Photodynamic Therapy with 3-(1′-Hexyloxyethyl) Pyropheophorbide a for Cancer of the Oral Cavity

This article is featured in Highlights of This Issue, p. 6333

The Surveillance, Epidemiology and End Results (SEER) report that the incidence rates of cancer of the oral cavity is 5.7 per 100,000 in the United States (1). In India, the incidence rate is as high as 20 per 100,000 population (2). Every year, more than 17,000 new cases of lip and oral cavity cancer are diagnosed in the United States.

Surgery and radiotherapy are the standard treatment modalities for T1 squamous cell carcinoma (SCC) of the oral cavity (3). Several studies demonstrated that surgery is the preferred treatment for these tumors, yielding superior 5-year survival rates when compared with radiotherapy (3, 4). However, effective surgical treatment requires wide local resection of the primary tumor with clear surgical margins. To secure tumor-free margins, excision of adjacent normal functional tissue is performed, often affecting speech and swallow function. On the other hand, radiotherapy can induce significant treatment-related adverse events such as xerostomia, chronic dental decay, and risk of mandibular osteonecrosis, which remain long after the patient is cured, and has shown to reduce patients' quality of life (QoL; ref. 5). Patients who are cured with standard therapies also have a significant life-long risk of developing second primary tumors in the oral cavity, which has been associated with poor prognosis (6–8). Although patients with superficially invasive tumors (equal or less than 4 mm in thickness) have a relatively low risk for local recurrence and metastasis (9–11), the treatment options have been limited to surgery or radiotherapy. There is a need to offer these patients a curative therapy that is safe, repeatable, and has no long-term toxicities.

Photodynamic therapy (PDT) is a minimally invasive treatment that involves the activation by light of a drug (photosensitizer) that generates cytotoxic reactive oxygen species, resulting in direct damage to tumor cells (12). PDT has proven to be an effective local treatment for a range of solid tumors (13). It has the potential to become an effective first-line treatment modality for early-stage SCC of the oral cavity because it is associated with minimal short-term side-effects, nominal scarring, and sparing of healthy vital structures such as nerves and major blood vessels (14–16). Importantly, PDT may be used with standard therapies.

The photosensitizers porfimer sodium (Photofrin), U.S. Food and Drug Administration approved for esophageal and endobronchial cancer, and mTHPC (Foscan), approved in Europe for the palliative use in squamous cell carcinoma of the head and neck (HNSCC), have shown promise for the treatment of oral cancers (17). Although Photofrin- or Foscan-mediated PDT is effective, the persistence of the photosensitizer in skin necessitates protection of patients from sunlight and other sources of bright light for long periods (30–90 days). Given this prolonged phototoxicity, there has been widespread interest in the development of newer photosensitizers with more favorable photophysical and pharmacokinetic properties. The chlorine-based compound, 3-(1′-hexyloxyethyl) pyropheophorbide a (HPPH) is one such photosensitizer (18) that has been shown to exhibit potent antitumor activity in a number of experimental tumor models (19). Clinical studies conducted in patients with lung and esophageal cancer have also revealed good responses (16, 20). We have shown that HPPH at clinically effective antitumor doses is associated with significantly reduced cutaneous photosensitivity that rapidly declines over several days (21). In addition, we developed an approach that allows the assessment of the PDT reaction, within 48 hours, through the analysis of PDT-induced cross-linking of the STAT3 (22). We have hypothesized that STAT3 cross-linking will serve as a molecular reporter for accumulated PDT dose in the treated oral lesions and have predictive value for treatment outcome. We have undertaken two phase I dose escalation trials to test this hypothesis, to determine the safety profile of HPPH-PDT for precancerous and early-stage SCC of the oral cavity, and to a gain preliminary indication of efficacy.

These are the first phase I, dose finding, open label, noncomparative studies of HPPH-PDT in high risk dysplasias, carcinoma in situ (CIS), and SCC of the oral cavity. The trials were carried out at Rosewell Park Cancer Institute (RPCI, Buffalo, NY) from April 2008 to June 2012. HPPH was used at a fixed, previously determined, dose of 4 mg/m² administered systemically 22 to 26 hours before light delivery (23). The studies had identical 3 + 3 dose-escalation scheme with an expanded cohort at the highest light dose. This design is a special case of the A + B design described by Lin and Shih (24). Rationale behind the design is nested in the assumption that both the probabilities of toxicity and efficacious response are continuous monotonic nondecreasing functions of the dose. Three courses of treatment were allowed for each oral lesion, with at least 6 weeks between treatments. Patients could have more than one lesion treated. In the first trial (NCI-2010-02401), the light dose was escalated from 50 J/cm² to 75, 100, and 125 J/cm². A second trial (NCI-2010-01493) with amended exclusion criteria was initiated with light dose ranging from 100 to 125 and 140 J/cm² for each course.

The primary objective was to establish the safety profile and to determine the dose-limiting toxicity (DLT) and the maximum-tolerated light dose (MTD). Secondary objectives were assessments of HPPH levels in blood and in the tumor tissue at the time of light treatment, the extent of HPPH-PDT–mediated STAT3 cross-linking in the treated tissue, and the pathologic treatment response as determined by biopsy at 3 months posttreatment and clinical follow-up.

Written informed consent was obtained from all patients and all protocol related procedures were approved by the RPCI Institutional Review Board and overseen by the RPCI Data and Safety Monitoring Board.

Patient eligibility criteria included: biopsy-proven moderate to severe dysplasia, carcinoma in situ or T1 SCC of the oral cavity, tumor thickness 4 mm or less, primary or recurrent, any type of prior therapy allowed, age at least 18 years, male or non-pregnant female using medically acceptable birth control, Eastern Cooperative Oncology Group (ECOG) score 0–2, signed informed consent.

Patients were excluded because of: T2 or greater SCC, primary tongue base tumors, porphyria or hypersensitivity to porphyrins or porphyrin-like agents, impaired hepatic alkaline phosphatase or SGOT >3 times the upper normal limits, minimal impairment of renal function (total serum bilirubin >2 mg/dL, serum creatinine > 2 mg/dL), concurrent chemotherapy or radiotherapy less than 4 weeks after the last dose of such therapies. In the second trial, severe preexisting trismus was added as exclusion criterion.

Patients underwent a pretreatment evaluation that included history and physical examination, palpation of the index lesion, baseline biopsy that was submitted to pathologic examination, performance status, and laboratory studies. If clinically indicated, patients received an electrocardiography, chest X-ray, and/or computed tomography (CT) scan of the neck to exclude the presence of nodal disease.

All patients received PDT in the operating room under general anesthesia to eliminate pain and to facilitate the collection of tissue biopsies, and the performance of noninvasive spectroscopic measurements, before and after treatment. Tissue outside the intended treatment field was shielded from the laser light. Each lesion was illuminated with light of 665 nm wavelength, delivered by a tunable dye laser via optical fibers with an in-house manufactured gradient index lens. The power output was measured with an integrating sphere. The treatment field included 1 to 2 cm margins around the index lesion. In patients with multicentric disease, more than one lesion could be treated during one treatment session. However, the total area of direct light exposure did not exceed 25 cm².

Following treatment, patients were monitored in the Ambulatory Center by a physician until ready for release, or hospitalized overnight for observation if deemed advisable. All patients were given prednisone for 9 days, starting 1 day after treatment, to control swelling. Pain was treated with oral narcotics and Fentanyl patch. All patients were instructed to avoid exposure to sunlight or bright indoor light for at least 7 days by wearing protective clothing and specific sunglasses provided by the RPCI PDT Center. They were also advised to expose small areas of skin to sunlight on day 8 for 10 minutes to detect any remaining photosensitivity.

Patients were seen at 1 week and 1 month after treatment to assess treatment-related toxicities and clinical response. At 3 months, patients had a 3 mm punch biopsy within the treated field for pathologic response verification. Thereafter, patients were examined at 3 to 6 months intervals.

Patients were monitored for systemic toxicity at the time of HPPH administration, laser treatment, and at each follow-up visit. Safety was determined by recording the occurrence of serious adverse events (SAE) during the first 30 days posttreatment using the revised NCI Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. At each clinic visit, patients were examined for local normal tissue toxicity, performance status, pain level, and skin phototoxicity. All adverse events and SAEs were documented as to onset and resolution date, classification of intensity, relationship to treatment, action taken, and outcome. Adverse events and SAEs were recorded as per MedDRA coding (Medical Dictionary for Regulatory Activities).

Responses reported here are based on the 3-month biopsy results and/or subsequent clinical observation. All biopsies were reviewed and interpreted by the study head and neck pathologist. Tumor and lesion response to therapy was graded as follows:

• Complete response (CR): complete absence of visible lesion and negative biopsy, or absence of visible lesion only (1 case).

• Partial response (PR): volume reduction of the lesion by 50% or more.

• Stable disease (SD): all responses <PR.

• Progressive disease (PD): any increase in lesion size or increase in grade of the treated lesion.

HPPH serum levels were determined on the basis of fluorescence (18). Coagulated blood was collected within 2 hours before light treatment, in anticoagulant-free tubes and centrifuged (Centrific 228; Fisher Scientific). Serum was analyzed by recording the amplitude of the fluorescence emission maximum (λ_ex = 412 nm, λ_em = 670 nm) followed by baseline correction.

In vivo fluorescence spectroscopy was used to assess HPPH tissue levels at the lesion surface. The details of the clinical spectroscopy instrument and the analysis method have been described previously (25, 26). Briefly, one source-detector pair (0.8 mm separation) of a custom, optical fiber–based probe was used for fluorescence detection and another pair (1.6 mm separation) for reflectance measurements to allow fluorescence normalization. Five independent measurements were carried out from the index lesion. The HPPH peak at approximately 670 nm of the normalized signal is reported as in vivo tissue fluorescence values.

One 3 to 5 mm punch biopsy was taken from the treatment site before treatment and one immediately after completion of treatment. Biopsies were immediately transferred to the Department of Pathology, where each sample was divided in two equal parts, one was submitted for routine processing for histopathologic examination, the other was snap-frozen on dry ice and processed for determining STAT3 cross-linking. Samples were thawed on wet ice in the presence of approximately 3-fold volume of radioimmunoprecipitation assay (RIPA) buffer containing protease and phosphatase inhibitors for 5 minutes and ultrasonicated with a micro steel tipped cell disruptor (Heat Systems-Ultrasonics, Inc). The homogenate was centrifuged at 15,000 rpm. Replicate aliquots of the supernatant containing 20 μg protein were boiled in SDS-sample buffer and electrophoresed on a 6% SDS-polyacrylamide gels. The separated proteins were analyzed by Western blotting for STAT3 and EGF receptor (EGFR) as described, previously (27, 28). Immunoblot signals for STAT3 were quantified and expressed as percentage conversion of monomeric into covalently cross-linked dimeric STAT3. In a subset of patients, we also assessed PDT-induced loss of EGFR expression (28).

The data from patient populations in both trials were combined for evaluation. Statistical analyses were primarily descriptive. Calculated P values were based on the unpaired t test, ANOVA, or Fisher exact test, as appropriate. For all tests, a P < 0.05 was considered significant. Statistical calculations and analyses were done using GraphPad InStat (ver. 3.10; GraphPad Software, Inc.). Data in Figs. 2, 3, and 5 are summarized as quartile (box-and-whisker) plots. Lower box boundary = 25th percentile, upper box boundary = 75th percentile, solid line within box = median, dotted line within box = mean, solid dot = outlying value, upper and lower whiskers = 90th and 10th percentile, respectively. Graphing and analysis were performed using SigmaPlot software (version 11.2.0.5; Systat Software).

Details of patient and lesion characteristics are presented in Table 1. A total of 41 patients and 51 lesions were accrued, with 40 patients receiving at least one session of PDT. One patient received HPPH, but was not treated because of unrelated cardiac complications encountered during anesthesia induction. For analysis, the patients' lesions were grouped into two categories: dysplasia and carcinoma in situ, including moderate, severe dysplasia, and carcinoma in situ (28 lesions), and SCC, including all invasive tumors as well as carcinoma in situ with microinvasion (23 lesions). In all, 5 patients presented with Lichen planus, 17 with leukoplakia, and 5 with erythroplakia. Six patients had more than one lesion treated. Five lesions were treated twice, one lesion was treated three times. Ten patients had been previously treated; 1 patient with SCC and 4 patients with dysplasia/carcinoma in situ had been treated with surgery alone, 1 patient with SCC had received radiation, and 4 patients with SCC had received surgery and concurrent chemoradiotherapy.

All patients reported the expected pain and edema at the treatment site, with pain peaking after approximately 1 week and lasting up to 4 weeks.

There was one DLT encountered at a light dose of 125 J/cm² with a patient developing grade 3 edema at the treatment site, which led to respiratory distress requiring a tracheostomy. This patient had been treated previously with radiotherapy and multiple surgical procedures and consequently suffered from severe trismus, which compromised accurate light delivery to the index lesion located on the hard palate. Although the patient recovered rapidly without sequelae, this event led to the closing of the protocol and addition of severe trismus as an exclusion criterion to a replacement protocol. Four more patients that also had prior radiotherapy had grade 1 edema. Three more patients who had grade 3 edema had no prior radiotherapy. Thus, our data suggest that there was no correlation between edema and prior radiotherapy.

No further airway obstruction due to edema was encountered at any of the light doses. Another patient in the 125 J/cm² cohort experienced a grade 3 edema of 12-hour duration after the third PDT treatment. Two patients in the expanded 140 J/cm² cohort experienced treatment site grade 3 edema that resolved within 24 to 36 hours.

Fifteen mostly elderly patients were hospitalized overnight for observation and released. Most patients took a full liquid/pureed diet due to pain for several days after treatment, but no medical alimentation support was needed. Four patients experienced mild sunburn reactions of short duration due to noncompliance with instructions.

An MTD was not reached at the highest planned light dose level of 140 J/cm², but further escalation was forgone due to the danger of unacceptable swelling and the risk of airway obstruction.

Details of outcomes, at 3 months posttreatment, by light dose and lesion type are shown in Table 2. Forty nine of 51 lesions treated, 26 dysplasia/carcinoma in situ and 23 SCC, were evaluable for response. All but one lesion (evaluated clinically) were evaluated at 3 months by biopsy. Two patients died from unrelated causes before the 3 months follow-up. Given the small numbers and heterogeneity of lesions in each light dose cohort, no light dose dependence is discernible.

Only the expanded 140 J/cm² cohort was large enough to give some insight into the response rates. In this cohort, a difference in outcome between dysplasia/carcinoma in situ and SCC is strongly suggested, although this difference is not quite significant (P = 0.0562) with a CR rate of 46% for dysplasia/carcinoma in situ and 82% for SCC. Of the 6 dysplasia/carcinoma in situ lesions with CRs, only 2 responses (33%) were durable with 15 and 9 months, respectively. One of these patients had leukoplakia, the other had erythroplakia. Retreatment of partially responding dysplasia/carcinoma in situ lesions did not achieve lasting CRs. Of the 14 SCC lesions with CR, 2 had complete disappearance of SCC, but some remaining minor focal dysplasia, which was locally excised. Two patients died disease-free of unrelated causes at 19 and 5 months. Eight SCC patients with CR are still disease-free (disease-free intervals 6–40 months). Of these, 3 were leukoplakia-free, 3 had leukoplakia, 1 had Lichen planus, and 1 had erythroplakia. Excluding the patients with SCC that had excision of minor focal dysplasia and the patients that died disease-free, the lasting CR rate for the 140 J/cm² cohort is 80%. Table 3 lists the CRs in the 140 J/cm² cohort by lesion location, suggesting that for both dysplasia/carcinoma in situ and SCC, those on the tongue and floor of mouth responded best.

Across all light dose cohorts, among the five SCC lesions that did not achieve CR, three were in patients who had received prior surgery plus chemoradiotherapy. All lesions that achieved less than CR were successfully treated surgically.

The majority of the lesions were nonconfluent. Hence, it was difficult to evaluate the lesion size. A representative example of a lesion that was treated in this study is shown in Fig. 1. This was a high grade dysplasia with microinvasion. The entire lesion and margins were illuminated with a 4 cm in diameter spot size. The PDT-induced necrosis was observed 7 days posttherapy (Fig. 1B). Complete clinical disappearance of the lesion, without scarring, was seen 1 month post-PDT (Fig. 1C). The complete clinical response was confirmed with pathologic evaluation at 3 months posttherapy.

Earlier in vitro and in vivo studies have shown that covalent STAT3 cross-linking is proportional to the extent of the photoreaction that is a function of the amount of photosensitizer and light received by the target tissue and the tissue oxygenation status. Hence, the extent of STAT3 cross-linking is a metric for the cumulative PDT reaction (22, 27).

A total of 46 lesions, having received 75, 100, 125, or 140 J/cm², were evaluable for STAT3 cross-linking, of which 26 were dysplasia/carcinoma in situ and 20 were SCC. None of the biopsies obtained before illumination showed any STAT3 cross-links. The percentage conversion of monomeric to cross-linked STAT3 for all lesions combined as a function of light dose is shown in Fig. 2. This nondiscriminatory comparison did not indicate any notable light dose dependence.

When the data analysis focused on lesion type at diagnosis, a significant pattern emerged–dysplasia/carcinoma in situ lesions had lower median and mean percentage STAT3 conversion than SCC (Fig. 3). Data were analyzed for the following groups: all light doses, doses <140 J/cm², and 140 J/cm². The difference between dysplasia/carcinoma in situ and SCC for all doses and 140 J/cm² is highly significant (P = 0.0006 and 0.0033, respectively). SCC samples for doses <140 J/cm² were too few for statistical analysis. Two instructive examples for low and high STAT3 cross-linking and corresponding biopsy pathology are shown in Fig. 4. It should be noted that not all of the samples for STAT3 cross-link analysis contained SCC histology but included various degrees of dysplasia, likely due to geographic misses in obtaining the biopsies. These samples contributed to the low STAT3 cross-link levels and large variations in the SCC groups in Fig. 2. Nevertheless, all histologically proven SCCs registered STAT3 conversion above 10%. Two samples in the SCC category from lesions initially characterized as carcinoma in situ with microinvasion showed very high STAT3 cross-link levels. All 10 SCCs (1 at 100 J/cm², 1 at 125 J/cm², 7 at 140 J/cm²) with STAT3 cross-link levels >10 (range, 12.3%–36%), except one previously chemo/irradiated lesion, achieved CR.

Twenty three dysplasia/carcinoma in situ and SCC lesions treated with 140 J/cm² were also analyzed for changes in EGFR expression. Of these, 20 showed an immediate loss of EGFR expression ranging from 15% to 98% (median, 81.5%; mean, 74.4%). A representative example is shown in Fig. 4. Three lesions showed increased EGFR expression of 9%, 28%, and 270%, respectively. No correlation with lesion type or response could be found.

HPPH serum levels, assessed by fluorescence assay of 40 evaluable patient samples revealed a normal distribution and there was no significant difference between HPPH levels from patients with dysplasia/carcinoma in situ and SCC (data not shown).

Noninvasive fluorescence spectroscopy assessed HPPH content of the lesions. Twelve dysplasia/carcinoma in situ and 14 SCC lesions could be analyzed (Fig. 5). Fluorescence values were significantly (P = 0.0431) higher in SCC (median, 9.73; mean, 10.91) than in dysplasia and carcinoma in situ (median, 6.38; mean, 6.80). Although this assessment does not provide the absolute concentration of HPPH in the tissue, it does suggest higher HPPH content in SCC compared with dysplasia and carcinoma in situ.

This report allows the following conclusions: (i) HPPH-PDT in the oral cavity can be safely delivered; (ii) early-stage oral SCC seems to respond better to HPPH-PDT in comparison with CIS/dysplasia lesions; (iii) the extent of PDT-induced STAT3 cross-linking was significantly higher in SCC when compared with carcinoma in situ/dysplasia lesions; (iv) the level of STAT3 cross-linking is a significant reporter for evaluating the extent of HPPH-PDT–mediated photoreaction.

Although no MTD was reached in these studies, pain and edema could be severe, as was also reported for other PDT agents (29, 30). We effectively relieve those symptoms with proactive administration of tapering dose of steroids, Fentanyl patch, and oral narcotics for breakthrough pain for 2 to 3 weeks. Interestingly, PDT of the tongue and floor of mouth, which produced the best responses, consistently was more painful and produced more tissue edema than other oral subsites. Special caution needs to be exercised in the treatment of patients with severe secondary trismus, particularly when accurate light delivery may be compromised. Scar formation leading to trismus, burns, and alimentation difficulties have been reported with the use of Foscan (30), but no such adverse events occurred with HPPH- PDT. General phototoxicity in this study was minimal and limited to erythema, which is in agreement with our previous studies (21).

The CR rate of 82% for SCC compares well with CR rates reported in the literature for Photofrin (17, 31, 32) and Foscan (33). Biel reported clinical CR rates of more than 90% (14) and Schweitzer reported clinical CR rates of 80% in retrospective studies with Photofrin-PDT in the treatment of carcinoma in situ and early-stage oral cancers (34). A recent retrospective study with Foscan-PDT reported a CR rate of 71% in 170 patients with oral cavity primary and recurrent tumors. They reported higher overall response in previously untreated neoplasms (30). In our study, we also observed a higher CR rate in previously untreated cancers. Interestingly, tumors located on the tongue or floor of mouth seemed to respond best, as was also reported by Karakullukcu and colleagues (30). The presence of leukoplakia or erythroplakia did not seem to affect outcome.

The CR rate of oral dysplasias and carcinoma in situ was less to that of SCC, as previously observed by Rigual and colleagues (29) with Photofrin-PDT. This result seems counterintuitive because of the limited tissue depth of these lesions. An explanation for this outcome might be found in lower HPPH levels in premalignant lesions suggested by noninvasive fluorescence measurements. The reasons for the higher HPPH accumulation in tumor lesions are likely 2-fold: better drug delivery due to a well-established vasculature and enhanced retention of HPPH by tumor cells. The importance of vascularity and photosensitizer uptake in small tumor nodules has been emphasized by Menon and colleagues (35). Also, earlier fluorescence spectroscopy studies with Photofrin (36) and HPPH (37) aimed at tumor diagnosis, have demonstrated increasing lesion fluorescence with progression from premalignant to malignant disease in a hamster buccal cheek pouch tumor model. Fluorescence values showed a high correlation with drug concentrations as determined by chemical drug extraction (38). Enhanced retention of HPPH by tumor cells has been clearly demonstrated for the first time by Tracy and colleagues (28) in primary human lung and head and neck tumor cells, maintained in the presence of corresponding stromal cells.

The better response of SCC to HPPH-PDT is proportionally reflected in the higher levels of STAT3 cross-links compared with dysplasia and carcinoma in situ. This held true whether data from all light doses or from only the 140 J/cm² cohort were examined. We had hypothesized that the light dose escalation would be manifested in the level of STAT3 cross-links detectable in biopsies collected following PDT. This was not the case. One could assume that this lack of dose response might be due to the cross-linking reaction having reached a saturation point. This, however, is unlikely as there was a large spread of STAT3 cross-link levels in all three cohorts (100–140 J/cm²) ranging from 0.5% to 36% STAT3 monomer conversion to cross-links. Interestingly, a dose analysis undertaken by Davidson and colleagues (39) in patients undergoing PDT with the photosensitizer TOOKAD for prostate cancer found that the threshold dose for PDT-induced damage was highly heterogeneous among patients. They speculated that this heterogeneity was due to varying amounts of photosensitizer or oxygen in the tissue.

Activated STAT3 has been implicated in the loss of growth control in HNSCC through antiapoptotic mechanisms (40). However, there is no evidence that cross-linking of a fraction of STAT3 per se affects cell survival (22). Earlier studies have demonstrated that PDT causes the loss of EGFR (22, 41). The current analyses confirmed the PDT-associated reduction of EGFR in oral lesion. Studies on cultured cells have indicated that restoration of full EGF responsiveness can require several days. This may contribute to a transient reduction in growth stimulation and delayed recovery.

STAT3 cross-linking was found only in biopsies collected after PDT. The level of STAT3 cross-linking was significantly higher (P = 0.0033) in SCC when compared with carcinoma in situ/dysplasia lesions for all levels of light doses, and the fluorescence values of HPPH were significantly higher (P = 0.0431) in SCC than in dysplasia and carcinoma in situ. Therefore, our data suggest that the level of STAT3 cross-linking is a significant reporter for evaluating HPPH-PDT–mediated photoreaction. Additional studies are required to assess the use of STAT3 cross-linking assay as a prognostic biomarker for evaluating tumor response.

Surgery and radiotherapy are currently considered effective treatment modalities for early-stage oral cancer (3, 42). However, surgery removes vital tissue and radiotherapy has lasting sequelae that have shown to impair QoL, long after the patients were cured. Therefore, targeted therapies which are repeatable and spare vital tissue are necessary. In this study, HPPH-PDT was found to be a promising therapy for the treatment of early-stage oral cancers in that the majority of patients had a pathologic CR or a PR in which invasive disease was downgraded to dysplasia and treated effectively with minimal surgery. This is of clinical importance, as the effective surgical treatment of dysplasia may be accomplished with a more limited resection of vital oral tissue than the surgical treatment of early-stage oral SCC. HPPH-PDT treatment also resulted in good tissue healing. This outcome is in agreement with observations in our recent intraoperative HPPH-PDT study, where we reported excellent secondary healing of skin burns in two patients due to phototoxicity (16). The good healing could be explained by the fact that HPPH is not retained in fibroblasts (28). Thus, the HPPH-PDT–induced damage cells are replaced by native tissue that regains its normal functions, significantly limiting function loss and minimized scar formation. These promising results warrant the design of a phase II study.

No potential conflicts of interest were disclosed.

Conception and design: N.R. Rigual, H. Baumann, D.A. Bellnier, G. Wilding, B.W. Henderson

Development of methodology: N.R. Rigual, H. Baumann

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N.R. Rigual, M. Cooper, H. Baumann, D.A. Bellnier, U. Sunar, E. Tracy, D. Rohrbach, M.A. Sullivan, M. Merzianu, B.W. Henderson

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): G. Shafirstein, H. Baumann, U. Sunar, D. Rohrbach, G. Wilding, W. Tan, M. Merzianu, B.W. Henderson

Writing, review, and/or revision of the manuscript: N.R. Rigual, G. Shafirstein, M. Cooper, H. Baumann, D.A. Bellnier, U. Sunar, G. Wilding, M.A. Sullivan, M. Merzianu, B.W. Henderson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N.R. Rigual, M. Cooper, H. Baumann, B.W. Henderson

Study supervision: N.R. Rigual, B.W. Henderson

The authors thank Drs. Thomas Foster and Sandra Gollnick for their critical review of this article.

This study was supported by NCI grants PO1CA55791 (to B.W. Henderson) and Roswell Park Cancer Institute Support Grant P30CA16056.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.