Outcome of relapsed or refractory childhood B-cell acute lymphoblastic leukaemia and B-cell non-Hodgkin's lymphoma treated with the UKCCSG 9003/9002 protocols


Dr Ayad Atra, Department of Paediatric Oncology, The Royal Marsden Hospital, Downs Road, Sutton, Surrey SM2 5PT, UK. E-mail: ayad.atra@rmh.nthames.nhs.uk


Twenty-six children with B-cell acute lymphoblastic leukaemia (B-ALL) or Murphy Stage III or IV B-cell non-Hodgkin's lymphoma (B-NHL) progressed or relapsed after first-line therapy with a short, intensive multiagent chemotherapy regimen [United Kingdom Childhood Cancer Study Group (UKCCSG) 9003] (n = 62) or a slightly less intensive regimen (UKCCSG 9002) (n = 112). Eight patients (4·6%) never achieved complete remission (CR) and 18 (10·3%) relapsed. Second-line therapy resulted in remission for eight patients (30%). All patients initially treated with the 9003 protocol died. Three patients (11·5%) in the 9002 group, including one who never achieved CR in the primary site, are alive after second-line therapy. This study confirms that the prognosis of relapsed or refractory B-ALL/B-NHL is poor and exceptionally so if relapse occurred less than 6 months from diagnosis. High-dose therapy with stem cell rescue was used in only seven patients; its role needs to be studied further.

The reduction in early toxic death and the increase in the intensity of systemic and central nervous system (CNS)-directed therapy dramatically improved the cure rate of children with B-acute lymphoblastic leukaemia (B-ALL) and B-non-Hodgkin's lymphoma (B-NHL) (Reiter et al, 1995; Atra et al, 1998). Relapse remains a major concern as the results of treatment following relapse and especially for primary resistant disease are disappointing, particularly in patients treated with intensive first-line therapy (Philip et al, 1993). High-dose systemic cytarabine and etoposide may salvage some relapsed or unresponsive cases (Gentet et al, 1990). High-dose chemo/radiotherapy with stem cell rescue may cure some patients, although its role is difficult to assess because of small numbers (Philip et al, 1993) and the value of adjuvant therapy using immunotoxins or monoclonal antibodies remains to be clarified (Grossbard et al, 1993). In this study, we reviewed all patients who had primary resistant disease or relapsed after first-line therapy with the United Kingdom Children Cancer Study Group (UKCCSG) 9003/9002 protocols to assess their curability after more intensive chemotherapy with or without megatherapy and stem cell rescue.

Patients and methods

Between 1990 and 1996, children with B-ALL (defined as the presence of more than 70% L3-type blasts in the bone marrow or more than 25% L3 blasts in the bone marrow with bone pain or myelosuppression) or stage IV B-NHL with up to 70% L3 blasts in the bone marrow and CNS disease were treated with the UKCCSG 9003 (n = 62) protocol (Atra et al, 1998). Those with Murphy Stage III or IV B-NHL and up to 70% L3-type blasts in the bone marrow (but with no CNS disease) were treated with the 9002 (n = 112) protocol (Atra et al, 2000). These protocols were identical to the French LMB 86 and 84 respectively. Patients' characteristics and details of treatment and follow-up are shown in Table I. Eight patients, two on 9003 and six on 9002, did not achieve complete remission (CR) and 18 patients relapsed after achieving CR, 10 in 9003 and eight in 9002 group. All patients progressed or relapsed in the first 6 months from diagnosis except six patients, two in the 9003 group who relapsed at 8 and 10 months and four in the 9002 group who relapsed at 8, 15, 24 and 37 months from diagnosis. Sites of relapse/progression after the 9002 protocol were bone marrow (n = 3), primary site (n = 6), CNS (n = 1) and multiple sites (n = 4), and after the 9003 protocol were bone marrow (n = 2), CNS (n = 6) and combined CNS and bone marrow (n = 3), and one patient relapsed in the bone marrow and jaw. In the 9002 group, treatment post relapse was palliative (n = 2), more intensive chemotherapy using the high-dose systemic cytarabine and etoposide regimen (CYVE) of the 9003 protocol (Atra et al, 1998) (n = 4) or cyclophosphamide, doxorubicin, vincristine and prednisolone (CHOP) (n = 3), or the CYVE regimen followed by high-dose chemo/radiotherapy and autologous stem cell rescue (n = 4), or cyclophosphamide/total body irradiation (TBI) and matched unrelated donor bone marrow transplantation (BMT) (n = 1). Treatment following relapse in patients initially treated with the 9003 protocol was palliative (n = 6), chemotherapy with or without radiotherapy (n = 4) or second-line chemotherapy, followed by cyclophosphamide/TBI and allogeneic bone marrow transplantation BMT (n = 2).

Table I.  Patients' characteristics and details of treatment and outcome.
First-line therapy
Time of relapse (months)
from diagnosis

Site(s) of relapse/Progression

Treatment post relapse

  1. CR, complete remission; BM, bone marrow; PBSC, peripheral blood stem cells; CNS, central nervous system; RT, radiotherapy; Cy\TBI, cylcophosphamide\total body irradiation; MS-BMT, matched sibling BMT; MUD-BMT, matched unrelated donor BMT; ABMT, autologous bone marrow transplant; HDT, high-dose therapy; BEAM, BCNU (carmustine), etoposide, cytarabine, melphalan.

190025BMCYVEDied of lymphoma, 1 month after relapse
290028BMCYVE + BEAM + ABMTDied of lymphoma 7 months after relapse
39002Never achieved CRPrimary site (abdomen)CHOP + etopsoside + ifosphamideDied of lymphoma 2 months later
49002Never achieved CRPrimary site (abdomen)CHOPDied of lymphoma 3 weeks later
5900224Primary site (abdomen)CHOP + RTAlive in CR, 6 years
69002Never achieved CRBM + primary site (abdomen)PalliativeDied one month after diagnosis
790026BM + primary site (abdomen)CYVEDied of lymphoma 2 months after relapse
8900237Primary site (abdomen)CYVE + BEAM + autologous PBSCAlive in CR, 3·5 years
99002Never achieved CRCNSCYVEDied of lymphoma 6 months after relapse
1090024·5BM + JawCYVEDied of lymphoma 4 months from relapse
1190026BM + CNS + Primary site (abdomen)CYVE + Cy\TBI + melphalan +
 autologous PBSC
Died of lymphoma 5 months post relapse
12900215BMCYVE + Cy\TBI + MUD-BMTRemained in CR for 3 years. Relapsed again in BM
Died of lymphoma 15 months later
139002Never achieved CRPrimary site (abdomen)PalliativeDied of lymphoma 6 months after diagnosis
149002Never achieved CRPrimary site (abdomen)HDT + ABMT resulted in CRAlive 4·5 years after diagnosis
159003Never achieved CRCNSPalliativeDied of lymphoma 9 months after diagnosis
1690036BM + CNS + AbdomenPalliativeDied of lymphoma 3 weeks after relapse
17900310BMCy\TBI + MUD-BMTDied of lymphoma 1 month after BMT
1890036CNSCYVEDied of lymphoma 3 months after relapse
1990035CNSCYVE + Cy\TBI + MS-BMTDied of lymphoma 3 months after relapse
2090036CNSPalliativeDied of lymphoma one month after relapse
2190036BMPalliativeDied of lymphoma 2 weeks after relapse
2290038BM + JawPalliativeDied of lymphoma 4 months after relapse
2390035BM + CNSPalliativeDied of lymphoma one week after relapse
2490035CNSCraniospinal RT and dexamethasoneDied of lymphoma 2 months after relapse
259003Never achieved CRBM + CNSCYVEDied of lymphoma 11months from diagnosis
2690035CNSCYVEDied of lymphoma 2 months after relapse


Seven of the eight patients who did not achieve CR died 1–11 months after diagnosis (median 5·5 months). The remaining patient continued to have residual lymphoma (biopsy proven) at the primary site. He went into CR after high-dose chemo/radiotherapy and autologous bone marrow rescue and is alive 4·5 years after diagnosis.

Eight of the 18 patients who relapsed after achieving CR entered second remission after second-line therapy, six in the 9002 group and two in the 9003 group. All patients who relapsed/progressed after initial treatment with the 9003 group died owing to recurrent lymphoma (n = 11) or treatment-related toxicity (n = 1), regardless of the time of relapse, the second-line treatment they received or whether they achieved CR or not. One patient in the 9002 group who relapsed in the bone marrow 15 months from diagnosis achieved a long second CR (2·5 years) after further chemotherapy (CYVE) and matched unrelated bone marrow transplantation following cyclophosphamide/total body irradiation (TBI) conditioning. He developed a second bone marrow relapse and died 15 months later. There were three survivors in the group initially treated with the 9002 protocol. Two of the survivors had relapsed at 24 and 37 months from the initial diagnosis and the third survivor achieved CR after high-dose chemotherapy and autologous BMT (details above). With a median follow-up of 3·6 years from diagnosis (range 2–6 years), three patients who were initially treated with the 9002 protocol remain alive (11·5%).


The cure rate of childhood B-ALL and B-NHL has dramatically improved in recent years (Reiter et al, 1995; Atra et al, 1998, 2000). This improvement is attributed both to a reduction in the rate of toxic death as well as the relapse rate following the use of increased intensity systemic and CNS-directed therapy (Patte et al, 1991; Cairo et al, 1996).

Relapse remains a major concern as the cure rate is markedly reduced post relapse, particularly in patients treated with intensive first-line therapy (Philip et al, 1993; Atra et al, 1998, 2000). Relapse can occur at any site, but most commonly in the bone marrow or CNS in patients with initially advanced disease. The increasing intensity of high-dose systemic cytarabine and intrathecal therapy had led to a marked decline in the risk of CNS relapse without the use of radiotherapy (Bowman et al, 1996; Atra et al, 1998, 2000). High-dose chemo/radiotherapy and stem cell rescue with or without monoclonal antibodies or immunotoxins may cure some patients, although the overall results remain disappointing (Grossbard et al, 1993; Philip et al, 1993). In addition to the stage of disease at diagnosis, patients with bulky abdominal lymphoma, CNS disease or high lactate dehydrogenase (LDH) level at diagnosis have a less favourable outcome (Reiter et al, 1995). Renal or infectious complications are not uncommon at the initial presentation or following induction of treatment and may delay the initiation or the delivery of chemotherapy. It is a concern that delay or modification of primary treatment could increase the risk of relapse, although this concern was not confirmed in a recent study (Atra et al, 1998). As the success rate in curing relapsed patients is very low, particularly in patients initially treated with intensive chemotherapy regimens, many patients were not offered second-line treatment. High-dose chemotherapy with or without radiotherapy and stem cell rescue may cure some relapsed patients (Frappaz et al, 1993; Philip et al, 1993) and the use of monoclonal antibodies and immunotoxins to control minimal residual disease may induce a durable response with acceptable toxicity (Grossbard et al, 1993).

In the UK, children with B-ALL and advanced stage B-NHL were successfully treated during the early 1990s with the UKCCSG 9003 and 9002 protocols, which are identical to the French LMB 86 and 84 regimens respectively (Atra et al, 1998, 2000). The main causes of failure were treatment-related toxic death and relapse of the underlying disease. Relapses during or after the 9002 protocol were treated with a more intensive chemotherapy using high-dose cytarabine and etoposide (CYVE) with or without high-dose chemo/radiotherapy and stem cell rescue. Historical experience shows that, without further intensive therapy, second remission is short lasting (Philip et al, 1993). It is of interest that one of the survivors in this study did not achieve CR at the site of the primary tumour until after megatherapy and continues to be in CR 41/2 years from diagnosis. High-dose therapy may be recommended for patients with biopsy-proven residual primary tumour after the first 10 weeks treatment. One patient had a long second remission after allogeneic bone marrow transplantation, which may be owing to a graft-vs.-leukaemia effect. He subsequently relapsed in the bone marrow and died of lymphoma. The survivors in this study relapsed late and there is no survivor among those who relapsed in the first 6 months from diagnosis. Another finding from this study, confirming that of others (Philip et al, 1993; Patte et al, 1999), is that patients who relapse after a relatively less intensive chemotherapy may achieve a durable remission with further more intensive treatment, particularly if the relapse occurs late. For the 9003 patients, the results were very disappointing with no long-term survivors, despite the achievement of CR in two patients. For this group, new modalities of treatment are urgently needed. Allogeneic BMT may be an option and the role of high-dose therapy in first CR and immunomodulation using interferon or monoclonal antibodies in children with relapsed lymphoma needs to be investigated. Sequential high-dose chemotherapy with autologous peripheral blood stem cell support may improve the long-term outcome of patients with relapsed NHL (Colombat et al, 1990).