Coecia, P. et al. Assessment of the interrelationship of prognostic factors in childhood acute lymphocytic leukaemia. Amer. J. Ped. Hem. Oncol. 1980: 2,5-13. 19. Acute lymphocytic leukaemia. U.K. Leukaemia Research Trials for Children and Adults. London: Leulaemia Research Fund 1987: 2032. Kay, H. 20. George, S., Fembach, D., Vietti, T. et al. Factots influencing survival in paediatric acute leukaemia. Cancer 1973: 32, 1542-53. 21. Evans, D., Morris-3ones, P., Hann, I. et al. Sex and prognosis in childhood acute lymphoctic leukaemia. Lancet 1978: ii, 522-523. 22. Revesz, T., Kavdos, G., Kajtar, P. et al. The prognostic significance of bone and joint manifestations in childhood leukaemia. Journal Rheum. 1982: 3,647-648. 23. Costello, P. B., Bleacher, M. L., Stair, J. C. et al, A prospective analysis of the frequency, course and possible prognostic significance of the joint manifestations of childhood leukaemia. Journal Rheumatol. 1983: I0, 753757. 24. Schaller, ]. Arthritis as a presenting manifestafiou of malignancy in children. 3. Paediatric 1972: 81,793-797. 25. Waiters, T. R., Buchare, M., Simona, J. et al.
Childhood acute lymphoblastic leukaemia 43 Poor prognosis in negro children with acute lymphoblastic leukaemia. Cancer 1972: 29, 210-214. 26. Hicosonmcz, G., Ozsoylu, S., Yetgin, S. et al. Poor prognosis of childhood acute lymphoblastic leukaemia. British Medical Journal. 1983: 286, 1437. 27. Michael, P. M., Garson, O. M., Ekert, H., Tauro, G., Rennis, G. C. and Pilkington, G. P. Prospective study of childhood acute lymphocytic leukaemia: Haemologic, immunologic and cytogeneticcorrelations. Medical and Paediatric Ontology. 1988: 16153-161. 28. Le Beau, M. M., Rowley, J. D. Recurring chromosomal abnormalities in leukaemia and lymphoma, pp. 371-394. Cancer Surveys, Vol 3, No. 3, 1984. 29. Secker-Walker, L. M., Sylvia, D,, Tawler, R. M. Hardisty. Prognostic implications of chrosomal findings in acute lymphoblastic leukaemia at ~dagnosis. British Medical Journal 1978: 2, 1529-1530. 30. Secker-Walker, L. M., Swansbnry, G. L, Hardisty, R. M. et al. Cytogenetics of acute lymphoblastic leukaemiain children as a factor in the prediction of long-term survival. Br. L Haematol. 1982: 52, 389-399. 31. Williams, D. L., Raimondi, S., River/a, G. et al. Presence of clonal chrosome abnommali-
ties invirtually all cases of acute lymphoblastic leukaemia. N. Eng. J. Med. 1985: 313, 640- 641. 32. Williams, D. L., Harber, J., Murphy, S. B. et al. Chromosomal translocations play a unique role in influencing prognosis in childhood acute lymphoblastic leukaemia blood. 1986: 68, 205-212. 33. Third International Workshop on Chrosomes in Leukaemia, 1981. 34. Bloomfield, C. D., Brunning, R. D., Smith, K.A. et al. Cancergenetics and cytogenetics 1,229-238 35. Therapeutic manoeuvres and the prolongation of remission in childhood acute lymphocytic leukaemia with poor prognostic factors. Preliminary report - S.W.O.G. Study 7420. Abstract 767. Proc. A.A.C.R. 1979: 20, 190. 36. Moe, P. G., Siep, M., Finn, P. H. et al. Intermediate dose methotrexate CI.D.M.) in childhood acute lynmphoeytic leukaemia in Norway. Acta Paediatric Stand. 1981: 70, 73-79. 37. Gaynor, P. S., Bleyer, W. A., Steirdaerg, p. G. et al. Modified BFM therapy for children with previously untreated acute lyrnphoblastic leukaemia and uafavourable prognostic features. The American Journal of Paediatric Haematology/Oncology 1988: 10(1), 42-50.