High Performance Visualization of Human Tumor Growth Software
Norma Alias (University of Technology Malaysia)Siti Nur Hidayah Khalid (University of Technology Malaysia)
Norfarizan Mohd Said (University of Technology Malaysia)
Dolly, Tien Ching Sin (University of Technology Malaysia)
Tau Ing Phang (University of Technology Malaysia)
Abstract:
The implementation of parallel algorithm for the simulation of human tumor
growth is a new invention nowadays. The research emphasised on the grand
challenge application of brain tumors growth. Based on the present
knowledge of the properties of tumor, some mathematical models have been
developed by researchers to quantify the proliferation and invasion
dynamics of tumor within anatomically accurate heterogeneous human tissue.
The human body is made up of many types of cells. Each type of the cells
has special functions. Most of the cells in the body grow and then divide
in an orderly way to form new cells as they are need to keep the body
healthy and working properly. The cells will divide too often and without
any order when they lose the ability to control their growth. The extra
cells from a mass of tissue are called tumours.
Brain cancer is the prevalent cancers in the world and they are leading
causes of death from cancer. Mathematical modelling of biomedical
phenomena (Murray, 2003) can be extremely helpful in analyzing factors
that may contribute to the complexity intrinsic in insufficiently
understood developmental process disease. Based on present knowledge of
the properties of tumors, a mathematical model has been developed to
quantify the proliferation and invasion dynamics of tumors within
anatomically accurate heterogeneous human tissue. The implications of the
model would be considerable interest to those interests in the study of
other diseases for which medical imaging plays a part of the assessment of
the disease (other cancer as well as developmental diseases). Breast
cancer seems to be predominant among Chinese women with an incidence of 25
per 100000 population and Malay women with an incident of 16 per 100000
populations (Malaysia Medical Association internet
site—http://www.mma.org.my/current_topic/woman.htm, March 2006). Breast
cancer happens when cells in the breast begin to grow out of control and
can then invade nearby tissue are called tumors. However, some tumors are
not really cancerous because they cannot spread or threaten someone’s
life. Medical ultrasound is essentially a means of producing visual images
based on echoes that occur in acoustic interface. This echoes contains
information that can be used t study the various breast tissues. A lesion
is localized if its echo pattern is different from that of the surrounding
medium.
This project focuses on the implementation of parallel algorithm for the
simulation of parabolic equation of human tumors growth. The platform for
high performance computing of the parallel algorithms run on a distributed
parallel computer system. Implementing parallel algorithm based on
parallel computing system is used to visualize the growth of human tumour.
Parallel Virtual Machine (PVM) is emphasized as communication platform in
parallel computer systems. The software system functions to enable a
collection of heterogeneous computers to be used as synchronize and
flexible concurrent computational resource.
The mathematical modelling framework is presented and numerical results
are carried out in a way to investigate some properties of the medical
modality. The numerical finite-difference method is focused to design the
discreatization of these partial differential equations. The result of
finite difference approximation (explicit, Crank-Nicolson and fully
implicit methods) will be presented graphically. The software will
visualize and predict human tumor growth based on real time and previous
time. Furthermore, it might be able to calculate the growth rate of cells
and predict the death or survival time for brain tumor patient. Robustness
and friendly to user are the main criteria of the software and it does
accept input data from user. Graph that being visualize are
multidimensional; 1D, 2D and 3D.
Yet, this kind of product is not in the market. And if yes, there are
built on sequential algorithm. Patient, doctors, scientists, engineers and
forecast people are the target user of this user friendly product. Its
computational platform is based on low cost supercomputer with open source
technology LINUX using a distributed parallel computer systems. High
performance computing is built on distributed parallel computer system.
Algorithm that being used in solving partial differential equation for
brain tumor growth is new instead of the high speed computation of
mathematical simulation and can resolve large path or unlimited memory.
The parallel performance measurement will be analyzed from the aspect of
speedup, efficiency, effectiveness and temporal performance.
Keywords:
Parallel and Distributed Computing, Computing in Healthcare and Biosciences, Numerical Algorithms for CS&E
