Conventional to Early Phase Standardized Uptake Value Ratio on 18F FDG PET/CT Could Reflect Prognosis in Patients with Non-small Cell Lung Cancer

Objective: A high metabolism/perfusion rate in the tumor is known to be factor indicating poor prognosis. The goal of this study was to investigate the prognostic value of the ratio of the conventional standard uptake value to early-phase imaging standard uptake value in patients with newly diagnosed non-small cell lung cancer. Materials and Methods: Early-phase imaging was obtained in the first 120 seconds and conventional imaging was taken after median 66 minutes. The ratio of the conventional standard uptake value to earlyphase imaging standard uptake value was calculated. Univariate and multivariate analyses were performed using the Cox proportional hazards regression model to assess the relationship between progression-free survival and ratio of the conventional standard uptake value to early-phase imaging standard uptake. Results: A total of 77 patients with non-small cell lung cancer were recruited. Progression-free survival analysis was performed in 52 inoperable patients. Progression-free survival was found to be related to conventional standard uptake value (p<0.001), tumor size (p<0.001) and ratio of the conventional standard uptake value to early-phase imaging standard uptake in univariate analysis (p=0.047). In the multivariate analysis, ratio of the conventional standard uptake value to early-phase imaging standard uptake (p=0.001, HR=1.145, 95CI%:1.145-1.719) and tumor size (p<0.001, HR=1.026, 95CI%:1.007-1.045) were found to be independent poor prognostic factors. Conclusion: The ratio of the conventional standard uptake value to earlyphase imaging standard uptake was found an indicator for prognosis. Evaluating the early phase imaging, without a need for kinetic program or software, without disrupting the routine functioning of the clinic, with conventional imaging may contribute to cancer patient management. There is a need for studies with larger groups of patients.


INTRODUCTION
F-fluorodeoxyglucose ( 18 F FDG) shows glucose metabolism in the tissue. Tumor glucose metabolism in non-small cell lung cancer (NSCLC) has been shown to significantly affect the biologic behavior of the tumor and a correlation between the maximum standardized uptake value (SUV max ) and tumor aggressiveness [1][2][3]. Decreased blood flow in the tumor leads to hypoxia [4,5]. Cancer cells develop an adaptation to hypoxia [6]. It is known that low oxygen levels reduce the distribution of chemotherapeutics [7] and that hypoxic tissues are more resistant to radiotherapy (RT) [8]. Some authors recommend to identify hypoxic regions within the tumor and escalate RT doses for these regions [9]. 18 F-FDG positron emission tomography/computed tomography ( 18 F-FDG PET/CT) early-phase (0-2min) images have been reported to reflect perfusion [10,11]. The mismatch between early-phase images and conventional images reflects hypoxia [12,13]. Combined perfusion and metabolism imaging is recommended in clinical studies evaluating the effectiveness of anti-angiogenic therapy [14]. If tumor blood flow could be monitored noninvasively, it might be possible to obtain more information about the effectiveness of the treatment [15].
The purpose of our study was to investigate the prognostic value of the ratio of early phase imaging SUV (SUV E ) values, which can be obtained without need for an expensive kinetic PET program and computer software, to conventional SUV values (SUV L ), for progression free survival (PFS) in patients with newly diagnosed NSCLC.

Ethical approval
The Clinical Research Ethics Committee of our institute reviewed and approved this study (no: 2019/50). Informed consent was obtained from all individual participants included in the study. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the Helsinki declaration and its later amendments or comparable ethical standards

Patients and Study Design
Patients in whom 18 F FDG PET/CT imaging was performed with indications of the characterization of lung mass/nodule and for lung cancer staging between December 2015 and April 2017 were included in the study. Lung cancer diagnosis was histopathologically confirmed in all patients. We evaluated early and conventional 18 F FDG PET/ CT images of 77 patients with NSCLC. We did not include 12 patients; who had brain metastasis at the time of diagnosis (n: 7), died due to non-disease related reasons (n= 2), and were lost to follow-up (n= 3). We excluded 13 operable patients who had surgical treatment. We included stage II patients who were not candidate for operation due to medical comorbidities and inoperable stage III-IV patients in PFS analysis (n=52).

Patient Follow-up
PFS was calculated from the date of diagnosis to the date of progression of the disease (new lesion or expansion of the previous existing lesion) or to the date of recurrence, or to the date of death associated with the disease. The period from the date of diagnosis to the last follow-up was calculated in patients who survived or in whom progression was not observed. Follow-up of the disease was conducted using conventional CT and/ or 18 F-FDG PET/CT.

F-FDG PET/CT Procedures
A PET/CT scanner Biograph mCT (Siemens Healthcare, Erlangen, Germany) was used. After at least 6 hours of fasting, patients with a blood glucose level of <200 mg/dl were administered a FDG injection at an approximate dose of 3.7 MBq/kg. After median 63 minutes [min-max 54-79 minutes] imaging was performed in the supine position with arms up. PET imaging was adjusted to 2 minutes per bed position. Low dose CT parameters: voltage, 120 kV; CARE Dose 4D mA tube current; and slice thickness, 5.00 mm.

Image Analysis
Two nuclear medicine specialists with 9 and 10 years' PET/CT experience (O.K, S.A) and a radiation oncologist with 10 years' experience (G.E) reviewed all 18 F-FDG PET/CT images on a dedicated Workstation (Syngo-via). Early-phase and conventional images were simultaneously evaluated and volumetric areas of the tumor were delineated. Early-phase imaging SUV (2 min. postinjection; SUV E ) and conventional imaging SUV (median 66 min. postinjection; SUV L ) values were calculated. Ratio values (SUV R ) were obtained by dividing conventional imaging SUV values by the early-phase imaging SUV (SUV L /SUV E ) ( Figure 1).

Statistical Analysis
Variables were compared using either Student's t-test or nonparametric tests, depending on the normality of the distribution. Univariate and multivariate analyses were performed using the Cox model to assess the relationship between survival and 18 F FDG PET/CT parameters. Any variable with p<0.2 in the univariate model was included in the multivariate Cox proportional hazards regression model. For multivariate analyses, the full model included all the variables that we studied, and the final model was constructed using the backward stepwise procedure. Differences between groups were investigated using the log-rank test. Receiver operating characteristic (ROC) curve analysis was used to determine the optimal cutoff value of the SUV R for PFS. Kaplan-Meier curves were drawn using defined cutoff values and PFS was interpreted using log-rank analysis. Correlation between groups was evaluated using Spearman correlation for non-parametric variables and the Pearson correlation coefficient for parametric variables. All analyses were performed using SPSS v.22 (SPSS, Inc., Chicago, IL), and two tailed p<0.05 was considered significant.

RESULTS
Early phase and conventional 18 F FDG PET/CT Images of 77 patients with newly diagnosed NSCLC were evaluated [7 women, 70 men, 47 patients (61%) had adenocarcinoma, 30 patients (39%) had SCC, mean age was 65.9±10.23]. The values of the SUV parameters and tumor size are given in Table  1. There was no statistically significant difference in SUV values between groups of patients with adenocarcinoma and SCC (p>0.05). In patients with SCC, the tumor size was statistically significantly larger (p=0.027).

Relationship between Early-phase Imaging and Conventional Imaging FDG uptake
Among the 77 patients, we found no statistically significant relationship between early-phase imaging and conventional imaging SUV values (r=0.177, p=0.125). Early phase and conventional SUV values of 77 patients are given in figure 2. When the patients were divided into two groups according to the primary tumor size, the group with a tumor size ≤3 cm comprised 21 patients (27.3%), and the group with a tumor size >3cm constituted 56 patients (72.7%). We found a statistically significant moderate correlation (r=0.443, p=0.044) between SUV E and SUV L values in the group with a tumor size ≤3 cm, whereas there was no correlation in the group with a tumor size >3 cm.

Relationship between Tumor Size and Conventional Imaging/Early-phase Imaging FDG Uptake Ratio
The images of 77 patients were evaluated. Conventional/early-phase imaging FDG uptake ratio increased as tumor size increased. The conventional imaging/early-phase imaging FDG uptake ratio was higher in the group with tumor size >3 cm (p=0.047). Then we divided the patients into 2 groups with tumor size <5 cm and ≥5 cm and found that the conventional imaging phase/ early-phase imaging FDG uptake ratio was higher in the group with tumor size ≥5 cm (p=0.003). Thus, as tumor size increased, the conventional imaging/ early-phase imaging FDG uptake ratio increased ( Table 2).

Progression-free Survival Analysis
Fifty-two NSCLC patients who were treated with chemotherapy and /or radiotherapy were included in the study [ Figure 3.

DISCUSSION
In this study, we investigated the prognostic value of conventional imaging/early phase imaging FDG uptake ratio and the relationship between early phase and conventional imaging SUV values, according to the tumor size, in patients with newly diagnosed NSCLC. This method has never been investigated previously in this group of patients. 18 F-FDG PET/CT early-phase (0-2minutes) images show perfusion, whereas conventional images show metabolism [10,11]. The mismatch between early-phase imaging and conventional imaging reflects hypoxia [12,16], which is a predictor for response to radiotherapy and chemotherapy. Low oxygen levels have been shown to decrease the effectiveness of radiotherapy and chemotherapy [7,17]. A high metabolism-perfusion ratio in head and neck tumors has been shown to be an independent poor prognostic factor for local control [18]. Komar et al. stated that the lack of tumor perfusion in pancreatic cancer caused resistance to chemotherapy, and that this led to a decrease in treatment effectiveness and a poor prognosis [14]. Park et al. showed that high perfusion in pancreatic cancer was associated with an increased CRT response [19]. Aoki et al. investigated the relationship between the average iodine density (AID) obtained from dual energy computerized tomography (DECT), SUV max value, and local control in 74 patients with NSCLC undergoing stereotactic body radiotherapy (SBRT). They found that local control of patients with high SUV and low AID values was worse, and proposed different fractionation schemes for this subgroup [20]. Mankoff et al. showed that patients with breast cancer with high glucose metabolism/perfusion ratio responded worse to treatment [21].
In this study, we obtained ratio values by dividing the conventional imaging FDG uptake value by the early-phase imaging (0-2 minutes) FDG uptake value. In univariate Cox proportional hazard regression analysis, SUV R , SUV L and Tsize were associated with PFS, meaning that the prognosis worsened as the conventional imaging/earlyphase imaging FDG uptake ratio increased. In this aspect, our result was consistent with studies   stating that the metabolism-perfusion mismatch was associated with poor prognosis.
Tissue perfusion and vascular permeability can be measured using dynamic contrast-enhanced MRI, 15 O-water PET, perfusion CT, and dynamic 18 F-FDG PET. It is assumed that 18 F FDG is transported from the blood into the tissue at a linear transfer rate K1 relative to the blood flow. K1 which is a measure of capillary permeability and perfusion has been shown to have prognostic value in the treatment of cancer [22]. In a study with patients with soft tissue sarcoma, Okazumi et al. suggested that SUV and K1 should be evaluated together for more accurate results [23]. Strauss et al. showed that there was a correlation between FDG K1, vascular endothelial growth factor (VEGF), and angiopoietin in patients with colorectal cancer, concluding that angiogenesis was a determining factor for FDG PET kinetics [24]. There is a strong correlation between the K1 value obtained from dynamic FDG PET earlyphase imaging (0-2 minutes) and the flow measured with 15 O water PET. Thus, tumoral perfusion would be evaluated with the FDG PET obtained in the firstpass FDG uptake (K1). A single 18 F-FDG injection, a method for measuring blood flow and metabolism in tumors could provide an important addition to the functional imaging of tumors with 18 F-FDG PET. A full, 1-h, dynamic 18 F-FDG PET acquisition is expensive making it impractical for routine practice. Accordingly, Mullani et al. developed a 2-min dynamic first-pass PET acquisition, based on a 1-compartment flow model to assess tumor blood flow BF [11].
In order to achieve K values, a dynamic PET application and computer software program are needed. However, in a study on lung cancer, it was shown that there was a strong association (r=0.83, p=0001) between K1 values obtained from dynamic FDG PET and early-phase imaging (0-2 minutes) SUV without the need for software [25]. A dynamic PET study in patients with soft tissue sarcoma also found a strong relationship between SUV obtained between 1.5-2.5 minutes and K1 values (r=0.79, p<0.05) [22]. In a study performed in patients with NSCLC, there was a strong correlation between 15 O water PET BF and FDG PET early-phase imaging (0-5 min) standardized uptake normalized to body surface area (SUVbsa), yet no correlation was found in 4-5-min SUV [15]. We obtained SUV E values that reflected perfusion and were highly correlated with K1, without the use of a kinetic PET program and without the need for software, and we investigated its prognostic value by calculating its proportion with conventional imaging SUV.
We also think that the combined visual evaluation of early-phase and conventional images may contribute to the evaluation of treatment response. Perfusion-metabolism imaging allows us to see the mismatched/hypoxic areas in the tumor. In some studies, perfusion and metabolism imaging was performed before and after treatment, and it was indicated that metabolism and perfusion did not show a balanced decline due to treatment. Miles et al. followed up the chemotherapy response with conventional CT, perfusion CT, and 18 F FDG PET in a patient with colon cancer and liver metastasis. After the treatment, the least change was observed in metabolism (metabolism 11%, perfusion 47%, area 43%), and they suggested that the tumor might have developed an adaptation to hypoxia during the treatment process. They stated that metabolism-perfusion combined imaging would contribute to determining adaptive treatment options [26]. Again, Miles et al. evaluated the tumor response with metabolic and perfusion imaging, and concluded that the best outcome was in the group that had the balanced response in perfusion and metabolism. They proposed adaptive treatment against neovascularization in the group with reduced metabolism and unchanged or increased perfusion, and adaptive treatment against hypoxia in the group with unchanged or increased metabolism and reduced perfusion. They emphasized that the ultimate goal in the management of patients with cancer was to adapt individual patient treatment to the vascular and metabolic response exhibited by their tumors [27]. A study of rectal cancer by Willett et al. using perfusion CT and FDG-PET reported significant falls in perfusion, but no change in glucose metabolism when the VEGF antibody, bevacizumab, was given alone. A reduction in glucose metabolism was only seen when bevacizumab was given in combination with radiotherapy [28]. A study that was conducted in patients with breast cancer revealed that although chemotherapy-related metabolism decreased, perfusion might not decrease and might even increase [29]. Herbst 18 F FDG PET increased [30].
Although not investigated in this study, we suggest that combined early-phase and conventional imaging as a developable idea can be used to evaluate metabolism and perfusion response in the follow-up of cancer treatment, similar to the abovementioned studies. Thus, the change in early phase imaging SUV and conventional imaging SUV values calculated in pre-treatment PET can be measured after treatment (%change) and may contribute to adaptive treatment planning according to metabolism predominant, perfusion predominant or balanced responses, as described in some of the above examples (Figure 4).
We found no statistically significant correlation between early-phase imaging and conventional imaging SUV values. Tateishi et al. studied the relationship between peak attenuation (Apa) and relative flow (RF) with SUV in patients with NSCLC (tumor size: 2.6±0.2; range, 0.9-3.4 cm) and showed that both perfusion parameters of CT had a correlation with SUV [31] . Miles et al. compared standardized perfusion value (SPV) with SUV in 18 patients with NSCLC and reported that no correlation existed in the entire group of patients (n=18), albeit there was a statistically significant relation only in the group of patients whose tumor size was less than 4.5 cm 2 (n=6, mean size 2.4 cm) [32]. The authors discussed that there was a relationship between tumor perfusion and tumor metabolism and tumor size; the biologic aspect of tumor differentiated as the tumor grew, and a metabolic-perfusion difference was more prominent in large and high-grade tumors. Likewise, we divided the patients into two groups according to tumor size, and we accepted 3 cm [≤ 3 cm (n=21), and >3 cm (n=56)] as a threshold. Based on the knowledge that early-phase images reflect perfusion, our result is consistent with the abovementioned two studies.
We found that the conventional imaging/earlyphase imaging FDG uptake ratio (SUV R ) increased as tumor size increased. Our findings are consistent with the literature showing that the metabolismperfusion difference increases as the tumor size increases [33][34][35].
Our study has limitations; first of all our study is a single-center study. The reason for male dominance in the patient group can be explained by the fact that the female population smoked less in this region, and that they had a relatively low incidence of lung cancer. The main limitation of the study was the lack of data about the treatment details which could be among the reasons for the results. In addition, a lack of comparison between the baseline characteristics of the patient groups with different SUV values was absent other than tumor size.
In conclusion, the ratio of conventional imaging SUV to early phase imaging SUV was found an indicator for prognosis. Evaluating the early phase imaging, without a need for kinetic program or software, without disrupting the routine functioning of the clinic, with conventional imaging may contribute to cancer patient management. There is a need for studies with larger groups of patients.