The cancer conundrum.

Thirty-four patients with previously treated, advanced, low grade NHL were treated with Fludarabine, a deamination-resistant analogue of adenosine arabinoside, at a dose of 25 mg m-2 intravenously, daily for 5 days (median number of cycles = 3, range 1-10). Complete remission (CR) was achieved in six and partial remission (PR) in a further seven. Overall, responses were seen in 11/23 patients (48%) with follicular lymphoma and in 2/11 (18%) with low grade, diffuse NHL. Fifteen patients with previously treated CLL and one patient with prolymphocytic leukaemia (PLL) were also treated as above (median no. of cycles = 3, range 1-6). A partial response was seen in three of the 11 evaluable patients with CLL and CR was achieved in the patient with PLL. There were four deaths due to infection and 19 further episodes requiring admission to hospital. No other significant toxicity was reported in a total of 164 cycles of Fludarabine. This agent is active in advanced low grade lymphoid malignancy. Further studies are required to assess its role in newly diagnosed patients.

The low grade lymphoid malignancies are characterised by responsiveness to single agent chemotherapy and by the virtual inevitability of recurrence. Thereafter, although repeated responses may be achieved, the majority of patients die as a consequence of the disease. The alkylating agents, chlorambucil and cyclophosphamide, when used alone, may induce response rates of up to 60% in previously untreated patients with CLL (Knospe et al., 1974;Huguley, 1977;Sawitsky et al., 1977) but lower response rates with a shorter duration of remission are the rule for subsequent therapy (Oken & Kaplan, 1979;Montserrat et al., 1986). Similarly, in follicular NHL, response rates to single agent chemotherapy at presentation and at first and second recurrence remain about 65-70% (Israels et al., 1958;Ezdinli & Stutzman, 1965;Jones et al., 1973;Anderson et al., 1982) but, thereafter, therapy becomes rapidly less satisfactory (Gallagher et al., 1986).
The Kiel classification (Gerard-Marchant et al., 1974) characterises certain diffuse lymphomas as 'low grade' (lymphoplasmacytoid, small cell centrocytic and diffuse centroblastic/centrocytic). The majority of patients present with advanced disease; many respond to alkylating agents, but responses are rarely complete or durable. Hence, fewer than 20% of those with stage III and IV disease survive for more than 5 years (Heinz et al., 1981;Swerdlow et al., 1983;Brittinger et al., 1984;Richards et al., 1989). While responsiveness to chlorambucil falls short of that seen in follicular NHL, the benefits of more intensive therapy in centrocytic lymphoma have yet to be convincingly demonstrated (Al-Katib et al., 1984;Meusers et al., 1989). Better therapy is urgently required.
Fludarabine is a fluorinated analogue of adenosine arabinoside (Ara-A), which retains cytotoxicity while being resistant in the rapid deamination and consequent inactivation which characterises the metabolism of Ara-A (Dow et al., 1980;Plunkett et al., 1980;Brockman et al., 1980;Tavoussi & Avramis, 1986). An active triphosphate metabolite is formed intracellularly which intereferes with DNA synthesis by inhibition of DNA polymerase and ribonucleotide reductase (Huang & Plunkett, 1986 murine models has been borne out in Phase I and II testing (Hutton et al., 1984;Spriggs et al., 1986;Champagne et al., 1987). At a dose of < 125 mg/m2/course, significant activity, with response rates of 60% and more, has been reported in CLL and follicular NHL, with some modest benefit in diffuse low grade lymphoma (Grever et al., 1986;Leiby et al., 1987;Redman et al., 1988' Hochster & Cassileth, 1990Keating et al., 1989a;Puccio et al., 1990). Fludarabine is reported to be well tolerated at such doses, the toxicity in these studies being predominantly infective episodes arising from myelosuppresion.
This report describes further experience of the efficacy and toxicity of Fludarabine in 50 patients with previously treated low grade lymphoid malignancy.

Patients
Thirty four patients with NHL, 15 with CLL, and one with prolymphocytic leukaemia (PLL) form the basis of this study. All had received previous chemotherapy. Their clinical characteristics at the time of treatment with Fludarabine are shown in Tables I and II. For those with NHL, the Kiel Classifiction (Gerard-Marchant et al., 1974), and a modification of the Ann Arbor staging system (Carbone et al., 1971) were used. aIncluding one patient with prolymphocytic leukaemia. bOne patient treated on two occasions. stances. In one patient, a dose reduction of 50% was made for the first two of six cycles because of an initial low performance status. (range  Thirty four patients with NHL received 115 cycles of Fludarabine (mean 3, range 1-10) and 49 cycles were given to 15 patients with CLL and one patient with PLL (mean 3, range 1-6). In general, two cycles of therapy were given beyond the best clinically evaluable response. If no response was seen after two cycles of therapy, no further Fludarabine (4-108) was given. Full supportive therapy with antibiotics and blood products was given as appropriate.

NHL
( 1-7) Response criteria NHL CR was recorded when there was no clinical or radiological evidence of disease and PR when an estimated reduction of at least 50% of tumour volume had occurred. Any lesser response was deemed a treatment failure.
CLL and PLL Complete remission was recorded if the peripheral blood lymphocyte count was <4.0 x IO' I1, bone marrow examination showed <30% lymphocytes and there were no lymph nodes > 0.5 cm or hepatosplenomegaly (Rai et al., 1975). Partial remission (PR) was recorded if a > 50% reduction occurred at all sites of disease i.e. circulating lymphocytosis, the degree of bone marrow infiltration, lymphadenopathy and hepatosplenomegaly. Again, a lesser response was deemed a failure of therapy.

Treatment
Fludarabine was reconstituted in 10 ml of sterile water and given as a slow intravenous injection. Twenty five mg m2 were given daily for 5 days and repeated every 21-28 days as dictated by the peripheral blood count and clinical circum-Response and remission rates according to histology are shown for the 34 patients in Table III.
The response rate (CR + PR) in follicular lymphoma was 11/23 (48%) including five patients in whom CR was achieved. The likelihood of response did not correlate with the number of previous treatments, prior exposure to anthracyclines (21/23 patients), or treatment of recurrent rather than refractory disease.
Two responses were seen in the 11 patients with diffuse histology, one of whom, a 64 year old man with a 5 year history of stage IV lymphoplasmacytoid lymphoma, entered CR. Two patients died of progressive lymphoma and another of a presumed pulmonary embolus. A fourth patient died of infection (Table IV).
No response was seen in 17 patients, including three of four patients who received two cycles of Fludarabine as treatment for 'minimal residual' follicular lymphoma (<10% bone marrow infiltration (two patients, <2 cm lymphadenopathy (one patient)) as the only sites of disease after conventional therapy. Subsequently, these patients received cyclophosphamide and total body irradiation supported by autologous bone marrow transplantation (Rohatiner et al., 1991).

CLL
Eleven patients are evaluable for response. No complete remissions were seen. A partial response was observed in three patients (27%). Three patients died whilst receiving treatment; causes of death are shown in Table IV. It is noteworthy that all three died of infection whilst responding to Fludarabine. In a further five patients Fludarabine was ineffective. Four patients are inevaluable for response, either because of a coexisting second malignancy (two), or due to death from unrelated causes (two). PLL Complete remission was achieved in a 73 year old man with prolymphocytic leukaemia, in whomr no response had been seen to chlorambucil and prednisolone (Smith & Mehta, 1990). Toxicity Four patients died of infection while receiving Fludarabine (Table IV). All were neutropenic at the onset of sepsis. In patients with CLL, there was nine episodes of infection requiring admission to hospital and 19 amongst those with NHL. Other toxicity related to Fludarabine was rare, with nausea reported on only one occasion. An urticarial rash occurred in one patient. There were no episodes of central nervous system or pulmonary toxicity.

Discussion
This report describes experience of the use of Fludarabine as a single agent in a heavily pretreated group of patients with advanced disease. Despite this unfavourable setting, undoubted activity has been demonstrated in both CLL and follicular NHL, in accordance with other results (Grever et al., 1986;Leiby et al., 1987;Redman et al., 1988;Keating et al., 1989a;Hochster & Cassileth, 1990;Puccio et al., 1990).
Since the use of Fludarabine in CLL was first described in 1986 by Grever et al., its promise has been confirmed in a number of studies in previously treated patients, with response rates of over 50% (Keating et al., 1989a;Puccio et al., 1990). Such response rates already compare favourably with established programmes of combination therapy such as CVP, the M2 protocol and CHOP (Kempin et al., 1982;Montserrat et al., 1986;French Cooperative Group, 1986). Although the addition of Prednisolone may confer little extra advantage (Keating et al., 1989b), the attainment of complete remission reported by Keating et al. (1989a), with the possibilities of immunophenotypic confirmation (Robertson et al., 1990) may provide an effective starting point for the development of treatment programmes that will prolong survival in CLL.
While the response rate in CLL reported here falls short of others' experience, direct comparison with a group exhibiting so many unfavourable features may be misleading. Also, the number of patients in this study is small and 4/16 patients died of causes unrelated to Fludarabine.
Prolymphocytic leukaemia is a rare form of leukaemia sharing many features with CLL but with a universally progressive course and an unsatisfactory response to therapy. The only other patient with PLL treated with Fludarabine reported in the literature also entered complete remission (Bouroncle et al., 1990). The place of Fludarabine in the therapy of PLL therefore demands further study.
The role of Fludarabine in the treatment of NHL has, to date, been less extensively studied. Leiby et al., in 25 patients with both aggressive and indolent histologies, reported responses in eight, of whom five had follicular lymphoma (Leiby et al., 1987). Larger series reported from the M.D. Anderson Hospital (Redman et al., 1988) and the Eastern Cooperative Oncology Group (Hochster & Cassileth, 1990) also suggest response rates of 45-50% in previously treated low grade lymphoma. This report confirms that Fludarabine may be of value in the treatment of refractory follicular lymphoma.
Myelosuppression, with attendant infections, represents the major toxicity of Fludarabine and demands careful surveillance. The incidence of infective episodes, complicating 17% of cycles of treatment, is slightly higher than in other series, but again to some extent reflects the adverse features of the patient population. The timing of Fludarabine-induced cytopenia is similar to that of other cytotoxic agents, and thus similar precautions, with consideration given to the state of the patient, details of previous therapy, and the degree of bone marrow infiltration, are essential during monitoring. Other toxicity is unusual.
In summary, this experience of the use of Fludarabine in previously treated patients with CLL and low grade NHL confirms that it is a relatively safe and well tolerated drug with considerable promise. Further studies will determine whether this drug will have a significant impact on the clinical course of these diseases.