Chronic myelogenous leukemia--progress at the M. D. Anderson Cancer Center over the past two decades and future directions: first Emil J Freireich Award Lecture.

The purpose of this study was to review the progress in clinical and translational research in chronic myelogenous leukemia (CML) over the past 20 years at M.D. Anderson Cancer Center. The CML database updating the clinical and basic research investigations was reviewed as the source of this report. Publications resulting from these investigations were summarized. The long-term results with intensive chemotherapy, IFN-alpha therapy alone or in combination, autologous stem cell transplantation, and new agents such as homoharringtonine and decitabine showed encouraging results. Biological studies related to the BCR-ABL molecular abnormality, other molecular events, and the detection of minimal residual disease were detailed. Future strategies with potential promise in CML were outlined. Significant progress in understanding CML biology and in treating patients afflicted with the disease has occurred. Several therapeutic and research tools are currently investigated, which should hopefully improve further the prognosis of patients with CML.

Acute lymphoid leukemia molecular phenotype in a patient with benign-phase chronic myelogenous leukemia.

The benign phase of chronic myelogenous leukemia (CML) typically is characterized by an overproduction of myeloid cells that eventually progresses to a more acute stage termed blast crisis. This latter stage can exhibit either myeloid or lymphoid blast clones. Our recent results have demonstrated the presence of the P210 BCR-ABL protein in blood cells from benign phase CML patients (Guo et al., Cancer Research 51:3048, 1991). This protein is the product of an 8.5 kb chimeric RNA encoded by fused BCR-ABL genes produced by the formation of the Philadelphia (Ph) chromosome. Using this new assay we have identified a patient with benign-phase CML who produces P190 BCR-ABL, the form of the BCR-ABL protein found in about 50% of cases of acute lymphocytic leukemia (ALL). This patient lacked detectable P210 BCR-ABL protein and did not contain a DNA rearrangement in the major breakpoint cluster region of the BCR gene. Consistent with this result, polymerase chain reaction (PCR) analyses detected a BCR-ABL mRNA with BCR exon 1 fused to ABL exon 2. No BCR-ABL mRNAs with 2'- or 3'-bcr exon to ABL exon 2 fusions were detected in these analyses. Blood cells from this patient lost P190 BCR-ABL after the patient underwent an allogeneic bone marrow transplant, but regained this protein although the patient was still in chronic phase after a subsequent autologous transplant as treatment for graft failure. These findings indicate that P190 BCR-ABL alone is not sufficient to induce a blast crisis phenotype in leukemia patients who are Ph chromosome-positive.

Cytogenetics for detection of minimal residual disease in acute myeloblastic leukemia.

Bone marrow samples collected from acute myeloblastic leukemia (AML) patients in complete clinical and hematological remission were studied for the persistence of cytogenetic abnormalities. AML patients from the three favorable cytogenetic categories [inv 16, t(8;21) and t(15;17)] and patients from the unfavorable cytogenetic categories (+8, -5, -7 and Philadelphia-positive) were studied. Seventy-one patients had evaluable metaphase spreads in remission marrows and 20 (28%) had one or more abnormal metaphases identical to that present in the pretreatment marrow. All 20 of these patients relapsed within 78 weeks, thus there were no false positive studies. Fifty-one patients had only diploid metaphases in their complete remission marrow, 25 relapsed, and 21 remained in continuous complete remission. Thus there was a 49% false negative rate of this study. These data indicate that the failure to detect residual chromosomally abnormal cells in the bone marrow does not guarantee continuous complete remission. Cytogenetic study was most useful in the favorable cytogenetic groups and least useful in the unfavorable groups. The persistence of normal metaphases in pretreatment marrows did not affect outcome or risk of recurrence. Twenty-five of 34 evaluable patients who relapsed after remission had either the identical cytogenetic abnormality present in the pretreatment marrow or showed the identical abnormality with additional chromosomal changes. Thus study indicates that cytogenetic examinations of complete remission bone marrow samples in patients with AML provides an objective method for detecting residual leukemia, and identifies patients with a potential for prolonged disease-free survival.

Treatment of hairy cell leukemia with 2-chlorodeoxyadenosine (2-CdA).

We administered one course of 2-chlorodeoxyadenosine (2CdA) at 4 mg/m2 daily for 7 days by continuous intravenous infusion to 46 patients with hairy cell leukemia. Complete remissions occurred in 36 patients (78%; 95% confidence limits, 63% to 89%), partial remissions in five (11%), and a minor response in one. One patient died of candida sepsis 3 weeks after beginning treatment and three patients were clearly resistant to therapy. These three either had morphologically atypical hairy cells, less than 20% of which expressed Ig light chain on the cell surface, or had failed prior treatment with deoxycoformycin and interferon-alpha. At a median of 37 weeks since discontinuation of therapy, recurrent thrombocytopenia has developed in one patient, whose marrow remains normal, while a bone marrow relapse has occurred in another patient, whose blood counts remain normal. Treatment produced a greater than 50% decrease in neutrophil count in 26 patients, which lasted 3 to 4 weeks and was associated with an increased incidence of febrile episodes. These episodes occurred in 21 patients but were associated with documented infection in only four patients. Decreases in the number of CD4+ lymphocytes appeared to occur regularly after treatment and have persisted for a median of 18 weeks without obvious clinical significance. Although years of follow-up will be needed, our results confirm Piro et al's observation (N Engl J Med 322: 1117, 1990) that 2CdA appears to be highly effective in the treatment of hairy cell leukemia.

Intensive chemotherapy induction followed by interferon-alpha maintenance in patients with Philadelphia chromosome-positive chronic myelogenous leukemia.

In a pilot study, 32 patients with Philadelphia chromosome-positive chronic myelogenous leukemia were treated with intensive chemotherapy induction followed by interferon-alpha (IFN-A) maintenance. Intensive chemotherapy consisted of three cycles of daunorubicin 120 mg/m2 on day 1, cytarabine 80 mg/m2 daily for 10 days, vincristine 2 mg on day 1, and prednisone 100 mg daily for 5 days (DOAP). Maintenance therapy with IFN-A at a doses of 3 x 10(6) to 5 x 10(6) units/m2 daily was adjusted according to counts and toxicity. The outcome of patients was compared with a matched historic population of 64 patients treated with IFN-A alone. Overall, 60% of patients had a cytogenetic response (partial or complete) with induction chemotherapy, but only eight (25%) had a sustained cytogenetic response with IFN-A maintenance. After a median follow-up of 67 months, the 6-year survival rate of the 32 patients was 58%, compared with 36% for the matched historic group (P = 0.084). The incidence of lymphoid blastic transformation in the two groups was 25% and 48%, respectively (P = 0.10) and durable cytogenetic responses, 25% and 19%, respectively (P = 0.48). In summary, the addition of intensive chemotherapy induction to IFN-A maintenance does not improve the survival rate, incidence of lymphoid blastic transformation, or incidence of durable cytogenetic response compared with the results achieved with IFN-A therapy alone.

Mitoxantrone and high-dose etoposide for patients with relapsed or refractory acute leukemia.

Among 35 patients with relapsed or refractory acute myelogenous leukemia (AML) who received salvage chemotherapy, 28 were treated with mitoxantrone (7.5 mg/m2/d intravenously [IV] over 1 hour for 5 days) and etoposide (VP-16) (2 g/m2 over 4 days either as a daily infusion or as two daily doses). Seven patients received mitoxantrone (6 mg/m2/d for 5 days) and VP-16 (1500 mg/m2 over 3 days). The median duration of the initial complete remission (CR) was 6 months and 83% of the patients had initial CR that lasted 12 months or less. Forty-six percent of the patients were undergoing a second or subsequent salvage attempt. Eight patients (23%) achieved CR; seven of these CR were obtained after one course of therapy. Twelve patients (33%) died and 15 patients (42%) had disease that was resistant to treatment. Patients undergoing a first salvage attempt had a higher incidence rate of CR than those undergoing a second or subsequent salvage attempt (37% versus 6%; P = 0.03). CR rates were also higher in patients with a favorable (translocation 8;21 or 15;17) or diploid karyotype compared with other patients (32% versus 8%; P = 0.10). The median survival time was 2 months for all patients and 8 months for patients achieving CR. Mucositis occurred in 74% of the patients and was severe in 32%. Diarrhea and rash occurred in less than 33% of the patients. Fever was noticed in all but 1 of the patients and documented infections occurred in 65% of the patients. Six patients had pancytopenia or thrombocytopenia that lasted more than 42 days from the initiation of treatment. Although mitoxantrone and high-dose VP-16 is an effective antileukemic regimen, it is associated with a high incidence of mucositis. Strategies that are used to limit mucosal damage may improve the tolerance of this combination.

Generation of cytotoxic NK cells in peripheral blood and bone marrow of patients with acute myelogenous leukemia after continuous infusion with recombinant interleukin-2.

We have studied the cytotoxic profile and distribution of lymphocyte subsets of patients with acute myelogenous leukemia in second remission, after continuous infusion with recombinant interleukin-2 (IL-2). The patients received repetitive cycles of 1-1.25 x 10(6) U/m2/day of IL-2, given as 4 days continuous intravenous infusion followed by a 3-day treatment-free interval for the first 4 weeks. Patients receiving greater than 4 cycles were treated with the same dose of IL-2 continuously for 4 days, followed by a 10-day treatment-free interval. These studies showed that IL-2 treatment resulted in the generation of peripheral blood cytotoxic activity against both NK-susceptible, K-562, and NK-resistant Daudi cell lines. In most patients, enhancement of lytic activity increased with the number of IL-2 infusions. The cytotoxicity in some patients increased as much as 700-fold and 830-fold against K-562 and Daudi cells, respectively. It is of importance that oncolytic activity was also induced in bone marrow compartment (up to 182-fold against K-562). Some decline in cytotoxicity was observed within 14 days after initiation of IL-2 infusion in peripheral blood, but high levels of lytic activity persisted at this time in bone marrow. It is of interest to note that the cytotoxicity of in vivo IL-2 primed lymphocytes was further potentiated by IL-2 in vitro. Importantly, the cytotoxic cells induced in vitro displayed lytic activity against fresh leukemic blasts. Phenotypic analysis demonstrated that CD3-, CD56+ NK cells were significantly increased by in vivo IL-2 treatment (34 to 47-fold in absolute numbers), while CD3-, CD56+ T-cell subset remained low. Characterization of cytotoxic cells using the complement-dependent assay and monoclonal antibodies indicated that both the in vivo-induced and ex vivo-potentiated lytic function was mediated by CD3-, CD56+, CD16- -NK cells.

Effects of low doses of recombinant human granulocyte-macrophage colony stimulating factor (GM-CSF) in patients with myelodysplastic syndromes.

There has been no previously published experience with granulocyte-macrophage colony stimulating factor (GM-CSF) doses less than 12 micrograms/m2 daily in patients with myelodysplastic syndromes, and most observations have been made at doses greater than or equal to 120 micrograms/m2 daily. We administered 5 micrograms/m2 daily by subcutaneous injection to 29 such patients increasing the dose in patients who did not show a haematologic response. Doses of 5 or 10 micrograms/m2 ('low-dose GM-CSF') produced an increase in neutrophils in 14/29 patients. Response was significantly (P = 0.03) more frequent in patients who had a higher pre-treatment neutrophil count (e.g. 11/16 in patients with greater than or equal to 0.5 x 10(9)/l). A rise in blasts followed administration of low-dose GM-CSF in five patients, all with either refractory anaemia with excess blasts (RAEB) or refractory anaemia with excess blasts in transformation (RAEBT). Platelets decreased in five patients, four of whom had no change in blasts, reverting to baseline when GM-CSF was discontinued. We and others have previously observed similar rises in blasts or decreases in platelets at doses of 120 micrograms/m2 daily. Low-dose GM-CSF produced no constitutional side effects. Our results suggest that low doses of GM-CSF might be initially employed in neutropenic patients with myelodysplastic syndromes who present with pretreatment neutrophil counts greater than 0.5 x 10(9)/l. Increasing the dose, and hence the risk of extramedullary toxicity, only in patients who do not respond to the low dose. Patients who present with lower pre-treatment neutrophil counts might begin treatment at doses above 10 micrograms/m2, but below the 120 micrograms/m2 commonly employed, which may be necessary in relatively few patients.

Fludarabine: a new agent with marked cytoreductive activity in untreated chronic lymphocytic leukemia.

Thirty-three patients with chronic lymphocytic leukemia (CLL) with advanced Rai stage (III-IV) or progressive Rai stage (0-II) disease were treated with fludarabine as a single agent. Eleven patients (33%) obtained a complete remission (CR), 13 (39%) a clinical CR with residual nodules as the only evidence of disease (nodular partial remission [PR]), and two patients (6%) achieved a PR for a total response rate of 79%. Response was rapid, usually occurring after three to six courses of treatment. The major morbidity was infection. Febrile episodes occurred in 13% of the courses (pneumonia 6%, minor infection 4%, and transient fever of undocumented cause 3%). Fludarabine appears to be the most cytoreductive single agent so far studied in CLL.

Effect of antileukemia chemotherapy on marrow, blood, and oral granulocyte counts.

This study was designed to elicit the effects of antileukemia chemotherapy on marrow production, blood carriage, and oral extravasation of granulocytes, and on the phagocytic activity of those harvested from the mouth. Fifteen adult patients with various morphologic forms of acute leukemia were followed through one to four courses of chemotherapy. Oral saline rinse samples were obtained thrice weekly and prepared for enumeration in a hemocytometer. The oral granulocyte counts were compared with concurrent counts in the bone marrow and peripheral blood. Phagocytic activity of the oral granulocytes was measured by the method of Smith and Rommel. The 15 patients were followed through 30 courses of chemotherapy and recovery. During each, there was a drug-induced decrease in marrow, blood, and oral granulocytes that was reversed when therapy was discontinued and bone marrow activity was restored. Phagocytic activity of the oral granulocytes was not perceptibly affected by the antileukemic drugs. Oral granulocyte counts provide a noninvasive method for monitoring the onset and recovery of chemotherapy-induced myelosuppression and granulocytopenia in patients with leukemia.

Highly oncolytic adherent lymphocytes: therapeutic relevance for leukemia.

We have generated and characterized a highly oncolytic adherent lymphocyte subset (A-LAK) from eight leukemic patients with non-lymphocytic leukemia (NLL) in remission and one NLL patient in relapse. Our studies demonstrated that A-LAK was superior in its oncolytic activity (tested in a 3-h 51Cr release assay) to conventionally prepared (LAK) and non-adherent (NA) IL-2 cultures. No activity was observed by this highly oncolytic subset against normal bone marrow (BM). A-LAK also displayed highest proliferative activity in 7-11 day cultures (5- to 58-fold expansion) in comparison to LAK (0.7- to 2.7-fold) or NA (1.0- to 2.6-fold) cultures. Analysis of phenotype of unseparated, NA and adherent (A-LAK) lymphocytes 24 h after IL-2 activation showed that the A-LAK was composed predominantly of high intensity (bright) CD11a+ (LFA-1) lymphocytes (75 +/- 4.8%) when compared to the other two populations (12 +/- 2.1%). Similarly, A-LAK contained higher proportion of CD11b (CR3 receptor)-positive lymphocytes (39 +/- 2.1%) than unseparated and NA lymphocytes (11 +/- 1.4%). Double marker phenotypic studies showed that A-LAK cultures were heterogeneous and distribution of individual lymphocyte subsets differed among NLL patients. While in A-LAK culture of some patients the CD56+, CD3- natural killer (NK) cell subset was predominant, CD3+, CD56- lymphocyte subset was prevalent in others. Highest A-LAK lytic activity was always correlated with highest NK cell content. Characterization studies (using the complement-depletion technique) showed that independently of the distribution of lymphocytes in A-LAK cultures, CD16+, CD56+, CD3- NK cell subset displayed highest oncolytic effect. CD5+ subset also participated in cytotoxic function. These observations indicated that A-LAK may represent a new therapeutic approach to treatment of leukemia.

The importance of cytogenetic studies in adult acute lymphocytic leukemia.

PURPOSE: The prognostic importance of pretreatment bone marrow cytogenetic studies in adults with acute lymphocytic leukemia treated at a single institution, with an identical treatment program, is described. PATIENTS AND METHODS: A total of 105 patients with a documented morphologic diagnosis of acute lymphocytic leukemia were reviewed for the purpose of this analysis. All patients had an extensive workup at presentation, and cytogenetic analysis was performed in 103 patients, using the Giemsa banding technique with trypsin pretreatment on 24-hour cultured cells. RESULTS: The specific cytogenetic classification in the 103 patients who had the karyotypic analysis was as follows: diploid 27%; Philadelphia chromosome-positive 13%; hyperdiploid 12%; B-cell karyotype 6%; 6q- and 14q+ abnormalities 4%; pseudodiploid 8%; hypodiploid 2%; and insufficient metaphases 28%. B-cell, 6q- or 14q+, and Philadelphia chromosome-positive karyotypes tended to correlate with other known negative prognostic factors. Patients with diploid, hyperdiploid, pseudodiploid, and hypodiploid karyotypes or with insufficient metaphases could be combined into one group with a favorable prognosis. In this group, the remission rate with induction chemotherapy was 89%, the median complete remission duration was 26 months, and the median survival was 25 months, with a 3-year survival rate of 45%. Patients with Philadelphia chromosome-positive, B-cell, and 6q- or 14q+ abnormalities collectively had an unfavorable prognosis. Their response rate to induction chemotherapy was 65%, the median response duration was 7 months, and the median survival was 8 months, with a 3-year survival rate of less than 10%. CONCLUSION: We conclude that the pretreatment bone marrow karyotype is an important part of the evaluation of adults with acute lymphocytic leukemia and provides significant prognostic information.

High-dose chemotherapy and unpurged autologous bone marrow transplantation for acute leukemia in second or subsequent remission.

The authors administered high-dose chemotherapy with cyclophosphamide, BCNU (carmustine) and VP-16 (etoposide) plus autologous bone marrow transplantation (ABMT) to 22 adult patients with relapsed acute leukemia in second or subsequent remission. The marrow was not treated ex vivo. The long-term, disease-free survival rate was 14%. Comparison of results with other treatments can be difficult because of patient selection biases. The concept of inversion (achievement of a longer remission with salvage therapy than with prior treatments) is proposed to compare treatment results. Three patients remain in complete remission beyond 4 years, with inversions. More intensive cytoreductive regimens will be needed to improve results.

Suitability of a new stable acetal analogue of aldoifosphamide for purging leukemic cells from human bone marrow.

The in vitro cytotoxic properties of acetaldoifosphamide, a new chemically stable bis-acetate analogue of aldoifosphamide that requires enzymatic activation by cellular carboxylate esterases, has been compared with that of 4-hydroperoxycyclophosphamide (4-HC). On a molar basis, acetaldoifosphamide was 8-10 times more potent than 4-HC against two different human leukemic myeloid cell lines, but only twice as potent as 4-HC against normal bone marrow granulocyte-macrophage colony-forming cells (GM-CFC). Acetaldoifosphamide retained its activity against leukemic cell lines that were highly resistant to the antileukemic drugs doxorubicin and m-AMSA. GM-CFC doubling times after exposure of bone marrow to high concentrations of acetaldoifosphamide in suspension cultures were 6-12 hours. Similar doubling times were obtained after incubation of marrow with 4-HC. Acetaldoifosphamide has a sparing effect on hematopoietic stem cells that is similar to that found for 4-HC; however, it is considerably more potent than 4-HC. Acetaldoifosphamide is different from 4-HC in its chemical stability and its unique requirement for carboxylate esterase activation. We conclude that acetaldoifosphamide may have advantages over 4-HC for in vitro purging of leukemic cells from human bone marrow.

Treatment of poor-prognosis, newly diagnosed acute myeloid leukemia with ara-C and recombinant human granulocyte-macrophage colony-stimulating factor.

We administered recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF) (120 micrograms/m2/d by continuous intravenous [IV] infusion) to 12 patients with newly diagnosed acute myeloid leukemia (AML) at relatively high risk of early death during remission induction. GM-CSF began 3 days after completion of induction chemotherapy (ara-C 1.5 g/m2 d x 4 days by continuous IV infusion after a 3 g/m2 bolus). Rates of fatal infection (42%), pneumonia and/or sepsis (83%), and CR (50%) did not differ significantly (P less than .05) from those observed after administration of the identical chemotherapy without GM-CSF to 53 historical controls with newly diagnosed AML at similarly high risk of early death. There were no significant differences between the GM-CSF-treated and the historical groups in the time required to reach neutrophil counts of 500 or 1,000/microL after administration of chemotherapy. Four patients died of infection before they could have benefited from the earliest recovery of neutrophil count observed in patients who entered CR. Growth of leukemia after GM-CSF administration was observed in only 1 of the 8 patients who survived long enough for response to induction therapy to be fully evaluated. This observation suggests that it might be safe to undertake larger, randomized studies, perhaps using earlier administration of GM-CSF, to definitively determine the role of GM-CSF added to chemotherapy in patients with newly diagnosed AML.

Fatal pulmonary failure complicating high-dose cytosine arabinoside therapy in acute leukemia.

One hundred three relapsed leukemia patients were treated with high-dose cytosine arabinoside (Ara-C); 3 g/m2 intravenously over 2 hours every 6 to 12 hours for a total of nine to 12 doses or 3 g/m2 intravenously over 2 hours for two doses 12 hours apart followed by a continuous infusion of 1.5 g/m2 over 24 hours daily for 3 to 4 days. Thirteen of them developed adult respiratory distress syndrome (ARDS) without having any recognized reason for the development of pulmonary edema. This problem showed no correlation with age or prior chemotherapy. Four of the patients recovered, but in nine this complication was fatal. The authors have reviewed the clinical course of these 13 patients and the postmortem findings of the seven patients who had an autopsy performed. The pulmonary tissue from six patients showed massive edema and one had diffuse alveolar damage. Histologic examination revealed a highly proteinaceous intraalveolar infiltrate without any inflammatory reaction in all cases. Intestinal tissue from all patients revealed changes compatible with cytotoxic damage, and pleura and/or pericardium from six of the seven patients showed an extensive fibrinous exudate suggestive of capillary leakage. The time sequence of the clinical events and the histologic findings indicate that high-dose Ara-C treatment in leukemia may cause a capillary leakage syndrome with ARDS that may progress to fatal respiratory failure.

Cellular ara-CTP pharmacokinetics, response, and karyotype in newly diagnosed acute myelogenous leukemia.

We evaluated relationships between ara-CTP pharmacokinetics in myeloblasts, response, and karyotype in 147 patients with newly diagnosed AML treated on three protocols each initiated by a 3 g/m2 ara-C dose given over 2 h. Area under the curve of ara-CTP concentration times time (AUC) or peak ara-CTP concentration after this dose did not predict response or response duration, suggesting that inability to form ara-CTP is an unlikely mechanism of ara-C resistance. Following high-dose bolus ara-C therapy patients with INV [16] or del [16q] had long remissions despite low AUC and peak values while patients with -5, del [5q], -7, or del [7q] were frequently resistant despite average AUC and peak values. In 55 patients treated with high-dose continuous infusion ara-C as a single agent, steady-state ara-CTP concentrations were significantly lower in resistant patients (who again were disproportionately those with -5, del [5q], -7, or del [7q]) although there was no correlation with CR duration.

Mitoxantrone and high-dose cytosine arabinoside for the treatment of refractory acute lymphocytic leukemia.

Twenty-five adult patients with refractory acute lymphocytic leukemia received salvage therapy with mitoxantrone 5 mg/m2 intravenously over 1 hour daily for 5 days and cytosine arabinoside 3 g/m2 intravenously over 2 hours every 12 hours for six doses. Overall, nine patients (36%) achieved complete remission, eight (32%) died during induction, and eight (32%) had resistant disease. No significant associations were found between pretreatment patient characteristics and remission. Remission durations toxic effects were related to myelosuppression. Febrile episodes requiring hospitalization occurred in 23 patients (92%), including five episodes of fever of unknown origin (20%) and 18 episodes of documented infections (72%). The authors conclude that the combination of mitoxantrone and high-dose cytosine arabinoside has significant activity in adults with refractory acute lymphocytic leukemia. The addition of colony-stimulating growth factors to the intensive chemotherapy, and the use of the combination regimen as part of front-line maintenance intensification therapy may further improve the prognosis in these patients.

Proposal for a simple synthesis prognostic staging system in chronic myelogenous leukemia.

PURPOSE: Several prognostic models or staging systems have been published that identify different risk groups in patients with Philadelphia chromosome (Ph)-positive chronic myelogenous leukemia (CML). The aims of this study were (1) to test, in an independent population, the prognostic reproducibility of these staging systems; and (2) to develop a synthesis staging system that could be easily applied in clinical practice. PATIENTS AND METHODS: A total of 406 patients with newly diagnosed Ph-positive CML were evaluated by the four published staging systems of Tura, Cervantes, Sokal, and our group. The proposed synthesis staging system was developed based on the most consistent prognostic characteristics, and the presence or absence of accelerated disease features at diagnosis. The staging systems were compared according to survival outcome, as well as by looking for differences of survival outcomes within a specific stage of a defined system, when patients in this stage were subclassified by a second staging system. RESULTS: Whereas the staging system of Cervantes et al identified only two prognostic groups (median survivals of 49 versus 40 months; p = 0.01), the remaining three staging systems were able to segregate patients into stage 1, 2, and 3 risk groups with respective median survivals of 56 to 57, 41 to 42, and 28 to 36 months, respectively (p less than 0.001 to p less than 0.002). The new proposed staging system, based on the existence of zero to one (stage 1), two (stage 2), or three or more unfavorable (stage 3) characteristics, or the presence of accelerated disease features (stage 4), categorized patients into four prognostic groups with median survivals of 56, 45, 30, and 30 months, respectively (p less than 0.001), the latter stage (stage 4) being associated with a higher one-year mortality rate (29%). The synthesis staging system was also able to subclassify patients within most of Tura's and Sokal's stages into significantly different prognostic groups by survival outcome. CONCLUSION: The predictive prognostic capacity of three of the four published staging systems was confirmed in this independent or test population. The new proposed staging system was superior to the staging systems of Tura and Sokal in identifying different prognostic subgroups.