Positron Emission Tomography for Basal Cell Carcinoma of the Head and Neck FREE

POSITRON EMISSION tomography (PET) is a unique imaging technique that detects positron release from radioactive substances and provides cross-sectional physiologic information. Positron emission tomographic imaging commonly uses 2-deoxy-2-[¹⁸F]-fluoro-D-glucose (FDG), a positron imaging agent, to measure the metabolic rate of tissue noninvasively. Tumors tend to metabolize more glucose than normal tissue, thus mobilizing the imaging tracer and being detectable by the PET scanner.^1- 3 This imaging has been proven to be effective for the staging, treatment planning, and monitoring of many extracutaneous cancers, including brain tumors,⁴ colorectal cancer,^3,5- 6 lung cancer,^7- 13 pancreatic cancer,^14- 15 breast cancer,^16- 18 esophageal and gastric cancer,^19- 21 bladder cancer,²² and prostate cancer.^23- 24

Use of the FDG-PET has also been successful in patients with metastatic melanoma for early detection of metastases,^25- 27 accurate staging,^28- 33 and close follow-up.^34- 35 To our knowledge, the use of PET in other cutaneous neoplasms has not been reported.

Basal cell carcinoma (BCC), while the most common human malignancy,³⁶ is the least studied with radiologic imaging techniques, and appropriately so given its usual limited soft tissue involvement. Patient 1 prompted us to further evaluate if PET imaging had any role in BCC evaluation. This patient had a multiple recurrent, infiltrative BCC of the periauricular region and presented with neurologic deficits of cranial nerves V and VII. Extensive tumor spread was suspected clinically, and anatomic localization before surgery was attempted. A computed tomographic (CT) scan was negative, followed by a negative PET scan. Despite these findings, extensive tumor spread was found involving the facial and trigeminal nerves in the infratemporal fossa to the base of the foramen ovale. This prompted the question as to the ability of PET scanning to detect BCC at all and, if so, whether this was affected by any clinical or histologic features.

In this pilot study, 6 patients are described. There were 2 infiltrative and 4 nodular BCCs. All tumors were larger than 1.0 cm. Positron emission tomographic scanning was positive in 3 patients and closely outlined the soft tissue extension of the tumor in one of the patients. Herein, we report our primary findings of the ability of PET imaging to identify the presence of BCC. We also discuss BCC tumor characteristics, such as recurrence, histologic subtypes, location, and size of the tumor, which may affect tumor detection by FDG-PET.

From January 1, 1997, to December 30, 1998, 6 patients with BCCs of the face and scalp underwent PET scanning of the anatomic area involved by the tumor (Table 1). The study was approved by our local institutional review board. Tumor sites included preauricular (1 patient), nose (2 patients), cheek (2 patients), and scalp (1 patient). Histologic subtypes included infiltrative (1 morpheaform and 1 micronodular) and nodular (4 patients). Three tumors were recurrent, and 3 were primary. All tumors were clinically apparent, with the nodular subtype the most noticeable clinically, with an exophytic growth pattern. The infiltrative subtypes were sclerotic, flat, and subtle appearing.

The FDG-PET imaging was performed on an ECAT 951/31 PET scanner (Siemens Medical Systems, Inc, Hoffman Estates, Ill). This tomograph has an axial field of view of 10.8 cm that is composed of 16 bismuth germanate rings producing 31 transaxial images (in-plane and cross-plane). It has an axial resolution of 5.7 mm full width at half maximum from a point source at a configuration of a 10.0-cm radius from the center. The tomograph has whole-body imaging capability.

The F 18 fluoride was produced by an on-site RDS 112 cyclotron (Siemens Medical Systems, Inc). The F 18 fluoride ions were transferred to an automated system for synthesis of FDG by the Hamacher method. The FDG was tested for sterility, pyrogenicity, and radiochemical purity on each production run.

Positron emission tomographic images of patients with BCC were obtained before Mohs surgery. A single 10.8-cm-bed position was acquired centered at the tumor and was performed 50 minutes after intravenous injection of 10 mCi (370 MBq) of FDG. Attenuation correction of the images was used in all patients.

The PET images were reconstructed using filtered backprojection with a Hann window of 5.0-mm width. Emission data were corrected for scatter, random events, and dead time losses using the manufacturer's software.

Interpretation of PET results was performed blinded from clinical and other imaging information. Scans were interpreted visually as to the presence or not of a malignancy in the region imaged, based on the presence of abnormal hypermetabolism. Axial, sagittal, coronal, and 3-dimensional surface projections were examined using vendor-provided computer software and were interpreted in gray scale.

A 61-year-old woman presented with a recurrent infiltrative BCC of the right temple and preauricular area. It had been excised 3 times and treated with radiation therapy. Motor (facial nerve frontal branch) and sensory (trigeminal nerve mandibular division) neurologic deficits were present. Computed tomographic and FDG-PET scans were negative. Histopathologically, perineural spread of BCC was found in the facial and trigeminal nerves into the infratemporal fossa to the foramen ovale. The patient was managed in collaboration with the Mohs surgery, head and neck oncology, and neurosurgical services. The tumor was unresectable, and the patient underwent postoperative stereotactic radiation therapy. The histologic subtype of the tumor was morpheaform.

A 56-year-old woman with no history of BCC presented with a 1.5 × 1.5-cm pink pearly nodule with crusting and telangiectasia involving the left cheek and nasolabial fold (Figure 1A). A PET scan demonstrated a hypermetabolic focus at the tumor region, with accurate detail (Figure 1B) that correlated well with the extent of the soft tissue invasion. The PET image showed a tumor mass in a figure-eight configuration, with a smaller mass of tumor extending in a deep fashion to just above the periosteum of the maxilla. Histopathologic examination of the tumor demonstrated a similar configuration, with a thin stalk of tumor connecting to a deeper nodular mass of tumor. The tumor did not involve the periosteum. After Mohs excision of the tumor, the wound size was 3.2 × 2.5 cm. The histologic subtype of the tumor was nodular.

A 63-year-old man with no history of BCC presented with a 3.1 × 3.0-cm ulcerated, pink pearly plaque with crusting, located at the left temporal scalp (Figure 2A). The PET scan demonstrated a hypermetabolic focus at the tumor region only involving soft tissues (Figure 2B). The wound size was 5.0 × 5.4 cm to the periosteum after Mohs excision. The histologic subtype of the tumor was nodular.

A 79-year-old man presented with a recurrent 4.0 × 2.8-cm BCC of the right cheek, which had been excised 10 years prior. It appeared as a sclerotic, depressed plaque, with an annular raised border. A PET scan demonstrated a hypermetabolic focus at the tumor region only involving soft tissues. The tumor was removed by Mohs surgery. The wound size was 6.0 × 6.0 cm to the level of the fascia. The histologic subtype of the tumor was nodular.

A 66-year-old woman presented with a recurrent 1.8 × 1.6-cm sclerotic BCC of the nasal dorsum. The PET scan was negative. After Mohs surgery, the wound size was 3.2 × 3.0 cm to the level of muscle and perichondrium focally. The histologic subtype of the tumor was micronodular.

A 95-year-old man with a primary history of BCC presented with a 1.2 × 1.2-cm pink pearly nodule located at the right nasal wall. The PET scan was negative. The tumor was removed by Mohs surgery. The wound size was 1.8 × 1.6 cm to the level of muscle. The histologic subtype of the tumor was nodular, and ulceration was present.

In patients 2, 3, and 4, there were hypermetabolic foci, which correlated well with the soft tissue distribution of the tumor. All 3 tumors were of the nodular subtype histologically and the noduloulcerative subtype clinically. All tumors were larger than 2.0 cm and clinically had obvious tumor volume that was exophytic and invading the soft tissue. Patient 2 demonstrates the degree of detail that can be seen with PET imaging.

In patients 1, 5, and 6, PET scans failed to demonstrate any foci with increased metabolic activity. Histologically, infiltrative patterns were present in 2 patients and nodular in 1 patient. Clinical features were subtler in the 2 infiltrative tumors. In patient 6, the nodular tumor was smaller (1.2 cm) than those that demonstrated increased metabolic activity and had less soft tissue involvement. The disease in patient 1 represents a histologically aggressive recurrent morpheaform BCC with extensive perineural spread. Despite the prominent microscopic extent of the tumor, there was no hypermetabolic focus identified on the PET images.

Tumor localization through FDG-PET is achieved by increased glucose uptake, a result of the higher rate of glycolysis in tumor tissue compared with surrounding normal tissue.³⁷ An increased hexokinase–glucose-6-phosphatase enzyme ratio is also found in many tumors. The FDG molecule can be phosphorylated by hexokinase, and in the relative absence of glucose-6-phosphatase, phosphorylated FDG is trapped within tumor cells, making FDG-PET a particularly attractive diagnostic tool for tumor location.³⁰

The increased metabolic rate, along with the changes in enzyme ratio in neoplastic tissue occurring before the increase in tumor size,¹ provides accurate tumor localization, extension, and early detection of metastases. This can occur before the tumor is clinically evident or of an abnormal size.²⁵ The efficiency of FDG-PET in the assessment and management of advanced melanoma is well reported, as is its improved specificity compared with the conventional cross-sectional imaging modalities of chest x-ray films, ultrasonography, and computed tomography.^{1,4,28,30,35,38- 42}

In our study, the objective was to determine if FDG-PET could detect the presence of BCC at all. To our knowledge, the use of this imaging technique for this tumor has not been previously reported. Although our study sample is small, we were able to show that FDG-PET can demonstrate metabolic tumor activity, with some degree of corresponding anatomic accuracy with regard to soft tissue invasion. With the study group being small, it is difficult to draw many conclusions, but a few inferences are made. First, the histologic subtype of the BCC affects FDG-PET detection, with the nodular histologic subtype more likely to be positive. This most likely reflects the larger individual tumor masses, which are more cohesive and create a dense tumor–surrounding stroma ratio. This is in contrast to the infiltrative subtypes of BCC, morpheaform and micronodular BCC. These are composed of much smaller islands and spikes of tumor often embedded in a more dense fibrous stroma, creating a smaller ratio of tumor–surrounding stroma, or a less dense and a more diffuse tumor pattern. This could result in small tumor aggregates taking up the FDG radionuclide, but not in a concentrated fashion to achieve the threshold necessary to display metabolic activity with the PET scanner. It may also reflect a differential in tumor metabolism and growth rate between the nodular and infiltrative histologic subtypes.

Tumor size appears to also affect PET scan results. All 3 tumors that showed metabolic activity were larger than 2.0 cm. They also had an obvious growth extension above the skin surface and a deeper soft tissue involvement. The tumor volume is greater and denser and can capture a more localized area of radionuclide uptake and metabolism, leading to greater amounts of radioactivity and easier detection.

Another factor that may affect the sensitivity of FDG-PET with BCCs is related to the vascularity of the tumor. Doppler perfusion imaging has revealed that BCC is significantly less perfused than melanoma,⁴³ and tumor microvessel density demonstrates that BCCs are less vascularized tumors than squamous cell carcinoma or keratoacanthomas.⁴⁴ The low level of blood flow through BCC compared with other skin neoplasms may lower the sensitivity of FDG-PET in tumor detection.

A competing factor that may contribute to the lower sensitivity of FDG-PET in our study is the fact that there is physiologic enhanced glucose uptake in the brain,^4,29 which may be imaged and obscure the increased glucose uptake in the tumor. It has been shown in patients with melanoma that FDG-PET imaging of the brain is not as sensitive as conventional imaging studies (CT or magnetic resonance imaging) as a result of this phenomenon.³⁵ This could also hinder detecting activity in tumors with perineural spread near the brain, as in patient 1.

New developments to increase tumor detection involve monoclonal antibody–PET imaging, such as in prostate cancer.⁴² The anti-human epithelial antigen has been developed specifically against the BCC membrane antigen.^45- 48 Its radiolabeled effectiveness combined with PET in detection of BCC is promising, and its use awaits future investigation.

Another new imaging technology is the fusion of anatomic and physiologic tomographic images.⁴⁹ In this way, PET studies can be interpreted with the corresponding anatomic images, such as magnetic resonance imaging or CT, to improve spatial accuracy. This has been used in patients with melanoma.

In summary, our study demonstrates that FDG-PET imaging can detect the presence of BCC. Its ability appears to be dependent on tumor density and volume, with the nodular histologic subtype more likely to be imaged. Use of FDG-PET imaging is of limited efficacy in evaluating infiltrative histologic subtypes of BCC, those with perineural spread, and those that have evidence of minimal tumor present clinically. Minimal tumor volume and low tumor density may contribute to negative FDG-PET imaging.

We realize that the use of PET imaging for BCCs will not be a common consideration, nor should it. Positron emission tomographic imaging is primarily used as a whole-body screening test for tumor metastasis. Most BCC is localized, with low potential of regional or distant spread. However, there are reports of BCC with locally aggressive behaviors, such as perineural spread or soft tissue involvement. Conventional imaging studies such as CT or magnetic resonance imaging scans are often nonrevealing in such situations. Defining the role of adjunctive imaging studies in such rare cases is useful.

Corresponding author and reprints: Scott W. Fosko, MD, Department of Dermatology, Saint Louis University Health Sciences Center, 1402 S Grand Blvd, St Louis, MO 63104 (e-mail: foskosw@slucare1.sluh.edu).

Accepted for publication April 24, 2003.

This study was presented as a poster at the American Society of Dermatologic Surgery–American College of Mohs Micrographic Surgery and Cutaneous Oncology Combined Annual Meeting; October 31-November 3, 2002; Chicago, Ill.