Prognostic model identifies patients with CLL at risk for ibrutinib failure
Click Here to Manage Email Alerts
A four-factor prognostic model identified patients with chronic lymphocytic leukemia who had poorer survival outcomes with ibrutinib, according to study results published in Journal of Clinical Oncology.
Because outcomes of patients with CLL are variable, this model may be used to identify patients at risk for ibrutinib (Imbruvica; Pharmacyclics, Janssen) failure who should instead be considered for more intensive therapy or treatment on a clinical trial, according to researchers.
“Being able to estimate risk for progression and outcomes in general can set expectations, help choose appropriate follow-up schedules and monitoring, and allows identification of patients who may benefit from different approaches,” Inhye E. Ahn, MD, staff clinician at the hematology branch of the NHLBI at NIH, told Healio. “Patients with favorable prognoses could be managed with a more expectant approach and less intensive therapy. On the other hand, patients with unfavorable prognoses will need more intensive approaches.”
Existing prognostic models for CLL mostly were developed in the context of chemoimmunotherapy or for patients treated with a variety of therapies, Ahn added.
“Our study stands out from other models in that it is developed in the context of ibrutinib, now one of the most common treatment agents used in CLL,” she said.
The study’s discovery data set consisted of 720 patients treated with ibrutinib, a Bruton tyrosine kinase inhibitor, in phase 2 or phase 3 trials. Researchers separated patients into training (n = 541; median age, 69 years; 65.8% men) or internal validation (n = 179; median age, 68 years; 68.2% men) cohorts.
Results of a multivariable analysis revealed four factors independently associated with poorer PFS and OS in the discovery cohort. These included relapsed/refractory CLL, TP53 mutations, beta-2 microglobulin of 5 mg/L or greater, and lactate dehydrogenase greater than 250 U/L. Researchers noted that the probability of PFS and OS decreased as the number of these factors increased among patients at baseline (P < .0001).
Researchers then used these factors to create a prognostic model, for which zero to one factor present before ibrutinib initiation indicated low risk, two factors indicated intermediate risk, and three to four factors indicated high risk.
An analysis of all patients in the discovery data set, plus an additional 84 patients (median age, 66.5 years; 57.1% men) treated with single-agent ibrutinib on an NIH trial, showed 3-year PFS rates of 47% for the high-risk, 74% for the intermediate-risk and 87% for the low-risk groups (P < .0001). Corresponding rates of 3-year OS were 63%, 83% and 93% (P < .0001).
Although one of the four factors in the model is relapsed/refractory CLL, researchers noted the model remained significant when analyzing patients receiving first-line treatment.
However, this may be an area for further research, Ahn said.
“Prior treatment is an important contributor to risk,” she said. “So, for treatment-naive patients, we likely need to refine the model.”
The third cohort of 84 patients underwent sequencing for BTK and PLCG2 mutations, which are common in patients who develop resistance to ibrutinib. Results showed cumulative incidence of these mutations of 50% among those deemed high risk per the prognostic model, 40% among those deemed intermediate risk and 17% among those deemed low risk.
Further, 17% of high-risk patients in that cohort developed Richter transformation, compared with 5% of those in the intermediate-risk group and none of those in the low-risk group.
Researchers plan to further analyze how complex karyotype may be an additional prognostic factor.
“Our group and others also have published data associating CD49d expression with shorter PFS,” she said. “Further research is needed to confirm additional data sets and, if validated, new parameters would need to be incorporated into the model.”
Ahn and colleagues have made the model available online at https://dir.nhlbi.nih.gov/lab/LLM/CLL4fxmodel/.
“This tool helps physicians decide what prognosis a patient might have when they were to start ibrutinib,” Ahn said. “For patients with good risk, single-agent ibrutinib can be expected to control the disease for a long time in most patients. For high-risk patients, doctors may want to do closer monitoring. [Although] likely to benefit initially, high-risk patients are more likely to develop progression and resistance to ibrutinib within a few years on therapy. These patients in particular could benefit from combination therapy, clinical trials or intensification of treatment.”
The four-factor model is a “handy tool” for clinicians and also “helps to formulate new questions and paves the way for further research,” Stephan Stilgenbauer, MD, professor of medicine at Saarland University Medical Center in Germany, wrote in a related editorial.
One area of further study would be the use of the model in the context of combination therapy vs. monotherapy, Stilgenbauer wrote.
“The relevant proportion of patients in their work was treated with ibrutinib plus obinutuzumab [Gazyva, Genentech/Roche], and the results with this combination seem to be valid,” he wrote. “However, other combinations (eg, ibrutinib with venetoclax [Venclexta; AbbVie, Genentech]) are promising, and the applicability of the model to this combination will need to be addressed in the future.”
References:
Ahn IE, et al. J Clin Oncol. 2020;doi:10.1200/JCO.20.00979.
Stilgenbauer S. J Clin Oncol. 2020;doi:10.1200/JCO.20.02685.
Perspective
Back to Top
Lindsey E. Roeker, MD
Although novel agents have transformed the therapeutic approach to CLL, some patients with CLL continue to experience more aggressive clinical courses with disease that is harder to control and shortens their survival. This article describes a four-factor model for patients treated with ibrutinib that predicts PFS and OS, development of ibrutinib-resistance mutations and Richter transformation. This model outperforms CLL International Prognostic Index (CLL-IPI), which serves as an important reminder that models are best applied to the population in which they were validated. Although the CLL-IPI is a robust predictor for chemoimmunotherapy patients, its performance in this novel agent-treated population was not as strong as this new model.
The authors point out that risk stratification will be useful in setting expectations for clinicians and patients, potentially guiding monitoring frequency and encouraging risk-adapted treatment approaches. Building this model was an important first step; the next step will be to design trials guided by this model.
Although novel agent combination therapy has certainly proved to be effective, toxicity is also a concern. The future of these combinations may lie in individualizing the therapeutic approach by identifying patients most likely to benefit. Whether we can change the natural history of CLL for patients with more aggressive disease remains to be seen and is a question of critical importance.
Collapse
Click Here to Manage Email Alerts