Chronic lymphocytic leukemia (CLL) is the most common adult leukemia and is currently incurable outside of stem cell transplantation. Chromosomal abnormalities are common in CLL, with up to 80% of patients possessing abnormalities detected by fluorescence in situ hybridization (FISH) analysis.1 The acquisition of cytogenetic abnormalities is also common, and the presence of complex karyotype (3 cytogenetic abnormalities) is associated with an adverse prognosis,2, 3, 4, 5 as are specific abnormalities. Deletion of 17p13.1, which results in loss of TP53, and 11q22.3, which results in loss of ATM, are found in 10% and 15–20% of patients, respectively. These are both associated with aggressive disease and shortened survival.1, 6, 7, 8, 9, 10 Thus, new therapies that are effective in patients with high-risk abnormalities represent a priority.
The cyclin-dependent kinase inhibitor flavopiridol has been shown to be effective in patients with relapsed and refractory CLL, including those refractory to fludarabine. In the initial phase I studies using an active dosing schedule of flavopiridol, approximately 40% of patients responded, including high response rates in patients with genetically high-risk disease.11 In a subsequent phase II trial, >50% of heavily pretreated patients responded.12 These two trials, OSU 0055 and OSU 0491, now have a longer follow-up (median 2.8 years), and herein we examine the contribution of interphase cytogenetic abnormalities and complex karyotype to response, progression-free survival (PFS) and overall survival (OS).
FISH was performed to detect ATM located at 11q22.3 and TP53 located at 17p13.1, as well as the chromosome 12 centromere and D13S319 located at 13q14.3, as previously reported.6 Dohner's hierarchical classification was used to form three groups of patients: those with del(17p13.1), those with del(11q22.3) and those with other cytogenetic abnormalities or normal cytogenetics. Metaphase cytogenetics was performed to determine complexity using pokeweed mitogen (Sigma-Aldrich, St Louis, MO, USA) or CpG stimulation; a complex karyotype was defined as 3 unrelated abnormalities. Estimates of PFS and OS were obtained by the Kaplan-Meier method, and the log-rank test was used to compare differences among survival curves. The proportional hazards model was used to model both PFS and OS as a function of cytogenetic group and presence of complex cytogenetics, controlling for age, Rai stage, bulky lymphadenopathy, number of prior treatments and treatment/dose schedule. The logistic regression model was used to model overall response rates (ORR) as a function of the variables listed above. Statistical significance was set at α=0.05.
A total of 112 patients are included in this analysis. Using Dohner's hierarchical classification, 40 patients (36%) had del(17p13.1), 37 (33%) had del(11q22.3) and 35 (31%) had neither of these abnormalities. These groups were not significantly different in terms of age, sex, Rai stage or percent fludarabine-refractory. The median number of prior treatments was 4 (range 1–11) and did not differ among cytogenetic groups (P=0.21). However, bulky adenopathy (5 cm) was more common in patients with del(11q22.3) (89%), and complex cytogenetics were more likely in patients with del(17p13.1) (63%) versus del(11q22.3) (32%) or neither abnormality (26%) (P=0.003).
The ORR was 46%, and was not significantly different among the cytogenetic groups (P=0.17). ORRs for patients classified with del(17p13.1), del(11q22.3) or without these abnormalities were 48%, 57% and 34%, respectively. Significant differences in ORR were not observed between those with and without complex karyotype (39 versus 52%, P=0.25).
PFS was not significantly different among the cytogenetic groups (P=0.52), with estimated PFS of 10.4 months (95% confidence interval (CI): 8.0–13.0) for those classified as del(17p13.1), 10.1 months (95% CI: 6.6–11.4) for those classified as del(11q22.3) and 8.1 months (95% CI: 4.2–18.1) for those without these abnormalities (Figure 1a). The risk of progression, however, changed significantly over time, and PFS at 24 months for patients with del(17p13.1), del(11q22.3) and without these abnormalities were, 4%, 5% and 24%, respectively. PFS was not significantly different in patients with or without complex karyotype (median PFS: 8.7 (95% CI: 6.1–12.1) versus 10.1 (95% CI: 8.0–12.6) months, respectively, P=0.16; Figure 1c), although in the subset of patients with del(17p13.1), PFS was inferior for those with complex karyotype (median PFS: 9.8 (95% CI: 6.2–12.6) versus 12.8 months (95% CI: 8.0–19.1), respectively, P=0.03; Figure 1d).
Similarly, OS was not significantly different among the cytogenetic groups (P=0.13, Figure 1b), with median OS of 19.8 months (95% CI: 13.0–32.9) for patients with del(17p13.1), 35.6 months (95% CI: 25.8–48.4) for those with del(11q22.3) and 25.8 months (95% CI: 10.3–33.6) for patients without these abnormalities. The presence of complex karyotype, however, was associated with a significantly shorter OS (median OS: 18.3 (95% CI: 11.0–27.7) versus 35.6 months (95% CI: 25.8–45.0), respectively, P=0.04, Figure 1e). Again, this difference could largely be attributed to those patients with del(17p13.1) and complex karyotype.
In a multivariable model that included both cytogenetic group and complex karyotype, no variable was significantly associated with ORR (P=0.21 and P=0.15, respectively). There were moderate differences in PFS between cytogenetic groups (patients with del(17p13.1)/del(11q22.3) versus others), when the hazard ratio was allowed to change over time (P=0.07; Table 1), with risk of progression for patients with del(17p13.1) or del(11q22.3) increasing over those without these abnormalities by 12 months of follow-up. The presence of complex karyotype did not add a significant amount of information in the model (P=0.29) and the only independent predictor of shorter PFS was the presence of bulky adenopathy (HR=2.0, 95% CI: 1.1–3.7; Table 1). No significant differences in OS were observed among the cytogenetic groups (P=0.24) in the multivariable model (Table 1). Complex cytogenetics were moderately associated with OS, independent of cytogenetic group (P=0.08), and number of prior therapies was significantly associated with OS (P=0.005). For each additional previous therapy received, the risk of death was 1.19 times higher (95% CI: 1.05–1.34).
Collectively, these data show that patients with genetically high-risk disease detected by FISH who are treated with flavopiridol in the relapsed setting have similar outcomes to those without these abnormalities. In contrast, complex karyotype is associated with inferior outcomes. As this is one of few studies to show an impact of karyotype complexity on outcome of a uniformly treated population, our data underscore the need for alternative treatment strategies, particularly in patients with del(17p13.1) and complex karyotype. It also identifies a group of patients in which early stem cell transplant should be considered.
We have also explored the relationship of the anti-apoptotic Bcl-2 family protein Mcl-1 to flavopiridol response, as high levels of Mcl-1 have been shown to correlate with adverse prognostic factors and predict non-response to chemotherapy.13 We see that Mcl-1 generally decreases during the first week of treatment then rises back to baseline by week 2. We have seen no correlation with either baseline Mcl-1 or change in Mcl-1 with flavopiridol treatment to response (Supplementary Tables 1 and 2). These exploratory findings will need to be confirmed with larger numbers of patients.
It has been demonstrated that patients with high-risk cytogenetic abnormalities do not respond well to traditional chemotherapeutic agents. Given the significant toxicities associated with nucleoside analogs and alkylating agents, patients who are unlikely to obtain long-term remissions due to high-risk genetic abnormalities are appropriate to refer for clinical trials investigating flavopiridol or other non-chemotherapeutic agents in the up-front setting. Additionally, we have treated 19 patients with flavopiridol who went on to receive reduced intensity conditioning allogeneic stem cell transplantation, and 47% of these patients were able to proceed directly to transplant after flavopiridol therapy.14 As flavopiridol is associated with a relatively low risk of infectious complications, given the highly refractory patient population,12 it may be an appropriate agent for debulking before stem cell transplantation.
This analysis shows that agents like flavopiridol that work independently of the p53 pathway should be actively investigated in patients with genetically high-risk disease. Flavopiridol and other cyclin-dependent kinase inhibitors are thus attractive agents to study in the first-line setting for patients with poor-risk cytogenetic abnormalities, as well as to use in further single agent and combination studies in relapsed or refractory disease.
This work was supported by Leukemia and Lymphoma Society SCOR Grant, the D. Warren Brown Foundation, NIH/NCI; P01 CA81534 and 5KL2RR025754-02, P50-CA140158, 5K12 CA133250-03, N01-CM-62207 and U01 CA 076576. AJ is a Paul Calabresi Scholar.
About this article
Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)
Insights into apoptotic proteins in chemotherapy: quantification techniques and informing therapy choice
Expert Review of Proteomics (2018)