Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Gene expression profiling in multiple myeloma – reporting of
entities, risk and targets in clinical routine
Tobias Meißner,1,2 Anja Seckinger,1,2 Thierry Rème,3,4,5 Thomas Hielscher,6 Thomas Möhler,1
Kai Neben,1 Hartmut Goldschmidt,1,2 Bernard Klein,3,4,5 and Dirk Hose1,2
1
Medizinische Klinik V, Universitätsklinikum Heidelberg, Heidelberg, Germany; 2Nationales
Centrum für Tumorerkrankungen, Heidelberg, Germany; 3INSERM, U1040, Montpellier, France;
4
CHU Montpellier, Institute of Research in Biotherapy, Montpellier, France;
Montpellier 1, UFR Médecine, Montpellier, France;
6
5
Université
Abteilung für Biostatistik, Deutsches
Krebsforschungszentrum Heidelberg, Heidelberg, Germany
Running title:
GEP-reporting in myeloma
Key words:
Gene expression profiling, multiple myeloma, molecular entities, risk adapted
treatment, targeted therapy
This work was supported in part by grants from the Hopp-Foundation, Germany, the University of
Heidelberg, Germany, the National Center for Tumor Diseases, Heidelberg, Germany, the Deutsche
Krebshilfe, Bonn, Germany, the Deutsche Forschungsgemeinschaft Bonn, Germany, the Ligue
Nationale Contre Le Cancer, Paris, France.
Correspondence:
Dr.
med.
Dipl.
phys.
Dirk
Hose,
MD,
Medizinische
Klinik
V,
Universitätsklinikum Heidelberg, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany;
e-mail: [email protected]; phone: +49 6221 56 6140; FAX: +49 6221 56 6238
Conflict-of-interest disclosure: The authors declare no conflict of interest.
Word count:
Statement of Translational Relevance: 147
Abstract: 242
Main Text: 437 + 670 + 1.104 + 918 = 3.129
Figures/Tables: 2/0 and supplementary data
References: 50
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Statement of Translational Relevance
Multiple myeloma is an example for a cancer entity characterized by a pronounced molecular
heterogeneity, transmitting into survival times ranging from months to over 15 years. Gene
expression profiling is an ideal tool to assess this heterogeneity in terms of gene expression based
biological classifications and prognosis and allows assessment of target genes. However, it is
almost never used prospectively in clinical routine, mainly due to the lack of an academic reporting
tool giving quality controlled, validated, and clinically digestible information.
We present here a gene expression based report (GEP-R). The GEP-R is a customizable opensource software that can be adapted to novel parameters or other cancer entities. It includes
molecular classification, risk stratification, and assessment of target gene expression. Furthermore,
the HM-metascore combines validated conventional and expression based risk factors into one
superior metascore, thus helping to overcome the confusing and increasing multitude of prognostic
factors.
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Abstract
Purpose
Multiple myeloma is an incurable malignant plasma cell disease characterized by survival
ranging from several months to over 15 years. Assessment of risk and underlying molecular
heterogeneity can be excellently performed by gene expression profiling (GEP), but its way
into clinical routine is hampered by the lack of an appropriate reporting tool and the
integration with other prognostic factors into a single “meta” risk stratification.
Experimental Design
The GEP-report (GEP-R) was built as an open-source software developed in R for gene
expression reporting in clinical practise using Affymetrix microarrays. GEP-R processes new
samples by applying a documentation-by-value strategy to the raw-data to be able to assign
thresholds and grouping-algorithms defined on a reference cohort of 262 multiple myeloma
patients. Furthermore, we integrated expression-based and conventional prognostic factors
within one risk-stratification (HM-metascore).
Results
The GEP-R comprises i) quality control, ii) sample identity control, iii) biological
classification, iv) risk stratification, and v) assessment of target genes. The resulting
HM-metascore is defined as the sum over the weighted factors gene-expression based riskassessment (UAMS-, IFM-score), proliferation, ISS-stage, t(4;14), and expression of
prognostic target-genes (AURKA, IGF1R) for which clinical grade inhibitors exist. The
HM-score delineates three significantly different groups of 13.1, 72.1 and 14.7% of patients
with a 6-year survival of 89.3, 60.6 and 18.6%, respectively.
Conclusion
GEP reporting allows prospective assessment of risk and target gene expression and
integration of current prognostic factors in clinical routine, being customizable regarding
novel parameters or other cancer entities.
3
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Introduction
Multiple myeloma is an incurable malignant disease of clonal plasma cells which accumulate
in the bone marrow causing clinical signs and symptoms related to the displacement of
normal hematopoiesis, bone marrow neovascularization, formation of osteolytic bone lesions,
and production of monoclonal protein (1, 2). On a molecular level, multiple myeloma is
characterized by a pronounced heterogeneity in terms of chromosomal aberrations and altered
gene expression (2-5). This molecular heterogeneity is thought to transmit into survival
ranging from months to 15 years or more (6-8). Within this time, almost all myeloma patients
relapse, necessitating a careful planning of subsequent treatment regimen, and an appropriate
selection of (further) compounds to be used, e.g. as an individualized or add-on-treatment.
Gene expression profiling (GEP) is an ideal tool to assess this molecular heterogeneity in
terms of gene expression based classifications of distinct biological entities (9-13).
Importantly, prognostic groups with highly different survival have been defined, namely using
the gene expression based risk scores of the University of Arkansas for Medical Sciences
(UAMS) and the Intergroup Francophone du Myélome (IFM) as well as a gene expression
based proliferation index (GPI) (14-16). Moreover, gene expression profiling allows the
assessment of target gene expression, e.g. aurora kinase A (AURKA), fibroblast growth factor
receptor 3 (FGFR3), insulin like growth factor 1 receptor (IGF1R) (3, 5, 9), for which
clinical inhibitors are available (e.g. VX-680, CHIR-258, NVP-AEW541) (3, 5, 17) and
potential targets of vaccination strategies (e.g. MAGE3) (18, 19). GEP can currently be
performed in about 80% of therapy requiring patients in terms of available material, i.e.
purified myeloma cells. It has been performed on large cohorts of patients within clinical
trials, but except in a very limited number of cancer centers, only retrospectively. This is
surprising in that the prerequisites, i.e. purification of malignant plasma cells, can be
performed at most academic institutions (20). The same holds true for gene expression
profiling using core facilities or commercial service providers. Thus, the main difficulty
4
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
remaining for prospective use of GEP in clinical routine is an affordable (i.e. academic)
reporting tool giving quality controlled, validated, and clinically digestible information that
can be used by clinicians without extensive bioinformatics training. We present here such a
gene expression based report (GEP-R) including molecular classification, risk stratification
and assessment of target gene expression. Within the GEP-R, an implemented meta-score
integrates current expression based and conventional prognostic factors into one superior
prognostic classification. The GEP-R is a non-commercial software framework adaptable to
other cancer entities that can be downloaded at http://code.google.com/p/gep-r. A “LiveDVD”
for
testing
without
the
need
to
install
the
program
is
available
at
https://gliserv.montp.inserm.fr (see Results for details).
5
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Materials and Methods
Software design. The GEP-R software is developed within the open source software
environments R (21) and Bioconductor (22) using a graphical user interface (GUI; see
Supplementary Fig. S1) based on the gWidgets package (23). It processes a single Affymetrix
U133 Plus 2.0 .CEL-file using a “documentation by value (docval)-strategy” (24) modified
for GC-RMA preprocessing (available at http://code.google.com/p/gep-r). This allows using
preprocessing information of our reference cohort of 262 quality controlled myeloma samples
(E-MTAB-372) and thus generated thresholds to be applied to a subsequent microarray, a
prerequisite for prospective use of a gene expression report based on GC-RMA preprocessed
data. A detailed description of the GUI and how to use it is provided in the Supplementary
Materials and Methods.
Quality control. QC implemented is based on the Bioconductor packages affyQCReport (25),
affyPLM (26-28), simpleaffy and yaqcaffy (29) adapted for GC-RMA preprocessing and
PANP-based (30) assessment of presence/absence of gene expression. To limit computation
time, six samples from the reference cohort have been chosen randomly as QC-reference.
Sample validation. Using prediction analysis for microarrays (PAM; (31)) predictors for sex,
the IgL- (lambda and kappa) and IgH-type (IgA, IgG, IgD) have been implemented using one
part of the reference cohort as training set and the other one as test set. For further validation,
the UAMS total therapy 2 (TT2-) cohort (GSE24080) and the Multiple Myeloma Research
Consortium data (MMRC; www.broadistitute.org/mmgp) were used.
Risk stratification. Two gene expression based risk scores and a risk delineating GPI are
implemented: i) The IFM 15-gene model (15) developed on one-channel DNA Unitè Mixte de
Génomique du Cancer (UMGC) microarrays transferred to respective genes represented on
Affymetrix U133 Plus 2.0 arrays. Cutoffs for low/high risk have been generated on the
reference cohort defining the 25% of patients with the highest risk score as high risk, showing
comparable results to the original publication (15). ii) The UAMS 70-gene-risk-score was
6
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
calculated as published on MAS5 preprocessed data using published cutoffs (14) iii) The
GPI(16) with cutoff-values for low/medium/high risk was re-generated on the reference
cohort. The attribution of the TT2-cohort to the respective risk stratification can be found in
Supplementary Table S1.
Assessment of gene expression. Expression height and presence/absence of expression by
PANP (30) is assessed for i) potential “target genes” for whose products clinical grade
inhibitors exist (currently AURKA, FGFR3 and IGF1R), ii) potential vaccination targets, and
iii) genes frequently aberrantly (i.e. not expressed in normal plasma cells) or differentially
expressed in myeloma (i.e. expression value greater/lower than the respective value from
normal plasma cells ± three standard deviations). In the report’s appendix, values are put in
comparison to normal plasma cells (n = 10) and myeloma cell samples (n = 262;
Supplementary Fig. S2).
Gene expression based classification of myeloma. Three gene expression based
classifications of multiple myeloma are assessed: i) The TC-classification by Bergsagel et al.
(10, 11) derived on Hu95Av2 DNA-microarrays (Affymetrix) using MAS5-preprocession
was implemented on U133 2.0 (Affymetrix) DNA-microarrays as described by
Chng et al. (32). ii) The molecular classification of multiple myeloma (9). We first performed
an unsupervised clustering of the reference cohort based on the 700 genes published (being de
facto 687 genes). The resulting 7 clusters could be attributed to the 7 groups of the molecular
classification. A “myeloid group” could not be delineated (see Results). A PAM-based
predictor using these 687 genes was calculated on 162 patients of the reference cohort and
validated on 100 patients, respectively. iii) The EC-classification (12) was re-implemented on
the reference cohort and assessed by a PAM predictor consisting of 188 genes. The attribution
of the TT2-patients to the respective gene expression based classifications can be found in
Supplementary Table S1.
7
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Chromosomal aberrations - presence of t(4;14). The presence of a translocation t(4;14) has
been assessed by iFISH as previously published (16) and a PAM-based predictor using parts
of the reference cohort as training set and the other part as test set has been generated
(36 genes).
Calculation of the HM-metascore. The calculation of the HM-metascore is described in the
Results section (see below).
8
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Results
GEP-R. After input of patient information (e.g. sex, IgH/L-type), the GEP-R software
performs MAS5 and GC-RMA preprocessing as detailed above and subsequently applies
i) quality control of gene expression data, ii) sample validation, iii) gene expression based
classification, iv) risk stratification, and v) target assessment. The physician preparing the
report comments respective findings within the report in an editable text field. A PDF-file
containing the final report is generated by the software using the Sweave package (33). The
GEP-R consists of two parts: i) the report given to the treating physician (Fig. 1), and ii) an
appendix containing detailed quality control and validation information, as well as details of
the assessment of gene expression (Supplementary Fig. S2). Analyzing a .CEL-file with
GEP-R takes about 20 minutes on a standard computer (2+ GB RAM required). The GEP-R
runs on Linux, MacOS and Windows systems. Optionally, results and gene expression data
can be written to a PostgreSQL database using PostgreSQL through pgUtils (34) and maDB
(35) packages, provided a running postgreSQL database. For metascore-generation (HMscore, see below), the patient’s ISS-stage needs to be entered.
Description of the graphical user interface. The GEP-R graphical user interface (GUI) is
divided into two parts (Supplementary Fig. S1). Left Part. Within the left part of the GUI, the
user 1) loads an Affymetrix U133 Plus 2.0 raw-file into the software, starts the analysis, can
save or load previous analyses, and creates the GEP-R as a PDF-document. The user enters
2) patient and 3) sample specific data (e.g. patient name, IgH-/IgL-type) necessary for
identification and validation. Afterwards, the .CEL file is analyzed by the GUI calling a Rscript performing MAS5 and GC-RMA preprocessing using preprocessed information of the
reference cohort. The user here 4) comments the parameters within the report and 5) includes
a concluding statement. Right part. Here, the results of the analysis are presented in three
tabs for 1) parameters analyzed (e.g. risk scores), 2) quality control parameters and plots, and
3) identity control parameters that appear if entered patient data do not match predicted values
9
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
(e.g. IgH-type), or a data field misses an entry (Supplementary Fig. S2-3). A detailed
description of the GUI and how to use it is included in the Supplementary Materials and
Methods.
Exemplary report. An example of a GEP-R (final PDF-based patient report given to
clinicians and patients) is depicted in Fig. 1, the full report including the appendix in
Supplementary Fig. S2.
Quality control. Quality control (QC)-metrics assessed are average background, percent
present calls as detected by the PANP-algorithm, 3‘, mid and 5‘ expression of β-actin and
GAPDH. Six plots allow the assessment of reproducibility, QC-statistics, spike-in
performance, normalized unscaled standard error (NUSE) and relative log expression (RLE),
pseudo images, e.g. allowing the detection of artifacts, and RNA degradation (Supplementary
Fig. S3). Supplementary Table S2 depicts samples from the initial cohort of 280 patients
failing these criteria. In all, 18 samples (6.4%) failed in at least one criterion and were
excluded. One criterion was failed by 10 (3.5%), two criteria by additional 4 (1.4%) and three
or more by four (1.4%) further samples, respectively.
Validation. The GEP-R software compares predicted sex, IgH- and IgL-type with user
entered patient information and gives a warning in case of inconsistency. For prediction errors
see Supplementary Table S3. Predictors were validated on the UAMS TT2 cohort and the
MMRC data set (Supplementary Table S3).
Risk stratification. Using the GEP-R based assessment, the gene expression based risk
scores of the UAMS and IFM significantly delineate a high-risk population of 29.2% and
22.9% of myeloma patients, respectively. The same holds true for the GPI significantly
delineating 6.3%, 47.9% and 45.8% of patients as high, intermediate and low risk
(Supplementary Fig. S4). The risk stratification according to the UMAS 70-gene score for the
UAMS patients is available as Supplementary Table S1.
10
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Assessment of target gene expression. On our reference cohort, the assessed target genes
AURKA, IGF1R and FGFR3 are expressed in 29.3, 28.2 and 13.4% of myeloma samples,
respectively. Expression of each gene is significantly associated with adverse survival on our
cohort (Supplementary Fig. S4). For the reporting, see Supplementary Fig. S1-2.
Gene expression based classification of myeloma. The TC-classification is implemented as
suggested by the authors (32) (see above). Grouping is validated by published data of the
TT2-dataset (Supplementary Table S1). The EC-classification implemented as stated above
can be predicted with an overall error rate of 3% (Supplementary Table S3B). In contrast,
grouping according to the “molecular classification” proposes a problem as there is no clear
delineation algorithm of the so-called “myeloid group” published. Nevertheless, the molecular
classification (7 groups, without the “myeloid group”) can be predicted with an overall error
rate of 11%. For details, see Supplementary Table S3B. The overlap of our prediction with the
grouping according to Zhan et al. (9) is visualized in Supplementary Table S1.
Chromosomal aberrations - presence of t(4;14). The presence of a translocation t(4;14) can
be assessed using a PAM-based predictor without error (Supplementary Table S3C). Its
expression is significantly associated with adverse prognosis (Supplementary Fig. S4).
Building a meta-score including current prognostic information. To integrate depicted
expression-based and clinical prognostic factors within one prognostic (meta-) score (HMscore), a weight of 0, 0.5 or 1 is given for each of the named prognostic factors. The HMscore is calculated as the sum over these weights for gene-expression based assessment of risk
(0, 0.5, 1 for none, one, or both of the UAMS- and IFM-scores depicting high risk,
respectively), proliferation (0, 0.5, 1 for GPI low, median, or high, respectively),
ISS (0, 0.5, 1 for ISS-stage 1, 2, and 3, respectively), t(4;14) (0 for absence, 1 for presence of
the aberration), and expression of prognostically relevant target genes (0 or 1 if none or at
least one of the genes AURKA or IGF1R is expressed, respectively). The model delineates
three significantly different groups of 13.1, 72.1 and 14.7% of patients with a HM-score of 0
11
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
(none of the risk-factors), 0.5 - 3, and >3, and a 6-year survival of 89.3, 60.6 and 18.6% of our
patients (HM-cohort), respectively. It is likewise prognostic for event-free survival (Fig. 2).
The HM-score is validated using published information on the respective parameters on the
TT2-cohort (Fig. 2) (3, 5, 16).
Downloadable version and files. The GEP-R software is open-source and can be
downloaded at http://code.google.com/p/gep-r. The source code can be freely modified and
customized to user requirements. The GEP-R is currently validated for use of Affymetrix
gene expression data obtained by double amplification protocol. A “LiveDVD” for direct
testing of the report containing three exemplary *.CEL-files of myeloma patients is available
at https://gliserv.montp.inserm.fr (user: gepr, password: review).
12
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Discussion
In this article, we introduce a reporting tool (GEP-R) allowing automated interpretation of
Affymetrix U133 Plus 2.0 gene expression profiles. It enables the prospective use of GEP
within clinical routine in terms of molecular classification, risk stratification, and assessment
of target genes. This represents the critical first step in a process that will eventually allow for
reproducible approaches and a standardization of the analyses across different centers as well
as for personalized treatment.
In 2009, the International Myeloma Working group stated: “Our group also recognizes that
the greatest prognostic ability for multiple myeloma resides in the comprehensive analysis of
GEP. At a minimum, all clinical trials should consider incorporation of GEP into the
correlative science studies to identify subgroups of high-risk disease. We also propose that
methodology to include GEP into the clinical testing is urgently needed and methods for
implementation should be identified.” (16). In the following, we discuss use, interest in, and
clinical applicability of GEP and the GEP-R, respectively.
Three gene expression based classifications delineate molecular groups in myeloma: The
“molecular classification” based on differential gene expression in which three of seven
groups (“proliferation”, MAF-expression, MMSET-overexpression) show different survival
(9); the TC-classification based on translocations and D-type cyclin without prognostic
relevance (10, 11) and the EC-classification based on chromosomal aberrations and resulting
changes in gene expression with only one of four groups (t(4;14) and FGFR3-expression)
showing adverse prognosis (12, 13). Biological classifications likely remain relatively stable
in contrast to prognostic factors prone to change with different treatment schedules (see
below).
The translocation t(4;14) represents a specific disease entity and is an independent risk factor
despite conventional or high-dose treatment (36-41). Treatment with bortezomib or
13
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
lenalidomide containing regimen seems to at least lessen the negative prognostic impact of
this aberration (6, 42-45) and can thus be selected. t(4;14) is present in 13% of patients in our
reference cohort and can be predicted without error by GEP (Supplementary Table S2C).
Gene expression based risk stratification is applied using four different strategies:
i) To investigate differences in survival between molecular groups (see above).
ii) To surrogate biological variables associated with prognosis, e.g. here proliferation as
strong adverse prognostic factor in myeloma (46-49) by GPI (16) (Supplementary Fig. S4).
iii) To build a score over a set of genes associated with survival, exemplified by the high-risk
scores of the UAMS (70 genes) and the IFM (15 genes) (14, 15). Both scores allowed
delineating a group of patients (13% and 25%, respectively) with very adverse prognosis in
the IFM and total therapy 2-dataset (not including bortezomib), whereas in the total therapy 3cohort (including bortezomib) only the UAMS-score remains significant (14, 15). In relapsed
patients treated with bortezomib within the APEX, SUMMIT and CREST trial (n = 188), both
scores significantly delineate different outcome, in patients treated with dexamethasone
(n = 76), only the UAMS score. In our cohort of patients, 32.4% and 25.2% of patients are
delineated to be of “high risk”. As implemented in the GEP-R, two groups of patients with
significantly different survival can be delineated on our cohort (Supplementary Fig. S4).
iv) To assess the presence of expression of a gene representing a potential target and
investigate its prognostic relevance, exemplified by AURKA (3), IGF1R (5) and FGFR3 (12,
13) in the current version of the GEP-R (Supplementary Fig. S4). As these genes are only
expressed in 13 - 28% of patients, they are ideal targets for personalized treatment, e.g. addon for current induction treatment regimen, and respective clinical trials are in preparation.
GEP likewise allows the screening for a large number of immunotherapeutic targets, e.g.
cancer testis antigens (18, 50), exemplified by MAGEA3 (18, 19). In both cases GER-R easily
allows assessment of expression of such target genes thus enabling personalized treatment.
14
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
A sufficient number of malignant plasma cells for GEP at a purity of 80%, i.e. 5x 104 - 105
cells if a double amplification protocol and Affymetrix U133 2.0 arrays are used, can be
obtained in about 80% of therapy requiring myeloma patients at most academic institutions
(20), as can be GEP. As global GEP allows simultaneous assessment of (almost) all genes
without any need for pre-selection, novel risk stratifications and target genes as well as
predictors for response can be applied using existing expression data and easily be included in
the GEP-R software framework.
A pressing practical problem for clinical risk assessment is how to integrate the various risk
factors into a single clear message given to physicians and patients. The HM-score here
combines main current expression-based and conventional prognostic factors and
prognostically relevant targetable genes reported in the GEP-R delineating three groups of
patients with highly significantly different event-free and overall survival (Fig. 2).
Therefore, the GEP-R here allows overcoming the main difficulty remaining for prospective
use of GEP in clinical routine: an affordable reporting tool giving quality controlled,
validated, and clinically digestible information that can be used by clinicians without
extensive bioinformatics training. It is used by the Universities of Heidelberg and Montpellier
in clinical routine as well as within the GMMG-MM5 multicenter trial of the GermanSpeaking Myeloma Multicenter group (GMMG) and will be validated prospectively in 600
evaluable patients within a multicenter trial funded by the Federal Ministry of Education and
Research.
In conclusion, gene expression profiling using GEP-R allows prospective reporting of
molecular classifications, risk, and therapeutic targets to clinicians in clinical routine in a
digestible manner, and thus represents a major step in translational oncology. It will foster
performing clinical trials using risk adapted and personalized treatment strategies.
15
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Disclosure of Potential Conflicts of Interest and Authorship
Conflict-of-interest disclosure: The authors declare no conflict of interest.
Contribution: T. Meißner performed programming and creation of the GEP-R and participated
in writing of the paper, A.S. participated in designing the research and writing of the paper;
T.R. participated in the creation of the GEP-R; T.H. participated in statistical analysis and the
modification of the docval-package; T. Möhler, N.K., B.K. and H.G. participated in the
analyzing of the data and in the writing of the paper, D.H. designed research, wrote the paper,
and participated in creating the GEP-R.
Grant Support
This work was supported in part by grants from the Hopp-Foundation, Germany, the
University of Heidelberg, Germany, the National Center for Tumor Diseases, Heidelberg,
Germany, the Deutsche Krebshilfe, Bonn, Germany, the Deutsche Forschungsgemeinschaft
Bonn, Germany, the Ligue Nationale Contre Le Cancer, Paris, France.
16
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
References
1.
Kyle RA, Rajkumar SV. Multiple myeloma. N Engl J Med. 2004;351:1860-73.
2.
Hose D, Moreaux J, Meissner T, Seckinger A, Goldschmidt H, Benner A, et al.
Induction of angiogenesis by normal and malignant plasma cells. Blood. 2009;114:128-43.
3.
Hose D, Rème T, Meissner T, Moreaux J, Seckinger A, Lewis J, et al. Inhibition of
aurora kinases for tailored risk-adapted treatment of multiple myeloma. Blood.
2009;113:4331-40.
4.
Seckinger A, Meissner T, Moreaux J, Goldschmidt H, Fuhler GM, Benner A, et al.
Bone morphogenic protein 6: a member of a novel class of prognostic factors expressed by
normal and malignant plasma cells inhibiting proliferation and angiogenesis. Oncogene.
2009;28:3866-79.
5.
Sprynski AC, Hose D, Caillot L, Reme T, Shaughnessy JD, Jr., Barlogie B, et al. The
role of IGF-1 as a major growth factor for myeloma cell lines and the prognostic relevance of
the expression of its receptor. Blood. 2009;113:4614-26.
6.
Barlogie B, Anaissie EJ, van Rhee F, Shaughnessy JD, Haessler J, Pineda-Roman M,
et al. Total therapy (TT) for myeloma (MM)--10% cure rate with TT1 suggested by >10yr
continuous complete remission (CCR): Bortezomib in TT3 overcomes poor-risk associated
with T(4;14) and DelTP53 in TT2. ASCO Meeting Abstracts. 2008;26:8516.
7.
Barlogie B, Tricot GJ, van Rhee F, Angtuaco E, Walker R, Epstein J, et al. Long-term
outcome results of the first tandem autotransplant trial for multiple myeloma. Br J Haematol.
2006;135:158-64.
8.
Harousseau JL, Moreau P. Autologous hematopoietic stem-cell transplantation for
multiple myeloma. N Engl J Med. 2009;360:2645-54.
9.
Zhan F, Huang Y, Colla S, Stewart JP, Hanamura I, Gupta S, et al. The molecular
classification of multiple myeloma. Blood. 2006;108:2020-8.
10.
Bergsagel PL, Kuehl WM. Molecular pathogenesis and a consequent classification of
multiple myeloma. J Clin Oncol. 2005;23:6333-8.
11.
Bergsagel PL, Kuehl WM, Zhan F, Sawyer J, Barlogie B, Shaughnessy J. Cyclin D
dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood.
2005;106:296-303.
12.
Hose D, Rossi J-Fo, Ittrich C, DeVos J, RŠme T, Benner A, et al. Molecular
Classification of Multiple Myeloma (MM) Based on Gene Expression Profiling (GEP) and
Fluorescence In Situ Hybridisation (FISH) Is an Independent Predictor for Event Free
Survival (EFS). [abstract]. Blood. 2005;106:150a-a.
13.
Hose D, Rossi J-F, Ittrich C, de Vos J, Reme T, Benner A, et al. A New Molecular
Classification Of Multiple Myeloma (MM) Using Gene Expression Profiling (GEP) And
Fluorescence In Situ Hybridisation As Predictor For Event Free Survival (EFS). [abstract].
Blood. 2004;104:25a-a.
14.
Shaughnessy JD, Zhan F, Burington BE, Huang Y, Colla S, Hanamura I, et al. A
validated gene expression model of high-risk multiple myeloma is defined by deregulated
expression of genes mapping to chromosome 1. Blood. 2007;109:2276-84.
15.
Decaux O, Lodé L, Magrangeas F, Charbonnel C, Gouraud W, Jézéquel P, et al.
Prediction of survival in multiple myeloma based on gene expression profiles reveals cell
cycle and chromosomal instability signatures in high-risk patients and hyperdiploid signatures
in low-risk patients: a study of the Intergroupe Francophone du Myélome. J Clin Oncol.
2008;26:4798-805.
16.
Hose D, Reme T, Hielscher T, Moreaux J, Messner T, Seckinger A, et al. Proliferation
is a central independent prognostic factor and target for personalized and risk-adapted
treatment in multiple myeloma. Haematologica. 2011;96:87-95.
17
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
17.
Trudel S, Li ZH, Wei E, Wiesmann M, Chang H, Chen C, et al. CHIR-258, a novel,
multitargeted tyrosine kinase inhibitor for the potential treatment of t(4;14) multiple
myeloma. Blood. 2005;105:2941-8.
18.
Condomines M, Hose D, Raynaud P, Hundemer M, Vos JD, Baudard M, et al.
Cancer/testis genes in multiple myeloma: expression patterns and prognosis value determined
by microarray analysis. J Immunol. 2007;178:3307-15.
19.
Szmania S, Gnjatic S, Tricot G, Stone K, Zhan F, Moreno A, et al. Immunization with
a recombinant MAGE-A3 protein after high-dose therapy for myeloma. J Immunother.
2007;30:847-54.
20.
Broyl A, Hose D, Lokhorst H, de Knegt Y, Peeters J, Jauch A, et al. Gene expression
profiling for molecular classification of multiple myeloma in newly diagnosed patients.
Blood. 2010.
21.
Team RDC. R: A Language and Environment for Statistical Computing; 2008.
22.
Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, et al.
Bioconductor: open software development for computational biology and bioinformatics.
Genome biology. 2004;5:R80.
23.
Verzani. gWidgets: gWidgets API for building toolkit-independent, interactive GUIs
2009.
24.
Kostka D, Spang R. Microarray based diagnosis profits from better documentation of
gene expression signatures. PLoS Comput Biol. 2008;4:e22.
25.
Parman, Halling, Gentleman. QC Report Generation for affyBatch objects.
26.
Bolstad B. Low Level Analysis of High-density Oligonucleotide Array Data:
Background, Normalization and Summarization. University of California, Berkeley. 2004.
27.
Bolstad B, Collin F, Brettschneider J, Simpson K, Cope L, Irizarry R, et al. Quality
Assessment of Affymetrix GeneChip Data in Bioinformatics and Computational Biology
Solutions Using R and Bioconductor. Gentleman R, Carey V, Huber W, Irizarry R, and
Dudoit S (Eds), Springer, New York. 2005.
28.
Brettschneider J, Collin F, Bolstad B, Speed TP. Quality Assessment for Short
Oligonucleotide Microarray Data. Technometrics 2008;50:279-83
29.
Wilson CL, Miller CJ. Simpleaffy: a BioConductor package for Affymetrix Quality
Control and data analysis. Bioinformatics. 2005;21:3683-5.
30.
Warren P, Taylor D, Martini PGV, Jackson J, Bienkowska J. PANP - a New Method
of Gene Detection on Oligonucleotide Expression Arrays.
Bioinformatics and
Bioengineering, 2007 BIBE 2007 Proceedings of the 7th IEEE International Conference on;
2007 14-17 Oct. 2007; 2007. p. 108-15.
31.
Tibshirani R, Hastie T, Narasimhan B, Chu G. Diagnosis of multiple cancer types by
shrunken centroids of gene expression. Proceedings of the National Academy of Sciences of
the United States of America. 2002;99:6567-72.
32.
Chng WJ, Glebov O, Bergsagel PL, Kuehl WM. Genetic events in the pathogenesis of
multiple myeloma. Best Pract Res Clin Haematol. 2007;20:571-96.
33.
Leisch F. Sweave: Dynamic generation of statistical reports using literate data
analysis. Compstat 2002 - Proceedings in Computational Statistics, in W Härdle und B Rönz,
Hrsg, Physica Verlag, Heidelberg. 2002:575-80.
34.
Rainer J. pgUtils: Utility functions for PostgreSQL databases.
35.
Rainer J. maDB: Microarray database and utility functions for microarray data
analysis.
36.
Fonseca R, Bergsagel PL, Drach J, Shaughnessy J, Gutierrez N, Stewart AK, et al.
International Myeloma Working Group molecular classification of multiple myeloma:
spotlight review. Leukemia. 2009;23:2210-21.
18
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
37.
Avet-Loiseau H, Attal M, Moreau P, Charbonnel C, Garban F, Hulin C, et al. Genetic
abnormalities and survival in multiple myeloma: the experience of the Intergroupe
Francophone du Myelome. Blood. 2007;109:3489-95.
38.
Chiecchio L, Protheroe RK, Ibrahim AH, Cheung KL, Rudduck C, Dagrada GP, et al.
Deletion of chromosome 13 detected by conventional cytogenetics is a critical prognostic
factor in myeloma. Leukemia : official journal of the Leukemia Society of America,
Leukemia Research Fund, UK. 2006;20:1610-7.
39.
Keats JJ, Reiman T, Maxwell CA, Taylor BJ, Larratt LM, Mant MJ, et al. In multiple
myeloma, t(4;14)(p16;q32) is an adverse prognostic factor irrespective of FGFR3 expression.
Blood. 2003;101:1520-9.
40.
Fonseca R, Blood E, Rue M, Harrington D, Oken MM, Kyle RA, et al. Clinical and
biologic implications of recurrent genomic aberrations in myeloma. Blood. 2003;101:456975.
41.
Moreau P, Facon T, Leleu X, Morineau N, Huyghe P, Harousseau J-L, et al. Recurrent
14q32 translocations determine the prognosis of multiple myeloma, especially in patients
receiving intensive chemotherapy. Blood. 2002;100:1579-83.
42.
Chang H, Sloan S, Li D, Zhuang L, Yi Q-L, Chen CI, et al. The t(4;14) is associated
with poor prognosis in myeloma patients undergoing autologous stem cell transplant. Br J
Haematol. 2004;125:64-8.
43.
San Miguel JF, Schlag R, Khuageva NK, Dimopoulos MA, Shpilberg O, Kropff M, et
al. Bortezomib plus melphalan and prednisone for initial treatment of multiple myeloma.
NEnglJMed. 2008;359:906-17.
44.
Avet Loiseau H, Moreau P, Mathiot C, Charbonnel C, Caillot D, Facon T, et al.
Induction with Velcade(R)/Dexamethasone Partially Overcomes the Poor Prognosis of
t(4;14), but Not That of Del(17p), in Young Patients with Multiple Myeloma. ASH Annual
Meeting Abstracts. 2009;114:957-.
45.
Knop S, Gerecke C, Liebisch P, Topp MS, Platzbecker U, Sezer O, et al.
Lenalidomide, adriamycin, and dexamethasone (RAD) in patients with relapsed and
refractory multiple myeloma: a report from the German Myeloma Study Group DSMM
(Deutsche Studiengruppe Multiples Myelom). Blood. 2009;113:4137-43.
46.
Reece D, Song KW, Fu T, Roland B, Chang H, Horsman DE, et al. Influence of
cytogenetics in patients with relapsed or refractory multiple myeloma treated with
lenalidomide plus dexamethasone: adverse effect of deletion 17p13. Blood. 2009;114:522-5.
47.
Boccadoro M, Gavarotti P, Fossati G, Pileri A, Marmont F, Neretto G, et al. Low
plasma cell 3(H) thymidine incorporation in monoclonal gammopathy of undetermined
significance (MGUS), smouldering myeloma and remission phase myeloma: a reliable
indicator of patients not requiring therapy. British journal of haematology. 1984;58:689-96.
48.
Greipp PR, Katzmann JA, O'Fallon WM, Kyle RA. Value of beta 2-microglobulin
level and plasma cell labeling indices as prognostic factors in patients with newly diagnosed
myeloma. Blood. 1988;72:219-23.
49.
San Miguel JF, Garcia-Sanz R, Gonzalez M, Moro MJ, Hernandez JM, Ortega F, et al.
A new staging system for multiple myeloma based on the number of S-phase plasma cells.
Blood. 1995;85:448-55.
50.
Condomines M, Hose D, Rème T, Requirand G, Hundemer M, Schoenhals M, et al.
Gene expression profiling and real-time PCR analyses identify novel potential cancer-testis
antigens in multiple myeloma. J Immunol. 2009;183:832-40.
19
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Meißner et al.
GEP-REPORTING IN MYELOMA
Legends
Figure 1. Gene expression report (GEP-R) as given to referring physicians. For the
complete report including validation-, and quality control-metrics as well as detailed depiction
of gene expression, see Supplementary Fig. S2.
Figure 2. Prognostic value of the HM-metascore. HM-metascore combining gene
expression based high risk scores, assessment of proliferation, ISS-stage, t(4;14), and
expression of prognostically relevant target genes (AURKA and IGF1R). (A) Significant
delineation of event-free survival within our (HM-cohort) and the TT2-cohort. (HM-cohort,
P<0.001, low versus medium P=0.02, medium versus high P<0.001, TT2-cohort, P<0.001,
low versus medium P<0.001, medium versus high P<0.001, respectively). (B) Significant
delineation of overall survival (HM-cohort, P<0.001, low versus medium P=0.03, medium
versus high P<0.001, TT2-cohort, P<0.001, low versus medium P=0.006, medium versus high
P<0.001, respectively).
20
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Figure 1.
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
Figure 2.
low risk
medium risk
high risk
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.
low risk
medium risk
high risk
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
low risk
r
medium risk
high risk
low risk
medium risk
high risk
P<0.001
P<0.001
O SH-group
OS
OS HM-group
B
P<0 001
P<0.001
P<0 001
P<0.001
E
EFS
SH-group
EFS HM-group
A
Author Manuscript Published OnlineFirst on October 10, 2011; DOI: 10.1158/1078-0432.CCR-11-1628
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
Gene expression profiling in multiple myeloma - reporting of
entities, risk and targets in clinical routine
Tobias Meißner, Anja Seckinger, Thierry Rème, et al.
Clin Cancer Res Published OnlineFirst October 10, 2011.
Updated version
Supplementary
Material
Author
Manuscript
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
doi:10.1158/1078-0432.CCR-11-1628
Access the most recent supplemental material at:
http://clincancerres.aacrjournals.org/content/suppl/2011/10/10/1078-0432.CCR-11-1628.DC1
Author manuscripts have been peer reviewed and accepted for publication but have not yet been
edited.
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from clincancerres.aacrjournals.org on June 15, 2017. © 2011 American Association for Cancer
Research.