2012 EDUCATIONAL BOOK - American Society of Clinical Oncology

Predicting Clinical Outcome in B-Chronic
Lymphocytic Leukemia
Overview: B-Chronic lymphocytic leukemia (CLL) is a relatively
common B-cell malignancy that has a very heterogeneous
clinical course, despite carrying the designation of
“chronic,” which is a gross oversimplification. Being able to
give some estimate of the rates of disease progression and
overall survival (OS) at first diagnosis is, therefore, important
in CLL. The ability to accurately predict response to therapy,
as well as subsequent duration of response to therapy, is
required given the variability of current therapies to induce
and sustain treatment responses. The holy grail of prognostics
THE MAJOR features of CLL that dictate need for
prognostics include the following points: it is a relatively
common leukemia (approximately one in 100,000
patients in North America), 1 the disease is incurable with
the exception of transplant, clinical courses are notoriously
variable, and the patients have considerable anxiety with
this diagnosis. The latter is now well documented and it is
attributable to the fact that we often do not treat these
patients with Rai stage 0–1 on diagnosis. This practice is
based on clinical trial data suggesting treatment is not a
benefit in the early stages of disease. 2 It is also partially
because we do not have a ready portfolio of relatively
nontoxic agents for treating our early-stage patients with
CLL. The anxiety found in CLL is pervasive and has been
quantified, with clear evidence that the anxiety may exist for
years postdiagnosis. 3,4 The cause of this psychologic distress
is multifactorial but one causal aspect is the current dogma
to watch and evaluate early-stage CLL as best as is possible
for a given patient. The case for prognostics, therefore, is
that it provides useful guides for initial patient assurance
(or not) and subsequent treatment. It is this need for
prognostic assistance in early-stage CLL that primarily
drove the research to develop more effective and powerful
methods that inform us. The result of this research is now
evident in that there are prognostic parameters used alone
or in model systems that assist us in counseling and/or
management of the majority of patients with CLL. These
models can incorporate either clinical-based factors or molecular
parameters that reflect the biology of CLL B cell and
its microenvironment. These prognostic features or models
are helpful for predicting time to first therapy, extent of
response to treatments, and duration of response. In some
cases, these features or models guide us in limited ways to
treatment choices. Although not ideal, the wise use of these
prognostics is and should be a major assist to us in CLL
practice.
In this article, I will survey the available maneuvers and
tests that can be used for most patients and that will provide
for the practitioner and the patient a guidepost to risk
stratification. These tests can tell us whether a given patient
has a lower or higher risk of progressing over time. In
addition, it is possible to use these tests to guide patient
counseling, including determining how often a patient
should be seen by the practitioner and potentially some
treatment selection.
394
By Neil E. Kay, MD
would be to state with accuracy which therapy or types of
therapy are best for a given patient. Although there is no
complete answer to prognostic counseling, there is a continued
development of markers specific to the CLL B cell and/or
to its environment, as well as of testing of prognostic models.
These models use both traditional and novel prognostic markers
that can aid in the dissection of outcome for early-stage
CLL in terms of progression risk and time to therapy. This has
resulted in significant enhancement of our ability to guide and
predict outcome for our patients with CLL.
Clinical Course and Prognostic Parameters
The current dogma is that for every 100 patients with
CLL, approximately one-third will not progress to treatment
even over decades, one-third will eventually progress, and
one-third will need more urgent treatment than the rest. 5 In
addition, because of intensive research on early-stage CLL it
is known that approximately 50% of those patients will have
high risk based on the presence of adverse prognostic features.
6,7 Indeed, we can now identify a cohort of patients
with high-risk CLL at diagnosis who will have rapid disease
progression, poor response to treatment, and poor survival
based on prognostic methods developed from an improved
understanding of the biology of CLL. The prognostic parameters
that are used to define this risk can be subdivided into
both traditional/clinical and novel prognostic factors. The
traditional and clinical factors are usually based on quantifiable
plasma factors (beta-2 microglobulin, lactic dehydrogenase
[LDH]), Rai stage, or hematologic features such as
bone marrow features (diffuse infiltration) or levels of blood
lymphocyte counts over time. The novel prognostic parameters
that are in routine practice are leukemic cell based.
Examples of the latter include the presence of membrane
proteins such as CD38 or CD49 days, cytoplasmic presence
of ZAP-70, and prognostic nuclear features including immunoglobulin
variable heavy chain (IgVH) gene mutation status
and cytogenetic abnormalities on fluorescent in situ
hybridization (iFISH). In the assessment of these prognostic
factors for CLL it is critical to consider and clarify the
clinical features that are studied for association with the
particular prognostic factor(s). Although the focus has been
on using prognostics for previously untreated CLL, there is
a special need for more information on subsequent course for
relapsed patients. The most helpful information in these
cases is the time of the patient’s relapse from initial therapy.
Here we also need better prognostic variable to predict their
subsequent clinical outcomes. However, for now, the treatment
responses following initial therapy can be best predicted
by clinical features that include extent of response
From the College of Medicine, Mayo Clinic, Rochester, MN.
Author’s disclosure of potential conflicts of interest are found at the end of this article.
Address reprint requests to Neil E. Kay, MD, College of Medicine, Mayo Clinic, Stabile
6-28, 200 First Street SW, Rochester, MN 55905; email: kay.neil@mayo.edu.
© 2012 by American Society of Clinical Oncology.
1092-9118/10/1-10