Prognostic Value of Quantitative 18F-FDG PET/CT Parameters Measured in Primary Tumors and Suspicious Lymph Nodes in Patients With Esophageal Carcinoma

Quantitative 18 F-FDG PET/CT parameters have been described as prognostic indicators in esophageal cancer. The objective of this study isto evaluate the prognostic value of the maximum standardized uptake value (SUVmax), metabolic tumor value (MTV) and total lesion glycolysis (TLG) measured in the primary tumor and suspicious lymph nodes. Methods: A cohort study was performed to assess the association of SUVmax, MTV and TLG measured prior to and post neoadjuvant therapywithoverall survival (OS) of patients with esophageal cancer who received trimodal therapy. The quantitative techniques were applied in the primary tumor and suspicious lymph nodes. The OS rates were analyzed. Results: 39 post-neoadjuvant PET/CT exams. Before neoadjuvanttherapy, PET/CT showed that all the variables of the evaluated lymph nodes were statistically signicant in predicting OS. Postneoadjuvanttherapy, none of the PET/CT variables of lymph nodes were related to prognosis. On the other hand, all primary tumor volumetric variables were related to overall survival. The MTV (HR: 4.66; 95% CI: 1.54-14.08) and TLG (HR: 4.86; 95% CI: 1.66-14.26) of the primary tumor post neoadjuvanttherapy and the variations in MTV (HR: 2.95; 95% CI: 1.01-3.52) and TLG (HR: 3.49; 95% CI: 1.01-3.52) of the primary tumor pre-to-post-neoadjuvanttherapy were prognostic variables. Conclusion:


Introduction
Since the CROSS group 1 reported good results after neoadjuvant therapy based on carboplatin and paclitaxel, preoperative chemoradiotherapy has become the mainstay treatment among most patients with potentially curable esophageal cancer. 1, 2 On the other hand, patients with poor response to neoadjuvant therapy present poor long-term survival rates and might not be the best candidates for surgical resection. 3 To reach the correct treatment decisions and to reduce therapeutic toxicity, the selection of patients who have favorable prognoses plays an important role.
provides the ability to functionally evaluate metabolic activity 4,5 , thus improving patient selection for surgical treatment. 6 The pretreatment standardized uptake value (SUV) is the most widely used parameter in 18 F-FDG PET/CT and is considered a prognostic factor for risk strati cation; however, this parameter does not re ect the heterogeneity of the primary tumor or lymph nodes. 7 Aggressive tumors grow rapidly and usually present intratumoral hypoxia 7 and necrosis, which leads to heterogeneous FDG accumulation in the tumor. 8,9 In this context, volume-based parameters such as metabolic tumor volume (MTV) and total lesion glycolysis (TLG) that re ect metabolic volume and activity in an entire mass, respectively, have been proposed as quantitative indexes of tumor metabolism. MTV is a volumetric measure of the tumor cells with high glycolytic activity, while TLG is de ned as the product of the tumor volume by the average SUV inside this volume. 10 These indexes are prognostic indicators for survival in several neoplasms, such as lung cancer, pleural mesothelioma, ovarian cancer, and head and neck cancer. [11][12][13][14][15][16][17][18] There have been a few studies on the prognostic value of MTV and TLG for esophageal cancer treated with trimodal therapy. [19][20][21][22][23][24][25][26] Nonetheless, these studies included heterogeneous neoadjuvant therapy regimens and did not present separate analyses of primary tumors and lymph nodes.
Therefore, the aim of this study is to assess the value of the maximum SUV (SUVmax), MTV and TLG of the primary tumor and suspicious lymph nodes, as measured with 18 F-FDG PET/CT prior to and post therapy, in predicting overall survival in a group of patients with esophageal cancer treated with neoadjuvant chemoradiotherapy using a platinum-and taxane-based regimen followed by curative intent esophagectomy.

Methods
A retrospective cohort study was performed in a group of patients with esophageal cancer, assessing the association of SUVmax and volumetric parameters (MTV and TLG) measured on 18

Patients
The patients originated from a single institute and completed neoadjuvant chemoradiotherapy using platinum-and taxane-based regimens, followed by curative intent esophagectomy. A transthoracic approach with two-eld lymph node dissection was performed for tumors extending proximally to the tracheal bifurcation. For tumors involving the esophagogastric junction, a transhiatal resection was preferred. Gastric tube reconstruction with cervical anastomosis was the preferred technique.
The recruitment period ranged from 2009 to 2019. Patients were staged with endoscopy, CT, and PET and classi ed according to the 8th edition of the UICC staging system 27 . The local ethics committee approved the study. 18 F-FDG PET/CT scans were acquired on a Discovery 690 with time-of-ight (General Electric, Waukesha, Wisconsin, USA). The patients were injected with ∼ 3.7 MBq 18 F-FDG per kg body weight. The patients were instructed to fast for at least 6 h and had a blood glucose level below 180 mg/dl before the radiopharmaceutical injection. Imaging was initiated 60 min after the injection, and the scans were acquired from the mid-skull to mid-thigh. The images were reconstructed with a standard iterative algorithm, and the burden of disease in the primary tumor and metastatic lymph nodes were evaluated using SUVmax, TLG and MTV. These variables were calculated by a nuclear medicine physician using AW VolumeShare 5® (General Electric, Waukesha, Wisconsin, USA). The SUV thresholds used to de ne the boundaries of the lesions were established by visual analysis, and the total volumes of interest that circumscribed the primary tumor and metastases were calculated automatically by the software.

18 F-FDG PET/CT acquisition and imaging analyses
The burden of the disease was measured on 18 F-FDG PET/CT scans acquired prior to and post neoadjuvant therapy.

Statistical analyses
Survival analysis was performed with Kaplan-Meier and log-rank tests. For long-term survival analysis, all patients who died within 30 days postoperatively were excluded from the analysis, and the cutoff was de ned as the threshold value of the continuous covariate determined by Lausen. 28 A signi cance level of 0.05 was adopted. Cox proportional hazard analysis was performed. In the multivariate analysis, the 18 F-FDG PET/CT parameters, age, sex, clinical stage, and grade of cellular differentiation were assessed.

Patients' baseline characteristics
One hundred seventeen patients underwent neoadjuvant chemoradiotherapy using platinum-and taxanebased regimens followed by curative intent esophagectomy and were included. Of these patients, 8 patients who died within 30 days postoperatively were excluded. One hundred and six patients underwent PET/CT before neoadjuvant therapy, and 39 of these patients also had post-neoadjuvant therapy PET/CT exams. Squamous cell carcinoma (SCC) comprised 70% of the cases, and adenocarcinoma comprised the remaining cases. The mean follow-up was 36.8 months (SD ± 24.9), with a mean age of 60.9 years (SD ± 8.2), and there was a male predominance (78.9%). There were 87 transthoracic (video-assisted thoracoscopic) procedures and 22 transhiatal procedures, all of which involved cervical anastomosis.

Overall survival
Prior to neoadjuvant therapy, 18 F-FDG PET/CT showed that all the variables of the evaluated lymph nodes were statistically signi cant in predicting long-term survival (See Fig. 2). The MTV of the primary tumor (HR: 1.89; 95% CI: 1.01-3.52; p = 0.04) was also able to predict overall survival. See Table 2. Table 2 The relationship between the 18 F-FDG PET/CT variables and overall survival. For lymph nodes, none of the post neoadjuvant therapy 18 F-FDG PET/CT variables or the variations preto-post neoadjuvant therapy were related to prognosis. On the other hand, all volumetric variables of the primary tumor were related to overall survival (See Fig. 3 In the subgroup analysis, none of the variables related to adenocarcinoma were statistically signi cant. On the other hand, for squamous cell carcinoma (SCC), the SUVmax and TLG of lymph nodes prior to neoadjuvant therapy; MTV of lymph nodes after neoadjuvant therapy; and the variations in the SUVmax of the primary tumor and TLG of lymph nodes were related to survival. See Table 2.
In the logistic regression, only age and the change in MTV value after neoadjuvant therapy were considered independent variables related to OS. See Table 3.

Discussion
The results of this cohort of esophageal cancer patients who underwent neoadjuvant chemoradiotherapy using a platinum-and taxane-based regimen followed by curative intent esophagectomy suggest that the absolute values of and changes in some of the quantitative 18 F-FDG PET/CT variables predict survival. The lymph node parameters pre neoadjuvant therapy and the primary tumor parameters post neoadjuvant therapy are prognostic factors, mainly for SCC.
MTV and TLG are parameters that re ect both tumor volume and metabolic activity. These parameters provide complementary information about disease burden. In this study, the change in metabolic volume after neoadjuvant therapy and age were unique independent variables related to survival. Although the SUV parameters were related to OS in the univariate analysis, they were not considered independent variables in the nal model. SUVmax and SUVmean are semiquantitative indexes that might fail to re ect the spatial distribution of lesions and could vary with different PET scanners, fasting duration, level of plasma glucose and region of interest (ROI). TLG, which is dependent on SUVmean, may also be in uenced.
A few studies have assessed SUVmax as a prognostic factor for survival. 21,29 In addition, several recent studies revealed that the pretreatment MTV and TLG have prognostic value in the prediction of survival in surgically resected or inoperable esophageal cancer. [19][20][21][22][23][24][25][26] However, these studies used heterogeneous neoadjuvant regimens or surgical approaches, and none of them evaluated metabolic volumetric changes in the lymph nodes. Additionally, these studies used different cut-off points for the logistic regression for survival analysis. The present study used the Lausen cut-off point, which is more suitable for survival analysis when patients are followed for different intervals. 28 The results of this study should be interpreted in the context of certain inherent limitations. This is a single-center retrospective study with a relatively small sample size and is mainly focused the period after neoadjuvant therapy. Larger and controlled prospective studies are warranted to clarify the predictive value of these 18 F-FDG PET/CT variables on prognosis.
Despite these limitations, this is the rst study to assess the relationship between changes in 18 F-FDG PET/CT variables and survival during neoadjuvant chemoradiotherapy with a platinum-and taxanebased regimen followed by curative intent esophagectomy to treat cancer.

Conclusion
PET/CT is a noninvasive imaging method that functionally evaluates metabolic activity, and the absolute values of and changes in SUVmax and volumetric variables provide important information on patient prognosis and may improve patient selection for surgical treatment. Measuring metabolic parameters offers an easy approach towards determining patient prognosis, as the majority of patients receive PET/CT during staging. Clinicians can predict which patients will respond favorably to neoadjuvant therapy and esophagectomy and customize the follow-up of each patient. Personalized medicine is a goal of modern cancer therapy and aims for individually optimized treatments that are dependent on the tumor characteristics of each individual patient.

Declarations
Ethics approval and consent to participate Local Ethics committee (CCEP) approved this study.

Consent for publication
Not applicable  Overall survival according to the 18F-FDG PET/CT parameters of lymph nodes prior to neoadjuvant therapy.Cut-off points were determined by Lausen.