TUGraz OPEN Library

Shekhovtsov, Oleksandr

Reinbacher, Christian

Graber, Gottfried

Pock, Thomas

Solving Dense Image Matching in Real-Time using Discrete-Continuous Optimization

Jean-Quartier, Claire

Jeanquartier, Fleur

Cemernek, David

Holzinger, Andreas

Tumor Growth Simulation Profiling
Claire Jean-Quartier(B), Fleur Jeanquartier, David Cemernek,
and Andreas Holzinger
Holzinger Group, Research Unit HCI-KDD, Institute for Medical Informatics,
Statistics and Documentation, Medical Informatics, Statistics and Documentation,
Medical University of Graz, Graz, Austria
c.jeanquartier@hci-kdd.org
Abstract. Cancer constitutes a condition and is referred to a group of
numerous diﬀerent diseases, that are characterized by uncontrolled cell
growth. Tumors, in the broader sense, are described by abnormal cell
growth and are not exclusively cancerous. The molecular basis involves
a process of multiple steps and underlying signaling pathways, build-
ing up a complex biological framework. Cancer research is based on
both disciplines of quantitative and life sciences which can be connected
through Bioinformatics and Systems Biology. Our study aims to provide
an enhanced computational model on tumor growth towards a compre-
hensive simulation of miscellaneous types of neoplasms. We create model
proﬁles by considering data from selected types of tumors. Growth para-
meters are evaluated for integration and compared to the diﬀerent disease
examples.
Herein, we describe an extension to the recently presented visual-
ization tool for tumor growth. The integration of proﬁles oﬀers exem-
plary simulations on diﬀerent types of tumors. The enhanced bio-
computational simulation provides an approach to predicting tumor
growth towards personalized medicine.
Keywords: Computational biology · Cancer types · Tumor growth ·
Simulation · HCI · Visualization · Systems biology · Kinetics · Data
visualization
1 Introduction
Tumors or neoplasms constitute an overgrowth of tissue, often called mass and
lump. Tumors can be non-cancerous (benign) or become cancerous (malignant).
The biological background involves a multistep development based on
genomic instability, metabolic reprogramming and evasion from the immune
system [1,2]. Underlying signaling events comprise proliferation and growth pro-
motion, cell death inhibition and replicative immortality, neoangiogenesis and
invasion [3].
Classiﬁcation and prediction of the diﬀerent types of cancer is crucial for ther-
apy and treatment strategies against the often deadly medical condition. Over
c© Springer International Publishing Switzerland 2016
M.E. Renda et al. (Eds.): ITBAM 2016, LNCS 9832, pp. 208–213, 2016.
DOI: 10.1007/978-3-319-43949-5 16
Tumor Growth Simulation Proﬁling 209
the last years, diagnostic classiﬁcation techniques have made progress advanc-
ing from microscopic examination of morphological tissue changes towards gene
analysis and biomarker discoveries [4,5]. Respective analysis methods are known
for the necessity of computerized support including statistical methods and
machine learning classiﬁers [6]. Yet, methods for predicting tumor growth and
cancer prognosis have not fully advanced from basic research to clinical investi-
gations [7]. Still, mathematical and computational models for cancer prediction
function as beneﬁcial resource in cancer research. There are several models each
on individual types of cancer including brain, ovarian, colon cancers, melanoma,
leukemia or head/neck tumors [9]. These models range from deterministic to
stochastic, from continuous population dynamics to agent-based individual cell
models, from ﬂuid dynamics to Monte Carlo simulations or energy minimization
models and are diﬀerentially used to describe the various phases of initiation,
growth, invasion and migration [10,11]. Moreover, so-called hybrid models sim-
ulate variable discrete and continuous aspects on intracellular and intercellular
processes [12].
Enhanced modeling tools can provide indications and quantitative criteria for
the prognosis of tumor progression as well as therapeutic strategies and support
decision making [7,8].
2 Approach
Our study is based on the recently presented simulation on tumor growth [13],
which is available at Github: https://github.com/davcem/cpm-cytoscape and as
Online Demo: http://styx.cgv.tugraz.at:8080/cpm-cytoscape/. Figure 1 shows a
view from the web-tool for tumor growth modeling and its graphical output.
The tool depicts an biocomputational approach to simulate biological cell sorting
using a two-dimensional extended Potts model [14,15] for the use case of tumor
growth on the cellular level.
By scanning through literature and web resources, we consider data from
selected types of tumors. These are scanned for usable growth parameters. We
create model proﬁles of exemplary pre-settings for an integration into the simula-
tion. The proﬁles are compared and evaluated to the diﬀerent disease examples.
In regard to ease of use for medical scientists, we further improve the graph-
ical user interface for choosing between the model proﬁles.
3 Results & Discussion
We choose three distinct types of cancer and integrated predeﬁned settings based
on recently presented studies on ovarian cancer [16], colorectal cancer [17] and
brain cancer [18].
Figure 3 demonstrates the option for choosing between the model proﬁles for
use cases of various types of cancer (Fig. 2).
Table 1 summarizes the selected parameters for the diﬀerent proﬁles. Exam-
ples are given as custom build as well as ovarian, colorectal and brain cancer
210 C. Jean-Quartier et al.
Fig. 1. Screenshot of the tool’s graphical view
Fig. 2. Screenshot of proﬁle selection via the tool’s user interface
Table 1. CPM parameter settings: comparison of settings for custom and pro-
ﬁle builds in case of two exemplary cell types corresponding to three diﬀerent cancer
diseases (ovarian, colorectal and brain, adopted from [16–18]).
Tumor Growth Simulation Proﬁling 211
time steps (au)
da
rk
/li
gh
tc
el
l-r
at
io
(%
m
ax
)
0 5 10 15 20
0
20
40
60
80
100
Custom
Ovarian
Colorectal
Brain
Fig. 3. Cell growth corresponding to various profiles: line chart showing repre-
sentative ratios between numbers of dark and light cells, given as % max, over computed
steps as time.
[13,16–18]. The underlying factors have been adapted in order to be compara-
ble within the tool’s framework. While testing identiﬁed growth parameters, a
key challenge is providing the possibility of dynamic adjustments of cell-related
parameters to meet the needs for handling peculiarities of the various use cases
of cancer diseases.
The integrated proﬁles depict distinct characteristics. The three selected
examples are compared to the default settings from the custom proﬁle. Rep-
resentative growth rates diﬀer in kinetic characteristics in terms of relative aver-
age and instantaneous growth rate as well as aggressivity and mortality. Figure 3
summarizes ratios between numbers of dark and light cells over time steps, calcu-
lated for the diﬀerent proﬁles. The use case for brain cancer exhibits the highest
growth rate of all selected cases and correlates to exponential growth succeeded
by a sudden saturation, which can be translated as the spatial limitation within
a biological compartment.
The simpliﬁed examples represent basic proﬁles suitable for likewise human
or animal models. Parameters have to be further reﬁned in order to reproduce
realistic scenarios for more detailed studies on various aspects of tumor growth,
in particular referring to the various cancer diseases.
Due to the variability within cancer classes, reﬁned models are of main inter-
est to supporting the evaluation of tumor aggressiveness and risk assessment.
212 C. Jean-Quartier et al.
They provide a resource to study the inter- as well as intracellular system in
correlation to temporal and spatial dynamics.
4 Conclusion
We believe that our approach oﬀers the possibility to study tumor growth
towards a comprehensive interpretation of tumor dynamics. Herein, we present
simulation examples for studying multiple cancer diseases in regard to compar-
ison of kinetic characteristics towards prediction. Results support the notion
that tumor growth follows a universal law at ﬁrst sight but has to be reﬁned and
reevaluated in detail.
References
1. Drake, J.W., Charlesworth, B., Charlesworth, D., Crow, J.F.: Rates of spontaneous
mutation. Genetics 148(4), 1667–1686 (1998)
2. Lodish, H., Berk, A., Zipursky, S.L., et al.: Molecular Cell Biology, 4th edn. W. H.
Freeman, New York (2000)
3. Hanahan, D., Weinberg, R.A.: Hallmarks of cancer. Cell 674(5), 646–646 (2011)
4. Corte´s, J., et al.: New approach to cancer therapy based on a molecularly deﬁned
cancer classiﬁcation. CA: Cancer J. Clin. 64(1), 70–74 (2014)
5. Rodr´ıguez-Enr´ıquez, S., Pacheco-Vela´zquez, S.C., Gallardo-Pe´rez, J.C., Marn-
Herna´ndez, A., Aguilar-Ponce, J.L., Ruiz-Garc´ıa, E., Ruizgodoy-Rivera, L.M.,
Meneses-Garc´ıa, A., Moreno-Sa´nchez, R.: Multi-biomarker pattern for tumor iden-
tiﬁcation and prognosis. J. Cell Biochem. 112(10), 2703–15 (2011)
6. Wang, Y., Tetko, I.V., Hall, M.A., Frank, E., Facius, A., Mayer, K.F., Mewes, H.W.:
Gene selection from microarray data for cancer classiﬁcation-a machine learning
approach. Comput. Biol Chem. 29(1), 37–46 (2005)
7. Vickers, A.J.: Prediction models in cancer care. CA: Cancer J. Clin. 61(5), 315–326
(2011). doi:10.3322/caac.20118
8. Li, X.L., Oduola, W.O., Qian, L., Dougherty, E.R.: Integrating multiscale modeling
with drug eﬀects for cancer treatment. Cancer Inform. 14(Suppl. 5), 21–31 (2016)
9. Enderling, H., Rejniak, K.A.: Simulating cancer: computational models in oncology.
Front Oncol. 3, 233 (2013)
10. Edelman, L.B., Eddy, J.A., Price, N.D.: In silico models of cancer. Wiley Interdisc.
Rev. Syst. Biol. Med. 2(4), 438–459 (2010)
11. Benzekry, S., Lamont, C., Beheshti, A., Tracz, A., Ebos, J.M., Hlatky, L., Hahn-
feldt, P.: Classical mathematical models for description and prediction of experi-
mental tumor growth. PLoS Comput. Biol. 10(8), e1003800 (2014)
12. Rejniak, K.A., Anderson, A.R.A.: Hybrid models of tumor growth. Wiley Interdis-
cip Rev. Syst. Biol. Med. 3(1), 115–125 (2011)
13. Jeanquartier, F., Jean-Quartier, C., Cemernek, D., Holzinger, A.: In Silico Mod-
eling For Tumor Growth Visualization. - Manuscript in revision (2016). https://
github.com/davcem/cpm-cytoscape/
14. Graner, F., Glazier, J.A.: Simulation of biological cell sorting using a two-
dimensional extended Potts model. Phys. Rev. Lett. 69, 2013–2016 (1992)
15. Szab, A., Merks, R.M.: Cellular potts modeling of tumor growth, tumor invasion,
and tumor evolution. Frontiers in oncology 3 (2013)
Tumor Growth Simulation Proﬁling 213
16. Giverso, C., Scianna, M., Preziosi, L., Lo Buono, N., Funaro, A.: Individual cell-
based model for in-vitro mesothelial invasion of ovarian cancer. Math. Model. Nat.
Phenom. 5(1), 203–223 (2010)
17. Osborne, J.M.: Multiscale model of colorectal cancer using the cellular potts frame-
work. Cancer Inform. 14(Suppl. 4), 83–93 (2015)
18. Rubenstein, B.M., Kaufman, L.J.: The role of extracellular matrix in glioma inva-
sion: a cellular potts model approach. Biophys J. 95(12), 5661–5680 (2008)


Tumor Growth Simulation Profiling

Ebner, M.

Mühlburger, H.

Google Glass in Face-to-face Lectures - Prototype and First Experiences

Hauser, Andreas W.

Horn, Paul R.

Head-Gordon, Martin

Bell, Alexis T.

A systematic study on Pt based, subnanometer-sized alloy cluster catalysts for alkane dehydrogenation: effects of intermetallic interaction

Ebner, Martin

Gailer, Christian

Khalil, Mohammad

Kopp, Michael

Lackner, Elke

Raunig, Michael

Potential of EPUB3 for Digital Textbooks in Higher Education

Karpenko-Jereb, Larisa

Rynkowska, Edyta

Kujawski, Wojciech

Lunghammer, Sarah

Kujawa, Joanna

Marais, Stephane

Fatyeyeva, Kateryna

Chappey, Corinne

Kelterer, Anne-Marie

Ab-initio study of cationic polymeric membranes in water and methanol

Interactive Machine Learning (iML)

Wiednig, Christopher Alois

Enzinger, Norbert

Beal, Coline

Elektronenstrahlschweißen einer artfremden Verbindung (A625/CB2) für Anwendungen im Kraftwerksbau

Zeppelzauer, Matthias

Poier, Georg

Seidl, Markus

Schulter, Samuel

Breiteneder, Christian

Bischof, Horst

Interactive 3D Segmentation of Rock-Art by Enhanced Depth Maps and Gradient Preserving Regularization

Geymayer, Thomas

Schmalstieg, Dieter

Collaborative Distributed Cognition Using A Seamless Desktop Infrastructure

Graz University of Technology

Not a member yet