Abstract
Aim: To investigate whether 18F-fluorodeoxyglucose uptake is associated with overall survival in patients with major salivary gland cancer using univariate and multivariate analyses after adjusting for pathological stage (eighth edition of the International Union Against Cancer). Patients and Methods: A total of 32 patients with major salivary gland cancer treated with curative surgery were enrolled. Parameters for 18F-fluorodeoxyglucose uptake were assessed by positron-emission tomography combined with computed tomography. Results: Using univariate and multivariate analyses after adjusting for pathological stage, a maximum standardized uptake value ≥26, peak standardized uptake value ≥20.3, metabolic tumor volume ≥9.7, and total lesion glycolysis ≥263 were significantly correlated with shorter overall survival. Conclusion: Parameters of 18F-fluorodeoxyglucose uptake in major salivary gland cancer are predictive of overall survival after adjusting for the pathological stage.
Major salivary gland cancer (MSGC) is a rare malignant tumor, accounting for <6% of all cases of head and neck cancer (1). The Tumor Node Metastasis (TNM) staging system has been broadly accepted as a predictor of overall survival (OS) in many types of cancer, including MSGC (2, 3). After the eighth edition of the International Union Against Cancer (UICC) was published in 2017 (4), restaging the pathological TNM stage from pathological reports using this version was expected to be a useful predictor for various cancer types, including MSGC (5-7). However, it was found to be difficult to predict OS for the same TNM stage in several types of cancer (3, 7).
Parameters of 18F-fluorodeoxyglucose (18F-FDG) uptake on preoperative positron-emission tomography with computed tomography (PET/CT) in head and neck cancer, including MSGC, were associated with survival outcomes such as overall survival (OS) by univariate and multivariate analyses by adjusting for pathological stage (3, 6, 8-15). Recently, we reported the significant association between 18F-FDG-uptake parameters of primary tumor and OS in oral cancer by multivariate analysis adjusting for pathological stage by the eighth edition of the UICC staging manual (UICC8) (3). However, to our best knowledge, no studies have investigated the association between 18F-FDG-uptake parameters and survival outcome in MSGC by multivariate analysis adjusting for pathological stage by UICC8.
In the present study, we investigated the possible correlation between OS and 18F-FDG-uptake parameters in MSGC, and examined whether 18F-FDG-uptake parameters predict OS by multivariate analysis after adjusting for the pathological stage by UICC8.
Patients and Methods
Patients. From March 2008 to March 2015, 33 patients underwent preoperative 18F-FDG-PET/CT at the East Nagoya Positron-Emission Tomography Imaging Center as well as primary tumor resection with/without neck dissection with curative intent for primary MSGC at the Aichi Cancer Center Hospital. Among 33 patients, one patient was excluded from the study due to a high preoperative glucose level (≥200 mg/dl). Therefore, in total, 32 patients were studied. This retrospective study was approved by the Institutional Review Board (approval number of institution: 2016-1-241), and all patients involved provided their informed consent for examinations and treatments. The pathological classifications of these 32 patients were as follows: mucoepidermoid carcinoma in eight; adenocarcinoma not otherwise specified in eight; adenoid cystic carcinoma in five; carcinoma ex pleomorphic adenoma in four; salivary gland carcinoma in three; acnic cell carcinoma in three; epithelial-myoepithelial carcinoma in one.
Clinicopathological parameters. Procedures regarding clinical staging, pathological examination, postoperative treatment, and follow-up have been described previously (16). In brief, clinical and pathological staging was initially based on the seventh edition of the UICC TNM classification (UICC7) (17). Postoperative treatment was administered for patients with positive surgical margins, extranodal disease, multiple lymph node metastasis, and high histological grade, if possible. Subsequent follow-up to treatment identified patients with early locoregional recurrence, and these patients then underwent salvage surgery. Restaging from the pathological report was determined using the eighth edition of the cancer staging manual of the American Joint Committee on Cancer (AJCC) and by UICC8, as described previously (3, 7). The clinicopathological parameters are shown in Table I.
18F-FDG-uptake parameters. The method regarding 18F-FDG-PET/CT and 18F-FDG-uptake parameter evaluation were described previously (3, 8). Biograph True Point PET/CT/40 with True V (Siemens Health Medical Solutions Inc., Malvern, PA, USA) as well as low-dose CT images for the RANDO Phantom (Alderson Research Laboratories Inc., Long Island, NY, USA) were utilized for attenuation correction of PET results. The evaluation of FDG-uptake parameter. for semiquantitative assessment using Advantage Workstation 4.6 program PET VCAR (GE Healthcare, Chalfont, UK) was performed using three-dimensional (3D) images based on PET/CT findings in standardized uptake value mode. The maximum standardized uptake value (SUVmax) from the primary tumor was automatically calculated using a volumetric region of interest (VOI) on the 3D images. In accordance with our previous report (3, 8), the mean standardized uptake value (SUVmean) as well as the metabolic tumor volume (MTV) from the VOI, which included the primary tumor only, were calculated using a 45% threshold fraction of SUVmax. The total lesion glycolysis (TLG) was calculated using the following formula: TLG=MTV×SUVmean. The peak standardized uptake value (SUVpeak) was defined as the average SUV within a 1 cm3 spherical VOI which contained the maximum pixel. The mean±standard deviation (SD) blood sugar level at staging was 100±17.0 mg/dl, and the mean±SD duration from preoperative 18F-FDG-PET/CT to surgery was 21.2±10.3 days.
Statistical analysis. Statistical analyses were conducted using the JMP software package (version 9; SAS; Cary, NC, USA). Among the 30 patients with a primary tumor detected on 18F-FDG-PET/CT, the association between the FDG-uptake parameters (SUVmax, SUVpeak, MTV, and TLG) and clinicopathological parameters [age, gender, anatomical location, histological grade, positive surgical margin, UICC7 clinical T and N classification, UICC7 clinical stage, UICC7 and UICC8 pathological T and N classification, and UICC7 and UICC8 pathological stage) were compared using the Mann–Whitney U-test. In all cases, the OS time based on the Kaplan–Meier method was defined as the period from 18F-FDG-PET/CT to death or the last follow-up. In accordance with our previous method, we determined the cut-off values for various FDG-uptake parameters using univariate OS analysis with log-rank test (3, 8). In the univariate OS analysis, patients were assigned into groups based on the SUVmax (SUVmax ≥26 or <26), SUVpeak (SUVpeak ≥20.3 or <20.3), MTV (MTV ≥9.7 or <9.7), TLG (TLG ≥263 or <263). For each group pa:ring, we compared the univariate cause-specific survival (CSS), locoregional recurrence-free survival (LRRFS), and distant metastasis-free survival (DMFS), as measured by the Kaplan–Meier method. Target events were defined as death from MSGC for CSS, local or regional recurrence for LRRFS, and distant metastasis for DMFS. Multivariate analysis was performed for OS with adjustment for the pathological stage on UICC8 (IVB/I-IVA), including hazard ratio (HR) and 95% confidence interval (95% CI), employing a Cox proportional hazards model. Statistical significance was defined as a p-value of less than 0.05.
Results
Clinicopathological parameters and 18F-FDG-uptake parameters. The sensitivity of detecting the primary tumor site using 18F-FDG-PET/CT was 93.8% (30/32). There were two false-negative results, which were tumors undetected by 18F-FDG-PET/CT; one patient had a pathological T1 adenoid cystic carcinoma and another had pathological T2 acinic cell carcinoma. Among the 30 patients with a primary tumor detected by 18F-FDG-PET/CT, SUVmax, SUVpeak, MTV, and TLG values (mean±SD) of the primary tumor were 23.9±14.6 g/ml, 16.2±10.4 g/ml, 10.6±10.3 cm3, 180±252 g, respectively. The association between 18F-FDG-uptake parameters and clinicopathological parameters are shown in Table II. Higher SUVmax was significantly associated with UICC7 clinical N1-N2b (p=0.01), as well as UICC7 clinical stage IV on (p=0.04). Higher MTV was significantly associated with UICC7 clinical T4 classification (p<0.01), clinical N1-N2b (p=0.01), and clinical stage IV (p<0.01); and with UICC8 pathological T4 classification (p=0.01). Higher TLG was significantly associated with UICC7 clinical T4 classification (p<0.01), clinical N1-N2b classification (p<0.01), and clinical stage IV (p<0.01); UICC8 pathological T4 classification (p=0.04); and the present of positive surgical margins (p<0.05). There were no significant association between SUVpeak and clinicopathological parameters.
Survival outcomes. The follow-up duration (mean±SD) after 18F-FDG-PET/CT was 1487±781 days. The mean follow-up duration of 22 patients who were still alive (68.8%), 10 patients who died (31.3%), and nine who died due to MSGC (28.1%) was 1,844±702, 702±467, and 776±428 days, respectively. Three patients (9.38%) developed local recurrence, one (0.03%) developed regional recurrence, four (12.5%) developed locoregional recurrence, and 12 (37.5%) developed distant metastasis. The duration (mean±SD) from 18F-FDG-PET/CT to local recurrence, regional recurrence, locoregional recurrence, or distant metastasis was 432±383, 430±0, 432±313, and 521±359 days, respectively.
Univariate survival analysis. The cut-off values for different 18F-FDG-uptake parameters were tested using log-rank tests in the univariate OS analysis. The cut-off values with the lowest p-values were SUVmax=26 (p<0.01), SUVpeak=20.3 (p=0.01), MTV=9.7 (p<0.01) and TLG=263 (p<0.01), as shown in Figure 1. The Kaplan–Meier curves from the univariate OS analysis are shown in Figure 2. Univariate survival analysis is shown in Table III. Patients with SUVmax ≥26 had significantly lower (p<0.01) OS, CSS, LRRFS, and DMFS than those with SUVmax <26. SUVpeak ≥20.3 was significantly associated (p≤0.02) with poorer OS, CSS, LRRFS, and DMFS compared to those with SUVpeak <20.3. Patients with MTV ≥9.7 had significantly lower (p<0.01) OS, CSS and LRRFS than those with MTV <9.7. Patients with TLG ≥263 had significantly lower (p<0.01) OS, CSS, LRRFS and DMFS than those with TLG <263.
Multivariate survival analysis. Results of the multivariate analysis of OS adjusted for pathological stage by UICC8 are shown in Table IV. SUVmax ≥26, SUVpeak ≥20.3, MTV ≥9.7, and TLG ≥263 were significantly associated (p<0.01) with shorter OS.
Discussion
This study demonstrated that higher SUVmax, SUVpeak, MTV, and TLG were significantly associated with shorter OS in patients with MSGC by univariate and multivariate analyses adjusting for UICC8 pathological stage.
In many studies for some types of cancer, including several meta-analyses and reviews, significant association has been reported between OS and 18F-FDG-uptake parameters such as SUVmax, MTV and TLG (3, 8-15). For example, 18F-FDG-uptake parameters were significant predictors of survival outcomes in a review of 941 patients with nasopharyngeal cancer (10). Moreover, we also reported significant association between pretreatment 18F-FDG-uptake parameters and survival outcomes in oral and hypopharygeal cancer (3, 8). Our present results showed significant correlation between high 18F-FDG-uptake parameters and shorter OS, and these data are in agreement with previous data (3, 8-15).
In MSGC, some authors reported the significant association between survival outcomes and 18F-FDG-uptake parameters (9, 13-15). For example, SUVmax ≥7.4 led to significantly lower 5-year LRRFS, DMFS, and OS in 46 patients treated by radiation therapy (9), and both MTV and TLG were significantly associated with OS and progression-free survival by univariate and multivariate analyses adjusting for UICC7 pathological stage in 49 patients with salivary gland cancer, including minor salivary gland cancer as well as that treated by surgery (13). Almuhaimied et al. investigated 75 patients with high-grade salivary gland cancer and found that SUVmax, SUVmean, SUVpeak, MTV and TLG were significantly associated with survival outcomes (14). Our present findings, which showed a significant association between high FDG-uptake parameters and shorter OS, are consistent with those in previous studies (8, 13-15).
Pathological stage by restaging based on using the eighth edition of the staging manual AJCC and UICC8 has been showed to be a useful predictor in several types of cancer (3, 6, 7). In 4,520 patients with salivary gland cancer from the National Cancer Database in United States of America, restaging stage such as pN3b was significantly associated with survival outcomes (6). We also showed that the pathological stage by restaging was significantly associated with OS in 543 patients with papillary thyroid cancer (7). Recently, we showed the significant association between 18F-FDG-uptake parameters of primary tumor and OS in 28 patients with oral cancer by multivariate analysis adjusting for UICC8 pathological stage (3). To our knowledge, there have been no studies for MSGC with multivariate analysis adjusting for 18F-FDG-uptake parameters and UICC8 pathological stage. The present study showed, for the first time, that higher 18F-FDG-uptake parameters in MSGC were significantly associated with shorter OS by multivariate analysis adjusting for UICC8 pathological stage.
The current study had certain limitations. This study was a retrospective analysis of a small sample due to the rarity of the cancer. Therefore, a prospective survey with a larger cohort could confirm our findings and lead to a more useful and precise result.
Conclusion
The present study demonstrated that higher 18F-FDG-uptake parameters (SUVmax, SUVpeak, MTV, TLG) are significantly associated with shorter OS in patients with MSGC treated by surgery by univariate and multivariate analyses adjusting for UICC8 pathological stage. These 18F-FDG-uptake parameters can be considered predictors for MSGC.
Acknowledgements
This study was supported by JSPS KAKENHI Grant Number 16K11253.
Footnotes
Conflicts of Interest
All Authors declared that they have no conflicts of interest in regard to this study.
- Received December 5, 2018.
- Revision received December 16, 2018.
- Accepted December 18, 2018.
- Copyright© 2019, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved