2.1.

Photodynamic Therapy System and Dosimetry

The system, SpectraCure P18 with IDOSE, was functionally similar to previously described instruments.^9,13,14 Eighteen individually controlled diode lasers, operating at 652 nm, were used as light sources; however, in the present study, only seven lasers were used because of the small size of the dog prostate. In monitoring mode, one fiber at a time was used as the emitter while the six closest surrounding fibers were used to collect and measure scattered light. With this setup, it was possible to monitor the optical properties of the tissue. The output power from each bare-end fiber was calibrated to 150 mW prior to each PDT session. Bare-end fibers were used because as point sources they provided both optimal light fluence measurement capability and a flexible means to shape the dose field.

A more detailed description of the dosimetry platform, IDOSE, is found in Refs. 15 and 16. Briefly, the three-dimensional (3-D) model of the prostate and surrounding tissues generated from transrectal ultrasound images was used as input for the calculation of optimum fiber positions, which were then presented to the user. Typical average optical properties of prostate tissue were used in the optimization, which was based on a stochastic method that positions the fibers in the 3-D prostate model with 2-mm spatial resolution, with constraints concerning organs at risk. All dose-related computations in IDOSE were based on the steady state diffusion equation of radiative transport; and thus, it was assumed that light propagation in the tissue takes place in the optically diffuse regime. When the optical fibers were in place the first monitoring sequence was performed. The optical properties were evaluated by assuming that the tissue was locally homogeneous around the seven optical fibers, where each fiber was sequentially used as emitter and the other fibers as receivers. The effective attenuation coefficient, ${\mu }_{\mathrm{eff}}$, was derived by fitting the data with the diffusion equation. The evaluated optical properties were used to compute a dose plan that was presented to the user, who then had to approve the start of the light delivery. The doses were determined by the illumination time of each fiber. The light delivery was interrupted at intervals for monitoring sequences, and the dose plan was updated in case significant changes had occurred. The dose planning algorithm aimed to deliver a minimum threshold light dose to all parts of the target volume, while attempting to avoid giving more than the threshold dose to surrounding organs at risk.

2.2.

Study Design

The study had a sequential design in which groups of dogs were subjected to one threshold light dose. The results were evaluated, and the threshold dose was increased for the next group to a value based on the results of the previous group. In the first group, online dosimetry was not used, which allowed some comparison between the results when using online dosimetry and not using online dosimetry.

The study comprised nine clinically healthy intact male research Beagles, aged 15 to 33 months and weighing 14 to 20 kg. They were divided into three groups (three dogs in each group). They were housed at the Department of Clinical Sciences, Swedish University of Agricultural Sciences in Uppsala, Sweden. Food was provided once a day and water was freely available to the dogs. The dogs were observed daily by the same veterinarian to monitor the animals’ general condition. Seven days after PDT, the dogs were euthanized with pentobarbital intravenously (IV) and subjected to autopsy.

The study was approved by the Uppsala Animal Ethics Committee, Sweden.

2.3.

Photodynamic Therapy

Three days before PDT, temoporfin (Foscan, Biolitec Pharma, Jena, Germany), $0.15\mathrm{mg}/\mathrm{kg}$, was administered by slow IV injection in vena cephalica during a 10 min span. Following administration of temoporfin, the ambient light was subdued due to the photosensitivity caused by temoporfin.

The dogs were premedicated 30 to 40 min before induction of anesthesia with acepromazine (Plegicil vet, Pharmaxim, Helsingborg, Sweden) $0.3\mathrm{mg}/10\mathrm{kg}$, administrated intramuscularly, and methadone (Metadon Recip, Solna, Sweden) $2\mathrm{mg}/10\mathrm{kg}$, administrated subcutaneously.

Anesthesia was induced using propofol (Rapinovet vet, Schering-Plough A/S, Skovlunde, Denmark), administrated IV. After intubation, the anesthesia was maintained with isoflurane inhaled in 2% mixture of oxygen and air. All dogs breathed spontaneously.

When the animals were anesthetized, the surgical field ventral of the anus (i.e., perineum) was aseptically prepared. They were placed in the operating room in a dorsal recumbency. In order to create the desired angle of the pelvis for access to the prostate, the dogs’ hind legs were stretched cranially and medially and tied together ventrally to the mid-abdomen.

Transrectal ultrasound (2101 Falcon EXL, B-K Medical, Herlev, Denmark), frequency range 5 to 7.5 MHz, was used to acquire images of the prostate and surrounding organs, and to create a pretreatment dose plan. Seven 18-gauge needles were inserted according to the dose plan transperineally under ultrasound guidance. Once the needles with optical fibers were in place, light delivery and dosimetry were performed automatically by the SpectraCure P18 system without any user intervention except approval before the start of the light delivery. The light delivery was interrupted by the system for measurements after 2, 4, and 9 min, and then every 5 min for the duration of the treatment.

The number of fibers, 7, was chosen as the minimum number of fibers to acquire useful measurements for the online dosimetry. More fibers could not be used due to the small size of the prostate of the dogs.

2.4.

Light Threshold Dose

The light threshold dose starting point was $5\mathrm{J}/{\mathrm{cm}}^2$, estimated from the lesion sizes reported by Chang et al.^24,25 In that study, typical lesion sizes resulting from PDT had radii of 9 to 11 mm for a light energy of 100 J (after selecting only the cases that were treated interstitially, were euthanized after 7 days, and showed clear, nonconfluent lesions from individual fibers). Assuming optical properties similar to human prostate,²⁶ this yields a threshold dose at the border of the lesions in the interval 3 to $18\mathrm{J}/{\mathrm{cm}}^2$. The value $5\mathrm{J}/{\mathrm{cm}}^2$ is in the lower end of this interval. This estimate was further verified by comparing with the recommended light dose for superficial illumination approved by the European Medicines Agency, $20\mathrm{J}/{\mathrm{cm}}^2$ ($\text{irradiance}\times \text{illumination time}$). Assuming typical tissue optical properties and a depth of necrosis of 5 mm, this yields a threshold dose of 6 to $10\mathrm{J}/{\mathrm{cm}}^2$ (fluence), which is in the same range. [Note the distinction between the physical properties “irradiance” (to describe the light power incident on a surface) and “fluence rate” (to describe the light power per unit area at a location inside a turbid medium) although they have the same unit, $\mathrm{W}/{\mathrm{cm}}^2$.] Based on these two independent estimates, $5\mathrm{J}/{\mathrm{cm}}^2$ was chosen as a conservative starting point.

The first group of three dogs was given a fixed light energy of $63\mathrm{J}/\text{fiber}$, which approximately corresponded to $5\mathrm{J}/{\mathrm{cm}}^2$ at the boundary of the prostate. Online dosimetry was used in the subsequent groups. After the results of group 1 had been evaluated, the $5\text{-}\mathrm{J}/{\mathrm{cm}}^2$ level was clearly too low to achieve the aim of ablating the entire prostate, and the target threshold dose was set to $10\mathrm{J}/{\mathrm{cm}}^2$ for group 2. Dog 6 was used as a control and received temoporfin and was subjected to fiber insertion but no light. The target threshold dose was finally increased to $20\mathrm{J}/{\mathrm{cm}}^2$ for the third group, based on the observation that even the $10\mathrm{J}/{\mathrm{cm}}^2$ target threshold used for the second group was not enough to achieve total ablation of the prostate.

2.5.

Magnetic Resonance Imaging

The dogs were premedicated with dexmedetomidin (Dexdomitor, Orion Pharma Animal Health, Sollentuna, Sweden) $0.3\mathrm{mg}/\mathrm{kg}$ and butorfanol (Dolorex vet, Intervet, Stockholm, Sweden) $0.2\mathrm{mg}/\mathrm{kg}$.

Magnetic resonance imaging (MRI) of the prostate was performed before administration of temoporfin and on day 7 after PDT. The dogs were examined in lateral recumbency in a 0.27 T permanent, C-shaped magnet (Hallmarq Veterinary Imaging, Guildford, United Kingdom). The sequences were T1-weighted spin echo (T1W SE) pre- and postcontrast and T2-weighted fast spin echo (T2W FSE). For the contrast examination, each dog received $0.2\mathrm{mL}/\mathrm{kg}$ gadodiamid (Omniscan, GE Healthcare, Chalfont St. Giles, United Kingdom) administered by IV injection. Directly after the second MRI examination, the dogs were euthanized with pentobarbital IV and subjected to autopsy.

2.6.

Histopathology

At necropsy, the prostate, adjacent rectum and bladder was measured and weighed and then fixed in 10% neutral buffered formalin. For macroscopic inspection, the prostate tissue was sectioned from the cranial part to the caudal end in 4- to 6-mm thick blocks transversally cut. Each block was evaluated macroscopically and then embedded in paraffin. Sections (5- to $7\text{-}\mu \mathrm{m}$ thick) from each block were processed for histological examination using hematoxylin and eosin staining and van Gieson staining. Tissues adjacent to the prostate, i.e., the bladder and colon, were collected and processed for histological examination.

2.7.

Blood Sampling

Blood samples were collected from all animals before administration of temoporfin and at $\sim 1$, 3, 6, 24, 48, and 72 h postadministration. The samples were shipped on dry ice for high-performance liquid chromatography (HPLC) analysis by Biolitec AG, Germany.

3.1.

Gross Pathology and Histopathology

Within the prostate tissue, there were changes of various degree and location in all dogs except dog 6 (control). Dogs 1, 2, 3, and 9 had lesions characterized by hemorrhages in the interstitium and only a few small necrotic foci within the prostate parenchyma. These necrotic areas were located in the center of the hemorrhagic zones. In dogs 4 and 5, several foci of necrosis ranging from a few mm up to 13 mm in diameter were found. Figures 1(a)–1(c) illustrate an example from dog 4. These necrotic areas were surrounded by hemorrhagic zones that in many areas extended into the capsule. In dogs 7 and 8, large necrotic lesions were found in the glandular and interstitial parts of the prostate. Figures 1(d)–1(f) show an example from dog 7. These lesions were surrounded by hemorrhagic zones.

Fig. 1

Necrotic areas and lesions. (a) Transverse slices from dog 4. The contours denote the regions of necrosis. (b and c) Dog 4. H&E-stained histological slides. (b) Low magnification shows large necrotic areas with hemorrhages that extend into the capsule. (c) High magnification, illustrating coagulation necrosis with hemorrhages and only few inflammatory cells. (d) Transverse slices from dog 7. The contours denote the remaining regions of healthy tissue. (e and f) Dog 7, H&E stained histological slides. (e) Low magnification, illustrating a distinct border between the necrotic areas to the right and the viable glandular tissue to the left. (f) High magnification, coagulation necrosis with hemorrhages and few inflammatory cells.

In all dogs except dog 6 (control), there were changes in the prostate capsule and the periprostatic tissue. These changes were characterized by hemorrhages, edema and necrosis of fat and loose connective tissue. In addition, there were fibrin accumulation and inflammatory cells (mainly lymphocytes and macrophages) admixed with few fibroblasts in these areas.

Dogs 1, 2, 3, and 7 had multifocal petechial hemorrhages in the urinary bladder mucosa. In the rectum/colon part, direct dorsally of the prostate lesions were found in dogs 1, 3, 7, 8, and 9. These lesions included necrosis of mucosal, submucosal, and muscular layers with vasculitis and thrombosis. Comparison of metric data before and after paraffin embedding showed that the shrinkage of the tissue was between 5% and 8%.

3.2.

Magnetic Resonance Imaging

Multiple intensity changes were seen in post-PDT scans both in the prostatic and the periprostatic tissues. All dogs showed similar intensity changes in the prostate compared to pre-PDT with a diffuse mixed intensity in T1W SE, marked contrast enhancement in T1W SE, and slight to moderate hyperintensity compared to surrounding fat in T2W FSE scans. The changes were generally distributed in the prostate in dogs 1 to 3 and 7 to 9, whereas dogs 4 and 5 had focal changes. See Fig. 2 for examples from dogs 4 and 7.

Fig. 2

Magnetic resonance T1W SE images (time to echo 18 ms, time to repetition 400 ms) of (a, b) dog 4, (c, d) dog 7, and (e) dog 3. (a and b) Dog 4, post-PDT. The precontrast image shows areas of mixed intensity in both left and right prostatic lobes that in postcontrast scans (b) are hyperintense with semicircular areas of signal void. (c and d) Magnetic resonance T1W SE contrast-enhanced images of dog 7 pre- and post-PDT. (d) The post-PDT image shows changes in the entire prostate with hyperintense parenchyma surrounding large areas of signal void. (e) Magnetic resonance T1W SE contrast-enhanced image of dog 3 post-PDT. Semicircular area of signal void representing hemorrhages and necrosis surrounds the right prostatic lobe. No areas of signal void in the prostatic tissue. SIN: sinister, left.

Areas of signal void in contrast-enhanced T1W SE scans were seen in dogs 1 (small focal area), 2 (three small focal areas), 4 (two large areas), 5 (two large areas), 7 and 8 (multiple large areas). No areas of signal void were seen in dogs 3 and 9.

In all dogs that had PDT, the prostatic capsules were irregular and diffuse in outline and the periprostatic fatty tissue contained multiple irregular hypointense linear structures. Dogs 1, 3, 7, and 8 also had semicircular areas of signal void in contrast-enhanced T1W SE scans along the periphery of the right prostatic lobe [see Fig. 2(e)].

Similar intensity changes in the ventral rectum were seen in dogs 1, 3, and 9, which were hypointense in T1W SE, hyperintense in T2W FSE and showed signal void in contrast-enhanced T1W SE scans. The border of the rectal lesion gradually became indistinguishable and on scans cranial to the lesion, it was not possible to distinguish between intensity changes in the periprostatic tissue and the lesion in the rectal wall. Dog 8 showed contrast enhancement in the ventral rectal wall and nearby muscular tissue but no signal void.

3.3.

Dosimetry

The performance of the dosimetry was assessed by analysis of light dose plans and measurement data. Isodose maps were calculated based on measurements and delivered light doses. The isodose maps were two-dimensional plots representing transverse slices taken from the tissue 3-D model, where isodoses represented the areas of the tissue that received above or below certain levels of light fluence dose. The isodose maps were compared with the pathology results to determine the light dose level at which tissue necrosis occurred. The dose maps were rescaled to fit the pathology slides to account for the shrinking of the tissue. The transverse planes were coregistered with the histopathology samples by observing the position of the histopathology cut in relation to the cranial and caudal ends, and selecting the corresponding plane in the 3-D tissue model.

As previously noted, in dogs 1 to 3, the online dosimetry was not active. Instead, a fixed light energy of $63\mathrm{J}/\text{fiber}$ was delivered. Post-PDT analysis based on the measurements was performed, and light dose maps were reconstructed and compared with the histopathology results. This comparison indicated that the light dose at the perimeter of the necrotic regions was $\sim 20$ to $30\mathrm{J}/{\mathrm{cm}}^2$.

In dogs 4, 5, 7–9, the online dosimetry was active. Examples of the resulting dose maps for dogs 4 (threshold dose $10\mathrm{J}/{\mathrm{cm}}^2$) and 7 (threshold dose $20\mathrm{J}/{\mathrm{cm}}^2$) are shown in Fig. 3. The dose maps were compared with the histopathology results [cf. Figs. 1(a) and 1(d) for dogs 4 and 7, respectively]. This again was consistent with a threshold dose to induce necrosis of $\sim 20$ to $30\mathrm{J}/{\mathrm{cm}}^2$, in all dogs except in dog 9, where the light dose needed to achieve necrosis was much higher at over $100\mathrm{J}/{\mathrm{cm}}^2$. There was no obvious difference between dog 9 and the other dogs from a dosimetry perspective, except that the prostate in dog 9 was smaller.

Fig. 3

Dose maps from (a) dog 4 and (b) dog 7, corresponding to the same transverse slices, as shown for the histopathology in Fig. 1. Axes scale: mm. Tissue model: light gray—prostate; medium gray—urethra; dark gray—rectum.

The light illumination times and delivered light doses are shown in Table 1, where $t_{\min }$ denotes the shortest illumination time of the seven fibers, $t_{\text{average}}$ the average time, and $t_{\text{long}}$ the longest. The total light dose delivered to each dog is also shown. The measured effective attenuation coefficients, ${\mu }_{\mathrm{eff}}$, of the prostate tissue are presented in Fig. 4. For each dog, the average ${\mu }_{\mathrm{eff}}$ value for every monitoring sequence is shown together with the standard deviation. The interfiber distances were in the range of 5 to 25 mm in all cases.

Table 1

PDT illumination times and light doses.

Fig. 4

The average effective attenuation coefficient at different monitoring times during PDT for each dog. The average and standard deviation of the effective attenuation coefficient at each monitoring time is calculated from the unique values of effective attenuation coefficient for each treatment fiber as assessed by the PDT instrument during therapy.

Dose–volume histograms for dogs 4 and 7 are shown in Fig. 5, as examples from one dog in group 2 and one in group 3. As a summary of the results for all dogs, the percentage volume of prostate tissue that received the target threshold dose is shown for all dogs in Fig. 6.

Fig. 5

Dose–volume histograms for (a) dog 4 and (b) dog 7. The threshold dose was set at $10\mathrm{J}/{\mathrm{cm}}^2$ for dog 4 and $20\mathrm{J}/{\mathrm{cm}}^2$ for dog 7.

Fig. 6

The volume fraction of the prostate that received the target threshold dose for all dogs.

3.4.

Pharmacokinetics

All plasma samples were investigated by HPLC analysis to specify the temoporfin content. The maximum plasma concentration in all animals was $\sim 0.6\mathrm{ng}\text{temoporfin}/\mathrm{mg}$ wet weight, measured 1 h after injection. The average elimination half-life of temoporfin was 18 h. This value is similar to that reported in dogs by Buchholz et al.²⁷