Abstract
Objective
The objective of the study is to examine the utility of positron emission tomography (PET) for staging and restaging after treatment of paranasal sinus carcinomas.
Study design
Retrospective data review was done.
Subjects and methods
Patients selected underwent PET for sinonasal neoplasms from 2003 to 2008 at a tertiary care referral center.
Results
Seventy-seven scans were reviewed from 31 patients. The pathologies included olfactory neuroblastoma (n = 9), squamous cell carcinoma (n = 6), sinonasal undifferentiated carcinoma (n = 6), sinonasal melanoma (n = 6), and minor salivary gland carcinomas (n = 4). The positive predictive value of studies performed for restaging at the primary, neck, and distant sites were 56%, 54%, and 63%; negative predictive values were 93%, 100%, and 98%, respectively. During restaging, 32% of patients were accurately upstaged secondary to neck or distant site involvement.
Conclusion
Positron emission tomography serves as a useful adjunct to conventional imaging in the management of sinonasal malignancies. Negative studies are effective in predicting absence of disease as seen in the consistently high-negative predictive values. Positive studies need to be viewed cautiously given the high rate of false-positive studies. When viewed in conjunction with clinical examination, endoscopic assessment, and focused biopsies, they may effectively result in a more accurate assessment of the extent of disease.
1
Introduction
Sinonasal malignancies are rare and account for less than 6% of all head and neck malignancies . As a group, they represent a heterogeneous group of malignancies with a range of biologic activity. Given the paucity of cases, the literature involves several institutional reviews of these heterogenous malignancies grouped together by the site of primary tumor in an effort to extract some clinically pertinent data . Irrespective of the pathology, most patients present with advanced local disease with infrequent distant metastases. In a combined series of 220 patients, 64% of patients with paranasal sinus cancers had locally advanced disease, and approximately 2% presented with distant metastases . Rates of regional disease tend to vary across studies with a range of 4% to 29% . Despite the variation in rates, the presence of cervical disease is generally regarded as a poor prognostic indicator . Given the importance in determining the status of the neck, conventional imaging modalities and nuclear medicine studies are commonly used to assist in the detection of disease.
Although fused positron emission tomography (PET)/computed tomography (CT) is emerging as an integral diagnostic tool in the management of head and neck cancers , its utility for sinonasal malignancies is largely undefined. A retrospective study demonstrated that PET/CT provided additional information to CT or magnetic resonance imaging (MRI) for restaging of locoregional disease in sinonasal malignancies but provided no additional value in the initial staging of patients . The aim of this study was to examine the role of PET/CT in the management of sinonasal malignancies, specifically in the initial staging and restaging of disease.
2
Material and methods
A retrospective chart review was conducted of patients who underwent PET/CT for malignancies of the paranasal sinuses at the Cleveland Clinic between September 2003 and April 2008. The patients were extracted from a larger database of all patients undergoing PET/CT at the Head and Neck Institute. The study was approved by the Cleveland Clinic Institutional Review Board. All patients had biopsy-proven malignancy of the paranasal sinuses. Only fused PET/CT images obtained at the Cleveland Clinic were evaluated; those from outside institutions and PET studies in the absence of coregistered attenuation correction CT were not included in the review.
Positron emission tomography/CT scans performed at the Cleveland Clinic were scanned after typical preparation, which involves fasting for 4 hours before the test. Patients were injected with 15 to 20 MCi FDG based on weight. They were imaged 60 minutes later from the skull base to midthigh with the arms raised with the exception of images for melanoma, which routinely extend down to patient’s feet. In addition, for these cancers, dedicated head and neck PET/CT images were also obtained with the arms down to limit attenuation artifacts. A Siemens Biograph 16 PET/CT scanner (Siemens Medical Systems, Hoffmann Estates, IL) was used with imaging time of approximately 3 minutes per bed position. Image reconstruction used the iterative ordered subset expectation maximization algorithm (2 iterations and 8 subsets) and 5-mm filter with the CT data used for attenuation correction. Positron emission tomography/CT studies were reviewed by nuclear medicine physicians using available coregistration software. Standardized uptake value (SUV) calculations were obtained using Siemens software for volumes of interest drawn around lesions of concern. Maximal SUV was recorded for analysis.
Positron emission tomography/CT scans were divided based on indications: staging and restaging. Scans performed for staging indications had biopsy-proven malignancy and were typically performed as a means of detecting regional or metastatic disease. Scans performed for restaging indications typically followed the completion of treatment or in the setting of subsequent surveillance and were performed as a means of detecting both recurrent disease and metastases. At our institution, there are no uniform protocols for obtaining PET/CT studies—the decision is left to the discretion of the surgical, medical, or radiation oncologist. Although the timing of these studies in relation to treatment is not uniform within the institution, they generally tend to occur at a minimum of 3 months after completion of radiation.
Patients with primary nasal vestibule tumors were excluded, as the tumors are more related to skin cancers than sinonasal malignancies. Nasopharyngeal carcinoma and cancers originating in the oral cavity with extension into the paranasal sinuses were also excluded. One patient with melanoma originating in the nasopharynx was included. Non–squamous cell primary sinonasal tumors were assigned TNM classification analogous to squamous cell carcinoma described by the American Joint Committee on Cancer—TNM Classification . All patients treated initially by surgery underwent gross total resection of disease. Two patients had residual microscopic disease both along the internal carotid artery at the level of the sphenoid sinus. One patient who presented with advanced stage metastatic sinonasal undifferentiated carcinoma did not undergo surgery after PET/CT, and biopsy confirmed the diagnosis, opting rather for palliative chemoradiation.
Data recorded included demographics, histology, tumor stage, treatment, additional imaging (CT or MRI), PET/CT results, maximum SUV, time of study in relation to conclusion of treatment, and how PET/CT altered treatment course.
Because of the lack of evidence to support the use of SUV in determining positivity , no SUV cutoff value was used while reviewing PET/CT images. Any increase in 18 F-fluorodeoxyglucose uptake above physiologic uptake as determined by the nuclear medicine physicians was considered positive. Each study was assessed for its ability to detect disease at the primary site, the neck, and distant sites. In determining the predictive value of PET/CT, histology was used as a criterion standard in determining the presence or absence of malignancy. In cases where histology or cytology was not available, clinical course was used as the criterion standard. If there was insufficient clinical data and follow-up to determine the true extent of disease, the scan was not used for analytic purposes.
2.1
Statistical analysis
Each PET/CT studied was evaluated for its ability to predict disease at the primary site, the neck, and distant sites. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated for each site using a series of 2 × 2 contingency tables. Ninety-five percent confidence intervals were calculated for each value. Given the metabolic changes of the local tissues that occur after treatment, analysis of staging and restaging (including surveillance) was performed independently.
2
Material and methods
A retrospective chart review was conducted of patients who underwent PET/CT for malignancies of the paranasal sinuses at the Cleveland Clinic between September 2003 and April 2008. The patients were extracted from a larger database of all patients undergoing PET/CT at the Head and Neck Institute. The study was approved by the Cleveland Clinic Institutional Review Board. All patients had biopsy-proven malignancy of the paranasal sinuses. Only fused PET/CT images obtained at the Cleveland Clinic were evaluated; those from outside institutions and PET studies in the absence of coregistered attenuation correction CT were not included in the review.
Positron emission tomography/CT scans performed at the Cleveland Clinic were scanned after typical preparation, which involves fasting for 4 hours before the test. Patients were injected with 15 to 20 MCi FDG based on weight. They were imaged 60 minutes later from the skull base to midthigh with the arms raised with the exception of images for melanoma, which routinely extend down to patient’s feet. In addition, for these cancers, dedicated head and neck PET/CT images were also obtained with the arms down to limit attenuation artifacts. A Siemens Biograph 16 PET/CT scanner (Siemens Medical Systems, Hoffmann Estates, IL) was used with imaging time of approximately 3 minutes per bed position. Image reconstruction used the iterative ordered subset expectation maximization algorithm (2 iterations and 8 subsets) and 5-mm filter with the CT data used for attenuation correction. Positron emission tomography/CT studies were reviewed by nuclear medicine physicians using available coregistration software. Standardized uptake value (SUV) calculations were obtained using Siemens software for volumes of interest drawn around lesions of concern. Maximal SUV was recorded for analysis.
Positron emission tomography/CT scans were divided based on indications: staging and restaging. Scans performed for staging indications had biopsy-proven malignancy and were typically performed as a means of detecting regional or metastatic disease. Scans performed for restaging indications typically followed the completion of treatment or in the setting of subsequent surveillance and were performed as a means of detecting both recurrent disease and metastases. At our institution, there are no uniform protocols for obtaining PET/CT studies—the decision is left to the discretion of the surgical, medical, or radiation oncologist. Although the timing of these studies in relation to treatment is not uniform within the institution, they generally tend to occur at a minimum of 3 months after completion of radiation.
Patients with primary nasal vestibule tumors were excluded, as the tumors are more related to skin cancers than sinonasal malignancies. Nasopharyngeal carcinoma and cancers originating in the oral cavity with extension into the paranasal sinuses were also excluded. One patient with melanoma originating in the nasopharynx was included. Non–squamous cell primary sinonasal tumors were assigned TNM classification analogous to squamous cell carcinoma described by the American Joint Committee on Cancer—TNM Classification . All patients treated initially by surgery underwent gross total resection of disease. Two patients had residual microscopic disease both along the internal carotid artery at the level of the sphenoid sinus. One patient who presented with advanced stage metastatic sinonasal undifferentiated carcinoma did not undergo surgery after PET/CT, and biopsy confirmed the diagnosis, opting rather for palliative chemoradiation.
Data recorded included demographics, histology, tumor stage, treatment, additional imaging (CT or MRI), PET/CT results, maximum SUV, time of study in relation to conclusion of treatment, and how PET/CT altered treatment course.
Because of the lack of evidence to support the use of SUV in determining positivity , no SUV cutoff value was used while reviewing PET/CT images. Any increase in 18 F-fluorodeoxyglucose uptake above physiologic uptake as determined by the nuclear medicine physicians was considered positive. Each study was assessed for its ability to detect disease at the primary site, the neck, and distant sites. In determining the predictive value of PET/CT, histology was used as a criterion standard in determining the presence or absence of malignancy. In cases where histology or cytology was not available, clinical course was used as the criterion standard. If there was insufficient clinical data and follow-up to determine the true extent of disease, the scan was not used for analytic purposes.
2.1
Statistical analysis
Each PET/CT studied was evaluated for its ability to predict disease at the primary site, the neck, and distant sites. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated for each site using a series of 2 × 2 contingency tables. Ninety-five percent confidence intervals were calculated for each value. Given the metabolic changes of the local tissues that occur after treatment, analysis of staging and restaging (including surveillance) was performed independently.
3
Results
Thirty-one patients fulfilled the inclusion criteria and underwent 78 PET/CT scans during the study period (56% men, 44% women; mean age, 58 years). One study was excluded because of lack of clinical information to ascertain the true nature of disease; thus, only 77 scans were reviewed for this report. Mean and median follow-up was 43.7 and 45 months, respectively. The time that elapsed between end of treatment (whether surgery or radiation ± chemotherapy) and the initial restaging PET/CT was a median of 3 months. Olfactory neuroblastoma was the most frequent diagnosis and was found in 9 patients (28%) ( Table 1 ). Most primary tumors were located in the ethmoid cavity (69%) with a smaller percentage occurring in the maxillary (22%), sphenoid sinuses (6%), and nasopharynx (3%). Most patients had advanced stage cancer (25% with T3 and 47% with T4). Of the patient population studied, 25% developed regional metastases, and 19% developed distant disease. All the patients in this series had radiation in the course of their management.
