Three-dimensional modeling advances photodynamic therapy for bladder cancer
Gary Boas
For many patients with aggressive bladder
cancer, surgical removal of the organ is the best treatment option. Researchers
are actively exploring viable alternatives to this procedure. Photodynamic therapy
(PDT), which relies on light-activated drugs to kill cancer cells, could permit
treatment of the disease while retaining organ function. However, treatments with
this method have resulted in a number of cases of fibrosis
and other complications — possibly because of overtreatment and inhomogeneous
light delivery due to nonuniform illumination in the bladder. For this reason, PDT
is rarely used in the bladder in the US.
To gain a better understanding of how the optical
properties of the bladder affect the delivery of light, researchers Chris Bonnerup,
Claudio Sibata and Ron Allison of East Carolina University
in Greenville, N.C., created a three-dimensional bladder model and analyzed it with
the Advanced Systems Analysis Program software package made by Breault Research
Organization of Tucson, Ariz. The work helped shed light on some of the causes of
light overtreatment and could ultimately advance PDT as a treatment option for bladder
cancer.
A novel software package helps create three-dimensional
models of the bladder, thus advancing photodynamic therapy of cancer in the bladder.
The software package enables modeling of the geometric
structure of complicated biological tissues by way of a computer-aided-design interface
as well as application of various optical coefficients to the model.
In addition, according to Bonnerup,
light sources of different powers and wavelengths can be made to interact with the
model. Subsequent analysis, therefore, can yield information about the ways in which
tissue type, source illumination patterns and patient-specific geometry affect the
light dosage for PDT. Thus, the software can contribute to treatment planning by
helping realize a photon dosimetry.
The software enabled researchers to model the ways in which different
light sources interact with the bladder. This is especially important because the
unique geometry of the bladder can lead to higher-than-intended fluences in photodynamic
therapy.
The program was developed for applications
such as automotive and industrial lighting design. The company began to explore
its potential for bio-optical applications and found that the software can combine
two previously disparate things: prototyping of optical designs and modeling of
light propagation through tissue. Thus, it allows users not only to develop and
test optical designs, but also to create virtual tissue structures to see how they
respond to light. As a result, users can optimize the designs for specific tissue
interactions without going through multiple rounds of prototyping with actual instruments.
In this case, the investigators used
it as a rudimentary treatment planning system. As described in a poster presented
at the 88th annual meeting of the American Radium Society in May, they produced a 3-D model
of the bladder from a series of 1-mm slices acquired with CT and used the software
to determine the optimum prescribed light dosage for the bladder.
After creating the model, they imported
it as a virtual tissue whose optical properties, wall thickness and internal media
could be defined and adjusted by the user. They simulated a standard treatment catheter
with a 2-mm isotropic spherical emitter model and placed it within the bladder model
in different locations and with different source output powers to see how the various
geometries affect bladder-wall fluence.
It was not entirely plug-and-play,
however. According to Bonnerup, the software has a very steep learning curve with
more than 500 commands and functions that can interact and provide erroneous results.
Therefore, he said, professional training on the software is paramount.
The simulations showed that PDT based
on incorrect assumptions about the optical properties of the bladder results in
higher than calculated fluences and doses. Scattering inside the bladder can produce
fluences up to nine times those calculated as well as a roughly twofold dosage variation,
possibly leading to spasms, nocturia, dysuria, suprapubic pain and fibrosis. Therefore,
rigorous control of the dosage is necessary, as is modeling of the bladder prior
to therapy.
The modeling software can help to this
end. “But software models are only as robust as the information used to generate
them,” Bonnerup said.
To create a realistic model of the
bladder, surgeons must have relatively accurate knowledge of all details of the
system, and they have to acquire this prior to modeling, making treatment planning
all the more complex.
Bonnerup added that “since [the
software] is proprietary, the source code and algorithms are not available for scrutiny
and examination.
As such, separate correlative measurements
and comparisons with phantoms are needed to verify the information [that it] provides.”
The researchers are developing tools with which
to measure specific tissue optical properties in vivo, prior to PDT. This data can
be employed, according to Bonnerup, to create an even more patient-specific model
with the software, obviating the need to use published data.
Also, they plan to use the software
to measure the fluorescence exhibited by photosensitizers during PDT. By modeling
this phenomenon, they may be able to quantify drug concentrations in normal versus
cancerous tissue without a sample — in effect, an ‘optical biopsy.’
More applications
The company
is working to extend the potential of the software, specifically by developing a
“realistic skin model” that encapsulates large amounts of information
about functioning tissue layers.
With the software alone, creating a tissue model
as robust as this involves a considerable amount of programming, explained Paul
Holcomb, who joined the company to develop the software for bio-optical applications.
“The realistic skin model asks questions that the normal scientist would be
able to answer fairly readily, and in the end spits out code for you.”
The company plans to introduce other
such add-ons, with the intention of building a library of tissues for biologists
who do not fully understand how light interacts with tissue, for example, and how
to model such interactions.
“Sometimes it is difficult —
from a biological standpoint — to grasp the optical syntax, and vice versa,”
Holcomb added. “We are trying to reverse that — trying to make things
as user-friendly as possible.”
Generally, the software will enable
users to import a wide range of complex, non-uniform geometries, acquired with a
variety of modalities. For example, the bladder model was built with CT images,
while other researchers have constructed a model of the brain derived from MRI scans.
Contact: Chris Bonnerup, East Carolina
University, Greenville, N.C.; e-mail: bonnerupc@ecu.edu and Paul Holcomb, Breault
Research Organization, Tucson, Ariz.; e-mail: pholcomb@breault.com.
LATEST NEWS