Experts argue that the failure to realize PDT's true potential is, at least in part, due to the complexity of the dosimetry problem. Unlike radiation therapy – where a century of clinical experience has shown that biological response is well correlated with the energy absorbed per unit mass of tissue – the PDT community has not come up with a widely accepted definition of dose.
Despite more than two decades of activity, it's only recently that PDT researchers have started to determine dose–response curves for normal and diseased tissue. For example, in 2005 it was demonstrated that reliable knowledge of photosensitizer uptake in animal tumours improved prediction of the treatment outcome and enabled the delivered light fluence to be adjusted to compensate for the level of drug uptake.
Further work in this field could yield big rewards: definition of a PDT dose and the design of tools to measure it in vivo will enable clinicians to optimize treatments for individual patients. Mike Patterson, head of Medical Physics at the Juravinski Cancer Centre in Hamilton, Ontario, Canada, talks to Tami Freeman about the complexities of PDT dosimetry.
TF: Why is PDT so far behind other modalities in terms of dosimetry?
MP: It's partly due to the relatively immature state of clinical PDT and the resources that have been invested in its development. But PDT dosimetry is also inherently a more difficult problem due to the complexity of the PDT mechanism itself.
How is clinical PDT dosimetry performed at present?
The common approach is to measure the amount of photosensitizer administered to the patient and the amount of light delivered to the treatment site, but this approach has several problems.
First, the local concentration of photosensitizer varies from site to site in the body and from individual to individual. Second, the penetration of light into the target depends on the specific optical properties of that tissue. Third, if the tissue is hypoxic, or becomes hypoxic as a result of the PDT treatment, the yield of singlet oxygen – the principal mediator of biological damage in PDT – will be lower than expected.
To complicate matters further, all of these parameters can change during treatment and each of the parameters can also influence the others. A successful dosimetry strategy must recognize and account for these variations.
What parameters would give a better indicator of likely clinical response?
Singlet oxygen has been implicated as the chief cytotoxic molecule in PDT since 1976, so it is not surprising that its direct measurement in vivo has received considerable attention.
How is this measurement performed?
The method used for direct dosimetry is the detection of the weak phosphorescence emitted at 1270 nm when singlet oxygen returns to the ground state. Attempts to measure luminescence in cells and tissues date back to the late 1980s but were limited by insufficient detector sensitivity. Efforts were renewed about a decade later when Hamamatsu introduced an IR photomultiplier tube with a quantum efficiency of around 1%.
This innovation resulted in two research groups reporting observations of the 1270 nm signal during PDT of animal tumours. One of these groups also demonstrated that the cumulative singlet oxygen signal measured during PDT of cell suspensions correlated well with cell survival over a range of treatment conditions. No measurements have yet been made during clinical PDT but this should be feasible during treatment of skin lesions.
If it's possible to detect singlet oxygen phosphorescence, why isn't this now a standard dosimetry method?
Singlet oxygen reacts rapidly with biomolecules in vivo so its ambient concentration is in the picomolar range, with emission of only about 108 photons cm-3 s-1. The instrumentation required to detect such weak luminescence is relatively complex and expensive.
It's not yet known to what extent the biological microenvironment influences the relationship between the amount of singlet oxygen generated and the actual 1270 nm emission. Also, the weak signal means that it's not possible to detect singlet oxygen using implanted optical fibres, due to their small effective detection volume. If measurements are restricted to an accessible surface, they can only be applied to a few disease sites.
So are there any other ways to measure singlet oxygen generation?
There are two alternative approaches: explicit dosimetry, in which the critical ingredients of light, photosensitizer and oxygen are measured and used to calculate the singlet oxygen yield; and implicit dosimetry, in which a "surrogate" for biological damage is measured.
Can you describe what's involved in explicit dosimetry?
The first vital step is determining the light fluence rate in tissue. This can be calculated using the radiation transport equation if the tissue's optical absorption and scattering coefficients are known. The absorption coefficient is determined by the concentration of light-absorbing molecules in the tissue. But the scattering coefficient is much harder to predict due to tissue's highly complex architecture. Some success has been achieved with fractal models, but most knowledge is based on experimental results.
The existing data on tissue optical properties are still relatively sparse, however, due to the difficulty of in vivo measurement. It's rarely possible to confine the light to tissue of one type, although if implanted optical fibres are acceptable it's possible to place optical sources and detectors within the organ of interest.
More generally, sources and detectors are placed on the external boundary of a "heterogeneous" tissue, and the geometry and identity of the constituent tissue types provided by an imaging modality such as MRI. Here, the goal is to estimate the optical interaction coefficients of each tissue type - this is really the only way that such coefficients can be measured in vivo.
Alternatively, the fluence rate can be measured directly, but only at discrete points, and such measurements are usually performed to support the calculations or to measure the fluence at critical locations. The measurement is usually made using optical fibres modified to collect light over a large solid angle.
How is photosensitizer concentration measured?
Most photosensitizers are fluorescent, with the emission from a small volume of tissue proportional to concentration. The downside of this approach is that the emission may also be affected by the sensitizer's local chemical environment. And if the light has to propagate some distance in the tissue this will affect the signal levels. Minimizing this distance – for example by delivering the excitation light and collecting the emission via a single optical fibre – can reduce this dependence.
For non-fluorescent sensitizers, absorption spectroscopy is an option. In general, the absorption spectrum is less sensitive to changes in the local chemical environment, although the tissue's optical properties will still affect the signal. Raman spectroscopy has been used to measure sensitizer concentration in excised tissue samples, but it's hard to justify its cost when most sensitizers are candidates for either fluorescence or absorption spectroscopy.
Oxygen, the third component, can be measured invasively using implanted oxygen electrodes. But, as demonstrated in radiotherapy studies, the spatial distribution can be highly non-uniform. Non-invasive methods based on MR or optical absorption spectroscopy can also provide indirect measures of oxygen concentration.
Clearly, the measurement and interpretation of all the components needed to perform explicit dosimetry is a challenging problem. Nonetheless, this approach has been successful in a few cases and is still being pursued.
What about implicit dosimetry; how does that work?
This approach uses a surrogate that can be readily measured and related to biological damage. For example, monitoring the distinct photoproduct that's produced when singlet oxygen reacts with the photosensitizer – or measuring the degradation of the sensitizer – can provide a real-time measure of singlet oxygen yield.
Has much work has been performed on this technique?
To date, implicit dosimetry has mainly been tested in simple biological systems. One experiment examined the bleaching of the sensitizer Photofrin in multicell tumour spheroids. The results implied the presence of two bleaching mechanisms – one mediated by singlet oxygen and the other by the sensitizer triplet state. The study also suggested that Photofrin's photoproduct is only produced by singlet-oxygen-mediated bleaching, making measurements of this molecule more reliable than measuring the reduction in Photofrin itself.
A separate study on a cell suspension system showed that the yield of one of the photoproducts of the sensitizer protoporphyrin IX was correlated with cell survival. Another experiment demonstrated that the concentration of a benzoporphyrin-derivative photoproduct was the most reliable indicator of response after PDT with this agent.
Another interesting hypothesis is that the photobleaching of endogenous fluorophores could be used as an indicator of biological damage. Although this autofluorescence bleaching has been measured in patients, it could not be correlated with eventual clinical outcome. While implicit dosimetry is appealing, much more work needs to be done, especially in vivo, to establish which photosensitizers are the best candidates and to demonstrate the links between implicit dose and relevant biological outcomes.
How do you see PDT dosimetry progressing in the near future – what needs to be done next?
It has taken many years of research to clarify what should be measured and to develop the tools to obtain these data in vivo. What is needed next – and this is already underway – is for measurements to be taken during PDT of animal and human tumours where the biological response is also evaluated. It's only by examining the correlation of response with dose that the importance of dosimetry can be established. Will it improve results and optimize individual treatments? The answers will probably depend on the choice of photosensitizer, the treatment site and the disease.
• See also Special report: seeing the light on PDT? on medicalphysicsweb.