Author's Accepted Manuscript

Prognostic Factors and Risk Stratification in

Chronic Lymphocytic Leukemia (CLL)
Sameer A. Parikh, Tait D. Shanafelt

www.elsevier.de/endend

PII: S0093-7754(16)00015-4
DOI: http://dx.doi.org/10.1053/j.seminoncol.2016.02.009
Reference: YSONC51913

To appear in: Semin Oncol

Cite this article as: Sameer A. Parikh, Tait D. Shanafelt, Prognostic Factors and Risk
Stratification in Chronic Lymphocytic Leukemia (CLL), Semin Oncol, http://dx.doi.
org/10.1053/j.seminoncol.2016.02.009

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting galley
proof before it is published in its final citable form. Please note that during the
production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 Prognostic Factors and Risk Stratification in Chronic Lymphocytic Leukemia (CLL)

Sameer A. Parikh, Tait D. Shanafelt

Division of Hematology, Department of Medicine

Mayo Clinic, Rochester, MN

Keywords: outcomes, cytogenetics, IGHV mutation, next generation sequencing

Corresponding Author:

Tait D. Shanafelt, MD

Division of Hematology, Department of Medicine

Mayo Clinic

200 First Street SW

Rochester, MN, 55905

Tel: 507-538-0591

Fax: (507) 266-4972

Email: shanafelt.tait@mayo.edu

1
ABSTRACT

There is considerable heterogeneity in the clinical outcome of patients with chronic

lymphocytic leukemia (CLL). While some patients live for decades without any therapy, others

die within years of diagnosis despite multiple treatments. To better counsel newly diagnosed

CLL patients about their disease course, the Rai and Binet staging systems were developed four

decades ago. A deeper understanding of the biologic and molecular aberrations contributing to

the pathogenesis of CLL led to identification of novel prognostic markers such as

immunoglobulin heavy chain variable gene (IGHV) mutation status, leukemia-cell expression of

CD38, ZAP-70 and CD49d, and cytogenetic abnormalities detected by fluorescent in situ

hybridization (FISH). The advent of next generation sequencing has provided unprecedented

insights into the sub-clonal architecture of CLL and its impact on disease progression and

survival. More recently, integrated prognostic scoring systems that incorporate clinical, biologic

and genetic characteristics into a single risk score have been developed and appear to improve

the accuracy of prognostication for individual patients. This review summarizes the state-of-the-

art prognostic factors and will guide the practicing clinician in their care of patients with CLL.

2
INTRODUCTION

Chronic lymphocytic leukemia (CLL) is a low-grade B-cell lymphoproliferative disorder,

with approximately 16,000 new cases diagnosed in the U.S. in 2014.1 There is considerable

heterogeneity in the disease course of CLL – some patients live for decades without therapy,

whereas others die due to disease progression within a few years from diagnosis despite

treatment.2 Given the chronic and incurable nature of CLL, it is important to develop effective

tools that accurately define prognosis for each individual patient at the time of diagnosis. Such

knowledge will allow the practicing clinician to develop effective strategies for i) patient

counseling; ii) determine the frequency of follow-up; iii) identify those most appropriate for

early intervention trials; and iv) inform therapy selection.

TRADITIONAL PROGNOSTIC FACTORS

The Rai3 and Binet4 staging systems, developed in the 1970s, represent one of the initial

efforts in segregating CLL patients into distinct prognostic groups. Both systems utilize simple

and readily available clinical and laboratory parameters that classify patients into three major

prognostic groups. The characteristics used to classify patient stage by the Rai classification, the

expected survival in the original publication3 and a large recent cohort seen at Mayo Clinic are

shown in Table 1. While these prognostic tools have proven incredibly useful over the last 40

years, there remains significant clinical heterogeneity among patients within each Rai and Binet

stage category. Additionally, approximately three-quarters of all newly diagnosed CLL patients

are now diagnosed at the Rai 0/Binet A stage, where these traditional staging systems are unable

to accurately stratify risk of progression. Between 1975 and 1995, a number of other prognostic

parameters such as lymphocyte doubling time (LDT),5 pattern of bone marrow involvement

3
(diffuse vs. non diffuse),6 and absolute lymphocyte count (ALC)7,8 were reported to be

associated with outcomes. Unfortunately, these markers do not provide significant insights into

the biological characteristics of the CLL clone and are therefore unable to fully account for the

heterogeneity in clinical outcomes.

NEWER PROGNOSTIC FACTORS

Serum-Based Markers

Serum parameters such as beta-2 microglobulin (β2M, a component of the HLA class I

complex on nucleated cells) and thymidine kinase (TK, an intracellular protein involved in DNA

synthesis) have been shown to correlate with disease burden and clinical outcome in patients

with CLL.9 These basic serum markers have proven robust prognostic parameters with

independent prognostic value even with the advent of sophisticated measures of CLL genetics

and biology.10 Patients with serum β2M of <3.5 mg/L have a substantially longer time to

progression compared to those with levels of ≥3.5 mg/L.7 Additionally, patients with an elevated

level of β2M have lower rates of complete remission (CR) and shorter overall survival (OS) to

frontline fludarabine-based chemoimmunotherapy.11 Similarly, patients with serum TK values of

>8.5 U/L have been associated with rapid LDT, shorter time to progression and lower rate of

response to therapy with fludarabine.12,13

Immunoglobulin Heavy Chain Gene Mutation Status

In 1999, two independent groups reported that patients with higher degrees of somatic

mutation in the immunoglobulin heavy chain variable gene (IGHV) of their CLL clone

experienced longer OS.14,15 These studies showed that CLL patients with mutated IGHV (defined

as a greater than 2% difference from the germline nucleotide sequence) had a median OS of

4
greater than 20 years whereas those with unmutated IGHV had median OS of ~8 years. Figure

1A shows the median survival of 1519 previously untreated CLL patients seen at Mayo Clinic

since 1995 with mutated IGHV was 17 years compared to 10 years for those with unmutated

IGHV genes (p<0.001).

Flow-cytometry Based Markers

Despite its robust prognostic value, assessing IGHV mutation status is labor intensive and

is not routinely available assay in many clinical laboratories. This led a number of groups to

search for surrogates of IGHV mutation status. These studies revealed that the expression of

CD38 and zeta-associated protein 70 (ZAP-70) as determined by flow cytometry, while

imperfect proxies for IGHV status, were nonetheless associated with clinical outcomes in CLL

patients. Both CD38 positivity (arbitrarily defined as a cut of ≥30% cells expressing CD38) and

ZAP-70 positivity (arbitrarily defined as a cutoff of ≥20% expressing ZAP-70) are also

associated with resistance to standard therapy, shorter time to first treatment and OS. 16-20 The OS

of previously untreated CLL patients cared for at Mayo Clinic since 1995 based on the

expression of CD38 (n=2371) and ZAP-70 (n=1816) are shown in Figures 1B and 1C,

respectively. More recent studies have explored alternative approaches to measuring ZAP-70 in

an effort to improve accuracy and reliability. For example, the CLL Research Consortium

reported that methylation in a single CpG dinucleotide 223 bp downstream of the transcriptional

start site of ZAP-70 was associated with a lack of ZAP-70 protein expression, longer time to first

treatment and improved OS.21 Determination of the methylation status of ZAP-70 by

pyrosequencing in a CLIA certified laboratory may eventually replace ZAP-70 analysis, pending

its assessment in a large prospective phase 3 trial (NCT01886872).22 Despite their value in

5
predicting outcome for populations of CLL patients, neither ZAP-70 nor CD38 are sufficiently

accurate in predicting the outcome of an individual CLL patient.

More recently, CD49d, an α-integrin subunit (α4) that binds fibronectin and VCAM-1,

and plays an important role in nurturing interactions between leukemic B-cells and the micro-

environment, has also been shown to predict outcomes in CLL patients.23 In an international

meta-analysis that evaluated the utility of all flow-cytometry-based markers in ~3000 patients,

CD49d emerged as the strongest flow-cytometry based predictor of time to first treatment and

OS, and provided prognostic value independent of IGHV and fluorescence in situ hybridization

(FISH) on multivariable analysis.24 The OS of newly diagnosed CLL patients seen at Mayo

Clinic according to the CD49d expression (n=1476) is shown in Figure 1D.

Cytogenetics, FISH and early sequencing efforts

Using conventional G-banding techniques, genomic aberrations detected on

chromosomes 12, 13 and 14 were shown to be associated with prognosis in CLL. However,

given the low mitotic rate in CLL, these were observed in only ~50% patients.25 With the use of

interphase FISH to evaluate common recurring genetic defects in non-dividing cells, Dohner et al

proposed a new prognostic model in 325 CLL patients. Using a hierarchical classification

scheme, they showed that the patients with del17p13 had the shortest survival (32 months),

followed by patients with del11q23 (79 months), trisomy 12 (111 months), normal karyotype

(114 month) and del13q14 as the sole abnormality (133 months).26 Figure 1E shows the survival

of previously untreated CLL patients seen at Mayo Clinic since 1995 according to their FISH

analysis determined within 36 months of diagnosis. Subsequent studies have suggested that the

genetic composition of the CLL clone changes over time with new cytogenetic abnormalities

6
(determined by FISH) acquired during follow-up in ~25% patients; this clonal evolution was

associated with worse outcomes.27-29

Although the study by Dohner and colleagues suggested that all patients with del17p13

have an aggressive disease and progress to requiring therapy within 2 years, ~75% patients had

received prior CLL therapy (which represents a high-risk group). In contrast, investigators from

M.D. Anderson Cancer Center (MDACC) and Mayo Clinic reported that of 99 treatment-naïve

CLL patients with del17p13, approximately 50% developed progressive disease requiring

therapy within 12 to 18 months. The other half had relatively stable disease extending out to 70

months of follow-up, suggesting that not all patients with del17p13 require therapy at the time of

diagnosis. Unfortunately, information about the TP53 mutation status on the remaining allele

was not available in this retrospective analysis.30

Using Sanger sequencing, a number of studies reported that monoallelic mutations in

TP53 is associated with poor prognosis in CLL and resistance to standard therapy.31,32 Rossi and

colleagues performed DNA sequencing of TP53 exons 2 to 10 and del17p13 interphase FISH at

diagnosis in 308 CLL patients and showed that TP53 disruption (either by mutation or deletion)

was present in 44 (15%) patients: of these, 18 (6%) had both TP53 mutation and del17p13 on

FISH; 10 (3%) had TP53 mutation but no del17p13 on FISH; and 16 (5%) had no TP53 mutation

but del17p13 on FISH analysis. These results suggest that ~3% patients with no evidence of

del17p13 by FISH testing have a mutated TP53 allele. Importantly, these patients had a short OS,

comparable to patients with del17p13 by interphase FISH.32 Data from the German CLL4 trial

(comparing fludarabine vs fludarabine and cyclophosphamide) and the UK LRF CLL4 trial

(chlorambucil vs fludarabine with or without cyclophosphamide) confirmed that TP53 mutations

7
occur in 3-5% newly diagnosed CLL patients in the absence of del17p13 on FISH testing, and

were associated with shorter OS and poor response to therapy.33,34

NOVEL PROGNOSTIC MARKERS

Next generation sequencing

Recent studies using next generation sequencing approaches have identified several

recurrently mutated genes in CLL patients. These include genes involved in DNA damage and

cell cycle control (ATM, TP53, RB1, BIRC3), Notch signaling (NOTCH1, NOTCH2, FBXW7),

inflammatory pathways (MYD88, DDX3X, MAPK1), spliceosome machinery (SF3B1), and

cytokine signaling (NRAS, KRAS, BRAF).35,36 Mutations in NOTCH1, a ligand-activated

transcription factor were present in ~10-15% of CLL patients at the time of diagnosis; and

preferentially occur in patients with unmutated IGHV genes and trisomy 12. Several studies have

reported that NOTCH1 mutations are associated with shorter time to progression and OS;37,38

some have suggested that patients with trisomy 12, NOTCH1 mutations and B-cell receptor

(BCR) subset #8 cooperate to increase risk of Richter syndrome.39 Finally, recent data from the

CLL8 trial suggest that patients with NOTCH1 mutations may not benefit from the addition of

rituximab.40 The mechanism by which NOTCH1 mutations influence response to rituximab

remains under investigation and must be replicated in other studies before it influences therapy

selection. It is also unclear if this decreased response will also be seen in patients with NOTCH1

mutations who receive other anti-CD20 monoclonal antibodies such as ofatumumab and

obinutuzumab. Mutations in SF3B1, part of the spliceosome machinery, are present in ~5-10% of

newly diagnosed CLL patients, preferentially occur in patients with del13q14, and are associated

with shorter time to progression and OS, independent of other prognostic markers.35,41,42

8
Subclonal mutations

Early studies using FISH and array comparative genomic hybridization (aCGH)

illustrated that the leukemic B-cell population in patients with CLL is comprised of various sub-

clones harboring distinct driver mutations.27 Braggio and colleagues performed a sequential,

longitudinal analysis of tumor samples in 22 CLL patients, where genome-wide genetic analysis

was performed by aCGH on at least two sequential tumor samples >6 months apart. Overall, 6 of

22 patients (27%) showed differences in copy-number abnormalities (CNA) obtained in paired

samples before and after therapy. Of the six cases with CNAs, two patients had linear evolution

characterized by the maintenance of the initial abnormalities and the subsequent acquisition of

additional CNAs in the same subclone. However, two or more subclones were identified prior to

therapy in the other four patients; these minor subclones emerged as dominant clones following

therapy for CLL (suggesting a branched evolution).43 Landau and colleagues extended this

finding by using whole-exome sequencing in 149 CLL patient samples, and found an average of

10.3±5.5 clonal mutations per sample and an average of 8.5±5.8 subclonal mutations per sample.

The mutations were defined as clonal if the cancer cell fraction harboring the mutation was

>0.95. These investigators demonstrated that after adjusting for IGHV mutation status, prior CLL

therapy, and high-risk cytogenetics, the presence of subclonal driver mutations was associated

with shorter time to progression.44 Rossi et al also reported similar findings in 309 CLL patients

where the presence of subclonal TP53 mutations (seen in ~9% patients) predicted for a poor

outcome.45 These studies also showed that the subclonal mutations became the dominant clone

after CLL therapy and predicted for a worse outcome, thereby supporting the notion that therapy

for CLL may increase the risk of chemo-refractory disease in the future.43-45

9
MicroRNA

Nearly a decade ago Calin and colleagues described the seminal role of microRNA’s in

the molecular pathogenesis of CLL. A unique microRNA (miR) expression signature composed

of 13 genes successfully differentiated patient samples with unmutated vs. mutated IGHV genes,

high vs. low expression of ZAP-70, and shorter time to disease progression.46 Since this

discovery, numerous reports have described the role of specific miRs in the prognosis of CLL

patients. Low miR-34a levels were strongly associated with impaired DNA damage response and

fludarabine-refractory disease in a study of 60 CLL patients (in the absence of del17p13 or TP53

mutations) seen at the University of Ulm.47 Ferrajoli et al reported that after adjusting for Rai

stage and IGHV mutation status among 228 CLL patients, a high expression of miR-155 in

plasma was associated with shorter OS. Additionally, patients with higher miR-155 levels

responded poorly to standard chemoimmunotherapy, thereby suggesting that it may act as both a

prognostic and a predictive marker in CLL.48 Mraz et al demonstrated that miR-150 was the most

abundant miR present in CLL; with differential expression based on IGHV mutation and ZAP-70

status. Plasma levels of miR-150 were associated with adverse outcomes independent of several

known prognostic markers in CLL.49 Visone et al showed that miR-181b targets the critical

TCL1 oncogene; differences in serum levels were able to predict disease progression in CLL

patients.50 Although several studies have reported other miRs (miR-331, miR-29a, miR-195, and

miR-29c) to be differentially expressed in CLL, the precise mechanism by which they affect

prognosis in CLL remains to be elucidated.51 Although novel drugs targeting aberrant miRs may

be developed for patients with CLL in the future,52 clinical testing of miRs for prognostic

purposes is not currently available.

10
BCR stereotypy

Immunogenetic analysis of the BCR surface immunoglobulins has identified that many

CLL patients share quasi-identical binding sites, so-called “stereotyped” BCRs.53 Although

mathematically this should be an exceedingly rare occurrence, large scale studies now indicate

that at least 1/3rd of all CLL patients have stereotyped BCRs and the remaining 2/3 rd have

heterogenous BCRs.54 Although ~200 different subsets of stereotyped BCR’s have been

described, subsets 1-8 occur most frequently.55 Several studies have suggested that specific

subsets of stereotyped BCRs are associated with distinct clinical features and outcomes,

independent of the IGHV mutation status.56 CLL patients in subset 4 appear to have younger age

at diagnosis and experience an indolent course, whereas patients in subset 1 and subset 3

experience an aggressive disease course.53,57 Although some reports suggest that patients in

subset 8 are at a higher risk of developing Richter syndrome,39 others have not confirmed this

observation.58 Finally, the association of specific subsets with distinct genetic aberrations (such

as SF3B1 mutations in subset 2, and NOTCH1 mutations in subset 8) may represent novel

methods to categorize CLL patients and consider individualized treatment options in the

future.39,59

Other Novel Prognostic Markers

In addition to the aforementioned prognostic markers, several others have been identified.

These include expression of antiapoptotic genes (Bcl2/Bax ratio, MCL-1),60 lipoprotein lipase A

expression,61 ADAM29 expression,62,63 telomere length and telomerase activity,64 CLLU1

expression,65 circulating microvesicles,66 and serum levels of vascular endothelial growth factor

(VEGF) and thrombopoietin,67 among others. Although these provide important insights into

11
disease biology and may be rational therapeutic targets, they are not presently available for

routine clinical use.

INTEGRATION OF PROGNOSTIC MARKERS

With the identification of several prognostic markers in the past 15 years, it has become

difficult for the practicing clinician to determine which factors are most important to counsel

individual patients and how to combine the results of multiple prognostic tests when some

suggest an aggressive clinical course and others a more indolent one. To address these issues,

several investigators have explored methods to combine these factors to develop prognostic

models/nomograms.

Early Attempts at an Integrated Prognostic Tool

Based on simple clinical and laboratory parameters, Wierda and colleagues reported a

prognostic nomogram for OS in 1674 previously untreated CLL patients seen at MDACC from

1981-2004. The following six factors were associated with OS in multivariable analysis: age,

β2M, ALC, sex, Rai stage and number of involved nodal groups. Using the results of these

factors, the authors developed a weighted nomogram (Table 2). Patients were assigned to one of

three groups with median OS of not reached in the low-risk group, 10.3 years in the

intermediate-risk group, and 5.4 years in the high-risk group.8 The index was subsequently

validated as a predictor of OS in an independent cohort and also proved to be a useful approach

to stratify time from diagnosis to first treatment.68 However, this index did not incorporate the

most robust molecular/genetic biomarkers such as IGHV mutation status and FISH, which limits

its clinical applicability.

12
 In an early attempt to combine the results of multiple bio-markers, Krober et al stratified

300 CLL patients into low-, intermediate- and high-risk groups using the combination of IGHV

mutation status and cytogenetic abnormalities detected by FISH. Patients with del17p13 were

classified as high-risk regardless of their IGHV mutation status, patients with del11q23 and/or

unmutated IGHV genes were considered intermediate risk, and those with mutated IGHV genes

in the absence of del17p13 or del11q23 were considered low-risk. The estimated median

survivals ranged from 30 months in the high-risk group to 152 months in the low-risk group.69 A

subsequent study by the CLL Research Consortium evaluated the role of ZAP-70, CD38 and

IGHV mutation status for predicting time to first therapy in ~1000 CLL patients. The

investigators found that the combination of ZAP-70 and IGHV were independent predictors of

time from diagnosis to first treatment, while CD38 lacked independent prognostic value after

adjusting for these factors.17 ZAP-70 positive patients represented a high-risk group with short

time to first treatment independent of their IGHV mutation status, while ZAP-70 negative

patients were divided into intermediate and low-risk groups based on the presence of unmutated

and mutated IGHV genes, respectively.

Recently Developed Integrated Prognostic Tools

In 2012, Rossi and colleagues developed a genetic based prognostic index that

incorporated the results of FISH and next generation sequencing. The authors classified 1274

patients into four risk groups: 1) high-risk, with TP53 and/or BIRC3 abnormalities; 2)

intermediate-risk, with NOTCH1 and/or SF3B1 mutations and/or del11q23; 3) low-risk, with

trisomy 12 or normal karyotype; and 4) very low-risk, with del13q14 as the sole abnormality.

The estimated 10-year survival for these 4 groups was 29%, 37%, 57% and 69% (p<0.001)

respectively. Although this model improved prognostication over genetic abnormalities detected

13
by FISH alone, it did not incorporate other established clinical (age, sex, Rai stage, Eastern Co-

operative Group [ECOG] performance status, number of involved nodal groups) and biological

variables (serum β2M and TK, IGHV mutation status and expression of CD49d).

The c-statistic (equivalent to the area under the Receiver Operating Characteristic curve)

is a standard measure of the predictive accuracy of a logistic regression model. A value of 0.5

indicates that the model is no better than chance at making a prediction of outcome. It has been

proposed that a c-statistic of 0.7 or higher is necessary to have value at the individual patient

level.70 Large studies suggest that the Rai and Binet staging systems have c-statistics of 0.56 and

0.58 for predicting OS, only slightly better than chance.10 While the prognostic index of Rossi

and colleagues was superior to the Rai and Binet systems, its c-statistic of 0.68 suggested it was

not precise enough to accurate predict the outcome of individual patients.

In another recent attempt to develop an integrated prognostic index, the German CLL

Study Group analyzed data from three large Phase III trials that collectively included ~2000

patients. The investigators evaluated the independent prognostic value of 23 clinical, biological

and genetic markers and identified 7 that were independently associated with OS: sex, age,

ECOG performance status, del17p13 on FISH, del11q23 on FISH, IGHV mutation status, serum

β2M level and serum TK level. These factors were then combined in an integrated index with

differential weighting of each factor based on its relative importance for predicting OS in the

multi-variable model (Table 3). Using a composite score that summed the risk of the 7

individual factors, four risk groups were identified: low (score 0-2), intermediate (score 3-5),

high (score 6-10) and very high risk (score >11), with significantly different 5-year survivals of

95%, 87%, 68% and 19%, respectively. The authors subsequently validated this prognostic

index in an external cohort of 676 previously untreated CLL patients seen at the Mayo Clinic.10

14
This prognostic index was also found to reliably stratify not only OS but also time from

diagnosis to progression requiring treatment. Although 6 of the 7 factors in this index are routine

clinical assays, serum TK is not currently available in the United States. This integrated CLL

prognostic index had a c-statistic of 0.75 in the training set and 0.83 in the validation cohort,

suggesting that it is the first prognostic tool with sufficient accuracy to predict clinical outcome

for individual patients.31

With longer follow-up of large cohorts of CLL patients that have novel prognostic

markers available, it is likely that several such “integrated prognostic nomograms” will be

published. Indeed, a collaborative international effort to develop a true international prognostic

index is currently underway. Such an approach would add value by i) providing a consistent

platform by which to counsel patients, ii) determining which assays are essential in routine

clinical practice and which can be eliminated, and iii) providing a benchmark by which to

evaluate new prognostic tests (which must prove they provide incremental prognostic value), and

iv) a consistent strategy by which to select patients for clinical trials evaluating the benefit of

intervention in asymptomatic patients who are at high risk for progression.

PREDICTIVE MARKERS

Although our ability to predict outcomes of newly diagnosed CLL patients has

remarkably improved with identification and integration of novel prognostic markers, we would

ideally like to have biomarkers that can help determine which therapy will work best for a given

patient. Such an approach could not only maximize treatment efficacy but also minimize toxicity

in patients who are not predicted to respond to a given therapy.

Although previous studies suggested that patients with del11q23 had a shorter OS, more

15
recent studies using chemoimmunotherapy regimens such as fludarabine, cyclophosphamide and

rituximab (FCR) suggest that this treatment approach may overcome its adverse impact.71

Currently, the only robust predictive marker in CLL is the presence of del17p13 (for both

treatment-naïve and previously treated CLL patients) on FISH testing or TP53 mutation on

sequencing. CLL patients with del17p13 or TP53 mutation have a very poor response to

chemoimmunotherapy regimens such as FCR.71 Among patients with del17p13 treated with FCR

(n=22) in the CLL8 trial, the CR rate was only 5% and the progression-free survival (PFS) at 3-

years was 18%, compared to 48% and 76%, respectively, for patients with del13q14 as the sole

abnormality (n=105).71 In the context of relapsed/refractory CLL, treatment with FCR led to no

CR’s and the median PFS was 5 months among 20 patients with del17p13 treated at the

MDACC.72 It appears that the prognosis of these patients may not be as dire with the advent of

new kinase inhibitors, such as ibrutinib. In a recent study (RESONATE™-17 trial), O’Brien and

colleagues reported the median PFS was not reached in 144 previously treated CLL patients with

del17p13 who received ibrutinib for a median of 13 months.73 These findings illustrate that new

therapies may alter the implications of even well-established prognostic factors.

RESPONSE TO THERAPY AS A PROGNOSTIC MARKER

Although somewhat of a self-fulfilling prophecy, response to therapy has long been

recognized as a predictor of adverse outcome. Although historically nearly all patients with CLL

had residual disease in the marrow or lymph nodes at the completion of therapy, the more

effective chemoimmunotherapy regimens developed over the last decade result in deep

remissions where no residual disease can be detected at the completion of therapy, even when

assessed using highly sensitive assays. Such approaches to evaluating for minimal residual

16
disease (MRD) employ either allele-specific oligonucleotide polymerase chain reaction for the

clonal IGHV gene or multicolor flow cytometry.74 The 2008 iwCLL guidelines suggest that

either of these techniques can achieve the 0.01% sensitivity in detecting residual leukemia in the

bone marrow.75 The presence of MRD after CLL therapy is one of the most powerful prognostic

markers, and several studies have demonstrated its utility in predicting outcomes.76-78 One of the

biggest drawbacks of using this prognostic marker is that it is not available until after therapy is

completed, and as such cannot be used to prognosticate outcomes in patients with newly

diagnosed CLL who are about to embark on therapy. Additionally, for those CLL patients who

do not need treatment and are in the “wait and watch” approach, this prognostic marker has no

utility at the time of initial diagnosis.

CONCLUSION

There has been an explosion in the discovery of novel prognostic markers in CLL over

the past 15 years. While a number of these markers have been validated across different patient

populations, several of the newer markers await confirmation and their value independent of

other widely available prognostic tools has not yet been established. The integration of clinical,

laboratory and molecular markers into a single prognostic index or nomogram has been a large

step forward and appears to hold the promise of a risk stratification system with sufficient

accuracy to allow application at the level of the individual patient in routine clinical practice.

Refining these tools and identifying disease characteristics that can inform therapy selection and

improve patient outcomes is the task now at hand. The importance of these prognostic markers

will also need to be re-examined in the context of newer treatments for CLL which may alter the

implications of well-established factors.

17
ACKNOWLEDGEMENTS

Sameer Parikh is the recipient of the Mayo Clinic Department of Medicine Career Development

Grant. Tait Shanafelt is a scholar of the Leukemia and Lymphoma Society.

We are grateful to Ms. Kari G. Chaffee, MS (statistician) for analysis of data in Table 1 and

Figures 1A-1E.

FINANCIAL DISCLOSURES:

Sameer A. Parikh: None.

Tait D. Shanafelt: Has received research funding from Genentech, Glaxo-Smith-Kline,

Cephalon, Hospira, Celgene, Jannsen, and Polyphenon E International.

REFERENCES

1. Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. CA: A Cancer Journal for

Clinicians. 2014;64(1):9-29.

2. Shanafelt TD. Predicting clinical outcome in CLL: how and why. Hematology / the

Education Program of the American Society of Hematology American Society of Hematology

Education Program. 2009:421-429.

3. Rai KR, Sawitsky A, Cronkite EP, Chanana AD, Levy RN, Pasternack BS. Clinical

staging of chronic lymphocytic leukemia. Blood. 1975;46(2):219-234.

18
4. Binet JL, Auquier A, Dighiero G, et al. A new prognostic classification of chronic

lymphocytic leukemia derived from a multivariate survival analysis. Cancer. 1981;48(1):198-

206.

5. Molica S, Alberti A. Prognostic value of the lymphocyte doubling time in chronic

lymphocytic leukemia. Cancer. 1987;60(11):2712-2716.

6. Rozman C, Montserrat E, Rodriguez-Fernandez JM, et al. Bone marrow histologic

pattern--the best single prognostic parameter in chronic lymphocytic leukemia: a multivariate

survival analysis of 329 cases. Blood. 1984;64(3):642-648.

7. Bergmann MA, Eichhorst BF, Busch R, et al. Early and Risk-Adapted Therapy with

Fludarabine in High-Risk Binet Stage A CLL Patients Prolongs Progression Free Survival but

Not Overall Survival: Results of the CLL1 Protocol of the German CLL Study Group

(GCLLSG). ASH Annual Meeting Abstracts. 2007;110(11):2038-.

8. Wierda WG, O'Brien S, Wang X, et al. Prognostic nomogram and index for overall

survival in previously untreated patients with chronic lymphocytic leukemia. Blood.

2007;109(11):4679-4685.

9. Hallek M, Wanders L, Ostwald M, et al. Serum beta(2)-microglobulin and serum

thymidine kinase are independent predictors of progression-free survival in chronic lymphocytic

leukemia and immunocytoma. Leukemia & lymphoma. 1996;22(5-6):439-447.

10. Pflug N, Bahlo J, Shanafelt TD, et al. Development of a comprehensive prognostic index

for patients with chronic lymphocytic leukemia. Blood. 2014;124(1):49-62.

11. Tsimberidou AM, Tam C, Wierda W, O' Brien S, Lerner S, Keating MJ. Beta-2

microglobulin (B2M) is an independent prognostic factor for clinical outcomes in patients with

19
CLL treated with frontline fludarabine, cyclophosphamide, and rituximab (FCR) regardless of

age, creatinine clearance (CrCl). ASCO Meeting Abstracts. 2007;25(18_suppl):7034.

12. Matthews C, Catherwood MA, Morris TC, et al. Serum TK levels in CLL identify Binet

stage A patients within biologically defined prognostic subgroups most likely to undergo disease

progression. European journal of haematology. 2006;77(4):309-317.

13. Di Raimondo F, Giustolisi R, Lerner S, et al. Retrospective study of the prognostic role

of serum thymidine kinase level in CLL patients with active disease treated with fludarabine.

Annals of oncology : official journal of the European Society for Medical Oncology / ESMO.

2001;12(5):621-625.

14. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status and CD38 expression as

novel prognostic indicators in chronic lymphocytic leukemia. Blood. 1999;94(6):1840-1847.

15. Hamblin TJ, Davis Z, Gardiner A, Oscier DG, Stevenson FK. Unmutated Ig V(H) genes

are associated with a more aggressive form of chronic lymphocytic leukemia. Blood.

1999;94(6):1848-1854.

16. Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as a surrogate for

immunoglobulin-variable-region mutations in chronic lymphocytic leukemia. The New England

journal of medicine. 2003;348(18):1764-1775.

17. Rassenti LZ, Jain S, Keating MJ, et al. Relative value of ZAP-70, CD38, and

immunoglobulin mutation status in predicting aggressive disease in chronic lymphocytic

leukemia. Blood. 2008;112(5):1923-1930.

18. Wiestner A, Rosenwald A, Barry TS, et al. ZAP-70 expression identifies a chronic

lymphocytic leukemia subtype with unmutated immunoglobulin genes, inferior clinical outcome,

and distinct gene expression profile. Blood. 2003;101(12):4944-4951.

20
19. Durig J, Nuckel H, Cremer M, et al. ZAP-70 expression is a prognostic factor in chronic

lymphocytic leukemia. Leukemia. 2003;17(12):2426-2434.

20. Ibrahim S, Keating M, Do KA, et al. CD38 expression as an important prognostic factor

in B-cell chronic lymphocytic leukemia. Blood. 2001;98(1):181-186.

21. Claus R, Lucas DM, Ruppert AS, et al. Validation of ZAP-70 methylation and its relative

significance in predicting outcome in chronic lymphocytic leukemia. Vol. 124; 2014.

22. Rituximab and bendamustine hydrochloride rai, or ibrutinib alone in treating older

patients with previously untreated chronic lymphocytic leukemia. Last accessed January 12,

2015; www.clinicaltrials.gov.

23. Shanafelt TD, Geyer SM, Bone ND, et al. CD49d expression is an independent predictor

of overall survival in patients with chronic lymphocytic leukaemia: a prognostic parameter with

therapeutic potential. British journal of haematology. 2008;140(5):537-546.

24. Bulian P, Shanafelt TD, Fegan C, et al. CD49d is the strongest flow cytometry-based

predictor of overall survival in chronic lymphocytic leukemia. Journal of clinical oncology :

official journal of the American Society of Clinical Oncology. 2014;32(9):897-904.

25. Juliusson G, Oscier DG, Fitchett M, et al. Prognostic subgroups in B-cell chronic

lymphocytic leukemia defined by specific chromosomal abnormalities. The New England

journal of medicine. 1990;323(11):720-724.

26. Döhner H, Stilgenbauer S, Benner A, et al. Genomic Aberrations and Survival in Chronic

Lymphocytic Leukemia. New England Journal of Medicine. 2000;343(26):1910-1916.

27. Shanafelt TD, Witzig TE, Fink SR, et al. Prospective evaluation of clonal evolution

during long-term follow-up of patients with untreated early-stage chronic lymphocytic leukemia.

21
Journal of clinical oncology : official journal of the American Society of Clinical Oncology.

2006;24(28):4634-4641.

28. Stilgenbauer S, Sander S, Bullinger L, et al. Clonal evolution in chronic lymphocytic

leukemia: acquisition of high-risk genomic aberrations associated with unmutated VH, resistance

to therapy, and short survival. Haematologica. 2007;92(9):1242-1245.

29. Janssens A, Van Roy N, Poppe B, et al. High-risk clonal evolution in chronic B-

lymphocytic leukemia: single-center interphase fluorescence in situ hybridization study and

review of the literature. European journal of haematology. 2012;89(1):72-80.

30. Tam CS, Shanafelt TD, Wierda WG, et al. De novo deletion 17p13.1 chronic

lymphocytic leukemia shows significant clinical heterogeneity: the M. D. Anderson and Mayo

Clinic experience. Vol. 114; 2009.

31. Zenz T, Krober A, Scherer K, et al. Monoallelic TP53 inactivation is associated with poor

prognosis in chronic lymphocytic leukemia: results from a detailed genetic characterization with

long-term follow-up. Blood. 2008;112(8):3322-3329.

32. Rossi D, Cerri M, Deambrogi C, et al. The prognostic value of TP53 mutations in chronic

lymphocytic leukemia is independent of Del17p13: implications for overall survival and

chemorefractoriness. Clinical cancer research : an official journal of the American Association

for Cancer Research. 2009;15(3):995-1004.

33. Zenz T, Eichhorst B, Busch R, et al. TP53 Mutation and Survival in Chronic

Lymphocytic Leukemia. Journal of Clinical Oncology. 2010;28(29):4473-4479.

34. Gonzalez D, Martinez P, Wade R, et al. Mutational status of the TP53 gene as a predictor

of response and survival in patients with chronic lymphocytic leukemia: results from the LRF

22
CLL4 trial. Journal of clinical oncology : official journal of the American Society of Clinical

Oncology. 2011;29(16):2223-2229.

35. Wang L, Lawrence MS, Wan Y, et al. SF3B1 and Other Novel Cancer Genes in Chronic

Lymphocytic Leukemia. New England Journal of Medicine. 2011;365(26):2497-2506.

36. Puente XS, Pinyol M, Quesada V, et al. Whole-genome sequencing identifies recurrent

mutations in chronic lymphocytic leukaemia. Nature. 2011;475(7354):101-105.

37. Rossi D, Rasi S, Fabbri G, et al. Mutations of NOTCH1 are an independent predictor of

survival in chronic lymphocytic leukemia. Blood. 2012;119(2):521-529.

38. Sportoletti P, Baldoni S, Cavalli L, et al. NOTCH1 PEST domain mutation is an adverse

prognostic factor in B-CLL. British journal of haematology. 2010;151(4):404-406.

39. Rossi D, Spina V, Bomben R, et al. Association between molecular lesions and specific

B-cell receptor subsets in chronic lymphocytic leukemia. Blood. 2013;121(24):4902-4905.

40. Stilgenbauer S, Schnaiter A, Paschka P, et al. Gene mutations and treatment outcome in

chronic lymphocytic leukemia: results from the CLL8 trial. Vol. 123; 2014.

41. Quesada V, Conde L, Villamor N, et al. Exome sequencing identifies recurrent mutations

of the splicing factor SF3B1 gene in chronic lymphocytic leukemia. Nature genetics.

2012;44(1):47-52.

42. Oscier DG, Rose-Zerilli MJ, Winkelmann N, et al. The clinical significance of NOTCH1

and SF3B1 mutations in the UK LRF CLL4 trial. Blood. 2013;121(3):468-475.

43. Braggio E, Kay NE, VanWier S, et al. Longitudinal genome-wide analysis of patients

with chronic lymphocytic leukemia reveals complex evolution of clonal architecture at disease

progression and at the time of relapse. Leukemia. 2012;26(7):1698-1701.

23
44. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal mutations in

chronic lymphocytic leukemia. Cell. 2013;152(4):714-726.

45. Rossi D, Khiabanian H, Spina V, et al. Clinical impact of small TP53 mutated subclones

in chronic lymphocytic leukemia. Vol. 123; 2014.

46. Calin GA, Ferracin M, Cimmino A, et al. A MicroRNA signature associated with

prognosis and progression in chronic lymphocytic leukemia. The New England journal of

medicine. 2005;353(17):1793-1801.

47. Zenz T, Mohr J, Eldering E, et al. miR-34a as part of the resistance network in chronic

lymphocytic leukemia. Blood. 2009;113(16):3801-3808.

48. Ferrajoli A, Shanafelt TD, Ivan C, et al. Prognostic value of miR-155 in individuals with

monoclonal B-cell lymphocytosis and patients with B chronic lymphocytic leukemia. Blood.

2013;122(11):1891-1899.

49. Mraz M, Chen L, Rassenti LZ, et al. miR-150 influences B-cell receptor signaling in

chronic lymphocytic leukemia by regulating expression of GAB1 and FOXP1. Blood.

2014;124(1):84-95.

50. Visone R, Veronese A, Rassenti LZ, et al. miR-181b is a biomarker of disease

progression in chronic lymphocytic leukemia. Vol. 118; 2011.

51. Fulci V, Chiaretti S, Goldoni M, et al. Quantitative technologies establish a novel

microRNA profile of chronic lymphocytic leukemia. Vol. 109; 2007.

52. Ling H, Fabbri M, Calin GA. MicroRNAs and other non-coding RNAs as targets for

anticancer drug development. Nature reviews Drug discovery. 2013;12(11):847-865.

24
53. Stamatopoulos K, Belessi C, Moreno C, et al. Over 20% of patients with chronic

lymphocytic leukemia carry stereotyped receptors: Pathogenetic implications and clinical

correlations. Blood. 2007;109(1):259-270.

54. Agathangelidis A, Darzentas N, Hadzidimitriou A, et al. Stereotyped B-cell receptors in

one-third of chronic lymphocytic leukemia: a molecular classification with implications for

targeted therapies. Blood. 2012;119(19):4467-4475.

55. Agathangelidis A, Darzentas N, Hadzidimitriou A, et al. Stereotyped B-cell receptors in

one-third of chronic lymphocytic leukemia: a molecular classification with implications for

targeted therapies. Vol. 119; 2012.

56. Agathangelidis A, Vardi A, Baliakas P, Stamatopoulos K. Stereotyped B-cell receptors in

chronic lymphocytic leukemia. Leukemia & lymphoma. 2014;55(10):2252-2261.

57. Murray F, Darzentas N, Hadzidimitriou A, et al. Stereotyped patterns of somatic

hypermutation in subsets of patients with chronic lymphocytic leukemia: implications for the

role of antigen selection in leukemogenesis. Blood. 2008;111(3):1524-1533.

58. Parikh SA, Rabe KG, Call TG, et al. Diffuse large B-cell lymphoma (Richter syndrome)

in patients with chronic lymphocytic leukaemia (CLL): a cohort study of newly diagnosed

patients. British journal of haematology. 2013;162(6):774-782.

59. Strefford JC, Sutton LA, Baliakas P, et al. Distinct patterns of novel gene mutations in

poor-prognostic stereotyped subsets of chronic lymphocytic leukemia: the case of SF3B1 and

subset #2. Leukemia. 2013;27(11):2196-2199.

60. Al-Harbi S, Hill BT, Mazumder S, et al. An antiapoptotic BCL-2 family expression index

predicts the response of chronic lymphocytic leukemia to ABT-737. Blood. 2011;118(13):3579-

3590.

25
61. Heintel D, Kienle D, Shehata M, et al. High expression of lipoprotein lipase in poor risk

B-cell chronic lymphocytic leukemia. Leukemia. 2005;19(7):1216-1223.

62. Maloum K, Settegrana C, Chapiro E, et al. IGHV gene mutational status and

LPL/ADAM29 gene expression as clinical outcome predictors in CLL patients in remission

following treatment with oral fludarabine plus cyclophosphamide. Annals of hematology.

2009;88(12):1215-1221.

63. Oppezzo P, Vasconcelos Y, Settegrana C, et al. The LPL/ADAM29 expression ratio is a

novel prognosis indicator in chronic lymphocytic leukemia. Blood. 2005;106(2):650-657.

64. Lin TT, Norris K, Heppel NH, et al. Telomere dysfunction accurately predicts clinical

outcome in chronic lymphocytic leukaemia, even in patients with early stage disease. British

journal of haematology. 2014;167(2):214-223.

65. Josefsson P, Geisler CH, Leffers H, et al. CLLU1 expression analysis adds prognostic

information to risk prediction in chronic lymphocytic leukemia. Blood. 2007;109(11):4973-4979.

66. Ghosh AK, Secreto CR, Knox TR, Ding W, Mukhopadhyay D, Kay NE. Circulating

microvesicles in B-cell chronic lymphocytic leukemia can stimulate marrow stromal cells:

implications for disease progression. Blood. 2010;115(9):1755-1764.

67. Molica S, Vitelli G, Levato D, Gandolfo GM, Liso V. Increased serum levels of vascular

endothelial growth factor predict risk of progression in early B-cell chronic lymphocytic

leukaemia. British journal of haematology. 1999;107(3):605-610.

68. Shanafelt TD, Jenkins G, Call TG, et al. Validation of a new prognostic index for patients

with chronic lymphocytic leukemia. Cancer. 2009;115(2):363-372.

26
69. Krober A, Seiler T, Benner A, et al. V(H) mutation status, CD38 expression level,

genomic aberrations, and survival in chronic lymphocytic leukemia. Blood. 2002;100(4):1410-

1416.

70. Hosmer D LS ed Applied Logistic Regression (ed 2nd). New York, NY: John Wiley &

Sons; 2000.

71. Hallek M, Fischer K, Fingerle-Rowson G, et al. Addition of rituximab to fludarabine and

cyclophosphamide in patients with chronic lymphocytic leukaemia: a randomised, open-label,

phase 3 trial. Lancet. 2010;376(9747):1164-1174.

72. Badoux XC, Keating MJ, Wang X, et al. Fludarabine, cyclophosphamide, and rituximab

chemoimmunotherapy is highly effective treatment for relapsed patients with CLL. Vol. 117;

2011.

73. O'Brien S, Jones JA, Coutre S, et al. Efficacy and Safety of Ibrutinib in Patients with

Relapsed or Refractory Chronic Lymphocytic Leukemia or Small Lymphocytic Leukemia with

17p Deletion: Results from the Phase II RESONATE™-17 Trial. Vol. 124; 2014.

74. Rawstron AC, Villamor N, Ritgen M, et al. International standardized approach for flow

cytometric residual disease monitoring in chronic lymphocytic leukaemia. Leukemia.

2007;21(5):956-964.

75. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of

chronic lymphocytic leukemia: a report from the International Workshop on Chronic

Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines.

Blood. 2008;111(12):5446-5456.

76. Strati P, Keating MJ, O'Brien SM, et al. Eradication of bone marrow minimal residual

disease may prompt early treatment discontinuation in CLL. Vol. 123; 2014.

27
77. Böttcher S, Ritgen M, Fischer K, et al. Minimal Residual Disease Quantification Is an

Independent Predictor of Progression-Free and Overall Survival in Chronic Lymphocytic

Leukemia: A Multivariate Analysis From the Randomized GCLLSG CLL8 Trial. Journal of

Clinical Oncology. 2012;30(9):980-988.

78. Tam CS, O'Brien S, Wierda W, et al. Long-term results of the fludarabine,

cyclophosphamide, and rituximab regimen as initial therapy of chronic lymphocytic leukemia.

Blood. 2008;112(4):975-980.

Table 1: Median survival of newly diagnosed CLL patients seen at Mayo Clinic by Rai

Stage (1995-present)

Rai Stage Characteristic Median survival, Median survival,

original publication by Mayo Clinic CLL
Rai et al,3 in months database*, in months
(n=125) (n=2559)
0 Lymphocytosis only 150 143
I Lymphadenopathy 101 125
II Organomegaly 71 100
III Anemia 19 57
IV Thrombocytopenia 19 63

*All patients with newly diagnosed CLL seen in the Division of Hematology from 1995-
present

28
Table 2: Prognostic Nomogram for Overall Survival (Adapted from Wierda et al8)

Characteristic Point contribution

0 1 2 3
Age, years - <50 50-65 >65
β2M, mg/L <ULN 1-2 x ULN >2 x ULN -
ALC, x109/L <20 20-50 >50 -
Sex Female Male - -
Rai Stage 0-II III-IV - -
No. of involved nodal groups ≤2 3 - -

The prognostic score is calculated by adding the risk score for each variable. An index score of
1-3 indicates low-risk, 4-7, intermediate risk; and ≥8, high-risk.

Abbreviations Used:
β2M: beta-2 microglobulin; ALC: absolute B-cell count; ULN: upper limit of normal

Table 3: CLL Prognostic Index for Overall Survival (Adapted from Pflug et al10)

Characteristic Risk Score

FISH del17p 6
Serum TK >10 U/L 2
Serum β2M >3.5 mg/dL 2
Serum β2M >1.7 and ≤3.5 mg/dL 1
IGHV mutation status Unmutated 1
ECOG PS >0 1
FISH del11q 1
Age, years >60 1
Sex Male 1

The composite risk score is obtained by adding the risk score for each variable. Based on the
composite score, four risk groups were identified: low (score 0-2), intermediate (score 3-5), high
(score 6-10) and very high risk (score ≥11).

Abbreviations:
FISH: fluorescence in situ hybridization; TK: thymidine kinase; β2M: beta-2 microglobulin; IGHV:
immunoglobulin heavy chain variable region; ECOG: Eastern Co-operative Oncology Group

29
Figure 1A: Overall survival of previously untreated CLL patients seen at Mayo Clinic
according to IGHV mutation status (1995-present), n=1519
Overall survival: IGHV, untreated
100

75
Survival (%)

50
p<0.001

25

Mutated
Unmutated
0
0 5 10 15 20
Years since diagnosis

Mutated 856 481 164 36 0

Unmutated 663 301 61 6 0

30
Figure 1B: Overall survival of previously untreated CLL patients seen at Mayo Clinic
according to CD38 expression (1995-present, CD38 expression of ≥30% considered
positive), n=2371
Overall survival: CD38, untreated
100

75
Survival (%)

50
p<0.001

25

Negative
Positive
0
0 5 10 15 20
Years since diagnosis

Negative 1722 864 294 50 0

Positive 649 320 85 9 0

31
Figure 1C: Overall survival of previously untreated CLL patients seen at Mayo Clinic
according to ZAP-70 expression (1995-present, ZAP-70 expression of ≥20% considered
positive), n=1816
Overall survival: ZAP-70, untreated
100

75
Survival (%)

50
p<0.001

25

Negative
Positive
0
0 5 10 15 20
Years since diagnosis

Negative 1150 542 157 33 0

Positive 666 282 67 12 0

32
Figure 1D: Overall survival of previously untreated CLL patients seen at Mayo Clinic
according to CD49d expression (1995-present, expression of ≥30% considered
positive), n=1476

Overall survival: CD49d, untreated

100

75
Survival (%)

50
p<0.001

25

Negative
Positive
0
0 5 10 15 20
Years since diagnosis

Negative 1047 467 129 30 0

Positive 429 166 25 3 0

33
Figure 1E: Overall survival of previously untreated CLL patients seen at Mayo Clinic
according to FISH obtained within 36 months of diagnosis (1995-present, hierarchical
FISH model), n=1639

Overall survival (years since diagnosis): FISH, untreated within 36 months of dx

100

75
Survival (%)

50
p<0.001

25
Normal
13q-
trisomy 12
11q-
17p-
0
0 5 10 15 20
Normal 435 186 56 7 0
.
13q- 663 291 59 6 0
trisomy 12 301 148 17 1 0
11q- 163 59 10 1 0
17p- 77 19 3 0 0

34