Use of Diffuser Fibres in the Treatment of Prostate Cancer by Interstitial Photodynamic Therapy

Abstract

SpectraCure AB is a start-up company aiming at treating internal solid tumours such as prostate tumours with an innovative technology called ‘Interstitial Photodynamic Therapy’. This treatment requires the use of fibres to deliver the amount of light required to activate a drug injected to the patient which leads to the destruction of the tumours. This treatment is patient-based which means that the amount of light, modulated by the irradiation time, is dependent on the characteristics of the tissue for each patient. In order to do so, the critical parameters to estimate are the optical properties and among them, a particular attention will be given to the effective attenuation coefficient μ_eff, which characterises the tissue to be treated. Those parameters are then used as inputs of an algorithm calculating the appropriate irradiation time for each light delivery fibre based on the minimal dose the tissue is supposed to receive to permit the activation of the drug. So far, SpectraCure AB has been using 400 μm bare end fibres assumed to be point sources to perform the light delivery; eighteen of those are required to have a good estimation of the optical properties of the tissue and enough light in all the region to be treated. The idea now is to introduce diffuser fibres, modelled as an array of point sources, to be able to increase the power delivered by the fibres and in the meantime reducing the number of fibres. The aim of this thesis is to evaluate the performance of diffuser fibres in interstitial photodynamic therapy. A finite element method (FEM) solver called Nirfast for near IR (Infrared) fluorescence and spectral tomography has been used. The study has shown that the properties of the tissue remain well determined for 5 and 7 mm diffuser fibres while they become less accurate for a higher fibre length. Furthermore, the diffusion approximation, which is the simplest approximation for the light propagation in homogeneous tissues, has been applied in the case we have a heterogeneous medium and has lead to correct enough results to enable the treatment with the P18 system developed by SpectraCure AB. We intended to change the power delivered by the fibre but no satisfying conclusion could be drawn.

Popular Abstract

Treating cancer with light? That has been possible since the early 1990, thanks to an innovative technology called photodynamic therapy involving the use of light and of a drug to destroy the tissues targeted. Spectracure AB, a Swedish start-up company, has chosen this solution for the treatment of prostate cancer. The device is made of eighteen point sources delivery fibres which are inserted in the prostate and deliver the amount of light required to activate the drug which was injected in the tissue and drives the death of the tumorous cells. Consequently, the fibres constitute an important part of the device and their optimization is a crucial point in the treatment. Spectracure AB has decided to substitute the eighteen point sources fibres by a lower number of diffuser fibres that do not only deliver light at their tip but all along a certain length. This would especially permit to increase the power delivered by the fibres without damaging the direct surrounding tissues, which would have happened with point sources fibres. The following article will expose briefly the theory used and the results found thanks to finite element simulations.
The treatment developed by Spectracure AB is patient based and so, requires to track tissues properties for each patient. It is also a real time system in the sense that the dose is adapted at different steps of the treatment, to avoid useless overexposure. The properties are evaluated through a coefficient called the absorption coefficient and denoted μa. It represents the attenuation of the light after its delivery and is due to the absorption by the medium in which it propagates. The propagation of light in tissue is governed by the radiative transfer equation, which after several simplifications, can enable to determine this attenuation coefficient thanks to the collection by each of the fibre of the signal coming from another fibre.
The success of the photodynamic therapy is determined by the combination of the light dose, the drug concentration and the presence of dioxygen in the tissue targeted. This complex interaction between those three factors quantifies the efficacy of the treatment. An approach called explicit dosimetry measures the light fluence rate, which corresponds to the amount of light produced by the fibres at a certain time in a tissue, to give an idea of the performance of the treatment. The master thesis presented here has the purpose of predicting the light fluence rate in a tissue, given the estimated optical properties of the tissues as mentioned in the previous paragraph. Therefore, the estimation of the absorption coefficients by the different types of fibres will be evaluated through the calculation of the light fluence rate in the tissue by using the radiative transfer equation in a homogeneous medium (this approximation will be used even for heterogeneous medium to see how well it performs in those conditions nonetheless). The ground truth for this problem was established by setting the absorption coefficient and estimating, with a finite element software called nirfast, what is the fluence rate obtained with such properties. This fluence rate serves as the ground truth and initial data since the optical properties are estimated based on this fluence rate. It is as if we were pretending that the light was coming from the fibres directly!
The first step in our case was to simulate the situation with a homogeneous prostate where we used point sources fibres to have a reference case to compare our results with. Then, we calculated the fluence rate for different lengths of diffusers fibres and secondly, we have treated the case with a heterogeneous prostate for the different diffuser length and point sources fibres. The final step was to change the delivery power of the fibres. All along this thesis, we have also compared different approaches concerning the presence of a dampening effect due to the phenomenon of blood pooling. Indeed, when inserting a fibre in the prostate, a blood accumulation might occur and modify the properties of the tissues. We have seen that the introduction of dampening coefficients to account for this effect leads to a better treatment of the patient.
The results have shown that the use of 5 mm or 7 mm diffuser fibres used in the same conditions as the point sources fibres are leading to as good results as with the first type of fibres. Another assumption that we made was that each diffuser fibre could be approximated as an array of point sources. We have finally concluded that the choice of the fibre resolution, i.e the number of point sources used to model the diffuser fibre, is dependent on the fibre length. If we do not want it to influence the results, the fibre resolution (number of points) should be twice the length in mm. Moreover, the treatment still gives good results in the case where the medium is not a homogeneous medium.
The change of the power did not lead us to conclusive results but permitted us to deepen our knowledge concerning the function of the algorithm which calculates the irradiation time for each
fibre. Indeed, the choice of parameters were found to be optimized for the case of fibres delivering a certain amount of power. Some attempts were made to change those parameters but a more thorough care should be applied in the future.

• author: Afir, Shaineze
• supervisor: Niels Christian Overgaard; Ida Arvidsson
• organization: Mathematics (Faculty of Engineering)
• course: FMAM05 20181
• year: 2018
• type: H2 - Master's Degree (Two Years)
• subject: Mathematics and Statistics
• keywords: photodynamic therapy, cancer treatment, matlab, diffuser fibres, biomedical
• publication/series: Master's Theses in Mathematical Sciences
• report number: LUTFMA-3360-2018
• ISSN: 1404-6342
• other publication id: 2018:E60
• language: English
• id: 8960135
• date added to LUP: 2018-10-11 16:14:13
• date last changed: 2018-10-11 16:14:13