Although CLL is the most common chronic leukemia, there remains no single agreed-upon standard of care for first-line treatment.
Ahmad Halwani, MD
Division of Hematology
and Hematologic Malignancies
Huntsman Cancer Institute
University of Utah Healthcare
ï¿¼Case Study: Mrs Smith
You’re in clinic and your next patient is Mrs Smith. She is a 66-year-old woman who was diagnosed with chronic lymphocytic leukemia (CLL) four years ago. Lately she has developed gradually worsening anemia and enlarging lymph nodes. She does not have deletion 17p by conventional karyotype analysis or FISH. You need to recommend a treatment option for her now symptomatic CLL.
Best Treatment Option Unclear
Although CLL is the most common chronic leukemia, there remains no single agreed-upon standard of care for first-line treatment. In Mrs Smith’s case, it would be quite reasonable to attempt to apply evidence from the recently reported German CLL Study Group randomized con- trolled trial (RCT) of fludarabine, cyclophosphamide, and rituximab (FCR) versus bendamustine rituximab (BR) for first-line treatment of CLL.1 At first glance, FCR seems like the better option, as it is associated with a longer progression-free survival (PFS).
However, FCR is more toxic with almost twice the rate of severe infections. In an unplanned subgroup analysis, there was no benefit in patients who were older than 65 years, or in those with multiple and/or moderate comorbidities. The trial excluded patients with impaired kidney function. Despite being quite active, Mrs Smith was diagnosed with diabetes more than 7 years ago, and has diabetic nephropathy. Her renal function is borderline normal. On account of her comorbidities and renal function, it’s quite likely that she would not have been eligible for this trial.
This case illustrates a common context in hematological malignancies. Despite having the results of a phase III RCT, the best treatment option for many patients who are older or who have comorbidities remains unclear.
Efficacy Established in Carefully Selected Populations
Treatment options for lymphoid malignancies have greatly increased over the past decade thanks to advances in cancer biology and therapeutics. For example, a clinician treating lymphoma or CLL can draw upon more than a dozen agents from a handful of modalities: chemoimmunotherapy,2 antibody-drug conjugates,3 checkpoint inhibitors,4 immunomodulators,5,6 and tyrosine kinase inhibitors.7—9
As a result, some patients with advanced-stage indolent lymphoma or CLL can live for decades, and a significant proportion of patients with aggressive lymphomas can be definitively cured of their disease.
Clinical research in oncology has typically relied on prospective, often randomized, clinical trials as the gold standard methodology to generate evidence of the benefits and risks of novel treatments. Prospective clinical trials rely on a number of elements to achieve a high level of internal validity: strict eligibility criteria, consistent treatment delivery, predefined endpoints, and a priori statistical power considerations to accept or reject the null hypothesis.10 Randomization ensures the exchangeability of patients in different treatment arms and allows causal inference.
The resulting high internal validity of prospective clinical trials is often achieved at the expense of external validity, or applicability of evidence to patients in real world settings. Stringent eligibility criteria often lead to exclusion of patients with comorbidities. As a result, trials often accrue patients who are younger and healthier than patients in actual clinical practice.11,12 Patients who belong to ethnic minorities are also underrepresented in clinical trials.13 The technical and regulatory challenges involved in the conduct of clinical trials often lead to a predominance of patients treated in academic settings as opposed to community settings.
The average patient enrolled in pivotal prospective trials in CLL over the last decade is typically in their early 60s, predominantly white, and has normal organ function and limited number and severity of comorbidities.14—16 At diagnosis, the average patient with CLL in real world settings is almost a decade older.17,18 About 15% of patients with CLL in the United States are nonwhite (including 7% black and 5% Hispanic.)19 Black and Hispanic patients have a worse overall survival compared with white patients with CLL.13,19,20 At the time of diagnosis, patients with CLL have a median of two comorbidities and take two prescription medications per day.21
Patients with CLL with two or more comorbidities who still qualify for participation in clinical trials have a worse PFS and overall survival compared with patients with a lower comorbidity burden.22 It is estimated that one in four patients with CLL is not eligible to participate in clinical trials on account of organ function per National Cancer Institute-Working Group 1996 guidelines.21
Evaluating Therapies in Real World Scenarios
As a result of these differences, the need to complement our understanding of the efficacy of cancer treatments as studied in RCTs with evidence of effectiveness in a real world setting is increasingly recognized.23 This type of research is typically referred to as comparative effective- ness research (CER), population research, or outcomes re- search. While not as common as explanatory RCTs, and at times lagging behind RCTs addressing the same question, CER studies are now starting to address the evidence gap.
In the example of chemo-immunotherapy for first-line treatment of CLL, three groups have so far examined the effectiveness of FCR in real world settings.18,24,25 Their findings are similar and point to frequent dose reductions, which lead to a lower PFS. In a study of 106 patients with CLL who were treated with FCR in the Oncomip healthcare network in France from 2005—2011, Bouvet el al reported that one in two patients had a dose reduction of more than 20%.24 About 60% of patients had grade 3-4 neutropenia, 16% had neutropenic fever, and 19% were hospitalized due to sepsis. Dose reductions were associated with a lower PFS.
Herishanu et al recently reported outcomes on 128 young patients treated in 10 Israeli medical centers, with similar results.25 About 50% of patients had dose reductions. Patients who received less than two-thirds the recommended dose had a fourfold increase in hazard of progression.
Knauf et al recently reported the characteristics of 806 patients with CLL treated by office-based hematologists in Germany.18 The group’s first report suggests that two regimens had comparable effectiveness; a more detailed report on the cohort’s survival outcomes is yet to be published.
Other groups have used Surveillance, Epidemiology, and End Results (SEER) data or SEER and Medicare databases to examine the impact of patient factors such as age, race, and ethnicity on outcomes of treatment as well as treatment practices in elderly patients with CLL in the United States.19,26
The Road Ahead
Going forward, various methodologies have been proposed to extend and validate the results of clinical trials in actual clinical practice settings. These include pragmatic trials, observational cohort studies, systematic reviews and meta-analyses.27 Widespread adoption of electronic health records offers the promise of observational effectiveness research using clinical information that was requested, recorded, and/or used by treating physicians at the point of care. This is a richer dataset than claims-based data, and provides greater longitudinal follow-up than registry-based data. The causal inference framework,28 applied to observational effectiveness research, offers the potential of explicitly specifying the assumptions made in the analysis, and minimizing and/or quantifying the biases inherent to the observational nature of the study.29
While prospective clinical trials will remain integral to our understanding of the efficacy of novel cancer treatments, there is a mostly unmet need for research that aims to further our understanding of the comparative effectiveness of cancer treatments in real world settings. Novel approaches in comparative effectiveness research, including cohort studies that leverage robust data from electronic medical records, have the potential to address these knowledge gaps.