E.VIO, 
li San 
ttuale 
i del 
de 
jrandi 
he e 
Marco 
onenti 
tale: 
Llano, 
Iro i= 
> the 
| the 
data 
‘erent 
osium 
rale, 
tical 
etric 
ice", 
3-7 
    
  
  
  
COMPUTER ASSISTED 3D-ANALYSIS OF PHOTODYNAMIC EFFECTS IN 
LIVING CANCER CELLS 
Leemann Th, Walt H**, Margadant F, Guggiana V, Anliker M 
Institute of Biomedical Engineering and Medical Informatics (IBTZ), University of Zürich and Swiss Federal Institute of 
Technology (ETH) in Zürich, Switzerland 
**Department of Gynaecology and Obstetrics, University Hospital, Ziirich, Switzerland 
ABSTRACT 
Photodynamic Therapy (PDT) [4] is a promising new procedure to treat tumors in a less invasive form. PDT has the potential 
to extirpate malignant cells by way of photo-activation of a photo-sensitizer located within a tumor cell. Upon activation, the 
initially inert sensitizer agent becomes toxic by producing singlet oxygen as well as free radicals which then destroy the cell. A 
crucial aspect for the success of this therapy is a better understanding of the basic mechanisms involved in PDT. In this study we 
have investigated dark exposure toxicity and the spatial density distribution of the sensitizer in living tumor-spheroids. The 
selected approach makes use of a in vitro model for the spheroids, a set of photo-sensitizers, confocal laser scanning microscopy 
(CLSM), and dedicated 3D-image processing and visualisation techniques. The results obtained with this approach document its 
capabilities to localize photo-sensitizers within living tumor cells prior and after light exposure in a time saving manner. 
Key Words: photodynamic therapy, photo-sensitizer, 3D-image processing 
1. INTRODUCTION 
Localisation and eradication of tumor cells without 
affecting or damaging normal tissue is a challenging target in 
today's cancer therapy. This objective may soon become 
reality thanks to a new method of cancer treatment named 
Photodynamic Therapy (PDT), which has the potential to 
selectively destroy malignant cells. With PDT the patient is 
treated with a photo-sensitizing substance which should 
accumulate in tumor cells only. This substance is nontoxic 
until it is exposed to light irradiation with the appropriate 
wavelength. As such it induces photoactivation and 
production of phototoxins like singlet oxygen and free 
radicals which initialize tumor necrosis by destroying the cell 
membrane and/or structures within the cytoplasm. 
In the current study our main interest was focused on the 
investigation of basic mechanisms involved in photodynamic 
processes such as the spatial distribution of the photo- 
sensitizer in normal and malignant cell populations and the 
inherent toxicity of the agent in the dark. For this purpose a 
system has been devised consisting of a set of photo- 
sensitizers based on derivatives of porphyrin (PD) as well as 
hematoporphyrin (HPD), a confocal laser scanning 
microscope and a suitable image workstation with 
appropriately devised 3D-image processing software. 
Detailed information about the spatial density distribution of 
the photo-sensitizer contributes to an improved understanding 
of the mechanisms involved in PDT and to the optimization 
of its clinical treatment modalities. 
2. CELLS AND PREPARATION 
For these investigations a human tumor cell line (mamma 
carcinoma - derived, MCF-7) was cultivated in soft agar- 
medium in the form of multi-cellular tumor-spheroids. These 
spheroids were incubated with photo-sensitizers (Photosan 
(HPD) and TPPS** (PS), for up to 24 hours. The location 
and distribution of the sensitizers in the cell population was 
then analysed with a confocal laser scanning microscope 
(Leica CLSM, F.R.G.) equipped with an Argon ion laser 
(excitation wavelength of 488nm or 514nm, respectively). 
The stack of CLSM image data stem from experiments 
performed during a medical thesis and for a poster 
presentation [1]. 
** Tetraphenylporphyrin sulfonate. This was a generous gift 
from Dr. A. Riick, Institut fiir Lasertechnologien in der 
Medizin an der Universitit Ulm, Ulm, F.R.G. 
3. IMAGE PROCESSING HARDWARE 
In general the processing of 3D-image data sets and their 
analysis within reasonable time frames call for extensive on- 
line computing power and large memory capacity. This is 
particularly true for the processing of stacks of medical 
images which are generated by confocal laser scanning 
microscopy and which often have a very low signal-to-noise 
ratio. To achive a cost-effective yet flexible way of 
processing and analysis, we have made use of a combination 
of commercially available workstations and of a custom