Intraoperative ¹⁸⁶Re-Liposome Radionuclide Therapy in a Head and Neck Squamous Cell Carcinoma Xenograft Positive Surgical Margin Model

The primary management of advanced-stage head and neck squamous cell carcinoma (HNSCC) relies on the complete resection of the tumor. However, the establishment of a histologically sound, margin-free resection is often impossible given the devastating side effects of aggressive surgery and the anatomic proximity to structures such as the carotid artery and the spinal cord. Severe consequences may include loss of voice, swallowing and speech problems, and anatomic deformities (1, 2). Unfortunately, positive margin status, commonly defined as any viable tumor (including carcinoma in situ or dysplasia) at the edge of the resected tumor on frozen section or permanent section, is associated with significantly decreased survival (3, 4) due to local or regionally persistent or recurrent disease (2). Traditionally, patients with unresectable HNSCC are treated with postoperative radiation therapy with less than 40% long-term survival (1). With the development of combination chemoradiotherapy, studies have reported a significant increase in overall survival (5); however, the dose-limiting local and systemic toxicity (6–8) of combination therapy leaves more to be desired in terms of better local control and more tolerable side-effect profile.

Given the problem of high local recurrence rate and unbearable toxicity with current therapeutic options for advanced unresectable HNSCC, advancements in treatment should be directed at delivering high local drug concentration while limiting normal tissue toxicity. Liposomes are well known to the medical community, particularly as drug carriers for cancer treatment (9). Most cancer chemotherapeutic agents have a large volume of distribution. Liposomes alter the pharmacokinetics and biodistribution of free drugs and/or function as a reservoir for sustained drug release (9). Moreover, in solid tumors, the leaky vasculature and the lack of a well-defined lymphatic system allow i.v. administered liposomes to achieve spontaneous intratumoral accumulation via the enhanced permeability and retention effect (9, 10). The unique properties of liposomes present a viable option for the locoregional delivery of encapsulated therapeutic agents in the positive margins of unresectable HNSCC.

One particularly attractive utilization of liposomes is the encapsulation of therapeutic radionuclides for local injection therapy. Harrington et al. (11) reported intratumoral retention of 15.0 ± 12.3% for ¹¹¹In-labeled diethylenetriaminepentaacetic-acid liposomes at 96 hours compared with 0.64 ± 0.42% in unencapsulated ¹¹¹In-diethylenetriaminepentaacetic acid following intratumoral injection in a HNSCC xenograft model. In a similar study, Bao et al. (12) reported a locoregional retention rate of 39.2 ± 10.6% at 20 hours via intratumoral injection of technetium-99m (^99mTc)–labeled liposomes in a HNSCC xenograft model with better intratumoral distribution and a higher local-retention profile than unencapsulated ^99mTc-N,N-bis(2-mercaptoethyl)-N′,N′-diethyl-ethylenediamine (BMEDA). These preclinical studies clearly indicate that liposomal radionuclides are capable of achieving sustained local retention via direct injection. To investigate tumor therapy with liposomal radionuclides, a method of encapsulating β-emission therapeutic radionuclides, rhenium-186 (¹⁸⁶Re) and rhenium-188 (¹⁸⁸Re), with high efficiency, good stability, and convenience has been introduced by Bao et al. (13).

Thus far, previous intratumoral studies have been conducted in an intact tumor model. In a surgical setting, positive margins are the result of maximum resection, and injecting liposomal radionuclides intraoperatively into residual tumor could alter the pharmacokinetics and biodistribution compared with an unresected tumor. Thus, to determine the retention characteristics in an intraoperative setting, we first did an intraoperative biodistribution and retention study, with ^99mTc liposomes of various liposome surface charges and particle sizes injected into nude rat HNSCC tumor model after partial tumor resection. In this previous biodistribution study, we achieved average locoregional and intratumoral retention rates of 65.1 ± 6.8% and 29.5 ± 18.5%, respectively at 44 hours, thus confirming the indication to proceed with therapy study. In the present study, we report the data from our ¹⁸⁶Re-liposome intraoperative therapy study in a nude rat HNSCC xenograft model.

Preparation of liposomes. Liposomes possessing both glutathione and ammonium (pH) gradients were prepared and characterized with the use of a modified protocol (13). Liposomes with neutral and cationic surface charges and 100-nm diameter were manufactured. Neutrally charged liposomes comprised distearoylphosphatidylcholine (Avanti Polar Lipids), cholesterol (Calbiochem), and vitamin E (molar ratio, 54:44:2). Positively charged liposomes comprised distearoylphosphatidylcholine, 1,2-distearoyltrimethyl ammoniumpropane (Avanti Polar Lipids), cholesterol, and vitamin E (molar ratio, 49:5:44:2). Lipids were co-dried from chloroform to form a lipid film and desiccated overnight. The dried lipid film was rehydrated with 300 mmol/L sucrose in sterile water at 60 mmol/L total lipid concentration and warmed to 55°C followed by an overnight lyophilization. The dried lipid-sucrose mixture was rehydrated with 200 mmol/L glutathione (Sigma) and 300 mmol/L ammonium sulfate (Sigma) in sterile water at 60 mmol/L total lipid concentration, then subjected to five freeze-thaw cycles, followed by sequential extrusion through polycarbonate filters with different pore sizes (2 μm, 1 μm, 400 nm, and 200 nm, two passes each; and 100 nmol/L, five passes) at 55°C (Lipex Extruder, Northern Lipids). After extrusion, liposomes were repeatedly centrifuged at 41,000 rpm for 50 min and washed with 300 mmol/L ammonium sulfate in sterile water thrice to remove any unencapsulated glutathione. Liposome pellets were resuspended in 300 mmol/L ammonium sulfate containing 300 mmol/L sucrose in sterile water at 60 mmol/L total lipid concentration and stored at 4°C until needed.

Following manufacture, liposome sizes were measured with a 488-nm laser light scattering instrument (Brookhaven Instruments) and found to be 109.0 ± 11.1 and 130.5 ± 18.1 nm in diameter for neutral liposomes and cationic liposomes, respectively. Phospholipid content was measured via the Stewart assay (14). Liposomes were also checked for endotoxin levels and bacterial growth (University Hospital Pathology Laboratory, San Antonio, TX). No growth of bacteria was detected within 14-day culture and endotoxin levels were <2.5 EU/mL.

Preparation of ¹⁸⁶Re-liposomes. BMEDA (3.0 μL) was added into a vial containing 50 mg glucoheptonate, to which 2.0 mL nitrogen-degassed saline was further added. The mixture was magnetically stirred for 20 min followed by the addition of 240 μL 15 mg/mL stannous chloride solution. After adjusting the pH to 5.0, a 1.0 mL aliquot of the glucoheptonate-BMEDA-stannous solution was placed in a new vial. The vial was then flushed with nitrogen and sealed. ¹⁸⁶Re-perrhenate solution (100-120 mCi; University of Missouri Research Reactor) was added to the glucoheptonate-BMEDA-stannous solution and incubated at 80°C for 1 h. After incubation, the ¹⁸⁶Re-BMEDA solution was cooled to room temperature and adjusted to pH 7.0. Immediately before radiolabeling, liposomes were eluted with PBS (pH 7.4) through a PD-10 column to remove free ammonium sulfate and create an ammonium (pH) gradient. Eluted liposomes were added into the ¹⁸⁶Re-BMEDA solution and incubated for 1 h at 37°C. Labeled ¹⁸⁶Re-liposomes were separated from any free ¹⁸⁶Re-BMEDA, with PD-10 columns eluted with PBS (pH 7.4). The labeling efficiency of ¹⁸⁶Re-liposomes was between 45% and 55%.

HNSCC xenograft model in nude rats. Animal experiments were done according to the NIH Animal Use Guidelines and approved by the University of Texas Health Science Center at San Antonio Institutional Animal Care Committee. During all procedures, the animals were anesthetized with 1% to 3% of isoflurane (Vedco) in 100% oxygen through an anesthesia inhalation machine (Bickford). SCC-4 cell line (American Type Culture Collection) was cultured as described previously (15). Each male rnu/rnu athymic nude rat (Harlan) at 4 to 5 wk of age (75-100 g) was inoculated s.c. with 5 × 10⁶ of SCC-4 tumor cells in 0.20 mL of saline on the dorsum at the level of the scapulae. Tumor size was obtained by measuring the length (l), width (w), and thickness (t) of each tumor with calipers. Tumor volumes were subsequently calculated with the ellipsoid volume formula, V = (π/6)lwt. When the tumor volume reached ∼2 to 3 cm³, which typically occurred between 15 and 17 d after tumor cell inoculation, the tumor-bearing animals were used for study.

Intraoperative therapy of HNSCC xenograft positive surgical margin. On the day of the surgery, the average tumor volume for the neutral liposome control group (n = 4) and cationic liposome control group (n = 4) were 2.6 ± 0.5 and 2.0 ± 0.9 cm³, respectively. The average preoperative tumor volume for the ¹⁸⁶Re-neutral-liposome group (n = 8) and the ¹⁸⁶Re-cationic-liposome group (n = 8) were 3.9 ± 0.6 and 3.2 ± 0.7 cm³, respectively. All instruments were autoclaved, and the sterility of the operating area was maintained. The positive surgical margin model was created via a transverse cutaneous incision directly superior to the tumor and by resecting a portion of the tumor while leaving a tumor remnant of 0.5 to 1 cm³. Each residual tumor was injected with either 0.5 mL ¹⁸⁶Re-neutral-liposomes (4.75 ± 0.38 mCi total activity/rat, 13.8 ± 1.3 mg total lipid/kg body weight) or 0.5 mL ¹⁸⁶Re-cationic-liposomes (3.95 ± 0.14 mCi total activity/rat, 14.6 ± 1.5 mg total lipid/kg body weight), or 0.5 mL unlabeled control liposomes (12.7 ± 0.7 and 12.8 ± 0.8 mg total lipid/kg body weight for neutral and cationic liposomes, respectively) with the use of 1 mL tuberculin syringes with skin retracted. The skin was closed immediately following surgery and injection with interrupted sutures (3-O silk).

Postoperative follow-up and hematology. Body weight and tumor volume were obtained daily for all rats. Control groups were followed to day 15 postoperatively and then euthanized due to excessive tumor burden by cervical dislocation under deep isoflurane sedation. Therapy groups were followed to day 35 postoperatively and then euthanized. To monitor changes in WBC and platelet levels, blood (500 μL) was collected from the tail vein into tubes containing acid-citrate-dextrose anticoagulant preoperatively and every 3 to 4 d posttherapy until day 35. Blood cell levels were determined by automated counting at the University Hospital Pathology Laboratory (San Antonio, TX).

Intratumoral distribution of ¹⁸⁶Re-liposomes via micro–single-photon emission computed tomography/computed tomography imaging. Planar γ-camera images, pinhole-collimator single-photon emission computed tomography (SPECT) images, and computed tomography (CT) images were acquired with a micro-SPECT/CT scanner (XSPECT, Gamma Medica). Static planar images in lateral view were acquired at baseline and at 4, 20, 44, 72, 96, 116, and 140 h after ¹⁸⁶Re-liposome administration with parallel hole collimators. During each static image acquisition, a standard ¹⁸⁶Re source representing ∼5% of the total injected activity per rat was positioned outside the animal but still within the field of view for image quantification.

At 20 h following the intratumoral injection and immediately following static image acquisition, 1-mm pinhole-collimator SPECT images were acquired with the center of the field of view focused on the tumor of each animal (dual detectors, 32 projections/detector, 60 s/projection). The radius of rotation was maintained at around 5.30 cm with a field of view of ∼5.00 cm.

SPECT imaging was followed by CT image acquisition (X-ray source: 80 kVp, 280 mA, 256 projections) while precisely maintaining the position of the animal. Software provided with micro-SPECT/CT was used for SPECT and CT image reconstruction, including the SPECT/CT image fusion. SPECT images were reconstructed to produce image sizes of 56 × 56 × 56 with an image resolution of 0.95 mm. CT images were also reconstructed, resulting in image sizes of 512 × 512 × 512 with 0.15-mm image resolution.

Histopathology studies. After the animals were euthanized, the tumor remnant, kidney, liver, and spleen were dissected and fixed in 10% formalin for 48 h and embedded in paraffin; 5 to 6 μm sections were then stained with H&E and assessed in a light microscope. H&E slides were scanned at low resolution with a high-resolution laser scanner, and ×4, ×10, ×20, and ×40 pictures were taken with a digital camera (Canon PowerShot S3 IS).

Data analysis and radiation-absorbed dose calculation. The percentage of injected ¹⁸⁶Re activity retained in the tumor and its surrounding tissues, kidneys, and liver over time with the standard source as point of reference was calculated from planar images created by drawing the region of interest with AsiPro software (Siemens Medical Solutions USA, Inc.). The radioactivity with decay correction versus the time curve in each organ for each animal was drawn, and the area from time 0 to 140 h was calculated to obtain the cumulative radioactivity (Ã; mCi-h) in each organ for radiation-absorbed dose calculation.

Most of the radiation energy emitted by ¹⁸⁶Re is nonpenetrative radiation mainly from β-ray emission; the nonpenetrative-to-penetrative energy ratio is 16.7 (16). Nonpenetrative radiation from ¹⁸⁶Re has an average path length of 1.8 mm (17). If we consider the above characteristics and that only a small fraction of the local dose deposition is due to penetrative radiation energy, we can rationally neglect the radiation-absorbed dose contributed by penetrative radiation from ¹⁸⁶Re and only account for the radiation-absorbed doses from nonpenetrative radiation. Because there is a short path length for nonpenetrative radiation, we can also rationally assume that all of the nonpenetrative energies were deposited locally. Based on the above dose calculation model, we can use the following equation to calculate the average radiation-absorbed dose in each organ, including tumor (16):

Here, D is the average radiation-absorbed dose in Gy, Ã is the cumulative radioactivity in mCi-h calculated as described above, and m is the weight of the organ in grams.

To perform radiation-absorbed dose calculation, the tumor weight was calculated from the measured tumor volume at day 3, assuming the tumor density is 1 g/cm³. For the normal organ weight, the previous organ data [kidneys, 1.974 ± 0.129 g (n = 8) and liver, 10.71 ± 0.67 g (n = 8)] in our ^99mTc-liposome distribution studies were used.

Statistics. Data were summarized and graphed with Origin 7.5 software (Origin Lab). All of the average data were expressed as average ± SD. Paired t tests were used to compare the differences at different time points. One-way ANOVA was used to compare the differences between each group. The correlation between radiation-absorbed dose and tumor size change used linear correlation algorithm. Statistical analysis of significant difference used the standard of 95% confidence interval (P < 0.05).

Tumor volume trend.Figure 1A shows the average tumor volume progression for the two ¹⁸⁶Re-liposome therapy groups and the two corresponding unlabeled liposome control groups. Final tumor volume is reported as a percentage of the initial tumor volume at the end of the study, and postoperative day 3 is used as the initial reference due to the significant tumor swelling on day 1 and day 2 (Fig. 1B). On day 35, the average tumor volumes of ¹⁸⁶Re-neutral-liposome and ¹⁸⁶Re-cationic-liposome groups decreased to 25.6 ± 21.8% and 28.5 ± 32.2%, respectively. In comparison, tumors in the neutral and cationic liposome control groups showed rapid growth, with tumor volumes of 288.0 ± 37.3% and 292.2 ± 133.7%, respectively, at day 15.

When individual animals were analyzed, seven of eight animals in the ¹⁸⁶Re-neutral-liposome group showed progressive reductions in tumor volume (Fig. 2A) without regrowth, and one of these seven animals had no measurable tumor at day 29. The remaining animal in this group followed a similar trend as the other seven animals for the first 15 days but began to show tumor growth thereafter. The average percent tumor volume change for the seven residual tumors with no regrowth was 20.0 ± 16.3% compared with 64.6% for the one animal with tumor regrowth. Individual tumor volume changes for the eight animals receiving ¹⁸⁶Re-cationic-liposomes are shown in Fig. 2B. Six of the eight rats showed consistent tumor volume reduction (12.0 ± 8.2%) without regrowth. Two of the eight rats showed tumor regrowth around day 15 with average tumor volume of 78.0 ± 20% on day 35. In contrast, the two control groups showed no response from the unlabeled liposomes and showed relentless tumor enlargement.

Local retention, distribution, and radiation-absorbed dose.Table 1 illustrates the locoregional level of radioactivity derived from planar images for both the ¹⁸⁶Re-neutral-liposome and the ¹⁸⁶Re-cationic-liposome groups over the period of study. Both groups showed high local retention rates over 140 hours. Local retention for the ¹⁸⁶Re-neutral-liposome group ranged from 86.2 ± 1.0% injected activity at baseline to 18.8 ± 5.3% injected activity at 140 hours. In comparison, the ¹⁸⁶Re-cationic-liposome group showed a range of local retention from 81.6 ± 2.7% injected activity at baseline to 28.7 ± 3.6% injected activity at 140 hours. Curve simulations for the locoregional retention of both therapy groups fit a first-order exponential decay model (R² = 0.999 and 0.995, respectively), with the following general formula:

where y₀ represents the Y offset (i.e., the percentage of constant retention activity other than the cleared components following the exponential decay mechanism) and A₁ represents the percentage of activity that was cleared with a constant of t₁ for each type of ¹⁸⁶Re-liposomes. Subsequently, the half-clearance times correlated with the percentage of radioactivity of A₁ is t₁ × 0.693.

Table 2 outlines the clearance characteristics of the two ¹⁸⁶Re-liposome therapy groups based on the simulated clearance function described above. There were 13.8% and 18.4% injected activity leaving the local area within a few minutes and distributing systemically to ¹⁸⁶Re-neutral-liposome and ¹⁸⁶Re-cationic-liposome groups, respectively. For both groups, the locally remaining radioactivity displayed a stable, long-term retention with a half-clearance time of >40 hours. ¹⁸⁶Re-cationic-liposomes showed greater constant retention, which may be due to their improved affinity for the tumor tissue. At 140 hours, the ¹⁸⁶Re-cationic-liposomes had significantly higher % injected activity compared with the ¹⁸⁶Re-neutral-liposomes (P < 0.001; Table 1).

Table 3 shows the initial activity (mCi) injected into each animal and the total radiation-absorbed dose (Gy) received by the tumor in each animal during 140 hours. The ¹⁸⁶Re-cationic-liposome group received an average absorbed dose of 801.9 ± 184.9 Gy in tumor compared with 617.9 ± 187.0 Gy for the ¹⁸⁶Re-neutral-liposome group over the period of 140 hours.

Figure 3 shows the micro-CT, pinhole-collimator SPECT, and fusion images of the HNSCC tumor–bearing rats acquired 20 hours after intraoperative injection of the ¹⁸⁶Re-liposomes into the residual tumor. In the ¹⁸⁶Re-cationic-liposome rats, the radioactivity was located almost exclusively around the area of the tumor with little peritumoral extravasation, whereas the rats injected with ¹⁸⁶Re-neutral-liposomes showed extensive subcutaneous spread of the activity along the lateral aspects of the torsal and forelimbs. The behavior observed with ¹⁸⁶Re-neutral-liposomes correlates to the observed recoverable skin ulceration in a few animals treated with ¹⁸⁶Re-neutral-liposomes. However, the diffusion of ¹⁸⁶Re-neutral-liposomes through the cavity may track the passage of tumor cell local metastasis, which could emerge as a potential advantage of applying ¹⁸⁶Re-neutral-liposomes in intraoperative therapy.

Systemic organ retention and dose.Figure 4 illustrates the % injected activity for kidney and liver based on the planar γ-camera images of the animals acquired from baseline until 140 hours after the initial injection of the ¹⁸⁶Re-liposomes. The highest average liver % injected activity for the ¹⁸⁶Re-neutral-liposome and the ¹⁸⁶Re-cationic-liposome animals were 1.7 ± 0.1% at 20 hours and 1.6 ± 0.2% at 4 hours, respectively. The total liver radiation-absorbed dose received by the ¹⁸⁶Re-neutral-liposome and the ¹⁸⁶Re-cationic-liposome animals were 4.3 ± 0.2 and 3.5 ± 0.1 Gy, respectively. The maximum average kidney % injected activity for the ¹⁸⁶Re-neutral-liposome and the ¹⁸⁶Re-cationic-liposome animals were 12.1 ± 1.5% and 8.4 ± 1.3%, respectively, at 92 hours. The radiation-absorbed doses received by the kidneys for the ¹⁸⁶Re-neutral-liposome and the ¹⁸⁶Re-cationic-liposome groups were 149.6 ± 11.3 and 101.6 ± 15.5 Gy, respectively. The radiation-absorbed dose in nude rats would translate to ∼1 Gy in humans whose kidneys are ∼150 times the size of rat kidneys. This estimated value will be significantly less than the kidney dose limit of 27 Gy in radiation therapy (18).

Body weight growth and hematology.Figure 5A shows the body weight growth trend for all study groups. The control groups showed steady body weight gain beginning on postoperative day 1. Both ¹⁸⁶Re-liposome therapy groups steadily lost weight between day 1 and day 4 and began to gain weight consistently from day 5 until the end of the 35-day study. When comparing the therapy groups with their corresponding control groups, the progression of their growth curves is almost identical. The short recovery period for the therapy groups supports the notion that the intraoperative injection of ¹⁸⁶Re-liposomes has minimal systemic side effects.

The ¹⁸⁶Re-liposome therapy groups and their respective control groups showed no signs of bleeding tendencies, and the platelet counts for the therapy groups lay mostly within the range of their respective control groups. Furthermore, no animals showed signs of infection following surgery. The WBC count for both control groups (Fig. 5B) showed a stable WBC count for 15 days. The therapy rats displayed a sustained increase in WBC count for 15 days. Their WBC count eventually decreased to within the reference range (19). However, it is important to note that the ¹⁸⁶Re-neutral-liposome group showed persistently higher WBC count beyond day 14 compared with its cationic counterpart. This is most likely due to the higher tendency of neutral liposomes for systemic spread.

Toxicities. The most common side effect was the poor wound healing at the incision site. Half of the animals from each ¹⁸⁶Re-liposome therapy group experienced wound dehiscence even with meticulously placed interrupted sutures. This is expected to be contributed by the prolonged irradiation of the skin. Furthermore, four of the eight ¹⁸⁶Re-neutral-liposome rats presented with superficial skin ulceration (<4 mm in diameter) on the lateral aspects of the upper forelimbs at about day 14 and healed by day 20, which was not observed in the ¹⁸⁶Re-cationic-liposome animals or the control groups.

In terms of gastrointestinal symptoms, animals from both therapy groups showed loose and watery stools during the 1st week after therapy, and most rats began to excrete normal stools by week 2.

Histopathology. From H&E histopathology slides, control tumors (Fig. 6A) showed basophilic-stained cancer cells partitioned into lobules by the stroma and extracellular matrix. Each lobule has a central necrotic region and a periphery of expanding cancer cells. In comparison, animals from the therapy groups had tumors with multiple regions of extensive necrosis and interweaving granulation tissue (Fig. 6B). Most tumors only maintained a thin periphery of viable cancer cells. Furthermore, most ¹⁸⁶Re-treated tumors showed disintegration of the lobules and stroma and extracellular matrix disruption. Although the remaining tumors from successfully treated therapy animals showed areas of basophilic tumor cells, it is difficult to assess whether these cells are still viable tumor cells or dead cells from the basic H&E stain. Special histochemical stains will be helpful in determining the status of these remaining cells. There were no identifiable histopathologic differences in the kidney, liver, and spleen between control and therapy rats.

Approximately 50% of the patients diagnosed with HNSCC present with stage III or IV disease (1). Despite aggressive surgery and postoperative radiation therapy, long-term survival remains low (2), and the majority of treatment failures are due to locoregional recurrence (20). Furthermore, the recurrence rate becomes even higher when the anatomic limitations of the head and neck and the existing comorbidities preclude complete surgical resection, leaving patients with positive surgical margins (21). For these patients, radiation and/or chemotherapy become their only chance for cure (20). The modern evolution of radiation therapy for head and neck cancer, either as a single modality or as combination therapy with chemotherapy, centers on providing the maximum amount of dose to achieve local control while minimizing normal tissue toxicity. Radiation therapy is typically delivered through high-energy external beam radiation to irradiate the cancerous tissue. Yet, due to the close proximity of critical organs to the tumor, delivering the therapeutic dose of irradiation is often impossible without causing serious complications even with technological advancements in radiation dose distribution, such as intensity-modulated radiation therapy (22–24).

An alternative to surgery and radiotherapy is the integration of chemotherapy as induction or concurrent therapy to utilize the synergistic effects of chemoradiotherapy (25–27). High rates of locoregional control and organ preservation were achieved in a number of studies; however, these intense combination therapies are hampered by their dose-limiting toxicity with systemic administration. Furthermore, in an intratumoral environment, such as HNSCC, high interstitial pressure and the presence of viable tumor cells in the highly necrotic and avascular areas of the tumor render the uniform spread of chemotherapy agents difficult (28).

With the above considerations in mind, direct intratumoral drug administration in advanced HNSCC with positive margins has clear appeal by immediately achieving high drug concentration at the target site and avoiding the associated side effects of external beam radiation or systemic chemotherapy. Direct injection can deliver drugs into areas of the tumor that have high interstitial pressure and large areas of avascular necrosis. However, if free chemotherapy drugs or radionuclides were to be injected into the tumor, they would be cleared immediately. Thus, liposomes are particularly useful as nanoparticle carriers of therapeutic agents for local administration in HNSCC because of the increased local retention rates of the therapeutic agents. Due to the enhanced permeability and retention effect and the lack of a developed lymphatic system in solid tumors, the use of liposomes can further improve the retention rate of the encapsulated agent (9, 10).

Current investigations on locoregional application of liposomes include encapsulating liposomes with chemotherapeutic agents (29), genes (30), and radionuclides (12, 13). These are all feasible options to treat HNSCC locally; however, chemotherapy and gene therapy share some fundamental weaknesses. First, they both require direct interaction with the cell and, second, it is difficult to assess the intratumoral distribution of these agents without an imaging technique. Liposomal radionuclides, on the other hand, do not require a homogeneous distribution because β-emitting radionuclides (such as ¹⁸⁶Re) can provide remote killing of the cancer cell, and imaging techniques allow full assessment of the biodistribution of these agents locoregionally and systemically.

Thus far, no therapy study has been done with liposomal radionuclides. Harrington et al. (11) and Bao et al. (12) have both shown high local retention rates in HNSCC animal models. However, to treat positive margins in an intraoperative setting in which the tumor is partially resected could lead to different pharmacokinetics compared with injecting liposomal radionuclides into an intact tumor. Generally speaking, intratumoral injection of liposomal radionuclide without surgery is associated with a high immediate clearance rate and a slower phase I or phase II (if applicable) clearance rate (12). However, the postoperative inflammation and the destruction of the vasculature during surgery should reduce the immediate clearance rate (31). Data from this study (Table 2) showed low immediate clearance for both ¹⁸⁶Re-neutral-liposomes and ¹⁸⁶Re-cationic-liposomes at 13.8% and 18.4%, respectively. The long phase I half-clearance times of 48.95 and 42.04 hours for the ¹⁸⁶Re-neutral-liposomes and ¹⁸⁶Re-cationic-liposomes, respectively, also translate into high locoregional retention rate once the immediate clearance has passed. These data support the notion that the intraoperative injection of liposomal radionuclide leads to higher retention compared with direct intratumoral injection without surgical resection.

The high locoregional retention rate led to excellent therapeutic results. In this study, we were able to deliver 801.9 ± 184.9 and 617.9 ± 187.0 Gy doses of radiation, respectively, over 140 hours for the ¹⁸⁶Re-cationic-liposome and the ¹⁸⁶Re-neutral-liposome groups. At the end of the 35-day study, only one animal injected with ¹⁸⁶Re-neutral-liposomes and two animals injected with ¹⁸⁶Re-cationic-liposomes showed apparent regrowth. The final tumor volumes for the ¹⁸⁶Re-neutral-liposome group and ¹⁸⁶Re-cationic-liposome group were only 25.6 ± 21.8% and 28.5 ± 32.2%, respectively, of the initial tumor volume. The discrepancy between the final tumor size and the number of recurrent tumors for the ¹⁸⁶Re-neutral-liposome group and the ¹⁸⁶Re-cationic-liposome group is most likely the result of the actual microenvironment and macroenvironment of the specific tumor.

The tumoricidal effect of ¹⁸⁶Re-liposomes can also be shown from the histopathology studies. Control tumor sections had intact stroma and extracellular matrix partitioning tumor cells into distinct lobules; however, treated tumor samples typically displayed disruption of the stroma and extracellular matrix that is vital for the survival and growth of tumor cells by releasing tumor growth factors and nutrients (32, 33). The combination of surgical resection and irradiation from liposomal therapeutic radionuclides synergistically disrupt the macroenvironment and microenvironment of the tumor. Furthermore, surgery reinforces the efficacy of radiation by exposing hypoxic tumor cells to oxygen; it is well known that hypoxia limits the tumoricidal effects of radiation (34). The above reasons make this intraoperative injection of liposomal therapeutic radionuclides a particularly effective treatment option in a setting of positive surgical margin in advanced HNSCC.

Liposomal radionuclide therapy cannot be considered a feasible treatment option if the associated side effects are unacceptable. In this study, all side effects were acute and transient, and no complications were observed. For the therapy group animals, there is initial weight loss and loose stools that occurred in the 1st week. At approximately the end of week 1, all the animals began to show consistent weight gain and normal stools, similar to the control animals. From the dermatologic standpoint, half of the animals from each therapy group showed poor wound healing. This mainly contributed to the prevention of epithelialization by radiation damage. With deeper-seeded tumor and thicker skin, humans should experience a less-pronounced effect. A therapy group–specific dermatologic side effect was superficial skin ulcerations on the lateral aspects of the upper forelimbs observed only in four of the eight ¹⁸⁶Re-neutral-liposome rats. The skin ulcerations presented at about day 14 and healed by day 20. From the SPECT/CT fusion images in Fig. 3, we noticed that the ¹⁸⁶Re-neutral-liposomes tended to spread into the lateral aspects of the torsal and forelimb area. This behavior emerges as a potential advantage in applying ¹⁸⁶Re-neutral-liposomes for the simultaneous treatment of microscopic disease around the locoregional area. From a hematologic standpoint, Fig. 5B illustrates that the ¹⁸⁶Re-liposome–treated animals only showed a slight increase in WBC count compared with the animals that received surgical resection only. WBC levels for all control and therapy animals eventually settled within their age-appropriate normal levels (19).

Overall, the data presented in this study open new avenues in the treatment of unresectable advanced HNSCC. For future clinical studies, patients will undergo surgery in an attempt to achieve maximal resection. When maximal resection has been achieved, the ¹⁸⁶Re-liposomes would be injected into the positive margin and/or around the tumor bed. For follow-up testing, we could use positron emission tomography, CT, or magnetic resonance imaging to check for recurrences.

No potential conflicts of interest were disclosed.