2.1.

Bladder Geometry and Light Fluence

The thickness of the rat bladder wall varies between 0.1 and $1.0\mathrm{mm}$ , depending on the absence or presence of a tumor. The relative thickness of the muscular layer compared to urothelium and submucosa is related to the degree of tumor infiltration. For the calculation of light fluence, the rat bladder is modeled as a sphere with a volume of $0.3\mathrm{ml}$ , corresponding to an internal bladder surface area of $2.2{\mathrm{cm}}^2$ .

During illumination, the light is multiply reflected inside the bladder, which causes a locally higher effective irradiance. This behavior is known as the integrating sphere effect, and might cause a significantly higher light dose than the intended dose without compensation it. The increased irradiance depends on the optical characteristics (reflection factor $\rho$ ) of the bladder tissue; however, there can be large variations of this parameter between different bladders.¹⁸

The total fluence rate $\varphi$ inside the bladder is the sum of the input light and the multiple reflected light:¹⁸

Light is necessary for the photochemical processes to produce ROS and cause cell damage. However, cell damage may also be caused by factors other than the ROS. Thermal damage or heat-induced necrosis (burns) can also contribute to the resulting cell damage after illumination. Thus, the temperature increase in the bladder during the illumination process must be considered. The temperature will typically reach a steady-state value after $5\text{minutes}$ of exposure. To avoid thermal damage (and pain), the temperature should not exceed $45°\mathrm{C}$ .¹⁹ It is therefore necessary to evaluate the temperature increase in the bladder during the illumination process. The steady-state temperature distribution can be found from the bioheat equation.²⁰

Worst-case calculations (all light absorbed, no blood perfusion) for the fluence rate used in this experiment give a less than $3°\mathrm{C}$ temperature increase, which does not exceed a tolerable level.

2.2.

Simulations

A mathematical model was used to simulate the measured spectra in order to extrude more information. The model is an analytical photon transport model based on the diffusion approximation²¹ and adapted to fit bladder parameters. The model consists of two flat layers to represent the urothelium and submucosa on top of a semi-infinite slab that represents the muscular layer and underlying tissue. Data for the optical properties of the bladder were found in the literature.^{22, 23} The boundary between the water and the urothelium was represented by refraction indices $n_{\text{water}}=1.33$ and $n_{\text{urothelium}}=1.36$ , and the anisotropy factor $g$ for the tissue scattering was set to $g=0.93$ at $630\mathrm{nm}$ . The blood hematocrit for rats was set to 47%.²⁴

Due to the short penetration length of light below $700\mathrm{nm}$ , most of the reflected light was from the bladder layers. With the dimensions kept fixed, the blood volume (BV) fraction and the oxy-, deoxy,- and methemoglobin content in the model were iterated to yield the best fit with the measured spectrum. The bladder dimensions used for analysis were $75\mu \mathrm{m}$ for the urothelium and submucosa, and $300\mu \mathrm{m}$ for the muscle layer.

3.1.

Chemicals

RPMI-1640 medium, L-glutamine, fetal calf serum (FCS), fungizone, penicillin-streptomycine, trypsin, and phosphate buffered saline (PBS) were obtained from Gibco BRL, Life Technologies (Inchinnan, Scotland). Gentamicin sulphate was obtained from Schering Corp (Kenilworth, NJ), and HAL was supplied by Photocure ASA (Oslo, Norway). Ultra-purified water, Milli-Q PLUS (Millipore, Molsheim, France) was used throughout the study.

3.2.

Animals and Tumor Cell Line

Female Fischer F344 rats (purchased from Möllegard, Skensvend, Denmark) with a body mass of $170\text{to}230\text{grams}$ were used for all experiments. The animals were anaesthetized with a mixture of Haldol (16.7%), Fentanyl (25%), and Midazolan (25%) in water, $0.4\mathrm{ml}∕100\mathrm{g}$ .

The AY-27 TCC cell line was grown in RPMI-1640 medium containing 10% v/v FCS, L-glutamine $(80\mathrm{mg}∕\mathrm{l})$ , penicillin $(100\mathrm{U}∕\mathrm{ml})$ , streptomycin $(100\mathrm{U}∕\mathrm{ml})$ , and fungizone $(0.25\mathrm{mg}∕\mathrm{ml})$ . The cancer cells were subcultured approximately twice a week and grown at $37°\mathrm{C}$ in a humidified atmosphere containing 5% $\mathrm{C}{\mathrm{O}}_2$ . A preliminary study on AY-27 cell suspensions confirmed that they produced PpIX endogenously when exposed to HAL. The cells were seeded in medium at a density of $0.30\text{to}0.45\times {10}^6$ in $75{\mathrm{cm}}^2$ cell culture dishes. After anaesthetization and catheterization, the rats received an acid/lye wash where the bladder was filled with HCl ( $0.3\mathrm{ml}$ , $0.1\mathrm{M}$ ) for $15\text{seconds}$ followed by NaOH neutralization ( $0.3\mathrm{ml}$ , $0.1\mathrm{M}$ , $15\mathrm{s}$ ), then washed with PBS $(3\times 0.6\mathrm{ml})$ before AY-27 cell instillation ( $4\times {10}^6∕\text{rat}$ , $1\mathrm{h}$ ) as described by Xiao ²⁵ To ensure homogenous distribution of the AY-27 cells inside the bladder, animals were rotated $90\mathrm{deg}$ every $15\text{minutes}$ during the one-hour instillation period.

To determine the optimal point in time for PDT treatment of AY-27 urinary bladder TCC, a pilot study was performed with histological examinations on days 7, 14, and 28 after cell instillation. The results showed that the thickness of the bladder wall was best suited for PDT treatment at $14\text{days}$ after instillation (data not shown).

All animals showed normal behavior (eating, urinating, activity) during the test period of $21\text{days}$ ( $28\text{days}$ for pilot study).

3.3.

PDT Treatment

On day 14 after AY-27 instillation, the animals were anaesthetized and catheterized with an $18\mathrm{G}$ catheter. The HAL solution ( $8\mathrm{mM}$ , $0.3\mathrm{ml}$ ) was instilled in the bladder for $1\text{hour}$ (animals were rotated $90\mathrm{deg}$ every $15\text{minutes}$ ). The light treatment was performed two hours after the end of HAL instillation, corresponding to the peak PpIX accumulation in the bladder tissue.¹⁷

The irradiation source consisted of an Argon-pumped (Innova 70, Coherent, Santa Clara, CA) dye laser (model 375, Spectra-Physics, Mountain View, CA) centered at $635\mathrm{nm}$ . Light was coupled into a $200\text{-}\mu \mathrm{m}$ fiber applicator with an isotropic diffuser tip (PDT Systems, Santa Barbara, CA). The applicator was inserted through the $18\text{-}\mathrm{G}$ catheter to illuminate the bladder, as shown in Fig. 1 . The fluence rate was set to $20\mathrm{mW}∕{\mathrm{cm}}^2$ for all animals (and corrected for the integrating sphere effect), then measured with an integrating sphere setup and a laser power meter (FieldMaxII-TOP, Molectron Corp., Sunnyvale, CA). Total light fluence was $20\mathrm{J}∕{\mathrm{cm}}^2$ . The bladders were filled with $0.3\mathrm{ml}$ saline to keep them distended during irradiation. All animals were kept anaesthetized and laid on a warming pad to maintain body temperature during the treatment period.

Fig. 1

Photograph showing rat with fiberoptic probe for measurement (left), and rat with fiber applicator during light irradiation (right).

3.4.

In Vivo Optical Spectroscopy

A forward-viewing fiber-optic probe was constructed to perform optical reflection and fluorescence spectroscopy of the rat bladders in vivo. The probe was constructed of three $200\text{-}\mu \mathrm{m}$ optical fibers (FG-200-LCR/FT-200-EMT, ThorLabs, Newton, NJ), two illumination fibers, and one collection fiber. Fibers were mounted side by side in a triangle configuration, inserted in a $20\text{-}\mathrm{G}$ needle as shown in Fig. 2 , and cemented in place using UV-curing glass adhesive.

Fig. 2

Illustration of the forward-viewing fiberoptic probe, showing the three fibers protruding from the $20\text{-}\mathrm{G}$ needle (not to scale).

The length of the probe was constructed to be $1\mathrm{mm}$ longer than the $18\text{-}\mathrm{G}$ catheter when fully inserted to ensure the same placement in the bladder and to standardize the measurements. However, the fiber placement and tip-to-wall distance was also dependent on the placement of the catheter and the geometry of the bladder. To ease handling, the fiber pigtails were placed inside protective jackets. The fibers were terminated with SMA 905 connectors compatible with the measurement equipment. A preliminary study showed that if the probe was placed in direct contact with the bladder wall, the increased perfusion due to the mechanical irritation was visible in the reflectance spectra. However, light contact between the probe and tissue did not cause injury to the bladder lining. Erroneous measurements due to practical difficulties in some animals (bladder contraction, blood in bladder) were excluded.

The measurement setup consisted of a fiber-coupled spectrometer (SD2000, Ocean Optics, Duiven, The Netherlands), a white light source (LS-1, Ocean Optics), and a pulsed $337\text{-}\mathrm{nm}$ ( $6\mathrm{mW}$ at $20\mathrm{Hz}$ ) nitrogen laser (VSL-337ND, LaserScience Inc., Franklin, MA). Each one was coupled to a fiber on the probe. Before the reflectance measurements, the fiber probe was calibrated by inserting it into a small cylinder that was placed in close contact with a Spectralon reflectance standard (Ocean Optics). The cylinder ensured that no ambient light entered the detection system, and it kept a fixed distance between the probe tip and the standard. For fluorescence measurements, the fiber probe was positioned to maximize the fluorescence. The animals were covered by black felt to avoid interference from other light sources during measurements. Due to the small size of the urethra, the fiber probe had to be removed when the illumination fiber applicator was inserted, then reinserted after the laser treatment process. Hence, the first spectra recorded after laser treatment were collected within two minutes after illumination was terminated. The total detected fluorescence was the sum of the fluorescence contribution from PpIX and that of tissue autofluorescence. The PpIX fluorescence intensity $\left(I_f\right)$ was calculated by normalizing spectra at $600\mathrm{nm}$ , and subsequently fitting a polynomial line between 600 and $660\mathrm{nm}$ as a baseline to represent tissue autofluorescence. The PpIX fluorescence was defined as the area above this baseline (see Fig. 3 ) in the interval between 615 and $655\mathrm{nm}$ (peak at $635\mathrm{nm}$ ).

Fig. 3

Representative fluorescence spectra from one animal in group VI showing both pre- (line) and post- (dash) laser treatment. The main PpIX fluorescence peak at $635\mathrm{nm}$ (vertical, dot) and the calculated baseline (dash-dot) for tissue autofluorescence are also shown.

3.5.

Oxygen Saturation

Hemoglobin is the most evident chromophore in perfused tissue. Various species of hemoglobin can be identified from the reflectance spectra, with oxygenated and deoxygenated hemoglobin being the two most common. Thus, knowing the absorbtion spectra of these chromophores makes it possible to calculate the oxygen saturation (OS) in the tissue,

3.6.

Treatment Protocol

The animals were divided into six groups, and each group received the treatment shown in Table 1 . Three additional groups—no treatment $(n=3)$ , only-HAL treatment $(n=3)$ , and only light treatment $(n=2)$ —were tested as controls.

Table 1

Treatment protocol, where + denotes treatment received.

Photosensitizer and light doses were kept equal for each animal: $8\mathrm{mM}$ of HAL combined with a light fluence of $20\mathrm{J}∕{\mathrm{cm}}^2$ , the amount shown to be optimal by Khatib for treatment of superficial bladder cancer.¹⁷

Reflectance and fluorescence spectra were recorded for all groups shown in Table 1.

3.7.

Statistical Analysis

Nonparametric methods were used to evaluate the fluorescence data due to a marked heterogeneity of variance between the different groups (the Levelene’s test for equality of variance). Comparisons between groups of interest were carried out using Kruskal-Wallis (several groups) and Mann-Whitney U (two groups) tests, respectively. The changes in fluorescence level before and after laser treatment were evaluated using the Wilcoxon test. Parametric tests were used for normally distributed variables (OS values) using an unpaired t-test to compare between two groups, and paired sample t-tests to analyze for significant changes before and after laser treatment.

Non-normally distributed variables were expressed as medians (25th and 75th percentiles) and normally distributed as means ± standard deviation (SD). Significance was defined as a p-value of less than 0.05 (two-tailed).

3.8.

Histological Examination of Bladders

The animals were sacrificed $7\text{days}$ after the PDT treatment. Each bladder was collected, instilled with buffered 4% formaldehyde for $24\mathrm{h}$ , and embedded in paraffin. Sections were cut through each tissue block at $4\text{-}\mu \mathrm{m}$ thickness and stained with hematoxylin, eosin, and saferin (H.E.S) for histological examination. Sections were evaluated at low magnification $(10\times )$ for the entire section, and at high magnification $(40\times )$ for details of tissue destructions. The profile area $\left({\mathrm{mm}}^2\right)$ of the tumor was measured by software (Nikon NIS-Elements Basic Research, Inter Instrument AS, Høvik, Norway). The evaluation was performed by a histologist blinded to the sample source.

4.1.

Fluorescence Spectra

Fluorescence measurements on group VI animals revealed that PpIX fluorescence decreased following laser treatment. A representative PpIX fluorescence spectrum from pre- and post-laser treatment for one group VI animal is shown in Fig. 3. Pre-laser treatment PpIX fluorescence at $635\mathrm{nm}$ was observed in all groups as shown in Fig. 4 (differences in number of animals $n$ from Table 1 are due to erroneous measurements that are excluded). PpIX fluorescence was also observed in HAL-treated animals despite the fact that they did not receive the AY-27 cells.

Fig. 4

Individual PpIX fluorescence values (%) before the laser treatment for groups I through VI.

Figure 5 represents the pre-/post-PpIX fluorescence for groups III through VI. There was a significant decrease in PpIX fluorescence before and after laser treatment for group IV $(p=0.031)$ , group V $(p=0.008)$ , and group VI $(p<0.0005)$ . Fluorescence measurements $24\mathrm{h}$ after PDT in group VI animals did not show any sign of PpIX in the bladders.

Fig. 5

PpIX fluorescence (%) pre- (dashed) and post- (solid) laser treatment for groups III through VI. The boxplot shows the 25th, 50th (median), and 75th percentiles, and the vertical lines outside the box represent the largest and smallest values.

4.2.

Reflectance Spectra

Figure 6 shows reflectance spectra from groups V $(\mathrm{HAL}+\text{light})$ and VI animals compared to a control animal (no treatment). The characteristic absorption peaks of oxyhemoglobin at 542 and $576\mathrm{nm}$ were clearly visible in the spectra recorded before the laser treatment. These peaks were not as evident in the spectrum recorded after laser treatment. Oxygen saturation of the bladder tissue was calculated from the reflectance spectra. No significant OS differences were observed in animals with only acid/lye washing before and after laser treatment. However, OS was slightly decreased in group IV animals (acid/lye $\text{wash}+\mathrm{HAL}+\text{light}$ ) after laser treatment, while it slightly increased after laser treatment for group V animals $(\mathrm{HAL}+\text{light})$ . None of the animals in these groups had carcinomas in the bladder. In contrast, all group-VI animals showed a clear decrease in tissue oxygenation following the laser treatment.

Fig. 6

Representative reflectance spectra of group V and VI animals.

Figure 7 shows the oxygenation data calculated from the reflectance spectra. There was a significantly lower OS in animals implanted with AY-27 cells (groups I through III and VI) compared to animals without AY-27 cells (groups IV and V). Mean values were $64.5\pm 13.4\%$ and $72.5\pm 11.6\%$ , respectively $(p=0.047)$ . Oxygen saturation significantly decreased in group VI after laser treatment $(p=0.0001)$ , but no significant difference was observed in the remaining groups. Methemoglobin (MetHb) was identified from the reflectance spectra of certain group VI animals. These animals also had the most marked deoxygenation of the bladder tissue after laser treatment.

Fig. 7

Oxygen saturation (%) from pre-laser treatment (groups I and II), and from pre- and post-laser treatment (groups III through VI), presented as $\text{mean}\pm \mathrm{SD}$ .

4.3.

Macroscopic Examination

Macroscopic inspection of the cancerous bladders revealed that the cells had attached to the bladder wall and cancer had developed as islands of tumor cells inside the bladders. In a preliminary experiment, fluorescence measurements were performed on excised bladders that had cancer and received HAL instillation. PpIX fluorescence was not detected outside the cancerous islands (data not shown), confirming the high selectivity of HAL.

There were variations in the cancer masses that developed between different animals. The mean mass of cancerous bladders was $0.179\mathrm{g}$ , compared to $0.098\mathrm{g}$ for healthy bladders—an increase of 82.3%.

4.4.

Histopathological Findings of the Urinary Bladder TCC

Tumors were found on the wall of the urinary bladder from all 6 rats in the AY-27 cell implantation group without PDT treatment [group I, Fig. 8a ], and in 3 of 6 in the PDT-treated AY-27 cell implantation group [group VI, Fig. 8b]. The tumors were characterized by displaying high mitotic activity, high nuclear-to-cytoplasmic ratio, and marked nuclear pleomorphism (see Fig. 8). Most tumors (from both groups) appeared intraepithelial, and sometimes papillary tumors and/or lymphocytic and mononuclear cell infiltration around the tumor could be observed. The tissue outside of the tumor islands appeared to be normal in the untreated group. In the PDT-treated AY-27 implantation bladders, healed necrosed urothelium with complete regeneration of urothelium was observed in all rats. In the 3 of 6 bladders with remaining tumors, partial tumor damage was seen and the tumor area was severely reduced (see Table 2 ). No tumors were found in the control or PDT-treated normal rats (group V).

Fig. 8

Representative micrographs showing transitional cell carcinoma in bladders 21 days after AY-27 cell implantation. (a) No PDT treatment; (b) 7 days after PDT treatment. (H.E.S. staining, photos taken at 10×.)

Table 2

TCC histological findings. Means±SD.

Differences in fluorescence intensity may partly explain the differences in tumor destruction of group VI. The 3 of 6 fully responsive animals in group VI corresponded to the animals that exhibited the largest prelaser-treatment PpIX fluorescence ( $I_f=206\%$ , 242% and 273%, mean $240\pm 33\%$ ). The other 3, not fully responsive animals did not exhibit as high of a pretreatment fluorescence ( $I_f=126\%$ , 198% and 148%, mean $157\pm 36\%$ ).

4.5.

Simulation Results

Simulated spectra suggested a large BV increase in the top layers of the model following HAL instillation, and this was largely independent of the various layer thicknesses. A measured simulated spectra pair from a representative group VI animal is shown in Fig. 9 . The dip in the reflection spectrum at $635\mathrm{nm}$ was only reproduced in the simulated spectrum by the introduction of MetHb.

Fig. 9

Measurement of group VI animal (post-laser treatment) and corresponding simulated spectrum ( ${\mathrm{OS}}_{\text{urothelium}}={\mathrm{OS}}_{\text{submucosa}}=20\%$ , ${\mathrm{OS}}_{\text{muscularis}}=50\%$ , ${\mathrm{BV}}_{\text{urothelium}}=26\%$ , ${\mathrm{BV}}_{\text{submucosa}}=40\%$ , ${\mathrm{BV}}_{\text{muscularis}}=20\%$ , $\mathrm{MetHb}=40\%$ ).