John Trachtenberg, Princess Margaret Hospital, 620 University Ave., 4th, 926 Toronto, Ontario, ON, Canada M5G2M9. e-mail: email@example.com
To report on the efficacy of TOOKAD® (WST 09; NegmaLerads, Magny-Les-Hameaux, France) vascular-targeted photodynamic therapy (VTP) as a method of whole-prostate ablation in patients with recurrent localized prostate cancer after the failure of external beam radiotherapy (EBRT).
PATIENTS AND METHODS
Patients received a fixed photosensitizer dose of 2 mg/kg and patient-specific light doses as determined by computer-aided treatment planning. Up to six cylindrical light-diffusing delivery fibres were placed transperineally in the prostate under ultrasonographic guidance. The treatment response was assessed by measuring serum prostate-specific antigen (PSA) levels, lesion formation (avascular areas of tissue) measured on 7-day gadolinium-enhanced T1-weighted magnetic resonance imaging (MRI) and a 6-month biopsy.
Treatment of the whole prostate was possible with minimal effects on surrounding organs. An increased light dose improved the tissue response, with MRI-detectable avascular lesions, encompassing up to 80% of the prostate in some patients. A complete response, as determined by the 6-month biopsy, required that patients received light doses of at least 23 J/cm2 in 90% of the prostate volume (D90 > 23 J/cm2). Of the 13 patients who received at least this light dose, eight were biopsy-negative at 6 months. In this group of eight patients, PSA levels decreased and did so to negligible levels for those patients with a baseline PSA level of <5 ng/mL. Side-effects were modest and self-limited in most patients; there were recto-urethral fistulae in two patients, one of which closed spontaneously.
TOOKAD-VTP can produce large avascular regions in the irradiated prostate, and result in a complete negative-biopsy response at high light doses. A response rate of more than half for those patients receiving the highest light doses shows the clinical potential of TOOKAD-VTP to manage recurrence of prostatic carcinoma after EBRT.
External beam radiotherapy (EBRT) is the primary treatment option for ≈30% of patients with prostate cancer, but the recurrence rates after EBRT are high. The disease-free survival for even the lowest-risk patient group ranges from 92% to 65% at the 3- and 4-year follow-up, decreasing to 64% at 6 years  and to 30.1% at 20 years . Considering these rates of recurrence after EBRT, safe and potentially curative salvage therapies are needed. Current curative treatment options, such as salvage radical prostatectomy, cryotherapy, and high-intensity focused ultrasound, all have the limitations of significant morbidity with variable efficacy. While hormonal therapy remains the mainstay for most of these patients, it does not offer a chance of cure and has incumbent significant morbidity. As yet, there is no salvage therapy option that offers the chance of cure with minimal side-effects.
We are investigating the use of light-activated drugs (photosensitizers)  as a possible minimally invasive treatment for localized EBRT-recurrent prostate cancer. There are currently few other reports on the use of photodynamic therapy (PDT) to treat patients with prostate cancer [6,7]. Apart from an early study  in which the photosensitizer porfimer sodium (Photofrin) was used in only two patients, there are two groups who reported phase I/II trials of PDT in patients with localized disease after radiotherapy. In 14 patients with recurrent prostate cancer after EBRT, Nathan et al. used meso-tetrahydroxyphenyl chlorine (mTHPC), a photosensitizer that is activated at 652 nm, to treat the whole organ. While there was some morbidity they suggested that with improved dosimetry, the whole organ could be treated, with no complications to surrounding organs. Du et al. treated 16 patients with motexafin lutetium at 732 nm using PSA levels as the primary indicator of response. The significant variability in outcome was the result of large variations in PDT dose, which they thought could be reduced with better planning and measurement of light and drug variables during treatment. Moore et al. also recently reported phase I results using mTHPC-PDT in six men with early prostate cancer, showing it was safe using a few treatment fibres directed at biopsy-derived regions of cancer.
TOOKAD (padoporfin, WST09; NegmaLerads, Magny-Les-Hameaux, France) [12–14], a palladium-bacteriopheophorbide molecule, is a novel photosensitizer. We recently completed a phase I clinical trial of TOOKAD-mediated vascular-targeted PDT (VTP) in which the primary goal was to show the safety of the procedure, with a secondary goal of determining the drug and light-dose responses [15–17]. In the phase 1 trial, TOOKAD-VTP was delivered to 24 patients with locally recurrent prostate cancer after EBRT . In that first trial, only two treatment fibres were placed in the prostate. Photosensitizer and light doses were escalated to maximum values of 2 mg/kg and 360 J/cm2, respectively. The treatment response was assessed by gadolinium-enhanced MRI measurement of the induced lesion at 1 week after treatment. There was minimal systemic and local toxicity and the treatment produced sizeable necrotic zones (up to 2.2 cm diameter) in those receiving the highest drug/light dose. Lesion formation was strongly drug- and light-dose dependent. Regions of avascularity seen on the MRI images at 7 days after treatment corresponded to regions of histopathological fibrosis in which there was no residual viable tumour, indicating the utility of MRI after treatment as an early marker of response . However, there was considerable variability among patients in the response, even at the same drug and light doses.
As part of our programme of clinical trials of progressively aggressive treatments, we report the results of an early phase II trial with TOOKAD-VTP in the same group of patients as that of the phase I trial. The goal in this new trial was to treat the whole organ by blocking the prostate blood supply while preserving surrounding tissue. Multiple interstitial light-diffusing fibres were used, with their position and power optimized by pretreatment planning and in situ dosimetry. Our strategy in this trial was to escalate the light dose received by the whole prostate until there was full and consistent treatment of the prostate; the results at up to 6 months after treatment are presented.
PATIENTS AND METHODS
Twenty-eight patients were enrolled in this phase II study of escalating light doses; the study was approved by the Research Ethics Boards of the three sites that participated in the trial. Inclusion and exclusion criteria were identical to the previous phase I study on the safety of TOOKAD-VTP.
Treatment-planning software developed at the University Health Network  was used to determine before treatment the optimum number of fibres, their light-diffusing lengths, and the power settings for each patient, using MRI of the pelvic region acquired 1–4 weeks before treatment. For each patient, the prostate and rectum were traced and the course of the prostatic urethra was defined. Cylindrical light-diffusing fibres were positioned virtually inside the prostate and a light-dose distribution was calculated. The positions, lengths and powers delivered to the fibres were iteratively changed until the dose distribution met the desired treatment requirements, i.e. that the minimum dose for tissue devascularization be delivered to the entire prostate while maintaining the dose delivered to the urethra, rectum and extraprostatic tissue below the level required to cause damage. The optical properties and threshold dose values used in the initial patients included in the phase II trial were obtained by analysing data collected in the phase I trial. Subsequently, the results of the analysis of patients treated earlier within the phase II trial were fed back into the treatment planning dose calculations for later phase II patients. The measurement of the prostate’s optical properties was applied as described by Weersink et al., while a full description of the treatment planning software and its role in calculating the threshold dose values used in this trial will be given in a future publication.
Previous publications described the clinical procedures [15,19], light dosimetry and monitoring techniques of TOOKAD-VTP , and thus only a brief description of the treatment delivery will be repeated here. The drug dose was fixed at 2 mg/kg, with light dose escalated iteratively based on the observed VTP-induced effect on MRI in the previous cases. Patients had general anaesthesia, were placed in the lithotomy position and the prostate visualized by standard TRUS. Using a standard brachytherapy stabilizing frame and template, translucent closed-ended catheters were placed in the prostate. Optical fibres were inserted in these catheters, including cylindrical light-diffusing optical fibres of varying lengths for light delivery, and isotropic point detectors for light detection. Light delivery fibres were connected to a six-channel laser emitting at 763 nm (Ceralase PDT/763 nm CeramOptec GmbH, Germany). In addition to the detection probes in the prostate, detection probes were also placed in the urethra, the space between the prostate and rectum, and on the rectal wall closest to the prostate (attached to the TRUS transducer). Once all catheters and fibres were in position, drug infusion was started, followed 6 min later by the treatment light. The infusion period lasted 20 min, while light was delivered for 30 min.
Treatment efficacy was assessed using three endpoints: 18-G needle core prostate biopsies taken 6 months after treatment at six sites in the prostate; the relative volume of devascularized prostate tissue as determined by gadolinium-enhanced MRI acquired at 1 and 24 weeks after treatment; and PSA levels at 4, 8, 12 and 24 weeks after treatment, compared with baseline values. Treatment safety was assessed using the IPSS and the validated quality-of-life Patient-Oriented Prostate Utility Scale questionnaires to monitor potency, bowel function, urinary function and rectal integrity at baseline and at the follow-up at 1, 3 and 6 months .
In the MRI analysis, the relative volume of treated prostate tissue was measured on pelvic MRI scans taken 1 week after treatment, using tracing tools in the treatment planning software. A T2-weighted series and a T1-weighted series acquired after an injection with gadolinium contrast enhancement agent were both loaded into the software and registered. The prostate was traced on the T2-weighted series while intraprostatic devascularized areas were identified by tracing dark areas on the gadolinium-enhanced T1-weighted series. Based on previous canine studies, the areas of devascularized tissue observed in the T1-weighted MRI, as regions without enhancement, correspond to regions of treatment response in whole-mount pathology . The prostate volume and the volume of treatment effect after treatment were both calculated by planimetry, where the areas of the tracings were added then multiplied by the thickness of the image slices .
All three measurements of treatment efficacy were compared against the light dose received by the prostate. Analysis of the treatment results focused on the dose received by the prostate tissue rather than the energy delivered by the treatment fibres, because the former metric accounts for the energy distribution within the tissue and allows comparisons that are independent of the energy delivery technique.
The light-dose distribution received by each patient was determined from virtual reconstructions of the treatment, using the treatment planning software with the exact positions of the fibres as determined by the TRUS imaging and patient-specific optical properties of the tissues measured during the treatment. Analogous to dose calculations in radiotherapy, dose/volume histograms (i.e. the relationship between the light dose and the percentage of tissue volume receiving at least that light dose) were calculated for all the key anatomical structures. The minimum light dose received by 90% of the prostate volume (D90) was the principal measurement of light dose received by the patient, as shown in Fig. 1, where the D90 light dose isocontour (determined from the dose-volume histogram) is shown for a single transverse slice through the prostate.
As described below, there was a large degree of variation in patient response; hence, patients were classified according to one of three responses based on both pathology and MRI effect: those not responding had at least one of six biopsies showing residual cancer at 6 months, and a poor MRI response (i.e. relative volume of devascularized tissue <25%); partial responders had at most half of their biopsy cores positive for residual cancer and a MRI response of >25%; and complete responders had all their biopsies negative and a MRI response of >25%. This grouping is best illustrated in Fig. 2, which shows patient outcomes based on these classifications.
In general, increased light doses resulted in a greater avascular effect, as measured in the 7-day MRI, and a smaller tumour burden, as monitored by 6-month biopsies. The dose delivered to each patient increased as the trial progressed, with fewer diffusers and lower treatment energies delivered to the earlier patients. The average D90 was 32 J/cm2, with minimum and maximum D90 values of 5.6 and 147 J/cm2. The difference in the measured light dose with light doses calculated in the reconstructions was ≈40%, due to inaccuracies in measuring the position of the optical sensors (±1 mm) and the optical property calculations used in the light-dose reconstructions. This difference directly translates into a measurement error in the D90 values of 40%. As a result of lower delivered doses, early recruits to the trial generally received lower doses. The dose was increased for subsequent patients based on observed MRI responses in previous patients. Planned and actual received doses were different because planning used the average optical properties of all previous patients, while the measured optical properties were different from these average values.
Figure 2 compares the D90 values to the 6-month biopsy results; at a D90 of <23 J/cm2, all patients had at least one positive biopsy, and so a D90 of 23 J/cm2 defined a threshold light dose for a successful treatment effect in the trial population. Of the 13 per-protocol patients who had a D90 of >23 J/cm2, eight had a complete biopsy response. These results indicate that, given the general population of patients with recurrent prostate cancer after EBRT, a complete response (i.e. all biopsies negative) of ≈60% might be achieved if a sufficient light dose is received throughout the prostate.
All patients had some MRI-detectable intraprostatic tissue damage but the treatment volumes were highly variable. In general there were three types of outcome: a partial response due to a light dose less than the threshold (23 J/cm2), a poor response despite a light dose above this threshold, and excellent response with a light dose above the threshold. Representative 7-day MRI results for these three types of outcome are shown in Fig. 3, with three-dimensional renderings showing the volume of intraprostatic and rectal necrosis, and the position of the delivery fibres within the prostate. The patient represented in Fig. 3a was treated with six fibres and received a D90 of 47 J/cm2. However, the treatment response was poor, with lesion volumes representing only 4% of the prostate volume and at least one positive biopsy at 6 months. In this patient the tissue immediately surrounding several of the source fibres was preserved, whereas in other regions of the prostate the lesions extended several millimetres from source fibres. In the patient shown in Fig. 3b, only five source fibres were used, and the light dose was less than the threshold value (D90 = 5.6 J/cm2). The lesion size at 7 days was about half of the prostate volume, surrounding the delivery fibres, but sparing the anterior prostate where the light dose was low. Despite the size of the lesion, the patient had a positive biopsy at 6 months. Figure 3c shows the optimum outcome, with a lesion volume covering 84% of the prostate volume, leaving only a portion of the apex untreated. Here, the D90 was 33 J/cm2, and the patient had no positive biopsies at 6 months.
The 7-day MRI was a reasonable surrogate for the 6-month biopsy, especially if there was a large avascular treatment response. All patients with a MRI response of >60% of devascularized prostate were complete responders. The converse was not always true; a few patients with a low percentage of prostate tissue devascularized on MRI still presented with negative or almost completely negative biopsies.
There was an MRI-observable treatment effect (devascularization) to a portion of the rectum, urethral wall, or extraprostatic tissues in all but three patients. However, while the treatment effect extended to the urethral wall in 18 patients, long-term clinical signs of damage to the urethra, as measured by persistent changes in the IPSS, were absent. (Fig. 4) There was a transient deterioration in urinary function in the period after treatment, but it returned to baseline by 6 months.
In all, 10 patients had rectal wall devascularization, with one of them not completing the study (see below). In most patients in whom there was some rectal devascularization detectable on MRI, no clinically adverse effects were reported. In many of these patients, devascularization occurred in the outer 1–3 mm of the rectal wall and did not extend into the rectal mucosa. At 6 months, MRI showed healing of rectal wall damage in most of these patients, indicated by restoration of blood flow to the devascularized section of the rectal wall. As observed in preclinical studies, VTP spares the collagen structure, providing a framework for tissue healing. Devascularization to the rectal wall was significantly more prevalent in patients who had a higher prostatic response to the TOOKAD-VTP. Of the 10 patients with some rectal devascularization, seven had a treatment effect in more than half of the total prostate volume. In the 14 patients with no rectal devascularization damage, only one had a prostatic necrosis volume of more than half. This suggests that the overall sensitivity to TOOKAD-VTP might be higher in some patients than in others. The previous EBRT delivered to these patients might have a role in determining this sensitivity, but analysis of the radiation dose received by these patients was inconclusive. This observation also suggests that while all patients received the same drug dose, the concentration of TOOKAD in the vasculature could vary among patients.
Two patients had rectourethral fistulae; one had had haemorrhagic proctitis after EBRT, and responded well to VTP (based on MRI) using six fibres, all at 200 J/cm2, but tissue devascularization also extended into the anterior rectal wall. Clinical signs of a fistula appeared at 2 months after the procedure. The light dose measured in the rectum for this patient during the TOOKAD-VTP was within the range measured in other patients, indicating that he had not received an unusually high light dose to the rectum. Drug concentrations in the rectal tissue were never measured, so this theory remains speculative. In the second patient no clear cause for the fistula was found, and his fistula closed spontaneously at 6 months.
The histological findings in some of the biopsies included complete fibrosis with no presence of tumour, which was associated with a complete TOOKAD-VTP effect, with a clear demarcation between tissue with a treatment effect and tissue unaffected by the treatment. A common finding within areas of VTP-induced fibrosis was the presence of entrapped benign, atrophic-appearing prostate glands. It is not clear whether these acini represent regenerating benign glands after VTP or glands that were somehow selectively spared during the VTP treatment.
Serum PSA levels decreased in the complete responders (Fig. 5); in those complete responders with baseline PSA levels of <5 ng/mL, the 6-month PSA levels were essentially undetectable, while patients with higher baseline PSA levels had no complete PSA response. Instead, in these patients the PSA levels decreased at 1 month but increased by 6 months. The treatment does not destroy all epithelial tissue, which is the source of PSA. It might be possible in those patients who had no complete PSA response that while the foci of tumour were destroyed, the remaining epithelial cells were stimulated to produce more PSA because of inflammatory effects of treatment. PSA levels in partial and nonresponders remained similar to baseline levels, regardless of light doses received.
This is the largest study to date of TOOKAD-VTP in the treatment of men with recurrent prostate cancer after EBRT. It shows the potential of this treatment to effectively eliminate residual prostate cancer, in a minimally invasive manner. This study clearly defines the unique opportunities of TOOKAD-VTP, as well as its present limitations.
Prostate tissue after irradiation has a definite and defined threshold to this treatment; the delivery of a light dose lower than the calculated threshold (23 J/cm2) in 90% of the prostate at a dose of 2 mg/kg of TOOKAD consistently results in a poor treatment outcome. Conversely, increasing the light dose beyond the threshold increases the likelihood of a complete response, although it does not guarantee it. Eight of 13 patients with a D90 of >23 J/cm2 had a complete response, with early MRI evidence of marked prostatic devascularization and no evidence of residual cancer at the 6-month biopsy. These results compare favourably with other treatments that are in their initial implementation phase in terms of anticancer efficacy.
Side-effects of the TOOKAD-VTP treatment were modest, generally self-limited, and compared favourably with other salvage methods. Most patients had an initial deterioration in voiding function, as measured by the IPSS, mostly due to storage symptoms that were tolerable and easily controlled with medical treatment. By 6 months most symptoms had spontaneously improved and the mean IPSS had reverted towards baseline. Two patients developed rectourethral fistulae, and the exact pathophysiology of their occurrence remains uncertain. One fistula closed spontaneously, needing only supportive care. The second remains patent at >6 months and continues to manifest itself mainly with intermittent loose stools but with no other significant side-effect. We can only speculate about the potential cause of developing the fistula in this patient. A hypervascularization of the rectal tissue secondary to proctitis after EBRT might have led to a higher drug concentration in blood pooling within the rectum in this patient, and hence a higher response (light × drug = TOOKAD-VTP response) than in the others.
The partial response to TOOKAD-VTP in some patients, despite the use of doses similar to those given to patients who responded completely, is difficult to explain. Drug and light were successfully delivered to all patients. A correlation between incomplete response at high TOOKAD-VTP doses and other biological variables, e.g. previous radiation dose or blood analytes, could not be found. Other possible causes of the interpatient variability in response include patient variation in drug pharmacokinetics. Development of a test that indicates a patient’s likely response to the treatment would enable the selection of patients eligible for TOOKAD-VTP. Treatment efficacy might also be improved by protecting vital organs from high drug × light doses, enabling higher delivered doses in the prostate. In particular, the rectum might be further protected by using a fixed ‘stand-off’ device like a percutaneously inserted balloon between the prostate and the rectum. This balloon might contribute passively by distancing the rectum from administered light, and/or actively after filling with an opaque liquid or cooling fluid, and thus reducing blood flow to the rectum.
TOOKAD-VTP offers a balance of cancer control and maintenance of quality of life in this difficult group of patients, and this seems better than most other forms of treatment in the early stages of development. Optimization of the process will depend on ensuring adequate and uniform light-dose delivery and distribution to the entire prostate, by careful treatment planning and execution, as well as enhanced protection of the rectum.
Gadolinium-enhanced MRI served as a good surrogate endpoint for treatment efficacy at 7 days and throughout the progression of the clinical trial, similar to the previous correlation of histology and MRI in the canine model using TOOKAD-VTP . This enabled us to increase the drug and light doses to subsequent patients without waiting for the 6-month biopsy results. Compared to radiation therapies or chemotherapy, this aspect of the TOOKAD-VTP provides excellent feedback on treatment outcome. In partial or nonresponders, the 7-day MRI allows possible re-treatment within a short delay after the first treatment, and could also allow the patient to choose other treatment options instead of waiting for much later signs of failure, such as biochemical or clinical progression.
Furthermore, the correspondence between TOOKAD-VTP cancer control and 7-day MRI results suggests that online measurement of the TOOKAD-VTP effect might be monitored and controlled by intraoperative methods, e.g. real-time ultrasonography-derived microbubble microvascular flow measurements. This concept might allow for both improved prostate cancer destruction and rectal protection.
TOOKAD-VTP is a novel minimally invasive treatment that offers the potential for an effective salvage therapy in men with recurrent prostate cancer after EBRT. Its current feasibility suggests that it could be delivered in an outpatient setting, with minimal morbidity. Clinical issues of light dosimetry, to adequately illuminate the entire prostate, and rectal sensitivity to the treatment, are being addressed. Further clinical trials have been initiated and will determine the exact role of this exciting and promising new treatment.
This work was supported by Steba Biotech, N.V., The Hague, the Netherlands, and by NIH grant ♯CA33894. R. Weersink is supported by the Ontario Centres of Excellence through the Laboratory for Applied Biophotonics. Further support was from the Muzzo Fund of the Princess Margaret Hospital Foundation.