Approval for this retrospective study was obtained from the Institutional Review Board. Patients were identified through a search of F-18 FDG PET/CT dictation reports stored in the University Hospital database at the Advanced Imaging Center in University Hospital (Newark, USA) between 01/01/2010 and 06/30/2018. Patients who underwent a preoperative F-18 FDG PET/CT for submandibular or sublingual salivary gland tumors with comparative CT or MRI were selected. In addition, 4 patients with postoperative F-18 FDG PET/CT scans were included with comparative CT or MRI in subsequent follow-up evaluations.

A total of 18 patients with either submandibular or sublingual salivary gland tumors were selected from the University Hospital database between January 2010 and June 2018. From the 18 subjects included in the analysis, a total of 43 F-18 FDG PET/CT scans with contrast-enhanced CT or MRI were obtained for this study (Table I). A total of 14 F-18 FDG PET/CT imaging studies were performed for initial diagnosis and cancer staging, while the remaining 29 imaging studies were performed to restage the disease and evaluate recurrence during the follow-up period. The 29 restaging F-18 FDG PET/CT scans were performed on 8 patients.

The histopathological diagnoses included 5 mucoepidermoid carcinomas, 3 adenoid cystic carcinomas, 3 lymphomas, 1 squamous cell carcinoma, 1 basal cell adenocarcinoma, 1 submandibular duct carcinoma, 1 synovial cell sarcoma, 2 pleomorphic adenomas and 1 lymphoid hyperplasia. The mean age of the patients was 58.1 years, with a standard deviation of 13.4 years. There were 6 men and 12 women. The majority of the tumors (72%) were from the submandibular gland, while 5 tumors (28%) were from the sublingual gland.

All patients without distant metastasis were treated with curative surgical resection. Patients diagnosed with malignant tumors were offered adjuvant or postoperative radiation. In addition, the 2 patients who presented with distant metastasis (lung, liver and/or lymph nodes) underwent chemotherapy.

Combined PET-CT was performed using a PET-CT scanner (Discovery LS; GE Healthcare) and standard techniques. The patients were fasted for at least 6 h prior to the examination and their blood glucose level was <250 mg/dl (normal, <200 mg/dl for non-diabetic and <250 mg/dl for diabetic patients). The patients received a barium sulfate oral suspension (Readi-CAT^® 2; Bracco Diagnostics Inc.). No intravenous contrast agent was used. Spiral low-dose CT (80 mA, 140 kV and 4 mm section thickness) was performed in the cranio-caudal direction covering the areas from the vertex to the mid-thigh for the purpose of attenuation correction and anatomic localization. Thereafter, an emission scan was performed in a reverse direction.

MIM imaging software, version 6.4 (MIM Software, Inc.) was used for image display and analysis. The whole-body maximum-pixel intensity projection was used for visual evaluation. The maximum standardized uptake value (SUVmax) of the lesions was recorded. SUVmax is reported to be a more accurate estimate of the true SUV compared with SUVmean (17). For this reason, the use of SUVmax is becoming increasingly more common. Studies have demonstrated that SUVmax is a robust metric for prognosis in certain cancers, and it may also be used to assess response to treatment (17-22).

CT and MRI scans were interpreted by a radiologist who specializes in head and neck imaging. Intravenous contrast material was used in both types of imaging. The scans were analyzed by the extent of the salivary gland lesion, as well as any significant cervical lymph nodes, without any prior knowledge of the clinical findings. Primary salivary lesions were noted on extent, enhancement pattern and location. The regional lymph nodes of the head and neck were also evaluated on the basis of size and SUV value. Both CT and MRI are considered as acceptable methods to be used interchangeably in the evaluation of salivary gland tumors and anatomic imaging. The current literature states that cross-sectional imaging, either CT or MRI, may be used to evaluate salivary gland tumors at the discretion of the attending physician (21). Our sample set included 41 CT and 2 MRI scans from our patients that were used to compare to PET/CT studies. As the 2 MRIs are too small a sample to perform a separate statistical analysis, CT and MRI scans were pooled together.

All PET/CT studies were reviewed by a nuclear medicine physician who is experienced in interpreting head and neck imaging. The images were reviewed without having any knowledge of the clinical findings. The scans were initially reviewed for any abnormal uptake of F-18 FDG in the salivary gland and neck lymph nodes by using the maximal uptake values with higher intensities compared with those of the surrounding tissues. Any abnormal location or asymmetry in the uptake of F-18 FDG was also noted. The interpretation of the nuclear medicine physician was then revised based on the anatomic information of the uptake based on the combined F-18 FDG PET/CT images.

Histopathological and/or cytopathological diagnosis was obtained for all cases of F-18 FDG PET/CT imaging that were included in the present study.

For initial lymph node staging and detecting locoregional recurrence, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of PET/CT, CT or MRI were calculated by comparison with pathological findings and expressed as percentages. The confidence intervals (CIs) for diagnostic accuracy are ‘exact’ Clopper-Pearson CIs. P<0.05 was considered to indicate statistically significant differences.

A total of 18 patient cases were analyzed in the present study, including 6 men and 12 women, with a mean age of 58.1±13.8 years (range, 38-85 years). The characteristics of the patients included in the study are summarized in Table I. The findings of F-18 FDG PET/CT and CT or MRI were characterized in groups of true positives, false positives, false negatives and true negatives. Detailed information on all the patients considered in the study are included in Table II.

For predicting primary tumors, CT or MRI had a sensitivity of 93.3%, specificity of 75.0%, PPV of 82.4%, and NPV of 90.0% (Table III). Comparatively, PET/CT scans had a sensitivity of 95.2%, specificity of 100.0%, PPV of 100.0% and NPV of 95.7% (Table III). The diagnostic accuracy of PET/CT imaging for primary tumor detection was 97.7% (95% CI: 87.7-99.9%), whereas the diagnostic accuracy for CT or MRI for primary tumor detection was 85.2% (95% CI: 66.3-95.8%). Thus, the diagnostic accuracy of both PET/CT and CIS for primary tumor detection was comparable and the difference was not statistically significant.

True negatives included patients who were scanned for staging or restaging purposes. Those patients underwent a CT or MRI within a 6 month time window of their PET/CT imaging. There were 20 true positives for primary tumor detection for PET/CT, whereas CT or MRI scans accurately detected the tumors in 14/20 scans. PET/CT confirmed or identified primary neoplasms, which were all corroborated by subsequent histopathological analysis.

A total of 22 PET/CT images were true negatives. These cases were follow-ups for restaging. A total of 29 restaging scans were performed. Restaging was performed between 1 and 8 years after treatment by clinical examination and imaging.

There was one false negative case on PET/CT: A case of submandibular adenoid cystic carcinoma, in which the lesion did not display increased SUV; however, the lesion was obvious on CT scan with an additional mass effect (Fig. 1A).

The false positive CT scans for primary tumor included one case with reported enlargement of the lymph nodes, which may be suggestive of malignancy, but a full evaluation was difficult without contrast; and another case of chronic sialadenitis, which was detected on PET/CT but was suggestive of malignancy on CT scan. The false negative CT scan was unable to distinguish between inflammation and invasive submandibular duct carcinoma, whereas PET/CT was able to discriminate between the two.

For predicting cervical lymph node metastasis, CT or MRI exhibited a sensitivity of 54.5%, specificity of 75.0%, PPV of 60.0% and NPV of 70.6% (Table III), whereas PET/CT had a sensitivity of 88.9%, specificity of 97.1%, PPV of 88.9% and NPV of 97.1% (Table III). The diagnostic accuracy of PET/CT imaging for detection of lymph node metastasis was 95.4% (95% CI: 84.2-99.4%), whereas the diagnostic accuracy for CT or MRI was 66.7% (95% CI: 46.0-83.5%). There is a statistically significant difference between PET/CT and CT for the evaluation of cervical lymph node metastasis, as the 95% confidence intervals do not overlap. PET/CT was able to detect lymph node metastasis 2 years postoperatively, whereas CT was unable to detect metastasis during surveillance in a 54-year-old man with synovial cell sarcoma of the left submandibular gland (Fig. 2).

In 9 of the 18 (50%) patients, PET/CT revealed F-18 FDG-avid lymphadenopathy suspicious for metastases to the cervical lymph nodes. These findings would affect treatment options regarding surgical and/or radiation fields.

In 4 of the 18 patients of the cohort, the PET/CT findings included avid lymphadenopathy suspicious for metastasis to distant regions. There was 1 case of submandibular synovial cell carcinoma with metastasis to the lung, and 1 case of submandibular squamous cell carcinoma with multiple metastases to the lung and liver (Fig. 3). A Table for distant metastases comparing PET/CT vs. CT or MRI was not included, as metastasis was only identified on PET/CT in 4 patients in our sample. No further imaging was performed (in terms of CT or MRI) for distant metastasis to specific regions, such as the liver, lung, or heart in our patients.

Benign and malignant lesions in our study did not differ significantly in terms of SUVmax. The mean SUVmax for malignant lesions was 5.88, with a standard deviation of 4.25 (Table IV). This may be due to the fact that some benign tumors also have high glucose metabolism. Both false positive cases were pleomorphic adenomas. There were 2 cases of false positive F-18 FDG PET/CT in primary tumor detection. Abnormally increased F-18 FDG uptake was suspicious of a malignant tumor, but subsequent biopsy and/or surgical resection and pathological examination did not identify malignancy. Both lesions displayed high SUVmax and were biopsied. They were found to be benign, but displayed high uptake of F-18 FDG, and were finally diagnosed as pleomorphic adenoma and benign reactive lymphadenopathy. The scan of a 38-year-old woman with a 2.5x2.7-cm mass with a SUVmax of 6.9, which was later pathologically diagnosed as a pleomorphic adenoma after excision, is shown in Fig. 1B.