| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
9
www.eprecisionmed.com/pmo
Received 8 July 2016; Revised 10 July 2016; Accepted 13 July 2016; Published 14 July 2016
*Corresponding author: [email protected], Tel.: +(52) 5553501970; Fax: +(52) 5553501990
THE PATH TO INTEGRATION IN COMPUTATIONAL
ONCOLOGY
Guillermo de AndaJauregui,´ Enrique Hernandez´Lemus* Computational Genomics Division, National Institute of Genomic Medicine, MEXICO´
ARTICLE
_______________________________
HIGHLIGHTS
This article provides a detail account of informatics and
computational biology and its application in precision oncology.
_______________________________
ABSTRACT
The complexity of cancer makes precision medicine a most
sought after goal. With each new technological
advancement a plethora of data is gathered that moves us
closer to it. However, integration of these efforts is still
needed. Computational power is needed in order to process,
manage, and integrate this information, in order to generate
models that are useful in the clinical setting. The following is
a review of the current computational oncology approaches
and its challenges.
KEY WORDS
Cancer, Precision Medicine, Computational Oncology, Systems Biology, Integrative Genomics
www.precision-medicine.in
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
10
INTRODUCTION The challenges for establishing precision medicine in oncology
In the years that follow the completion of the first draft of the human genome, high throughput
technologies gave rise to the idea that therapeutic options could be tailor made, based on the individual
characteristics of each patient at the genomic level. It took, however, more than a dozen years until the
Precision Medicine Initiative was able to brought into public attention the possibilities of incorporating
these technologies to current clinical settings [1]. Precision medicine (PM) –as it is presently understood–
should aim at being able to improve prediction, treatment, and outcomes.
Cancer, being an archetypal complex disease, is an obvious candidate for PM efforts due to its highly
heterogeneous, individual wise character. Cancer events involve multiple alterations at different levels:
molecular, tissue, and systemic. Each of these levels present its own challenges to study, many of them
related to inter-individual variability. PM in oncology intends to pave the way to the high level personalized
diagnostic, prognosis and treatment of cancers.
Of course, achieving success in this area is not an easy endeavour. Here we will consider some of the
challenges that contemporary research in oncology has to face and also the way in which several
disciplines have been developing strategies to cope with these. The key to success, in our opinion, is
integration of the strengths and capabilities of all the different approaches.
Genetic and environmental factors in cancer
The genetic contribution to cancer has been known for a long time. Since the discovery of the role of
oncogenes and tumor suppressors in the development of cancer [2–5], many genetic alterations have
been identified in different neoplasms [6], with genomic instability being now considered a hallmark of
cancer [7, 41]. Epigenetic effects in cancer [8–10] add another layer of complexity to the study of the
disease; however, the use of genetic information has led to success stories in both the diagnostic [11] and
treatment [12, 13] areas. Genetic cancer risk assessment and its use for therapeutic decision making is
also steadily gaining traction [14–18].
The contribution of environmental factors to carcinogenesis has also been known for a long time now [19–
23]. Gene-environment interactions are involved in the susceptibility to carcinogenic effects [24–27] and
have been a sought after subject of study [28–31]. In a similar fashion, genetic variants are associated to
differences in therapeutic susceptibility [32–35], again having an impact in clinical decision making [36–
39]. It is precisely this complex interplay between genomic susceptibility traits and environmental factors
which lies behind the poor results obtained by straightforward approaches to treat cancer. At the heart of
these interactions is the emergence of new phenomena giving malignant tumors a rather unpredictable
character.
Emerging factors in cancer
The study of individual factors in cancer has gradually given way to studies beyond the single molecule
level. Cancer has indeed been characterized as a pathway based disease [40, 41], with pathological
alterations being associated to deregulation in groups of molecules acting together. This approach has
been useful in bridging molecular data with functional information and clinical applications [42–46]. In a
similar way, integration of multiple molecular data types has led to increased efforts in studying the major
role that metabolic deregulation has in cancer. While the idea of energetic deregulation in cancer is not
new, dating back to Warburg [47], there is a renewed interest in metabolic phenomena in order to
understand the disease and perhaps, as a source of novel therapeutic targets [48–50].
Cancer is a molecular disease. However, it is now known that not only intracellular molecular deregulations
are involved in the disease, but a complex set of dynamic interactions between tumor cells and also
between the tumor and its surrounding microenvironment [51]. Cells of the immune system, adipocytes,
fibroblasts, vascular endothelial cells, among others with different tumor cells [52], include cancer stem
cells [53, 54]. Nowadays we recognize the need to understand not only the nature of tumors themselves,
but also their interplay with other cells.
The challenging path from knowledge to precision medicine, and the role of computational biology
We have witnessed a shift in the understanding of neoplastic diseases. The challenge now lies in being
able to incorporate each new finding in an intelligent, and most importantly, useful manner in order to
complete the much desired bench-to-bed goal of translational medicine. Computational oncology must
play two roles in order to accomplish this goal: To help collect, process, and compile the vast quantities of
data; and to develop and implement new models that combine the available information and eventually
become useful in the clinical decision making process.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
11
What follows is a brief description of the state of computational oncology, in the context of systems
biology. The two roles of computational oncology today are described, both as a data managing tool and as
a tool for modelling. We discuss the two main avenues for these research lines: the data-driven,
mathematical oncology approach, and the physical oncology approach. We present integrative frameworks
and current computational tools available, and finally we elaborate on what are the current areas in need
for further development of the discipline.
The systems biology paradigm
The Systems biology paradigm considers that complex phenomena behind the functions of biological
systems are the result of interactions between different constituting elements [55, 56]. Systems biology
attempts thus, to understand biological mechanisms behind phenotypes using large scale data and
mathematical models that use said data [57]. Cancer can be thought of as a systems biology disease [58,
59]. Nonlinear interactions between the many elements that are altered in cancer may drive the diseased
phenotype [60, 61]. The systems biology paradigm is useful then, in bringing together information in order
to generate both descriptive and predictive models [62].
Model generation is central to systems biology [63]. Biological, systems level models provide new insights
in the biological phenomena [64] something that is extremely important for oncology research. These
insights are useful in the generation of new hypotheses that may drive experimental research [65].
Furthermore, the contribution of systems biology must be understood as iterative [66]: models are
constructed based on data available, generating new hypotheses to be tested; new experiments are used
to refine the model, generating new hypotheses to be retested and so on.
The goals for the integrative, systems biology study of cancer should be directed at generating new models
regarding diagnostic and therapeutic choices, in order to gain improvements in prognosis, diagnostics and
ultimately therapeutics [67]. In order to do so, it is necessary to a) manage omics level data and b)
implement appropriate models based mathematical and physical principles using said data. It is evident
that for this, the use of computational approaches is mandatory.
COMPUTATIONAL ONCOLOGY Computational oncology is the study of cancer with computer-implemented tools from the quantitative
sciences [68]. As we have previously discussed, there are two main computational oncology branches as it
currently stands: first, the one related with processing, storage, retrieval and analysis of data generated
with analytical technologies; and second, the development of descriptive and predictive models that
incorporate said data, and are able to drive experimental research by generating new questions. More
importantly, computational oncology develops applications in the clinical setting in order to improve
diagnosis, treatment selection and prognosis.
In a similar fashion, computational oncology can be roughly divided on the basis of the disciplines from
which its applications are derived: Mathematical oncology originates in the mathematical and
computational sciences. It is cantered on the development and implementation of algorithms that enable
the analysis and management of biological data. Physical oncology comes from the application of ideas
derived from physical models to oncology problems.
Overlap between approaches exists, as this division is mostly descriptive, and in many cases the use of
language is more defined by previous training and self identification. Nonetheless, let us use this simple
classification in order to further discuss current uses of computational methods in oncology [Figure 1].
Mathematical Oncology
Mathematical oncology can be thought of as the use of mathematical techniques in order to extract
information from omics level datasets. Sufficient (that is, large) computational resources are needed in
order to be able to manage this data [69]. Some applications of mathematical oncology include database
management (that is, storage and retrieval of data), data processing, and data analysis. Let us examine
some of these applications in more detail.
Database management
Current large scale research deposits the data generated in readily accessible databases, with a rapid
growth in the number of databases and their sizes [70]. Several databases specialized in oncological data
exist, both from general studies (TCGA [71], COSMIC [72]) as well as related to particular types of cancer
(for instance, the METABRIC database [73]). Public repositories of experimental data, such as the Gene
Expression Omnibus and ArrayExpress [74] for microarray data are also largely used by researchers.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
12
As applications continue to be developed, the use of these platforms will be more common in the clinical
setting.
As technologies keep evolving, a major concern is the size of data. The difference in size between RNAseq
data and microarray data, for instance, is about an order of magnitude. Incorporating experimental data
from other levels of description (proteomics, metabolomics) directly increases the size of data. IT
requirements in the clinical setting shall then move forward to either a) the acquisition of large storage
media or b) the implementation of cloud-based storage, with dedicated infrastructure to rapidly and
efficiently access data.
Fig: 1. The roles of computational oncology. Analytical technologies generate large amounts of data that can only be
manipulated through information technologies. The development of tools for the processing, storage, retrieval and data analysis is one branch of computational oncology. Computational models can be developed, based on mathematical or physical principles, that uses said data in order to describe or predict the pathology. These models can be used to guide new highthroughput research. Finally, both data and models can be integrated into the clinical setting by using computational platforms. …………………………………………………………………………………………………………………………………………………
Imaging data processing and analysis
Imaging data has been for long, an important part of comprehensive cancer management, with a role in
screening, diagnosis, treatment selection and follow-up [75–77]. The traditional approach of expert
clinicians examining individual images is time consuming, and limited by the availability of qualified
professionals. Computer aided detection (CAD) techniques take advantage of image processing and
machine learning algorithms in order to assist said clinicians in data interpretation [78].
Computer aided interpretation of imaging data is currently used, for instance, in breast cancer screening
[79]. Routine mammography is used for breast cancer screening: in the United States, biennial screening
mammography is recommended for women age 50 to 74 years [80]. Studies have found evidence of
benefits in the use of CAD technologies, leading to increased diagnostic sensitivity [81–83]. A recent study
shows that the use of CAD for screening mammography increases the rate of detection of ductal
carcinoma in situ and early stage invasive cancer; however, it also shows an increased risk in false
positives [84]. The role of CAD for breast cancer screening has detractors, for instance [85]; however, the
authors of said work do consider that current limitations in CAD may be overcome with further research,
and describe the potential of CAD as ”limitless” [86]. Complementation between the use of new
technologies and the development of proper computational algorithms is needed in order to advance their
clinical relevance. Development of said algorithms is currently a productive area of research [87– 91]. Sequence data
Next generation sequencing (NGS) is expected to revolutionize the clinical setting [92, 93], allowing the
simultaneous determination of any genetic variant related to disease. Cancer is a disease that naturally
benefits from such approach, allowing to move away from the reductionist approach that former
technological limitations imposed [94] [Figure-2]. Currently, the main practical limitation for NGS in the
clinical setting is the cost. However, as the technology becomes more accessible, so will its uses [95, 96].
NGS has particular bioinformatics requirements. These have been described as primary, secondary and
tertiary analysis [97]. In primary analysis, instrument signals are processed in order to generate a
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
13
sequence (for instance, a FASTQ file). In secondary analysis, a comparison with a reference sequence is
made. Finally, tertiary analysis involves the contextualization of variant information in order to enhance
clinical decision making.
NGS methodologies have been used in order to identify potentially pathogenic variants in patients [98,
99]. A goal is to be able to use NGS information to identify genomic biomarkers [100]. There have been
successful efforts in this regard, with pipelines that are able to identify potentially actionable alterations
[101–103]. The identification of useful, reproducible genomic biomarkers poses a considerable technical
feat [104]. It also presents computational challenges: Biomarker reproducibility for instance, is highly
sensible to the particular processing pipeline. Furthermore, algorithms do not always take into account the
intra-tumor heterogeneity present in cancer. Data analysis pipelines should be improved to take into
account these factors in order to have better clinical value [105].
Fig: 2. Computational applications in the analysis of Next Generation Sequencing data. By its very nature, the analysis
of data generated by NGS technologies can only be achieved through the use of computational tools. The processing and storage of raw NGS data requires considerable computational resources. Later, NGS analysis pipelines need to incorporate algorithms for sequence alignment, variant calling and finally, association of sequence data to known biological functions. …………………………………………………………………………………………………………………………………………………
As with other genomic data, the functional annotation of identified biomarkers, and its public sharing, is
vital. Again, the development of databases [106] will allow clinicians the ability to work with this
information. Meanwhile, data oriented scientists may use the information and implement computer based
approaches (such as data mining) to identify new biomarkers [107], an example of the feedback expected
in the development of precision medicine.
An important consideration for the use of these technologies is the need of validation for the clinical
setting [108]. Current guidelines stress the importance not only of platform validation, but also of the
validation of bioinformatics pipelines used for data analysis [109]. Identifying driver and passenger genes
A current problem in cancer genomics is the identification of driver mutations from passenger mutations,
that is, to identify a mutation that is causally implicated in oncogenesis from one that does not confer
clonal growth advantages [110]. For this, computational oncology approaches are necessary, and
constitute indeed a central area of research. Discerning between driver and passenger status usually
takes into account a) mutation frequency, b) mutual exclusivity of mutations between gene sets, or c)
information derived from pathway or other biological network annotation [111, 112].
Development and refinement of the computational strategies to identify driver from passenger genes is
ongoing [113]. In [114], a framework based on network enrichment analysis is proposed, able to properly
identify driver genes in gliobastoma and ovarian carcinoma data. In [115] evolutionary population dynamic
models are used in order to find driver genes in colorectal tumors. Meanwhile, [112] uses a methodology
capable of identifying driver genes with low mutational frequency due to epistatic interactions.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
14
Expression data
Gene expression data has notably impacted cancer research. Microarray, and more recently, RNAseq data
have been thoroughly used in order to profile the expression of several oncological diseases [116–124].
Evidently, the analysis of these high throughput technologies requires suitable bioinformatics techniques
able to go from raw experimental data all the way to expression quantification. Computational pipelines
suitable for specific technologies are available [125, 126].
Gene expression patterns associated to disease manifestations have been codified into gene signatures.
These have been extensively used in the characterization of neoplastic diseases [127–131]. Collections of
these signatures are available, for instance in the MSIGdb and geneSigdb databases [132]. Gene
signatures are increasingly being implemented as clinical indicators of drug susceptibility [133–135].
Another common use of molecular signatures is as prognostic tools. There is evidence that a expression
signatures can be linked to clinical outcome [136]. However, there exists a degree of controversy
regarding their actual usefulness in the clinic, particularly whether there is a unique signature with
predictive power [137–139]. Nonetheless, commercial applications of gene signatures for the clinical
setting are currently available [140]. Computational oncologists have two roles in this regard: first, using
robust informatical methods to identify the validity of signatures in the research setting; and second, to
implement adequate computational platforms in order to use gene signature profiling in the context of
molecular testing [141].
A closely related subject and a major triumph of mathematical oncology is the identification, using data
driven approaches, of breast cancer molecular subtypes. Molecular subtypes are different manifestations
of a disease, with particular clinical and pathological characteristics, that can be identified based on gene
expression profiles. Molecular subtypes were first identified in breast cancer [142–144]. Soon, PAM50, a
molecular classification strategy (based on nearest shrunken centroids [145]) was developed in order to
classify samples [146]. Commercial applications [147] make this classification strategy widely available,
with validated, predictive clinical value [148, 149]. Following the initial success of breast cancer sub-
typification, molecular subtypes have been identified for other cancers [150, 151]
Finally, transcriptional networks can be inferred from whole genome expression experiments [60, 152].
These networks provide a data driven model of relationships between genes based on the statistic relation
between them. Transcriptional networks can be as simple as co-expression networks (ie. Spearman
correlations) or use more sophisticated, information theoretical measures such as mutual information (for
instance, the ARACNE algorithm [153]).
Physical Oncology Physics has a long standing relationship with oncology: imaging technologies and radiotherapy have been
part of the clinical setting for a long time now. However, in the computational setting, physical oncology is
oriented to the generation of integrative, mechanistic-driven models based on molecular cancer
information [154]. An advantage of this approach is that models are able to refine experimental work.
Furthermore, several problems in current oncological research can be mapped to equivalent physical
problems for which solutions have already been proposed: Problems such as cancer initiation/growth,
temporality, the role of microenvironment, metastases, etc. The role of quantitative physical oncology will
become increasingly important, as new technologies allow the measurement of quantitative parameters in
the preclinical and clinical setting [155].
Tumor growth models
The development of tumor growth models, based on quantifiable clinical observations is an important area
of research. The history of tumor growth modelling goes back to the early 20th century, with kinetic models
of cell proliferation [156]. Tumor growth can be thought of as a deregulated interaction between
environmental constraints and the cell program, and so models have moved forward by considering the
interactions between tumor, microenvironment and vascularization [157, 158].
Tumor growth modelling is generally divided in two classes: discrete and continuum modelling. Discrete
modelling uses experimentally derived rules to define the stepwise or discrete interactions between
individual cells and provide biological insights. In continuum modelling tumors are seen continuous tissue,
described in terms of morphology and nutrient distribution, using partial differential equations [159–161].
Hybrid models are gaining traction for the modelling of tumor growth. This systems couple continuous (ie.
reaction diffusion) and discrete (ie. Agent-based) formalisms for the description of tumor evolution [162].
The number of cells that these models can handle is inversely proportional to the detail of individual
cellular geometry [163]. The hybrid approach lends itself to multi-scale modelling: integrating data from
the subcellular, cellular, and tissue level in order to increase the relevance of the prediction.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
15
Tumor growth models may be exponential linear, Gompertzian, Bertalanffyan, etc. Independent of their
nature, tumor growth models shall be validated [164]; For instance, mathematical modelling [165]
identified unstable tumor morphologies associated to nutrient spatial variations in an in vitro spheroid
model. However, just as not all cancers conform to the same biological model [166], each mathematical
model may be valid only for a certain pathological manifestation [167].
Growth models are constantly developed [168–171] in which tumor morphologies are driven by nutrient
spatial variations. Particularly noticeable is recent work [172], which describes near criticality in a model of
early tumor growth. The authors of said work argue that this result supports proposed roles of cancer stem
cells in tumor growth.
Modeling drug efficacy in tumors
Models able to predict pharmacological response have a self-evident clinical importance. These models
need to take into account the cellular heterogeneity of tumors, interactions with the microenvironment,
interactions with immune system, reactions to therapy, and the rise of therapeutic resistance [173].
Pharmacokinetic/Pharmacodynamic (PK/PD) models have been used for the study of cancer drugs. Drug
concentration over time is coupled with response information in order to produce said models. The
simplest approach is a direct model, simply a pharmacokinetic model coupled to dose response equations
(for instance, agent-based models [174]). For drugs that do not exhibit an immediate effect, Indirect
models are generally more appropriate. [175]
Anticancer drugs generally act upon processes in a delayed time frame, making the use of special indirect
models necessary. Cytokinetic models consider the progression of cell cycle. This type of model is
particularly useful when studying agents that affect cells in a particular stage of the cycle: DNA damaging
agents, tubulin polymerisation inhibitors, apoptosis inducing drugs, etc. [175]. Drugs that act upon
signalling pathways, in the other hand, may benefit from a pathway-based model approach. Signalling
pathways are usually part of a large biological network with several interactions and complex mechanisms
of regulation, some of which may have impact on drug efficacy [176, 177]. Oncogenic singling networks
are complex and present features such as modularity, redundancy, adaptation, heterogeneity, which make
them robust (that is, resistant) to targeted perturbation [178]. Detailed pathway models [179] are useful
tools to model anticancer drug effects.
In the clinical setting PK/PD models can be used for therapeutic drug monitoring and dosing optimization
[180] in order to maximize therapeutic efficacy and reduce toxicity. A multi-scale computational model
[181] incorporates local drug, oxygen and nutrient concentrations within the three-dimensional tumor
volume to quantify the diffusion barrier effect on doxorubicin efficacy. Another use of coupled PK/PD
models is found in [182], where the effects of drugs on tumor progression in an experimental model of
tumor bearing mice are modelled using Gompertztype growth law and a pharmacokinetic
pharmacodynamic approach, and control algorithms are used to propose optimal therapeutic patterns. In
[183] a multi-scale model, with a set of stochastic differential equations to describe pharmacokinetics,
cellular dynamics, and progression free survival at the patient level, was able to identify different synergy
patterns in combination therapies. Intelligent dosing regime design remains an important factor for
anticancer therapeutic efficacy [173], and the successful implementation of rational, computer aided
systems may soon have a clinical impact.
Tumor metabolism models
Metabolism has been at the centre of many cancer studies recently [50, 184, 185]. It is natural to look for
ways to model metabolic reactions in cancer. Unlike other omics, the technology to generate large
metabolomic data is still emerging [186], but quickly evolving [187]. Metabolic cantered studies have
found, however, important results with potential clinical application [188, 189].
Biochemical processes can be modelled as serial reactions, with concentrations of reactants and products
related by reaction rates and equilibrium constants. These reactions, at the genome scale, can be
abstracted and represented as a metabolic reconstruction [190], using top-down (data-driven), bottom-up
(manually curated) or mixed approaches [191]. A stoichometric matrix is a mathematical object that
relates each chemical species with a particular reaction.
Two schools of metabolic modelling exist: kinetic modelling and constraint based modelling. Kinetic
modelling aims to characterize fully the mechanics of each enzymatic reaction. The major limitation of
these approaches is in the parameterization of mechanistic models, which is costly in resources and time
[192].
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
16
Constraint based modelling (CBM) is largely used in the context of metabolic analysis [193]. In a CBM,
feasible phenotypic states compose a solution space. These are limited based on assumptions on the
metabolic reconstruction, such as mass balance, energy balance, thermodynamics and flux limitations,
etc. This approach requires less experimental data, but such methods are unable to give insight into
cellular substrate concentrations [194]. Flux based models are constructed under a steady state
assumption, with constrains imposed by the metabolic network’s stoichiometry, thermodynamics, and
measured rates [195]. [196] Reconstruction of a generic metabolic network model of cancer cells. Model
is reconstructed by collecting the available data on tumor suppressor genes. Notably, we show that the
activation of oncogene related reactions can be explained by the inactivation of tumor suppressor genes.
Metabolic modelling is being used to identify molecules and reactions of importance to cancer cells. In
recent work [196], a constraint-based model was able to identify a relation between oncogene related
reactions and the inactivation of tumor suppressor genes. In [197] the role of five metabolic genes
susceptible to pharmacological modulation in clear cell renal carcinoma was identified. Another use of
constraint-based models is found in [198], to study a mutually beneficial coupling between aberrant and
non-aberrant cells through the recycling of metabolic products. In [199] use results of flux-based analysis
to support an origin in metabolic adaptation to the Warburg effect.
Fig: 3. Biological levels in computational oncology. Life (and dis-ease) is comprised of different description layers.
Therefore, computational oncology has applications in a range that comprises genetic information at the sequence level, gene expression and corresponding transcriptional networks, protein interactions and functional pathways, and the interactions between cells and tissues. In addition, Computational approaches to drug therapeutics involve the modeling of drug absorption, distribution, metabolism and excretion, as well as the molecular interactions of drugs with their therapeutic targets. …………………………………………………………………………………………………………………………………………………
Integrative frameworks
Integration in computational oncology must include both, conceptual and theoretical formalism to consider
all the available information at a glance in order to catch and unveil hidden interactions and to discover
emergent phenomena (say, for instance, hormone mediated drug resistance mechanisms), but also
integration of data in computational platforms by developing algorithms to deal with multilevel and multi-
scale data.
We have already talked about the former integration scheme. The goal in the latter (let us call it
bioinformatics) issue is to be able to connect all information of a patient: her/his electronic medical
records, biochemical tests, imaging materials, as well as the available genomic, proteomic and
metabolomic data into a single, easy to use computational platform. In this regard, there has been a lot of
development in programming languages and environments such as: biopython, bioPerl, Biolinux,
bioconductoR, and so on. Some of these approaches have been largely used in the biomedical resarch
setting but important efforts are being carried out to make their way to the translational and clinical
applications.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
17
Even more pertinent for the advancement of computational oncology than the above mentioned software
tools is the development of suites or even initiatives that may nourish on these algorithms to built
comprehensive strategies for the treatment of cancer [200] comprising the molecular, clinical, and even
social/epidemiological dimensions of the disease, firmly founded on the use of all the computational
resources already mentioned [201].
Ironically, the rise of personalized, precision medicine has to be built upon a large scale collective effort.
An effort that should comprise the collaboration of basic scientists, translational experts and clinicians
working side by side with quantitative researchers to create such all inclusive platforms.
CONCLUDING REMARKS
In the preceding review article we outlined some of the major challenges that biomedical and clinical
research in oncology face at the light of the nascent paradigm of personalized medicine. We also
mentioned how a number of quantitative disciplines are drawn together to face the computational side of
the issue (which is no minor part at all). There are a number of short term goals that have been
accomplished in the past few years.
However, the major obstacles consist not only in solving the individual parts of the cancer puzzle (a task
that is in fact daunting for most of these parts) but also how to put all the pieces of information together in
a coherent form (i.e. how to solve the puzzle) that at the same time is comprehensive enough to deal with
the gigantic heterogeneity that cancers have at the individual level (a goal of precision medicine) but also
is able to bring answers in times reasonable enough to be useful in the clinical setting [202] and is
compliant with ethic considerations regarding confidentiality (most of the proposed approaches to date
have a component of cloud-based solutions), and the differences in criteria for research and clinical use of
patient information. Even at the light of such an enormous task, what we have accomplished in the past
give us some reasons to be optimistic.
AUTHOR CONTRIBUTION GDJ and EHL write down the manuscript; GDJ designed the figures; EHL outlined the contents for the review article; both authors (GDJ and EHL) read and approved the submitted version of the manuscript.
CONFLICT OF INTEREST The authors declare no competing interests. ACKNOWLEDGEMENT The authors are grateful to the funding institutions for support provided. FINANCIAL DISCLOSURE The present research program has been funded by the National Council on Science and Technology (CONACYT) under grant number 179431/2012, as well as from federal funding provided by the National Institute of Genomic Medicine (both Mexican-based Institutions). The founders had no role in the design of the research.
REFERENCES
[1] G. Knudson, Mutation and cancer: statistical study of retinoblastoma, Proceedings of the National Academy of
Sciences 68 (4) (1971) 820–823.
[2] H. Oppermann, A. D. Levinson, H. E. Varmus, et al. Uninfected vertebrate cells contain a protein that is closely
related to the product of the avian sarcoma virus transforming gene (src), Proceedings of the National Academy
of Sciences 76 (4) (1979) 1804–1808.
[3] C. M. Croce, Oncogenes and cancer, New England Journal of Medicine 358 (5) (2008) 502–511.
[4] W. Sun, J. Yang, Functional mechanisms for human tumor suppressors, J Cancer 1 (2010) 136–140.
[5] P. A. Futreal, L. Coin, M. Marshall, et al. A census of human cancer genes, Nature Reviews Cancer 4 (3)
(2004) 177–183.
[6] S. Negrini, V. G. Gorgoulis, T. D. Halazonetis, Genomic instability an evolving hallmark of cancer, Nature reviews
Molecular cell biology 11 (3) (2010) 220–228.
[7] P. Feinberg, B. Tycko, The history of cancer epigenetics, Nature Reviews Cancer 4 (2) (2004) 143–153.
[8] P. Tung, P. Knoepfler, Epigenetic mechanisms of tumorigenicity manifesting in stem cells, Oncogene 34 (18)
(2015) 2288–2296.
[9] M. Montenegro, L. SanchezDelCampo, M. FernandezPerez, et al. Targeting the epigenetic machinery of cancer
cells, Oncogene 34 (2) (2015) 135–143.
[10] K. L. Smith, C. Isaacs, Brca mutation testing in determining breast cancer therapy, Cancer journal (Sudbury,
Mass.) 17 (6) (2011) 492.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
18
[11] B. J. Druker, S. Tamura, E. Buchdunger, et al. Effects of a selective inhibitor of the abl tyrosine kinase on the
growth of bcrabl positive cells, Nature medicine 2 (5) (1996) 561–566.
[12] D. J. Slamon, B. LeylandJones, S. Shak, et al. Use of chemotherapy plus a monoclonal antibody against her2 for
metastatic breast cancer that overexpresses her2, New England Journal of Medicine 344 (11) (2001) 783–
792.
[13] H. Dohner, V. I. Gaidzik, Impact of genetic features on treatment decisions in aml, ASH Education Program Book
2011 (1) (2011) 36–42.
[14] J. E. Dancey, P. L. Bedard, N. Onetto, et al. The genetic basis for cancer treatment decisions, Cell 148 (3)
(2012) 409– 420.
[15] S. J. Katz, A. W. Kurian, M. Morrow. Treatment decision making and genetic testing for breast cancer:
Mainstreaming mutations, JAMA 314 (10) (2015) 997–998.
[16] C. VillarrealGarza. Steps to address the profound disparity in access to genetic cancer risk assessment in
Mexico: Documenting hereditary breast and ovarian cancer and the beginning of a prevention program, Cancer
Epidemiology Biomarkers & Prevention 24 (10 Supplement) (2015)
http://dx.doi.org/10.1158/15387755.DISP14IA23
[17] K. R. Blazer, T. Slavin, J. N. Weitzel, Increased reach of genetic cancer risk assessment as a tool for precision
management of hereditary breast cancer, JAMA oncology.
[18] K. Hari, et al., Environment and cancer: Who are susceptible?, Cancer Res 49 (1989) 7057.
[19] F. P. Perera, Environment and cancer: who are susceptible?, Science 278 (5340) (1997) 1068–1073.
[20] L. A. Loeb, C. C. Harris. Advances in chemical carcinogenesis: a historical review and prospective, Cancer
research 68 (17) (2008) 6863–6872.
[21] C. B. Ambrosone, C. C. Harris. The development of molecular epidemiology to elucidate cancer risk and
prognosis: a historical perspective, International journal of molecular epidemiology and genetics 1 (2) (2010)
84.
[22] J. Hyndman. The contribution of both nature and nurture to carcinogenesis and progression in solid tumours,
Cancer Microenvironment 9 (1) (2016) 63–69.
[23] P. Brennan. Gene–environment interaction and aetiology of cancer: what does it mean and how can we
measure it?, Carcinogenesis 23 (3) (2002) 381–387.
[24] S. Nickels, T. Truong, R. Hein, et al. A. Vrieling, et al., Evidence of gene–environment interactions between
common breast cancer susceptibility loci and established environmental risk factors, PLoS Genet 9 (3) (2013)
e1003284.
[25] B. A. Hocevar, L. M. Kamendulis, X. Pu, et al. Contribution of environment and genetics to pancreatic cancer
susceptibility, PloS one 9 (3) (2014) e90052.
[26] A. Rudolph, J. ChangClaude, M. K. Schmidt, Gene–environment interaction and risk of breast cancer, British
journal of cancer.
[27] Liu, J. Huang, Y. Zhang, et al. Identification of gene–environment interactions in cancer studies using
penalization, Genomics 102 (4) (2013) 189–194.
[28] C. M. Hutter, L. E. Mechanic, N. Chatterjee, et al. Geneenvironment interactions in cancer epidemiology: A
national cancer institute think tank report, Genetic epidemiology 37 (7) (2013) 643–657.
[29] O. Fletcher, F. Dudbridge, Candidate geneenvironment interactions in breast cancer, BMC medicine 12 (1)
(2014) 1.
[30] N. I. Simonds, A. A. Ghazarian, C. B. Pimentel, et al. Review of the geneenvironment interaction literature in
cancer: What do we know?, Genetic Epidemiology.
[31] N. Bozina,ˇ V. Bradamante, M. Lovric.´ Genetic polymorphism of metabolic enzymes p450 (cyp) as a
susceptibility factor for drug response, toxicity, and cancer risk, Archives of Industrial Hygiene and Toxicology
60 (2) (2009) 217–242.
[32] M. H. Diekstra, J. J. Swen, E. Boven, et al. Cyp3a5 and abcb1 polymorphisms as predictors for sunitinib
outcome in metastatic renal cell carcinoma, European urology 68 (4) (2015) 621–629.
[33] W.J. Gong, J.Y. Yin, X.P. Li, et al. Association of wellcharacterized lung cancer lncrna polymorphisms with lung
cancer susceptibility and platinumbased chemotherapy response, Tumor Biology (2016) 1–10.
[34] Y. Zhang, S. D. Somtakoune, C. Cheung, et al. Therapeutic application of pharmacogenomics in oncology, The
AAPS Journal (2016) 1–11.
[35] B. ElKhoueiry, H. J. Lenz, The role of pharmacogenetics in making treatment decisions in colorectal cancer,
Current Colorectal Cancer Reports 2 (3) (2006) 109–115.
[36] K. K. Filipski, L. E. Mechanic, R. Long, et al. Pharmacogenomics in oncology care, Front. Genet 5 (73) (2014)
10– 3389.
[37] D. L. Hertz, J. Rae, Pharmacogenetics of cancer drugs, Annual review of medicine 66 (2015) 65–81.
[38] M. V. Relling, W. E. Evans, Pharmacogenomics in the clinic, Nature 526 (7573) (2015) 343–350.
[39] D. Hanahan, R. A. Weinberg, The hallmarks of cancer, cell 100 (1) (2000) 57–70.
[40] D. Hanahan, R. A. Weinberg, Hallmarks of cancer: the next generation, cell 144 (5) (2011) 646–674.
[41] M. L. Gatza, J. E. Lucas, W. T. Barry, et al. A pathwaybased classification of human breast cancer, Proceedings
of the National Academy of Sciences 107 (15) (2010) 6994– 6999.
[42] Y. Drier, M. Sheffer, E. Domany. Pathwaybased personalized analysis of cancer, Proceedings of the National
Academy of Sciences 110 (16) (2013) 6388–6393.
[43] Y.H. Chang, C. M. Chen, H. Y. Chen, et al. Pathwaybased gene signatures predicting clinical outcome of lung
adenocarcinoma, Scientific reports 5.
[44] D. Lee, G. K. Lee, K. A. Yoon, et al. Pathwaybased analysis using genomewide association data from a korean
nonsmall cell lung cancer study, PloS one 8 (6) (2013) e65396.
[45] S. Kim, M. Kon, C. DeLisi, et al., Pathwaybased classification of cancer subtypes, Biol Direct 7 (1) (2012) 21.
[46] O. Warburg, et al., On the origin of cancer cells, Science 123 (3191) (1956) 309–314.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
19
[47] P. Hainaut, A. Plymoth, Cancer as a metabolic disease, Current opinion in oncology 24 (1) (2012) 56–57.
[48] T. N. Seyfried, R. E. Flores, A. M. Poff, et al. Cancer as a metabolic disease: implications for novel therapeutics,
Carcinogenesis 35 (3) (2014) 515–527.
[49] D. S. Wishart, Is cancer a genetic disease or a metabolic disease?, EBioMedicine 2 (6) (2015) 478–479.
[50] F. Chen, X. Zhuang, L. Lin, et al. New horizons in tumor microenvironment biology: challenges and
opportunities, BMC medicine 13 (1) (2015) 1.
[51] F. R. Balkwill, M. Capasso, T. Hagemann, The tumor microenvironment at a glance, Journal of cell science 125
(23) (2012) 5591–5596.
[52] Albini, A. Bruno, C. Gallo, et al. Cancer stem cells and the tumor microenvironment: interplay in tumor
heterogeneity, Connective tissue research 56 (5) (2015) 414–425.
[53] K. Kise, Y. KinugasaKatayama, N. Takakura, Tumor microenvironment for cancer stem cells, Advanced drug
delivery reviews.
[54] E. Hernandez´Lemus. Data integration, systems approach and multilevel description of complex biosystems, in:
Journal of Physics: Conference Series, Vol. 475, IOP Publishing, 2013, p. 012006.
[55] E. Hernandez´Lemus, W. Li, P. Meyer, Advances in systems biologynew trends and perspectives, Computational
Biology and Chemistry 59 (PB) (2015) 1–2.
[56] C. E. H. Patino, G. JaimeMunoz, O. ResendisAntonio, Systems biology of cancer: moving toward the integrative
study of the metabolic alterations in cancer cells, Frontiers in physiology 3 (2012) 481.
[57] J. J. Hornberg, F. J. Bruggeman, H. V. Westerhoff, et al. Cancer: a systems biology disease, Biosystems 83 (2)
(2006) 81–90.
[58] J. Gentles, D. Gallahan, Systems biology: confronting the complexity of cancer, Cancer research 71 (18) (2011)
5961–5964.
[59] E. Hernandez´Lemus. Further steps toward functional systems biology of cancer, Frontiers in physiology 4
(2013) 256.
[60] E. Hernandez´Lemus. Systems biology and integrative omics in breast cancer, in: Omics Approaches in Breast
Cancer, Springer, 2014, pp. 333–352.
[61] L. Hood, J. R. Heath, M. E. Phelps, et al. Systems biology and new technologies enable predictive and
preventative medicine, Science 306 (5696) (2004) 640–643.
[62] Schmid, L. M. Blank, Systems biology: Hypothesisdriven omics integration, Nature chemical biology 6 (7) (2010)
485–487.
[63] R. A. Hillmer, Systems biology for biologists, PLoS Pathog 11 (5) (2015) e1004786.
[64] F. Villaverde, J. R. Banga. Reverse engineering and identification in systems biology: strategies, perspectives
and challenges, Journal of The Royal Society Interface 11 (91) (2014) 20130505.
[65] H. M. Werner, G. B. Mills, P. T. Ram. Cancer systems biology: a peek into the future of patient care?, Nature
reviews clinical oncology 11 (3) (2014) 167–176.
[66] W. Du, O. Elemento. Cancer systems biology: embracing complexity to develop better anticancer therapeutic
strategies, Oncogene 34 (25) (2015) 3215–3225.
[67] T. Lefor. Computational oncology, Japanese journal of clinical oncology 41 (8) (2011) 937–947.
[68] A. H. Chauviere, H. Hatzikirou, J. S. Lowengrub, et al. Mathematical oncology: how are the mathematical and
physical sciences contributing to the war on breast cancer?, Current breast cancer reports 2 (3) (2010) 121–
129.
[69] Pavlopoulou, D. A. Spandidos, I. Michalopoulos, Human cancer databases (review), Oncology reports 33 (1)
(2015) 3–18.
[70] The cancer genome atlas. URL http://cancergenome.nih.gov/
[71] S. A. Forbes, D. Beare, P. Gunasekaran, et al.Cosmic: exploring the world’s knowledge of somatic mutations in
human cancer, Nucleic acids research 43 (D1) (2015) D805– D811.
[72] C. Curtis, S. P. Shah, S. F. Chin, et al. The genomic and transcriptomic architecture of 2,000 breast tumours
reveals novel subgroups, Nature 486 (7403) (2012) 346– 352.
[73] N. Kolesnikov, E. Hastings, M. Keays, et al. Arrayexpress updates implifying data submissions, Nucleic acids
research (2014) gku1057.
[74] L. Fass, Imaging and cancer: a review, Molecular oncology 2 (2) (2008) 115–152.
[75] T. E. Yankeelov, D. A. Mankoff, L. H. Schwartz, et al. Quantitative imaging in cancer clinical trials, Clinical
Cancer Research 22 (2) (2016) 284– 290.
[76] C. Hunter, G. Brown, Preoperative staging of rectal cancer: A review of imaging techniques, Expert review of
gastroenterology & hepatology (justaccepted).
[77] H. Lee, Y. P. P. Chen, Image based computer aided diagnosis system for cancer detection, Expert Systems with
Applications 42 (12) (2015) 5356–5365.
[78] Redman, S. Lowes, A. Leaver, Imaging techniques in breast cancer, Surgery (Oxford) 34 (1) (2016) 818.
doi:10.1016/j.mpsur.2015.10.004.
[79] H. Nelson, A. Cantor, L. Humphrey, R et al. Screening for breast cancer: A systematic review to update the 2009
us preventive services task force recommendation.
[80] J. C. Dean, C. C. Ilvento, Improved cancer detection using computeraided detection with diagnostic and
screening mammography: prospective study of 104 cancers, American Journal of Roentgenology 187 (1)
(2006) 20–28.
[81] M. Gromet, Comparison of computeraided detection to double reading of screening mammograms: review of
231,221 mammograms, American Journal of Roentgenology 190 (4) (2008) 854– 859.
[82] A. Jalalian, S. B. Mashohor, H. R. Mahmud, et al. Computeraided detection/diagnosis of breast cancer in
mammography and ultrasound: a review, Clinical imaging 37 (3) (2013) 420–426.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
20
[83] J. J. Fenton, G. Xing, J. G. Elmore, et al. Shortterm outcomes of screening mammography using computeraided
detection: a populationbased study of medicare enrollees, Annals of internal medicine 158 (8) (2013) 580–
587.
[84] C. D. Lehman, R. D. Wellman, D. S. Buist, et al. Diagnostic accuracy of digital screening mammography with and
without computeraided detection, JAMA internal medicine (2015) 1–10.
[85] C. Lehman, Potential benefits of computeraided detection for cancer identification and treatmentreply, JAMA
internal medicine 176 (3) (2016) 411–411.
[86] D. O. T. Bruno, M. Z. do Nascimento, R. P. Ramos, et al. Lbp operators on curvelet coefficients as an algorithm
to describe texture in breast cancer tissues, Expert Systems with Applications 55 (2016) 329–340.
[87] M. U. Dalmıs, A. GubernMerida, S. Vreemann, et al. A computeraided diagnosis system for breast dcemri at
high spatiotemporal resolution, Medical physics 43 (1) (2016) 84–94.
[88] GubernMerida,´ S. Vreemann, R. Martı, et al. Automated detection of breast cancer in falsenegative screening
mri studies from women at increased risk, European journal of radiology 85 (2) (2016) 472–479.
[89] K. Maitra, S. Bhattacharjee, D. Bhattacharyya, et al. Adaptive edge detection technique towards features
extraction from mammogram images, Journal of Cancer Research Updates 5 (2) (2016) 47–58.
[90] G. Ertas, S. J. Doran, M. O. Leach, A computerized volumetric segmentation method applicable to multicentre
mri data to support computeraided breast tissue analysis, density assessment and lesion localization, Medical
& biological engineering & computing (2016) 1–12.
[91] C. Meldrum, M. A. Doyle, R. W. Tothill. Nextgeneration sequencing for cancer diagnostics: a practical
perspective, Clin Biochem Rev 32 (4) (2011) 177–195.
[92] B. Rabbani, M. Tekin, N. Mahdieh. The promise of wholeexome sequencing in medical genetics, Journal of
human genetics 59 (1) (2014) 5–15.
[93] Gagan, E. M. Van Allen. Nextgeneration sequencing to guide cancer therapy, Genome medicine 7 (1) (2015) 1–
10.
[94] F. S. Collins, M. A. Hamburg. First fda authorization for nextgeneration sequencer, New England Journal of
Medicine 369 (25) (2013) 2369–2371.
[95] Ku CS, Cooper DN, Iacopetta B, et al. Bringing nextgeneration sequencing oncology tests into the diagnostic
setting, Next Generat Sequenc & Applic 01 (S1). doi:10.4172/24699853.s1004.
[96] G. R. Oliver, S. N. Hart, E. W. Klee, Bioinformatics for clinical next generation sequencing, Clinical chemistry 61
(1) (2015) 124–135.
[97] E. Toland, L. Durham, T. S. Ross. Use of whole genome sequencing for diagnosis and discovery in the cancer
genetics clinic.
[98] S. NikZainal, H. Davies, J. Staaf, et al Landscape of somatic mutations in 560 breast cancer wholegenome
sequences, Nature.
[99] M. Farwell, V. A. Joshi. Dna sequencing of cancerrelated genes for biomarker discovery, Tumor Biomarker
Discovery: Methods and Protocols (2009) 205–220.
[100] S. Roychowdhury, M. K. Iyer, D. R. Robinson, et al. Personalized oncology through integrative highthroughput
sequencing: a pilot study, Science translational medicine 3 (111) (2011) 111ra121–111ra121.
[101] H. Beltran, K. Eng, J. M. Mosquera, et al. Wholeexome sequencing of metastatic cancer and biomarkers of
treatment response, JAMA oncology 1 (4) (2015) 466–474.
[102] M. Schwaederle, G. A. Daniels, D. E. Piccioni, et al. On the road to precision cancer medicine: analysis of
genomic biomarker actionability in 439 patients, Molecular cancer therapeutics 14 (6) (2015) 1488–1494.
[103] J. D. Brooks. Translational genomics: the challenge of developing cancer biomarkers, Genome research 22 (2)
(2012) 183–187.
[104] P. C. Boutros. The path to routine use of genomic biomarkers in the cancer clinic, Genome research 25 (10)
(2015) 1508–1513.
[105] R. Dienstmann, I. S. Jang, B. Bot, et al. Database of genomic biomarkers for cancer drugs and clinical
targetability in solid tumors, Cancer discovery 5 (2) (2015) 118–123.
[106] D. G. Covell. Data mining approaches for genomic biomarker development: Applications using drug screening
data from the cancer genome project and the cancer cell line encyclopedia, PloS one 10 (7) (2015) e0127433.
[107] N. C. Bennett, C. S. Farah. Nextgeneration sequencing in clinical oncology: next steps towards clinical
validation, Cancers 6 (4) (2014) 2296–2312.
[108] S. Gargis, L. Kalman, M. W. Berry, et al. Assuring the quality of nextgeneration sequencing in clinical laboratory
practice, Nature biotechnology 30 (11) (2012) 1033–1036.
[109] M. R. Stratton, P. J. Campbell, P. A. Futreal. The cancer genome, Nature 458 (7239) (2009) 719–724.
[110] J. R. Pon, M. A. Marra. Driver and passenger mutations in cancer, Annual Review of Pathology: Mechanisms of
Disease 10 (2015) 25–50.
[111] T. Sakoparnig, P. Fried, N. Beerenwinkel. Identification of constrained cancer driver genes based on mutation
timing, PLoS Comput Biol 11 (1) (2015) e1004027.
[112] B. J. Raphael, J. R. Dobson, L. Oesper, et al. Identifying driver mutations in sequenced cancer genomes:
computational approaches to enable precision medicine, Genome Med 6 (5).
[113] S. K. Merid, D. Goranskaya, A. Alexeyenko. Distinguishing between driver and passenger mutations in individual
cancer genomes by network enrichment analysis, BMC bioinformatics 15 (1) (2014) 1.
[114] J. Foo, L. L. Liu, K. Leder, et al. An evolutionary approach for identifying driver mutations in colorectal cancer,
PLoS Comput Biol 11 (9) (2015) e1004350.
[115] P. F. MACGREGoR, J. A. SqUIRE, Application of microarrays to the analysis of gene expression in cancer, Clinical
Chemistry 48 (8) (2002) 1170–1177.
[116] O. Modlich, H. B. Prisack, H. Bojar. Breast cancer expression profiling: the impact of microarray testing on
clinical decision making, Expert opinion on pharmacotherapy 7 (15) (2006) 2069–2078.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
21
[117] V. Trevino, F. Falciani, H. A. BarreraSaldana. Dna microarrays: a powerful genomic tool for biomedical and
clinical research.
[118] MOLECULAR MEDICINECAMBRIDGE MA THEN NEW YORK13 (9/10) (2007) 527.
[119] PerezDiez, A. Morgun, N. Shulzhenko. Microarrays for cancer diagnosis and classification, in: Microarray
Technology and Cancer Gene Profiling, Springer, 2007, pp. 74–85.
[120] L. Hoopes, Genetic diagnosis: Dna microarrays and cancer, Nature Education 1 (1) (2008) 3.
[121] J. Beane, A. Spira, M. E. Lenburg. Clinical impact of highthroughput gene expression studies in lung cancer,
Journal of Thoracic Oncology 4 (1) (2009) 109–118.
[122] D. Rusinek, S. SzpakUlczok, B. Jarzab. Gene expression profile of human thyroid cancer in relation to its
mutational status, Journal of molecular endocrinology 47 (3) (2011) R91–R103.
[123] P. E. Colombo, F. Milanezi, B. Weigelt, et al. Microarrays in the 2010s: the contribution of microarraybased gene
expression profiling to breast cancer classification, prognostication and prediction, survival 5 (5).
[124] R. Kumar, A. Sharma, R. K. Tiwari, et al., Application of microarray in breast cancer: An overview, Journal of
Pharmacy And Bioallied Sciences 4 (1) (2012) 21.
[125] R. Gentleman, W. Huber. Processing affymetrix expression data, in: Bioconductor Case Studies, Springer, 2008,
pp. 25–45.
[126] M. I. Love, S. Anders, V. Kim, et al. Rnaseq workflow: genelevel exploratory analysis and differential expression,
F1000Research 4.
[127] J. Y. Cho, J. Y. Lim, J. H. Cheong, et al. Gene expression signature–based prognostic risk score in gastric cancer,
Clinical Cancer Research 17 (7) (2011) 1850–1857.
[128] R. SanzPamplona, A. Berenguer, D. Cordero, et al. Clinical value of prognosis gene expression signatures in
colorectal cancer: a systematic review, PloS one 7 (11) (2012) e48877.
[129] T. P. Lu, E. Y. Chuang, J. J. Chen. Identification of reproducible gene expression signatures in lung
adenocarcinoma, BMC bioinformatics 14 (1) (2013) 1.
[130] F. Kottakis, N. Bardeesy. Gene signatures from pancreatic cancer tumor and stromal cells predict disease
outcome, Nature genetics 47 (10) (2015) 1102–1103.
[131] D. HandkiewiczJunak, M. Swierniak, D. Rusinek, et al. Gene signature of the postchernobyl papillary thyroid
cancer, European journal of nuclear medicine and molecular imaging 43 (7) (2016) 1267–1277.
[132] A. C. Culhane, M. S. Schroder, R. Sultana, et al. Genesigdb: a manually curated database and resource for
analysis of gene expression signatures, Nucleic acids research (2011) gkr901.
[133] J. J. Chen, S. Knudsen, W. Mazin, et al. A 71gene signature of trail sensitivity in cancer cells, Molecular cancer
therapeutics 11 (1) (2012) 34–44.
[134] J. Liang, P. Tong, W. Zhao, et al. The rest gene signature predicts drug sensitivity in neuroblastoma cell lines
and is significantly associated with neuroblastoma tumor stage, International journal of molecular sciences 15
(7) (2014) 11220–11233.
[135] Y. Zheng, J. Zhou, Y. Tong. Gene signatures of drug resistance predict patient survival in colorectal cancer, The
pharmacogenomics journal 15 (2) (2015) 135–143.
[136] L. J. Van’t Veer, H. Dai, et al. Gene expression profiling predicts clinical outcome of breast cancer, nature 415
(6871) (2002) 530–536.
[137] J. Subramanian, R. Simon. Gene expression–based prognostic signatures in lung cancer: ready for clinical
use?, Journal of the National Cancer Institute 102 (7) (2010) 464–474.
[138] Kela, L. EinDor, G. Getz, D. et al. Outcome signature genes in breast cancer: is there a unique set?, Breast
Cancer Research 7 (Suppl 2) (2005) 1–1.
[139] D. Venet, J. E. Dumont, V. Detours. Most random gene expression signatures are significantly associated with
breast cancer outcome, PLoS Comput Biol 7 (10) (2011) e1002240.
[140] M. Buyse, S. Loi, L. Van’t Veer, et al. Validation and clinical utility of a 70gene prognostic signature for women
with nodenegative breast cancer, Journal of the National Cancer Institute 98 (17) (2006) 1183–1192.
[141] N. Chen. Incorporate gene signature profiling into routine molecular testing, Applied & Translational Genomics
2 (2013) 28–33.
[142] C. M. Perou, T. Sørlie, M. B. Eisen, et al. Molecular portraits of human breast tumours, Nature 406 (6797)
(2000) 747–752.
[143] T. Sørlie, C. M. Perou, R. Tibshirani, et al. Gene expression patterns of breast carcinomas distinguish tumor
subclasses with clinical implications, Proceedings of the National Academy of Sciences 98 (19) (2001) 10869–
10874.
[144] T. Sørlie, R. Tibshirani, J. Parker, et al. Repeated observation of breast tumor subtypes in independent gene
expression data sets, Proceedings of the National Academy of Sciences 100 (14) (2003) 8418–8423.
[145] R. Tibshirani, T. Hastie, B. Narasimhan, et al. Diagnosis of multiple cancer types by shrunken centroids of gene
expression, Proceedings of the National Academy of Sciences 99 (10) (2002) 6567–6572.
[146] J. S. Parker, M. Mullins, M. C. Cheang, et al. Supervised risk predictor of breast cancer based on intrinsic
subtypes, Journal of clinical oncology 27 (8) (2009) 1160–1167.
[147] T. Nielsen, B. Wallden, C. Schaper, et al. Analytical validation of the pam50based prosigna breast cancer
prognostic gene signature assay and encounter analysis system using formalinfixed paraffinembedded breast
tumor specimens, BMC cancer 14 (1) (2014) 1.
[148] B. J. Caan, C. Sweeney, L. A. Habel, et al. Intrinsic subtypes from the pam50 gene expression assay in a
populationbased breast cancer survivor cohort: prognostication of shortand longterm outcomes, Cancer
Epidemiology Biomarkers & Prevention 23 (5) (2014) 725–734.
[149] M. C. Liu, B. N. Pitcher, E. R. Mardis, et al. Pam50 gene signatures and breast cancer prognosis with adjuvant
anthracyclineand taxanebased chemotherapy: correlative analysis of c9741 (alliance), npj Breast Cancer 2
(2016) 15023.
[150] I. B. Tan, T. Ivanova, K. H. Lim, et al. Intrinsic subtypes of gastric cancer, based on gene expression pattern,
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
22
predict survival and respond differently to chemotherapy, Gastroenterology 141 (2) (2011) 476–485.
[151] P. Roepman, A. Schlicker, J. Tabernero, et al. Colorectal cancer intrinsic subtypes predict chemotherapy benefit,
deficient mismatch repair and epithelialtomesenchymal transition, International journal of cancer 134 (3)
(2014) 552–562.
[152] BacaLopez,´ E. Hernandez´Lemus, M. Mayorga. Informationtheoretical analysis of gene expression data to
infer transcriptional interactions, Revista mexicana de f´ısica 55 (6) (2009) 456– 466.
[153] A. Margolin, I. Nemenman, K. Basso, et al. Aracne: an algorithm for the reconstruction of gene
regulatory networks in a mammalian cellular context, BMC bioinformatics 7 (Suppl 1) (2006) S7.
[154] H. Hatzikirou, A. Chauviere, A. L. Bauer, et al. Integrative physical oncology, Wiley Interdisciplinary Reviews: Systems
Biology and Medicine 4 (1) (2012) 1–14.
[155] P. M. Altrock, L. L. Liu, F. Michor. The mathematics of cancer: integrating quantitative models, Nature Reviews
Cancer 15 (12) (2015) 730–745.
[156] R. P. Araujo, D. L. S. McElwain. A history of the study of solid tumour growth: the contribution of mathematical
modelling, Bulletin of mathematical biology 66 (5) (2004) 1039–1091.
[157] BrU, M. A. Herrero. From the physical laws of tumor growth to modelling cancer processes, Mathematical
Models and Methods in Applied Sciences 16 (supp01) (2006) 1199–1218.
[158] P. Tracqui. Biophysical models of tumour growth, Reports on Progress in Physics 72 (5) (2009) 056701.
[159] S. Lowengrub, H. B. Frieboes, F. Jin, et al. Nonlinear modelling of cancer: bridging the gap between cells and
tumours, Nonlinearity 23 (1) (2009) R1.
[160] T. S. Deisboeck, Z. Wang, P. Macklin, et al. Multiscale cancer modeling, Annual review of biomedical
engineering 13.
[161] Z. Wang, J. D. Butner, R. Kerketta, et al. Simulating cancer growth with multiscale agentbased modeling, in:
Seminars in cancer biology, Vol. 30, Elsevier, 2015, pp. 70–78.
[162] Stephanou,´ V. Volpert. Hybrid modelling in biology: a classification review, Mathematical Modelling of Natural
Phenomena 11 (1) (2016) 37–48.
[163] A. Rejniak, A. R. Anderson. Hybrid models of tumor growth, Wiley Interdisciplinary Reviews: Systems Biology and
Medicine 3 (1) (2011) 115–125.
[164] S. Benzekry, C. Lamont, A. Beheshti, et al. Classical mathematical models for description and prediction of
experimental tumor growth, PLoS Comput Biol 10 (8) (2014) e1003800.
[165] H. B. Frieboes, X. Zheng, C. H. Sun, et al. An integrated computational/experimental model of tumor invasion,
Cancer research 66 (3) (2006) 1597–1604.
[166] T. VargoGogola, J. M. Rosen. Modelling breast cancer: one size does not fit all, Nature Reviews Cancer 7 (9)
(2007) 659–672.
[167] H. Murphy, H. Jaafari, H. M. Dobrovolny. Differences in predictions of ode models of tumor growth: a cautionary
example, BMC cancer 16 (1) (2016) 1.
[168] Hartung, S. Mollard, D. Barbolosi, et al. Mathematical modeling of tumor growth and metastatic spreading:
validation in tumorbearing mice, Cancer research 74 (22) (2014) 6397–6407.
[169] S. A. Mazlan, N. Rosli, N. S. Azmi. Modelling the cancer growth process by stochastic delay differential
equations under verhults and gompertzs law, Jurnal Teknologi 78 (32).
[170] M. Le, H. Delingette, J. KalpathyCramer, E. Gerstner, et al. Mri based bayesian personalization of a tumor
growth model. IEEE Trans Med Imaging. 2016 Apr 29.
[171] Tariq, T. Chen, N. F. Kirkby, et al. Modelling and bayesian adaptive prediction of individual patients tumour
volume change during radiotherapy, Physics in medicine and biology 61 (5) (2016) 2145.
[172] G. Remy, P. Cluzel. Nearcriticality underlies the behavior of early tumor growth, Physical biology 13 (2) (2016)
026005.
[173] F. Michor, K. Beal. Improving cancer treatment via mathematical modeling: Surmounting the challenges is
worth the effort, Cell 163 (5) (2015) 1059–1063.
[174] Z. Wang, J. D. Butner, V. Cristini, et al. Integrated pkpd and agentbased modeling in oncology, Journal of
pharmacokinetics and pharmacodynamics 42 (2) (2015) 179–189.
[175] R. C. Jackson. Pharmacodynamic modelling of biomarker data in oncology, ISRN pharmacology 2012.
[176] M. V. Karamouzis, A. G. Papavassiliou. Tackling the cancer signal transduction labyrinth : A combinatorial use of
biochemical tools with mathematical models will enhance the identification of optimal targets for each
molecular defect, Cancer 120 (3) (2014) 316–322.
[177] G. de AndaJauregui,´ R. A. Mej´ıaPedroza, J. EspinalEnr´ıquez, et al. Crosstalk events in the estrogen signaling
pathway may affect tamoxifen efficacy in breast cancer molecular subtypes, Computational biology and
chemistry 59 (2015) 42–54.
[178] D. Kirouac, M. Onsum. Using network biology to bridge pharmacokinetics and pharmacodynamics in oncology,
CPT: pharmacometrics & systems pharmacology 2 (9) (2013) 1–7.
[179] S.Y. Shin, O. Rath, A. Zebisch, et al. Functional roles of multiple feedback loops in extracellular signalregulated
kinase and wnt signaling pathways that regulate epithelialmesenchymal transition, Cancer research 70 (17)
(2010) 6715–6724.
[180] D. Barbolosi, J. Ciccolini, B. Lacarelle, et al. Computational oncology [mdash] mathematical modelling of drug
regimens for precision medicine, Nature Reviews Clinical Oncology.
[181] H. B. Frieboes, M. E. Edgerton, J. P. Fruehauf, et al. Prediction of drug response in breast cancer using
integrative experimental/computational modeling, Cancer research 69 (10) (2009) 4484–4492.
[182] M. M. Hadjiandreou, G. D. Mitsis. Mathematical modeling of tumor growth, drugresistance, toxicity, and optimal
therapy design, Biomedical Engineering, IEEE Transactions on 61 (2) (2014) 415–425.
[183] X. Sun, J. Bao, Y. Shao. Mathematical modeling of therapyinduced cancer drug resistance: Connecting cancer
mechanisms to population survival rates, Sci Rep. 2016 Mar 1;6:22498. doi: 10.1038/srep22498.
ARTICLE ISSN: 2456-2262 PRECISION MEDICINE
IN ONCOLOGY….
| de AndaJauregui and Hernandez´Lemus | PRECISION MEDICINE IN ONCOLOGY | VOL 1| NUMBER 1| 2016 | 9-23
23
[184] J. Serkova, K. Glunde. Metabolomics of cancer, Tumor Biomarker Discovery: Methods and Protocols (2009)
273–295.
[185] E. G. Armitage, C. Barbas. Metabolomics in cancer biomarker discovery: current trends and future perspectives,
Journal of pharmaceutical and biomedical analysis 87 (2014) 1–11.
[186] A. Vermeersch, M. P. Styczynski, et al., Applications of metabolomics in cancer research, Journal of
carcinogenesis 12 (1) (2013) 9.
[187] D. B. Liesenfeld, N. Habermann, R. W. Owen, et al. Review of mass spectrometry–based metabolomics in
cancer research, Cancer Epidemiology Biomarkers & Prevention 22 (12) (2013) 2182–2201.
[188] M. Tomita, K. Kami. Systems biology, metabolomics, and cancer metabolism, Science 336 (6084) (2012) 990–
991.
[189] D. A. Gaul, R. Mezencev, T. Q. Long, et al. Highlyaccurate metabolomic detection of earlystage ovarian cancer,
Sci Rep. 2015 Nov 17;5:16351. doi: 10.1038/srep16351.
[190] Thiele, B. Ø. Palsson. A protocol for generating a highquality genomescale metabolic reconstruction, Nature
protocols 5 (1) (2010) 93–121.
[191] ResendisAntonio, C. Gonzalez´Torres, G. JaimeMunoz et al. Modeling metabolism: a window toward a
comprehensive interpretation of networks in cancer, in: Seminars in cancer biology, Vol. 30, Elsevier, 2015, pp.
79–87.
[192] Smallbone, E. Simeonidis, D. S. Broomhead, et al. Something from nothing bridging the gap between
constraintbased and kinetic modelling, Febs Journal 274 (21) (2007) 5576–5585.
[193] Bordbar, J. M. Monk, Z. A. King, et al. Constraintbased models predict metabolic and associated cellular
functions, Nature Reviews Genetics 15 (2) (2014) 107–120.
[194] Smallbone, E. Simeonidis, N. Swainston, P. et al. Towards a genomescale kinetic model of cellular metabolism,
BMC Systems Biology 4 (1) (2010) 1.
[195] W. Dubitzky, O. Wolkenhauer, H. Yokota, et al. Encyclopedia of systems biology, Springer Publishing Company,
Incorporated, 2013.
[196] Hadi, S. A. Marashi. Reconstruction of a generic metabolic network model of cancer cells, Molecular
BioSystems 10 (11) (2014) 3014–3021.
[197] F. Gatto, H. Miess, A. Schulze, et al. Flux balance analysis predicts essential genes in clear cell renal cell
carcinoma metabolism, Scientific reports 5.
[198] F. Capuani, D. De Martino, E. Marinari, et al. Quantitative constraintbased computational model of
tumortostroma coupling via lactate shuttle, Sci Rep. 2015 Jul 7;5:11880. doi: 10.1038/srep11880.
[199] Y. Asgari, Z. Zabihinpour, A. SalehzadehYazdi, et al. Alterations in cancer cell metabolism: The warburg effect
and metabolic adaptation, Genomics 105 (5) (2015) 275–281.
[200] A. Garraway. Genomicsdriven oncology: framework for an emerging paradigm, Journal of Clinical Oncology 31
(15) (2013) 1806–1814.
[201] V. G. LeBlanc, M. A. Marra, Nextgeneration sequencing approaches in cancer: Where have they brought us and
where will they take us?, Cancers 7 (3) (2015) 1925–1958.
[202] S. W. Gray, E. R. Park, J. Najita, et al. Oncologists’ and cancer patients’ views on wholeexome sequencing and
incidental findings: results from the canseq study, Genet Med. 2016 Feb 11.