Department of Internal Medicine, West German Cancer Center, University of Essen Medical School, Hufelandstrasse 55, 45122 Essen, Germany

Correspondence to: O Kloke, Fax: +49 201 723 5924

The purpose of this study was to evaluate the long-term outcome of interferon (IFN) alfa treatment in patients with Philadelphia chromosome-positive chronic myeloid leukemia (CML). Between 1984 and 1990, a total of 71 patients with newly diagnosed CML had been enrolled into two consecutive IFN trials at our institution. Follow-up extended to December 1998, resulting in a median observation period for surviving patients of 11.4 years. The median survival time from diagnosis was 5.9 years. A plateau in the actuarial survival curve was found from 8.2 to 12.3 years following diagnosis with a projected 10-year survival rate of 32%. 'Landmark' studies showed a significant survival advantage for patients with karyotype responses. Of 68 patients accessible to calculation of the Hasford score, three were in the high risk group, 24 belonged to the medium risk group, and 41 had low risk features. The majority of cytogenetic responders including all eight assessable patients in complete cytogenetic remission were in the low risk group. Achieving a cytogenetic remission was found to provide a survival advantage also for patients with low risk disease. Of the seven patients surviving more than 11 years, six were in continuous complete cytogenetic remission. Their favorable outcome appears to translate into an out-flattening of the survival curve for the 71 single center patients presented. It will be of interest to see whether prolonged follow-ups of the large multicentric randomized trials will similarly show a subset of long-term surviving patients with ongoing IFN-induced remission. Leukemia (2000) 14, 389-392.

CML; cytogenetic remission; interferon alfa

Introduction

Different to conventional cytoreductive chemotherapy, interferon (IFN) alfa treatment is capable of inducing suppression of the neoplastic, Philadelphia chromosome (Ph)¾positive cell clone and enabling the restoration of normal hematopoiesis in a minority of patients with chronic myeloid leukemia (CML).¹ Achieving a cytogenetic remission with IFN alfa therapy has been shown to be independently associated with a survival advantage.² Moreover, a recent meta-analysis of randomized trials comparing IFN alfa with standard chemotherapeutic agents has demonstrated a significant survival benefit for treatment regimens that included IFN alfa.³ With the exception of the outcome of subgroups of patients exhibiting major cytogenetic remissions, however, survival analyses performed on unselected IFN-treated patient groups have all revealed, so far, a steady decline in the proportion of surviving patients. This also appears to hold true for the subpopulation of IFN-treated patients whose distinct disease features constitute a low risk profile.⁴

In a previous study, we have assessed the association of survival expectancy and cytogenetic response of patients with newly diagnosed CML treated at our department with IFN alfa. Results of this analysis were confirmatory to the above-mentioned studies as to the prognostic advantage associated with IFN-induced cytogenetic improvement.⁵ The purpose of the present study was to evaluate the long-term outcome of this patient population.

Methods

Patients and treatment

From 1984 to 1990, a total of 71 patients (40 males and 31 females, median age at diagnosis 38 years) with newly diagnosed, Ph-positive, chronic-phase CML received IFN alfa-2b as primary treatment within two consecutive trials. The initial daily dose of IFN alfa in both trials was 4 ´ 10⁶ U/m². In the first study which included 12 unpretreated patients, the IFN dosage was reduced in parallel to the decrease in WBC and doses minimally effective to keep the WBC counts within the normal range were administered as maintenance therapy. In the second trial, 59 patients were randomized to receive IFN alfa either alone or in combination with low-dose IFN gamma. In both arms of this study, the maximally tolerable dose of IFN alfa was used for maintenance. Detailed information on patient characteristics, baseline evaluations, follow-up schedules and early clinical results have been given in two previous reports.^6,7

Criteria for cytogenetic response

Patients were categorized by their best cytogenetic response which was graded as proposed by Talpaz et al:⁸ no response (NR) if more than 95% of bone marrow metaphases remained ph-positive; minimal response (MR) if the proportion of Ph-positive cells was reduced to 35-95%; partial response (PR) if Ph-positive metaphases were suppressed to levels of 5-34%; and complete response (CR) if no Ph-positive metaphases were detectable. Patients were considered not evaluable (NE) if no results of any cytogenetic follow-up were available.

Statistical analysis

Survival estimations were based on the product-limit technique.⁹ The log-rank test was used for comparing survival distributions.¹⁰ The key date was December 1998. Surviving patients were censored at the time of last contact. Patients who had allogeneic bone marrow transplantation (allo-BMT) or were lost to follow-up were handled as censored observations at the time of their referral to the transplant unit and of their last visit, respectively.

To circumvent any bias due to the time dependence of cytogenetic remission, the landmark method was used to assess the impact of karyotype improvement on survival. For this analysis, only patients who were alive and uncensored at the landmark time of 2 years from the start of IFN therapy were considered. They were categorized and compared according to their cytogenetic response achieved at that time point.

Results

Response and follow-up

Of the 71 patients treated, five were not evaluable for cytogenetic response. Discontinuing IFN after 1-23 months (median 10) of treatment, these five patients had not achieved a normalization of WBC and differential counts. They were grouped together with 20 patients who had no cytogenetic improvement. A total of nine patients have attained a CR; cytogenetic remission was partial in 12 patients and minimal in 25 patients (Table 1).

Of the 47 patients neither referred to allo-BMT nor lost to follow-up, 35 have died up to December 1998. The median follow-up from diagnosis of the 12 surviving patients was 11.4 years (range, 8.2-12.3 years). Seven of these 12 long-term survivors were patients in complete cytogenetic remission. The remaining five surviving patients had all experienced incomplete karyotype improvement which, however, had disappeared in the further course of treatment. In two patients, loss of Ph-positive cells was noted at only one time point during the observation period and could not be confirmed by follow-up analysis. Both patients later progressed to the blastic phase of disease and succumbed. In the remaining seven CR patients, the remission proved to be durable (Table 1).

Survival

As depicted in Figure 1, the median survival time of the 71 patients was 5.9 years. There was a steady decline in the proportion of surviving patients until 8.2 years from diagnosis was reached. From 8.2 to 12.3 years, we found a plateau in the actuarial survival curve with an estimated probability of 32% to remain alive 10 years after diagnosis. Of note, six of the seven patients at risk after 10 years from diagnosis were in continuous CR.

Prognostic impact of cytogenetic response

To confirm the survival advantage by achieving a cytogenetic remission, a landmark analysis was carried out on the 54 patients who were alive and uncensored at 2 years from the start of treatment. Patients who had attained a complete or partial cytogenetic response at this time point lived longer (P = 0.026) than patients with minimal response. The projected 10-year survival rates of these two patient groups were 75% and 34%, respectively (Figure 2). Patients exhibiting a minimal remission after 2 years from the start of IFN therapy had a survival advantage (P = 0.008) over patients without cytogenetic improvement.

Impact of disease-related prognostic factors

From 68 (96%) of the 71 patients studied, all data required for the calculation of the Hasford score⁴ were available. Forty-one (58%) patients had a low risk profile, 24 (34%) patients were assigned to the medium risk group, and three (4%) patients had features of high risk disease. The latter three patients died at 2.7, 3.6 and 6.9 years from diagnosis. Survival estimates of the low and medium risk patients are given in Figure 3. Comparison of the two cohorts revealed a significantly better survival outlook (P = 0.013) for low risk patients.

As shown in Table 2, the majority of cytogenetic responders including all eight CR patients accessible for calculation of the Hasford score had low risk disease. Vice versa, only nine out of the 41 low risk patients failed to achieve cytogenetic improvement.

To address the issue of whether the favorable prognosis of patients belonging to the low risk group merely reflects an 'intrinsically more favorable' disease,¹ a separate landmark analysis was performed on low risk patients. In this patient subgroup, there was no significant survival benefit (P = 0.2) from attaining a complete or partial remission as compared to a minimal response. However, low risk patients achieving even a minimal remission within the first 2 years of treatment had a survival advantage (P = 0.002) over patients with low risk disease yet lacking cytogenetic improvement (Figure 4).

Discussion

This long-term follow-up of IFN-treated CML patients provides further support for the notion that cytogenetic improvement on IFN therapy translates into improved survival expectancy. In addition, results from this study confirm the power of the newly developed Hasford score⁴ to identify patients with a particularly favorable prognosis on IFN treatment. Obviously, our attempts to further discern the relative prognostic impact of cytogenetic improvement and of favorable disease-related features were hampered by the small number of patients with medium and high risk disease in our study population. However, the accumulation of cytogenetic responders in the low risk group in conjunction with the dismal outcome of low risk patients lacking cytogenetic improvement appears to suggest that at least part of the merit of the prognostic score is to recognize patients likely to achieve a favorable cytogenetic response to IFN.

The sample of 71 single-institution patients reported here differs from the large patient population evaluated by Hasford et al⁴ as to the distribution of low risk and high risk disease. Accordingly, the median survival duration and the proportion of long-term survivors obtained from the present study reflect the outcome of a patient cohort with a comparatively favorable risk profile. It should be noted, however, that similar figures for both median and 10-year survival have been reported in the recent long-term follow-up of the Italian multicentric randomized trial that included 218 IFN-treated patients.¹¹

In the Italian study, the median observation period of living patients was 9.3 years from randomization. Based upon a median follow-up of 11.4 years from diagnosis, the present evaluation shows a plateau in the actuarial survival curve from 8.2 to 12.3 years. Clues as to the significance of this finding may be provided by considering the present status of the long-term surviving patients. Four of the five patients without on-going cytogenetic remission were off IFN at the time of follow-up. Given the dubious prognosis of patients discontinuing IFN,¹² some further decline in the proportion of surviving patients may be expected to occur with longer follow-up. The majority of long-term surviving patients, however, have achieved and maintained complete cytogenetic remission. We can see little reason at present to doubt the ongoing favorable status of this patient group which included six of the seven patients surviving more than 11 years.

Taken together, these data suggest that the durability of complete cytogenetic remissions as present in the majority of long-term survivors is translating into an out-flattening of the survival curve for the patient population reported here. With regard to its relatively small size, it remains to be seen whether future follow-ups of the large randomized trials launched in the late 1980s will similarly show a subset of long-term surviving patients with ongoing IFN-induced remission.

Acknowledgements

We are grateful to R Prittwitz and I Hawig for assistance in the collection and management of clinical data.

1 Kantarjian HM, O'Brien S, Anderlini P, Talpaz M. Treatment of chronic myelogenous leukemia: current status and investigational options. Blood 1996; 87: 3069-3081, MEDLINE

2 Kantarjian HM, Smith TL, O'Brien S, Beran M, Pierce S, Talpaz M. Prolonged survival in chronic myelogenous leukemia after cytogenetic response to interferon- therapy. Ann Intern Med 1995; 122: 254-261, MEDLINE

3 Chronic Myeloid Leukemia Trialists' Collaborative Group. Interferon alfa versus chemotherapy for chronic myeloid leukemia: a meta-analysis of seven randomized trials. J Natl Cancer Inst 1997; 89: 1616-1620, MEDLINE

4 Hasford J, Pfirrmann M, Hehlmann R, Allen NC, Baccarani M, Kluin-Nelemanns JC, Alimena G, Steegmann JL, Ansari H. A new prognostic score for patients with chronic myeloid leukemia treated with interferon alfa. J Natl Cancer Inst 1998; 90: 850-858, MEDLINE

5 Kloke O, Niederle N, Qiu JY, Wandl U, Moritz T, Nagel-Hiemke M, Opalka B, Seeber S, Becher R. Impact of interferon alpha-induced cytogenetic improvement on survival in chronic myelogenous leukaemia. Br J Haematol 1993; 83: 399-403, MEDLINE

6 Kloke O, Wandl U, Opalka B, Moritz T, Nagel-Hiemke M, Franz T, Hirche H, Seeber S, Niederle N. A prospective randomized comparison of single-agent interferon (IFN)-alpha with the combination of IFN-alpha and low-dose IFN-gamma in chronic myelogenous leukaemia. Eur J Haematol 1992; 48: 93-98, MEDLINE

7 Niederle N, Kloke O, Wandl UB, Becher R, Moritz T, Opalka B. Long-term treatment chronic myelogenous leukemia with different interferons: results from three studies. Leuk Lymphoma 1993; 9: 111-119, MEDLINE

8 Talpaz M, Kantarjian HM, McCredie KB, Trujillo K, Keating MJ, Gutterman JU. Clinical investigation of human alpha interferon in chronic myelogenous leukemia. Blood 1987; 72: 1280-1288,

9 Kaplan EM, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457-481,

10 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719-748,

11 The Italian Cooperative Study Group on Chronic Myeloid Leukemia. Long-term follow-up of the Italian trial of interferon- versus conventional chemotherapy in chronic myeloid leukemia. Blood 1998; 92: 1541-1548, MEDLINE

12 Hehlmann R, Heimpel H, Hasford J, Kolb HJ, Pralle H, Hossfeld DK, Queisser W, Löffler H, Hochhaus A, Heinze B, Georgii A, Bartram CR, Griesshammer M, Bergmann L, Essers U, Falga C, Queisser U, Meyer P, Schmitz N, Eimermacher H, Walther F, Fett W, Kleeberg UR, Käbisch A, Nerl C, Zimmermann R, Meuret G, Tichelli A, Kanz L, Tigges FJ, Schmid L, Brockhaus W, Tobler A, Reiter A, Perker M, Emmerich B, Verpoort K, Zankovich R, von Wussow P, Prümmer O, Thiele J, Buhr T, Carbonell F, Ansari H, and the German CML study group. Randomized comparison of interferon- with busulfan and hydroxyurea in chronic myelogenous leukemia. Blood 1994; 84: 4064-4077, MEDLINE

Figure 1  Kaplan-Meier plot of the probability of survival. The number of patients at risk is given below the time points. Asterisks denote patients in continuous CR.

Figure 2  Kaplan-Meier plots of the probabilities of survival from start of IFN treatment for patients still alive and uncensored at 2 years, according to cytogenetic response achieved within the first 2 years of treatment.

Figure 3  Kaplan-Meier plots of the probabilities of survival from diagnosis of low risk and medium risk patients as defined by the Hasford score.

Figure 4  Kaplan-Meier plots of the probabilities of survival from start of treatment for low risk patients still alive and uncensored at 2 years, according to cytogenetic response achieved within the first 2 years of treatment.

Table 1  Follow-up according to best cytogenetic response

Table 2  Best cytogenetic response according to risk status of the patients accessible to calculation on the Hasford score