Performance of a non-laser light source for photodynamic therapy

Colin Whitehurst', Karen T Byrnel, Colin Morton2, J V Moore'

I Laser Oncology Programme, Cancer Research Campaign
Department of Experimental Radiation Oncology, Paterson Institute for Cancer Research, Christie Hospital NHS 
Trust, Wilmslow Road, Manchester, M20 9BX, UK, 2 Department of Dermatology, Western Infirmary, 
Glasgow, G11 6NZ, UK.

ABSTRACT

Advances in short arc technology and optical filter coatings led to the design and construction of a table-top light 
source in 1989; the first viable and cost-effective alternative to a laser. The device can deliver over 3 W within a 
30 nm band centred at any wavelength from 200 nm to 1200 nm at fluence rates of over 1 W cm~2. Its relative 
biological effectiveness (RBE), in vitro, has been proven alongside an argon pumped dye laser and a copper 
vapour pumped dye laser. These in vitro tests showed an efficiency of haematoporphyrin derivative (HPD) 
induced cellular photoinactivation close to that of the argon/dye laser (RBE 100%), with a mean RBE for the 
lamp of 87 + 3% (p < 0.05). The lamp proved to be superior to that of the copper/dye laser system with an RBE 
of up to 150% at fluence rates above 50 mWcm~2. In vivo tests show that the extent and depth of tumour 
necrosis to be comparable to that of an argon/dye laser. An in situ bioassay using tumour regrowth delay is 
currently underway. Early clinical trials show clearance of Bowen's disease and actinic keratosis using the same 
light fluences as costly PDT lasers.

Keywords: PDT, light source, dose response characteristics, in vitro, clinical, HpD

1. INTRODUCTION

Presently, lasers are used as the light source in PDT because of their particular beam and spectral qualities. 
Unfortunately, the uniqueness of any laser can be inhibiting as each system is commonly limited to just one 
medical application. More often that not the laser itself is a far more sophisticated piece of equipment, reflected 
by high costs, complex operation and large size, than is actually required for some biological/clinical tasks and is 
therefore an over-engineered solution. Previous attempts at phototherapy using non-laser light sources have 
resulted in devices prone to thermal damage and long exposure times. Militating against all of these sources has 
been the fact that insufficient optical power could be delivered to the target within the critical wavelength band 
and overspill into ineffective wavelength regions have reduced their relative biological efficiency (RBE) and 
increased undesirable side-effects. All previous nonlaser sources have used exposed delivery beams and suitable 
shielding was necessary to protect patient and operator alike. Coupling the light into fibres has not thus far been 
achieved satisfactorily and is essential for convenience, accessibility and safety. We report here the development 
of an alternative and the first viable table-top light source which embodies those qualities of a laser appropriate 
for PDT, i.e. high brightness, high photoactivation efficiency and fibre deliverable, whilst avoiding high costs 
and complexity. The in vitro performance of this prototype was compared with two existing PDT laser systems, 
a continuous wave (cw) argon ion pumped dye laser and a pulsed copper vapour pumped dye laser. The 
programme sets out to prove its RBE, in terms of clonogenic cell kill in vitro, and subsequently in vivo and in 
the form of clinical trials.

2. MATERIALS AND METHODS

2.1 In vitro study

The prototype light source in Fig. 1, measures 30 cm x 15 cm x 15 cm and incorporates a short arc discharge 
producing a cw broadband flat spectral output. Custom filters select the appropriate wavelength and bandwidth 
before the output is focused into a light guide via a programmable shutter. The prototype delivers over 3 W of 
photoactivating light directly or 1W via a light guide within a bandwidth of 30 nm which can be tuned to any 
wavelength from 300 nm to 1.2 ,um. The lamp is currently subject to a worldwide patent. The light was 
delivered to the irradiation site via a 5 mm fibre bundle. A collimating lens assembly broadened the field and 
random scrambling of the bundle produced uniformity. The performance of this prototype was compared with a 
cw argon pumped dye laser and a 10 kHz pulsed copper vapour pumped dye laser tuned to a 630 nm. Both laser 
beams were delivered to the irradiation site via a fibre and collimating lens assembly. All three sources were 
equipped with programmable shutters to deliver precisely timed light doses or fluences. The various light 
fluence rates were measured with a thermopile and checked for uniformity (+ 3%3 over the cell culture plates. 
The wavelength of delivered light was measured to :t 1 nm with a handheld monochromator. 7

Exponentially growing Chinese hamster ovary fibroblasts (CHO) were used in all experiments. 10 ,ug ml~' of 
HpD was added to CHO cells and incubated in the dark for 24h. One ml samples of identical concentration of 
cells were placed into alternate 16 mm diameter wells of a 24 multi-well plate. These cells were protected from 
light until they were ready for irradiation. Clonogenic cell assays were carried out following light treatment. The 
cell survival fraction for the three light sources was plotted as a function of light dose. Measurement of the 
action spectra for HPD, Iwhich defines the wavelength efficiency for HPD-induced photoinactivation of CHO 
cells, was achievedlby measuring the surviving
fraction of cells exposed in different wells to a range of wavelengths from 615 to 645 nm in steps of 5 nm at a 
fluence rate of 20 mW cm~2 for four different light fluerlces from the copper/dye laser. Measurement of the 
RBE was achieved by measuring the surviving fraction of cells exposed to various light fluences. The cells were 
irradiated by light centred at 627 nm for the three different light sources. The various PDT treatment groups 
were exposed to light fluences ranging from 0 to 2.5 Jcm -2 delivered at fluence rates ranging from 10 to 200 
mW cm~2. Cells that received no irradiation were used as drug-only controls. A light-only control was taken for 
light fluences of 1-100 Jcm -2 at a fluence rate of 975 mW cm2.

2.2 In vivo pilot studs

Nude mice with subcutaneously implanted mammary tumours on the flank were injected with 5 or 10 mg/kg 
HpD, 24 hrs pretreatment. The mice were restrained and the tumour irradiated externally via a 5 mm fibre at 150 
mW/cm2 for a total dose of 180 J/cm2. 4 ,um histological sections were taken 24 hrs post treatment and 
comparison made in tumour damage between the prototype and PDT laser. Currently, an in situ bioassay using 
tumour regrowth delay on the same animal model is underway.

2.3 Clinical studv

Pilot PDT treatment of skin lesions eg Bowen's disease, actinic keratosis, and basal cell carcinoma using 
topically applied SALA (Glasgow lnfirmary and Leeds General Infirmary), recurrent breast tumour (Cookridge, 
St James's Hospital) and superficial vulva carcinoma (Leeds General) is currently underway. The irradiation 
geometry is simiIar to that of the in vitro study.

3. RESULTS

3.1 In vitro

There was no observable cytotoxic effect from HPD or light alone. Initial experiments were designed to 
determine the wavelength range of HPD phototoxicity in the red region of the spectrum, this region being known 
to penetrate tissue well compared to other visible wavelengths(l), and to evaluate the range of efficiencies of 
phototoxicity in this region.

Figure 2 shows the action spectrum for HPD-induced photoinactivation of CHO cells in the region of 615645 
nm. Curves are shown for cells incubated with 10 ,ug ml~l of HPD for 24 hrs and exposed to increasing light 
fluences, 0.5-2.0 Jcm 2 in steps of 0.5 Jcm~2. The maximum biological effectiveness was observed at 
wavelengths between 626 and 628 nm + lnm. Cell photoinactivation was drastically reduced at 615 and 645 nm 
with a reduction in efficacy of three orders of magnitude from the peak value indicating an effective 
phototoxicity width of 30 nm, narrower than the 45 nm absorption band of HPD in the region(2). The action 
spectrum illustrates a rapid reduction in cell response away from the peak activation wavelength even within 5 
nm.

Figure 3 illustrates the superposition of the prototype spectral output using a bandwidth of 30 nm with the 
measured CHO/HPD action spectrum.

Figure 4 shows the shouldered cell survival curves for CHO cells incubated with HPD for 24 hrs and exposed to 
the three light sources. Curves were fitted using Ethe maximum likelihood technique ). The light fluences 
ranged from 0 to 2.5 Jcm-2 in steps of 0.5 Jcm~2 for a set of light fluence rates, 10, 20, 50, 100 and 200 mW cm 
2 which cover the range of typical PDT treatments.

At low fluence rates, the RBE of the two lasers was the same (p = 0.35 and 0.94 at 10 and 20 mW cm~2, 
respectively, variance ratio f-test). The lamp prototype was comparable, but showed a small but statistically 
significant difference from the lasers in its survival curve (p = 0.01 and 0.02 for 10 and 20 mW cm~2, 
respectively). Comparison of the light fluences from the lamp and the argon laser required to achieve the same 
surviving fraction of 10-l, 10-2, and 10-3 showed an average RBE of 87 + 3% (p <0.05) right up to the maximum 
output available from the lamp. About 50 mW cm2 the copper/dye laser became progressively less effective than 
both the argon/dye laser and the prototype with an RBE as low as 67%. This was due to transient photobleaching 
of the photosensitiser.

3.2 In vivo

The in vivo pilot tests show the extent and depth of tumour necrosis to be comparable to that of an argon laser. 
The equivalent diameter of necrosis for treatment by the lamp was 5.4 + 0.4 mm and 6.3 + 0.4 mm for 5 and 10 
mg/kg of injected drug respectively, and by the Ar/dye laser it was 4.9 + 0.1 mm and 4.9 + 0.3 mm for 5 and 10 
mg/kg of drug respectively.

3.3 Clinical

To date 27 skin lesions have been treated using the lamp plus topical 5 ALA (20 Bowen's, 6 actinic keratosis and 
1 amelanotic melanoma). Every lesion has shown improvement and indeed most of them beyond the two month 
review period have shown clinical clearance confirrned by histology and clinical
observation.

4. DISCUSSION

A convenient, low cost portable alternative has been produced incorporating features that are appropriate to the 
field of PDT and other phototherapies. It was capable of delivering, via a fibre bundle, up to 1 W of narrow 
bandwidth, activating light, sufficient for the PDT treatment of tumours. The performance of the prototype was 
assessed by comparing the cell survival curves of the argon/dye laser with the
prototype and an RBE of 87 + 3% was measured. In addition, the prototype comparedlvery favourably with a 
pulsed copper pumped dye laser at higher influence rates, with an RBE of up to 150%. It has many obvious 
advantages over PDT lasers, namely its optical output may be centred at any wavelength from the ultra-violet to 
the near infra-red, its simplicity to use and maintain, and it is small and portable. Thermal problems and long 
treatment times due to inefficient sensitizer activation, which were associated with earlier non-laser light sources 
have been overcome by removing all wasteful, minimally-activating wavelengths from the source. The filtering 
within the device was crucial and the criticality of activating bandwidth was confirmed by the measurements 
taken of the HPD action spectrum. The action spectrum of the response of CHO cells to PDT indicated the most 
efficient wavelength of red light to be between 626 and 628 nm, and had a phototoxicity width of 30 nm which 
is narrower than the associated absorption band of HPD. The rapid fall-off in phototoxic efficiency shown by the 
action spectrum illustrates the high specificity of the PDT mechanisms.

For safety and convenience the light was delivered via a fibre optic or light guide. This produced a much more 
manipulable and precise instrument for delivering PDT to the target tissue without fear of exposing cells or 
patients to stray light. Prior to this work, this had not been achieved satisfactorily for non-laser sources. This has 
now been overcome using recent advances in optical and short arc technology by imaging such sources directly 
onto the input face of an optical fibre. The prototype delivers PDT via a 5 mm light guide which can be replaced 
by a monofilament delivery system suitable for interstitial in vivo and clinical work.

The in vivo pilot study has shown the lamp to produce comparable tumour penetration and damage to that

using laser irradiation. The ongoing clinical trial has shown clearance of skin lesions mentioned above, using 
non-invasive PDT following simple application of topical ALA. In addition, a second prototype is being 
constructed, capable of delivering over 1 W of narrow bandwidth light via a monofilament fibre of < lmm 
diameter, which should be sufficient to treat larger tumours interstitially using multifibre insertions.

Finally, this device can photoactive at any wavelength for any of the PDT drugs, in contrast to the single 
wavelength output of lasers. It is also capable of optical hyperthermia and photocoagulation, which although 
requiring higher fluence rates than PDT, are not as demanding in their bandwidth restriction or wavelength 
range.

5. REFERENCES

1. M.L. Pantelides, C. Whitehurst, J.V. Moore et al. "Photo-dynamic therapy for localised prostate cancer: light 
penetration in the human prostate gland". J. Urol 143: 398-401, 1990.

2. C.J. Gomer, D.R. Doiron, N. Rucker et al. "Action spectrum (620-645 nm) for hematoporphyrin derivative 
induced cell killing". Photochem. Photobiol. 39: 365-8, 1984.

3. S.A. Roberts. "DRFIT: a program for fitting radiation surviving models". Int. J. Radiat. Biol. 57: 12436, 1990.