1.

Introduction

Photodynamic therapy (PDT) is based on photochemical reactions involving photosensitizer molecules (e.g., porphyrin derivatives) to treat cancerous and noncancerous target tissue. ^{1, 2, 3, 4, 5} Transported by lipoproteins after intravenous injection, photosensitizer molecules accumulate primarily in fast-growing tissues—most likely cancer cells—and after a specific interval; they are activated by a light dose of a specific wavelength. The activated photosensitizer interacts with local oxygen, creating a highly reactive singlet oxygen form, which is considered to be the main cause of cell death.⁶ Conventional light delivery systems consist of a nonthermal light source and a light diffuser, which is responsible for homogeneous illumination of the target cell area. Many types of light diffusers have been developed recently to address the need for topical or body cavity applications. For topical use, suitable reflectors can be used to obtain a homogeneous light distribution using a noncoherent light source⁷; for spot irradiations, optical fibers with microlenses have been described in the literature.⁸ Applications to deliver light inside body cavities have led to a multiplicity of different illuminator and diffuser devices and techniques, dependent on size, geometry, and accessibility of the solid or hollow organ to be treated. Frontal illuminators (microlens, spherical tip fiber), and cylindrical and balloon diffusers of virtually any shape are presented in the recent literature. ^{9, 10, 11, 12, 13, 14, 15}

It is well known that successful and safe PDT depends mainly on correct irradiance,^{16, 17} as well as control of local photosensitizer concentration, tissue oxygenation, and depth of light penetration into the tissue. The disadvantage of many diffusers mentioned previously is their limited ability to adjust to complex surfaces. They are relatively inflexible or even rigid, and the applied light dose is difficult to predict and control during treatment due to insufficient distance control and unintentional movements. The main goal of this work was to produce a 2-D laminar diffuser with superior flexibility for complex superficial, intraluminal, and interstitial PDT applications. We investigated the construction and performance of this novel light diffuser, which is based on a thin textile support populated with a plurality of plastic optical fibers (POF) delivering the necessary light dose.¹⁸ Combinations of textiles with optical fibers already exist for a variety of functions in fashion design, security devices, sensorics, data communication, or automotive applications. ^{19, 20, 21, 22, 23} In most cases, woven textile structures that incorporate POF are utilized. For our design we used two-thread shuttle embroidery technology for the POF instead of weaving (see Fig. 1 ). Light emittance was achieved by macrobending loops^{24, 25} in the optical fibers automatically through the embroidery process during diffuser production (there is no additional step to damage the cladding in certain locations through mechanical abrasion or chemical solvents). The advantages of embroidery are numerous: 1. POF patterns of every character (single spots, lines, curves, zigzag) can be stitched onto many substrate materials (e.g., textile, nonwoven, membrane, film, paper); 2. embroidery allows manufacturing of every 2-D form that adjusts easily even to complex organ or skin topologies; 3. 3-D structures can be produced; 4. light delivery can be positioned according to diffuser design; 5. a flat textile requires significantly less space; and 6. safety is enhanced.

Fig. 1

Schematics of textile fabrics with integrated POF (shown in black): (a) woven structure with woven POF, and (b) woven structure as a substrate with embroidered POF (equivalent to one luminous fiber element).

In this work we summarize our measurements of flexibility, radiance, and temperature rise for textile-based diffusers. The results suggest that PDT treatment by means of luminous textiles is feasible, and therefore offers an alternative light delivery with respect to common diffusers, especially for complex surfaces and body cavities.

2.

Materials and Methods

2.3.

Radiation Measurement Setup

A sketch of the light delivery system is shown in Fig. 2 . For all measurements, we used a 2-W diode laser, wavelength $\lambda =652\mathrm{nm}$ (Ceram Optec/Biolitec, Bonn/Jena, Germany). The emitted light was coupled into the 178-fiber bundle of the textile diffuser using a frontal light distributor (Medlight, Ecublens, Switzerland). For all measurements, a power output of $100\mathrm{mW}$ was adjusted at the distal frontal light distributor end as measured by a calibrated radiometric detector head (sensing area $1{\mathrm{cm}}^2$ , Gigahertz-Optik, Puchheim, Germany), assembled at the calibrated integrating sphere ( $150\mathrm{mm}$ in diameter, Gigahertz-Optik). All experiments were executed under low-light conditions in free space, in a room climate of $22°\mathrm{C}$ and 50% relative humidity.

Fig. 2

Schematic diagram of the light delivery system for irradiation measurements consisting of laser, frontal light distributor, connector, and textile diffuser, including an approximately circular array of luminous fiber elements.

3.

Results

3.1.

Flexibility

The forces to bend the samples through angles of 5 and $15\mathrm{deg}$ (front and back) are shown in Fig. 3 . We found all samples to be more flexible at the lower bending angle. In comparison to the optical lighting film, the textile diffusers were similarly easy to bend through $5\mathrm{deg}$ lengthwise and $15\mathrm{deg}$ crosswise, and easier to bend through $5\mathrm{deg}$ crosswise and $15\mathrm{deg}$ lengthwise. The results at the 15-deg bending angle (lengthwise and crosswise) for the textile diffusers show the highest variations, because stronger bending increases the effect of the easy-to-bend distances between the luminous elements, and the strong-to-bend POF, which both varied among the samples.

Fig. 3

Measured bending torque of the indicated systems at the given angles (white is the textile diffuser, black is the film).

3.3.

Brightness Distribution

The brightness of the pixels vary in the range of millimeters [Fig. 4a ]. The mean values of the distribution of dark and bright pixels of the six textile diffusers measured using the CCD camera are listed in Table 2 . On a brightness scale of 0 to 255, the peak in the distribution is at 65 on average, the median (Q50) averaged at 69, and the 95% quantile (Q95) averaged at 124 (standard deviations are 7.3, 8.5, and 8.5, respectively). The average full-width at half maximum (FWHM) is 21 with a standard deviation of 4.6. The median (Q50) lies close at the peak, but constantly higher.

Fig. 4

(a) 3-D image and (b) brightness distribution (histogram) of the calculated luminous textile surface.

Table 3

Measured mean and maximum (max) temperatures of one textile diffuser (luminous area of 11cm2 , without aluminum foil) and connector, respectively, at the end of radiation for several minutes at different power settings. The irradiance of the textile diffuser surface was calculated.

Power settings (mW)	Irradiance (calculated) (mW/cm2)	Temperature measurement (°C)				Time (minutes)
		Textile diffuser		Connector
		Mean	Max	Mean	Max
100	3.6	24.4	25.4	23.1	25.8	5
100	3.6	25	26	23.5	26.2	10
1000	36	32	39.5	23.1	25.4	1
1000	36	35.2	43.1	24	29.5	5
1000	36	35.7	43.7	24.5	31.4	7
1250	45	38.4	48	25.8	33.9	5

Table 2

Intensity level of maximum, 50% quantile (Q50), full width at half maximum (FWHM), and 95% quantile (Q95) of the brightness distribution histogram of the circular center area (brightness scale 0 to 255) of six textile diffusers.

Diffuser number	Intensity level	Q50	FWHM	Q95
2	69	75	26	135
4	55	57	24	116
9	69	75	24	130
10	61	64	16	117
11	61	63	15	116
13	75	78	20	129
Average	65	69	21	124
SD	7.3	8.5	4.6	8.5

The skewed brightness distribution, as seen in Fig. 4b, and for all textile diffusers in Table 2, shows a bell-shaped distribution with maximum and FWHM. The range from zero upward and the left steepness at 50 implies neither dark nor numerous faint pixels. At the other end of the scale, 5% of the pixels are at least twice as bright as the maximum level.

4.

Discussion

Our aim was to develop a light applicator that would irradiate an area of almost any surface shape. With the present approach, it is possible to stitch the relatively brittle POF into arrays of intact, bent POF using an established embroidering method. Embroidery uses the flexible woven substrate, and with a small number of stitches, allows fixation of the POF, which in turn allows maximum flexibility in all directions of the textile diffuser, and therefore gives the textile haptic—a major advantage of the present development. Furthermore, our bending measurement results demonstrate low bending forces of the textile diffuser, i.e., high flexibility. The relatively high variation is likely due to the disordered bends and loops of the luminous fiber elements. However, compared with commercially available optical film, the textile diffuser is equally flexible or easier to bend.

In comparison to linear light diffusers, which in air preferentially radiate toward the cut off fiber end²⁶ or laser beam profiles, which show microvariations of local irradiance²⁷ but need some distance to reach the treatment surface, the present textile diffuser radiates randomly out of the surface. These results demonstrate that the emission of the required radiation power achieved on the textile diffuser is sufficient for PDT. These findings were confirmed by successful biological experiments using this kind of textile diffuser,²⁸ and follow the trend to use low fluence rates for effective PDT.²⁹ In our measurements, the textile diffuser achieved an irradiance of $3.6\mathrm{mW}∕{\mathrm{cm}}^2$ , corresponding to a mean efficiency of 40%. According to our calculations, a maximum efficiency of 60% would be possible, whereas the maximum obtainable for many other systems is less than or equal to 60%.⁹ As expected, reflected radiation with the aluminum backing enhances the irradiance to the face (treatment side). The backing serves also as a reflecting coverage, which is useful to shield the treatment personnel. Furthermore, we observed steady light emission during bending of the textile diffuser by hand.

While the bend-induced light loss mechanism of POF is well known, ^{24, 30, 31, 32, 33} the radiant exitance of complex structures like textiles is to the best of our knowledge underexplored. Connected to the laser, each of the stitched POF bends radiate in a unique manner. Our brightness distribution results indicate that most pixels show a specific homogeneous distribution, accomplished through a mixture of mainly scattered and redirected light. This multiple scattering saves space and allows irradiation of a full 2-D surface. The several percent (5 to 10%) bright pixels, so-called hot spots, indicate powerful emission of the light mainly directed out of bent POF, and could result in overdosing during photodynamic treatment. As the scattering and absorbing properties of the treated tissue will also determine the light distribution in the tissue,²⁶ the spatial variability of the textile diffuser emission on tissue needs to be better characterized, especially for the treatment of superficial and thin tissues. Nevertheless, a homogeneous 2-D irradiation of malignant cells is of essential importance for successful PDT.⁹ It was therefore an important confirmation of the present approach to find that the brightness distribution showed only a few isolated dark pixels.

It is difficult to compare the light distribution of the textile diffuser with previously reported systems, because not all relevant parameters are available. For comparison with the integrating balloon device, we did an approximate calculation on the uniformity of the irradiation as described by Dwyer, ⁹ who analyzed images of PDT devices in a similar manner carried out here. Because the pixel brightness distribution of the textile diffuser is skewed, we employed the FWHM and Q50 (see Table 2) instead of the standard deviation and mean value, respectively. We found a slightly lower homogeneity of the textile diffusers (average uniformity 85%, variation of standard deviation of $\pm 11\%$ ) than that of the investigated balloon device of Dwyer’s group (average uniformity of 93%, standard deviation of $\pm 2\%$ ). Several other groups used radiometric measurement setups to investigate the variations of irradiation of PDT devices. It is not clear that their results are directly comparable with the Dwyer algorithm. Nevertheless, these results seem to lie within a similar range and help to illustrate the state of the art. For laser beams (frontal light distributor),^{15, 27} the percentage of variation of the light has been characterized as “very homogeneous” or shows variations up to 17%. Variations of irradiance of different types of cylindrical light distributors (e.g., fiber optic, longitudinal balloon devices)^{10, 15} can achieve very homogeneous results $(\pm 10\%)$ along the main axis, and while showing good polar isotropy, the azimuth isotropy can vary by more than a factor of 5. The indwelling balloon applicator¹¹ has been characterized as well to be “very homogeneous.”

Regarding the potential for tissue warming at the illuminated areas of the skin, a combination with hemoglobin oxygenation³⁴ or hyperthermia³⁵ was found. We measured the temperature increase of the isolated textile diffuser with a low irradiation dose ( $100\mathrm{mW}$ , associated with common PDT) over a period of $10\min$ , and could show that heating would not be a concern for a potential patient nor for the device itself. We found that higher laser settings $\left(1\mathrm{W}\right)$ could lead to temperatures over $37°\mathrm{C}$ (normal body temperature is $37°\mathrm{C}$ ) at the textile diffuser surface within a short time $(<5\min )$ . For higher laser settings in clinical practice, it is indispensable to use a good heat conductor like the aluminum backing, in good contact to the POF, to reduce temperature rise because the skin is a poor heat conductor.

5.

Conclusion

The results of the radiation and homogeneity measurements presented here show that a textile diffuser using bent plastic optical fibers is feasible, and that the embroidered textile, together with an aluminum backing, supplies useful irradiation power for low-fluence-rate PDT treatment on an area of several ${\mathrm{cm}}^2$ . In the future, the irradiation of uneven surfaces has to be investigated. Compared with existing irradiation systems, i.e., laser beams, frontal light distributors (microlens fiber), or balloon catheters, the new textile diffuser is thinner and more flexible and the variations of irradiance should be improved, especially for superficial applications.

Beyond these technical advantages, the present design offers easy handling, resembling an adhesive plaster, with a constant distance to the area to be treated. This allows considerable flexibility of movement (e.g., breathing), which is especially helpful for time-intensive PDT, as suggested by others,¹¹ and avoids the need for complex energy computations.

The growing number of PDTs for dermatological cancer³⁶ suggests a parallel trend with applications with textile diffusers in the future. The textile diffuser was designed for applications to complex body surfaces and can also be used advantageously in body cavities for both nonmalignant disease and for cancer. Therefore the new flexible textile diffuser could revolutionize a number of well-established, effective PDT modalities³⁷ such as treatment of early dermato-oncologic stages like actinic keratoses, Bowen’s disease, and superficial basal cell carcinoma. Besides these fields, the textile diffuser can offer potential advantages for PDT of body cavities prevalent in the head and neck area, such as for leukoplakia and more advanced stages of squamous cell carcinoma in the oral cavity.