Correlation between hematological parameters and PET/CT metabolic parameters in patients with head and neck cancer

Systemic inflammation is predictive of the overall survival in cancer patients and is related to the density of immune cells in the tumor microenvironment of cancer, which in turn correlates with 18F -fluorodeoxyglucose (FDG)-positron emission tomography/computed tomography (PET/CT) metabolic parameters (MPs). The density of tumor-infiltrating lymphocytes (TILs) in the microenvironment has the potential to be a biomarker that can be used clinically to optimize patient selection in oropharyngeal head and neck squamous cell carcinoma (HNSCC). There is little to no data regarding the association of systemic inflammation with PET/CT-MPs, especially in HNSCC. This study aimed to evaluate the correlation between markers of host inflammation, namely blood neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR), with the PET/CT-MPs standardized uptake value (SUV), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) of the primary tumor, derived from FDG-PET/CT in patients with nonmetastatic (cM0) HNSCC before treatment. We hypothesized that NLR and PLR at baseline are positively correlated with PET/CT-MPs. A retrospective review of consecutive patients with HNSCC with a pretreatment PET/CT was performed. NLR and PLR were computed using complete blood counts measured within 10 days before the start of any treatment. The correlation between NLR and PLR with PET/CT-MPs was evaluated with Spearman's rho test. Seventy-one patients were analyzed. Overall survival (OS) at 1, 2, and 3 years was 86%, 76%, and 68%. PLR was found to be correlated with MTV (rho = 0.26, P = .03) and TLG (rho = 0.28, P = .02) but not with maximum SUV or mean SUV. There was no correlation between NLR and the analyzed PET/CT-MPs. TLG was associated with worse survival in uni- and multivariable analysis, but no other PET/CT-MPs were associated with either OS or disease-specific survival (DSS). NLR and PLR were associated with OS and DSS on uni- and multivariable analysis. In patients with HNSCC before any treatment such as definitive radio (chemo)therapy or oncologic surgery followed by adjuvant RT, baseline PLR correlated with MTV and TLG but not with SUV. NLR was not correlated with any PET/CT-MPs analyzed in our study. Confirmatory studies are needed, and a potential interaction between tumor microenvironment, host inflammation, and FDG-PET/CT measures warrants further investigation.


Introduction
Radiotherapy (RT) is the standard of care for most patients with head and neck squamous cell carcinoma (HNSCC) treated definitively or adjuvantly. Besides the effects of RT through direct or indirect deoxyribonucleic acid damage, RT can also induce antitumor immune responses that contribute to indirect tumor cell killing [1,2]. Several studies have demonstrated that tumor-infiltrating lymphocytes (TILs) are present in the tumor microenvironment in various malignant tumors [3][4][5]. In oropharyngeal carcinoma, TILs can identify stage I human papillomavirus (HPV)-associated patients likely to be poor candidates for treatment de-escalation [6]. In non-small-cell lung carcinoma (NSCLC), studies have shown that the tumor microenvironment correlates with 18 F-fluorodeoxyglucose (FDG) positron-emission tomography/computed tomography (PET/CT) metabolic parameters (MPs) [7]. The therapeutic potential of radioimmunotherapy in HNSCCs is as vast as the tumor microenvironment of HNSCCs is complex [8]. This field awaits results from ongoing clinical trials providing immunological and genomic analyses from radioimmunotherapy trials, hopefully revealing information on tumor behavior and therapeutic options [8]. The expression of programmed cell death ligand-1 (PD-L1) by tumor cells and TILs have been proposed as promising prognostic biomarkers for response to immunotherapy in HNSCCs [9,10]. In NSCLCs, Wang et al. showed an association between FDG-PET/CT-MPs and immune cell expression in the tumor microenvironment [7]. This association suggests that PET/CT-MPs could be potential predictors for selecting immunotherapy candidates for treatment with anti-programmed cell death-1/PD-L1 antibodies [7,11]. Systemic inflammation is also correlated with cancer prognosis. Many lines of evidence indicate that markers of systemic inflammation are independent prognostic factors for survival in many malignancies [5,[12][13][14][15][16][17]. The neutrophil-lymphocyte ratio (NLR) is one such commonly used marker that has been reported as a prognostic factor [16][17][18]. Since the correlation between systemic inflammation and tumor microenvironment has been described [7] and owing to the scarcity of retrospective data as well as relevant literature concerning interactions between the tumor microenvironment and MPs derived from FDG-PET/CT in HNSCC, we aimed to evaluate correlations between readily available markers of host inflammation with PET/CT-MPs in patients with nonmetastatic HNSCC before definitive treatment such as surgery followed by adjuvant radio(chemo)therapy or definitive radio(chemo)therapy alone. As described in other tumor entities, e.g., breast [19] and lung cancer [7], we hypothesized that NLR and platelet-to-lymphocyteratio (PLR) at baseline are positively correlated with PET/CT-MPs such as standardized uptake value (SUV), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) of the primary tumor.

Patients and complete blood count
We identified consecutive patients with HNSCC in our institutional database treated with primary or adjuvant intensity-modulated RT with or without concomitant systemic treatment between January 2007 and December 2010 at the Department of Radiation Oncology, Inselspital, Bern University. Treatment for these patients was with curative intent in all cases. Exclusion criteria were history of another malignancy within 5 years of diagnosis, prior radiation to the head and neck, non-squamous cell carcinoma histology, distant metastases, lack of differentiated complete blood count within 10 days before oncologic surgery or starting RT, and lack of PET/CT before beginning of any treatment. Pretreatment complete blood counts with differential values were used to calculate NLR and PLR. NLR was calculated by dividing absolute neutrophil count by absolute lymphocyte count; PLR was calculated by dividing absolute thrombocyte count by absolute lymphocyte count. Patients with leukocytosis caused by other reasons (e.g., infection, steroid use, etc.) were excluded from the analysis. This study was approved by the local ethics committee (289/2014) and was conducted in full accordance with ethical principles, including the World Medical Association Declaration of Helsinki (version 2002) and local legislation.

PET/CT acquisition
All patients fasted for at least 6 h before intravenous administration of FDG. All examinations were performed on one of two cross-calibrated BIOGRAPH-mCT PET/ CT scanners (Siemens, Erlangen, Germany). Image acquisition started 60 min after tracer injection, and patients were studied as follows. First, a non-contrastenhanced CT scan was performed from the skull base to mid-thigh with arms elevated using the following parameters: slice thickness of 5 mm; increment of 3.0 mm; soft tissue reconstruction kernel; maximum of 100 keV and 90 mAs by applying CARE kV and CARE Dose. Immediately after the CT scan, a whole-body PET (pelvis to vertex) was acquired in three dimensions (3D; matrix 200 × 200) with a zoom factor of 1. For each bed position (16.2 cm, overlapping scale 4.2 cm), a 2-min acquisition time with a 15.5 cm field of view was used. The emission data were corrected for randoms, scatter, and decay. Reconstruction was conducted with an ordered subset expectation maximization algorithm with four iterations/21 subsets and Gauss filtered to a transaxial resolution of 5 mm at full width at half maximum. Attenuation correction was performed using the low-dose non-enhanced CT data.
After the whole-body PET/CT imaging, a dedicated head and neck acquisition was performed from the cranial vertex to the thoracic inlet in an arm-down position for use in the SUV and MTV calculations. PET images were obtained in two table positions (10 min per table position). The PET data were reconstructed iteratively (four iterations/21 subsets) and an image matrix of 512 × 512 pixels was used. The head and neck CT scan parameters were: 120 kV, 80 mA, helical thickness of 3 mm, and field of view of 78 cm.

PET/CT analysis
FDG-PET/CT data were analyzed using an appropriate workstation and software (SyngoVia; Siemens, Erlangen, Germany). PET/CT images were visually interpreted for increased FDG uptake by the first author (BB) and a board-certified nuclear medicine physician with more than 15 years (JW) of clinical experience in oncological PET/CT. Compared to the FDG uptake of background and liver, the primary tumor with focally increased FDG uptake was considered positive for carcinoma. Semiquantitative analysis was based on manually placed regions of interest around the primary tumor with increased FDG uptake to calculate the MTV. The mean and maximum SUV values (SUVmean and SUVmax, respectively) of the lesions were calculated according to the formula: SUV = tissue concentration (Bq/mL)/(injected dose (Bq)/body weight (kg)).
The region of interest was delineated on the primary tumor site. SUVmax, SUVmean, and MTV were measured from 3D isocontour at 40% of maximal pixel value [20]. TLG was calculated by multiplying SUVmean by MTV for the lesion [21].

Treatment and follow-up
As previously published, the standard treatment was based on institutional policies following the multidisciplinary tumor board decision [22,23]. The standard treatment for oral cavity cancer was to perform surgery followed by adjuvant RT [22,24,25]. In oropharyngeal, hypopharyngeal, and laryngeal cancers, the joint recommendation of the multidisciplinary meeting was primary RT [26]. The standard concomitant therapy consisted of cisplatin 100 mg/m 2 on day 1 in 3-week intervals for all patients. Patients not deemed medically fit for cisplatin chemotherapy because of pre-existing comorbidities were evaluated for weekly treatment with the monoclonal antibody cetuximab [27] or carboplatin every 3 weeks. In a few cases of induction chemotherapy, the triplet of cisplatin, docetaxel, and 5-fluorouracil was used. Patients were regularly followed, and toxicities were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 4.03) [28].

Statistical analysis
NLR was calculated by dividing absolute neutrophil count by absolute lymphocyte count measured in peripheral blood. PLR was calculated by dividing absolute thrombocyte count by absolute lymphocyte count. Thresholds were set as described by De Felice et al. [29]. Frequencies and percentages were reported for categorical variables; continuous variables were reported as medians with range or interquartile range. The endpoints of the study were the correlation between NLR and PLR with PET/CT-MPs, overall survival (OS), and diseasespecific survival (DSS). The correlation between NLR and PLR with PET/CT-MPs (i.e., SUVmax, SUVmean, MTV log, TLG log) was examined using Spearman's rho test. Time to event was calculated for OS and DSS from the start of RT to death, with censoring patients without such events at the last follow-up. The median time to event was estimated using the Kaplan-Meier method. The prognostic value of NLR and PLR and other variables (i.e., age, sex, tumor localization, SUVmax, MTV, TLG) was assessed by univariable Cox regression-analysis. Subsequently, a multivariable analysis with backward elimination was performed, including all variables with a P-value < 0.05 in the univariable analysis. Analyses were carried out using JMP ® version 14.3.0 (SAS Institute Inc., Cary, NC, USA). The threshold for statistical significance was set at P < 0.05, and no correction for multiple testing was performed.

Results
A total of 189 patients were identified, of whom 118 patients (62%) were excluded. Of these, 105 were excluded because no pretreatment PET/CT was available. In 12 patients, the PET/CT scan was obtained after surgery. The remaining patient received induction chemotherapy before his PET/CT scan. Thus, a total of 71 patients with a pathologically proven diagnosis of oral cavity, oropharyngeal, hypopharyngeal, or laryngeal cancer were eligible and included for analysis. Eight patients underwent postoperative treatment, 6 of them being OCC patients having the standard of care treatment with primary surgery and adjuvant RT. The other 2 patients received a postoperative RT after an oncologic tonsillectomy could be performed. In 4 patients, due to the locally advanced stage and the inoperability of the primary tumor, it was not possible to distinguish between an oropharyngeal and an hypopharyngeal origin. Therefore we defined these primary tumors as multicompartmental. The median follow-up was 41 months (range 6-71 months). Patient and disease characteristics are presented in Table 1. The majority of patients were male (86%) and had good performance status (Karnofsky performance status > 70). Median NLR and PLR were 3.5 (IQR 2.1-4.7) and 195 (IQR 133-249), respectively.

Discussion
In this retrospective analysis of 71 patients with HNSCC with a relatively long follow-up of 41 months treated with definitive or adjuvant radio(chemo)therapy after oncologic surgery, we found a statistically significant correlation between pretreatment PET/CT-MPs and hematological parameters (HPs; i.e., NLR and PLR). To our knowledge, the present study results are the first to show in patients with HNSCC that MTV and TLG correlate with pretreatment systemic inflammatory parameters, i.e., PLR. These findings align with the few other studies analyzing relationships between PET/CT-MPs and HPs in different tumor entities (Table 4). A positive correlation between HPs and PET/CT-MPs has previously been demonstrated in carcinomas of the rectum [21], esophagus [20], cervix [30], breast [19], and lung [7,31]. In colorectal cancers, NLR was correlated with SUV, MTV, and TLG [21], except in the recurrent setting [32]. In esophageal carcinomas, NLR correlated with MTV but not with SUV [20]. In cervical carcinomas, a correlation was described between NLR and MTV with PLR and TLG [30]. In lymph node-positive breast carcinomas, a correlation between NLR and SUV and TLG of regional  [19]. Similarly, correlations between NLR, PLR, SUV, and MTV could also be demonstrated in small cell [31] and non-small cell [7] lung carcinomas.
Overall, the limited number of studies published on this topic have mainly described a correlation between NLR and SUV and MTV, and less frequently between PLR and SUV. Interestingly, we found no correlation between NLR and SUV in our analyses. In contrast, the analysis of PLR with MTV and TLG shows a significant positive correlation, albeit not strong. The same results are provided by Sürücü et al., showing no correlation between NLR and SUV [20] in esophageal cancer. The investigators of this work attribute this to the relatively small cohort (n = 52) they analyzed. How the interaction of the tumor microenvironment and FDG-PET/CT could influence treatment decisions has also been evaluated. In NSCLCs, Wang et al. described a correlation    [7]. Several studies have demonstrated that TILs are present in the tumor microenvironment in various malignant tumors [3][4][5]. In oropharyngeal carcinoma, TILs can identify stage I HPV-associated patients likely to be poor candidates for treatment de-escalation [6]. An interplay between systemic inflammatory parameters and tumor microenvironment has also been shown. In gastric cancer, systemic inflammation is associated with the density of CD4 + lymphocytes in the tumor microenvironment [33]. We investigated whether systemic inflammatory parameters correlate with PET/CT MPs to potentially provide indirect information about the tumor microenvironment with easily accessible blood tests in combination with PET/CT-MPs. Hitherto, such analyses have not been performed in HNSCC using FDG-PET/CT. We   propose that such findings would be particularly prescient in oligometastatic disease. For example, it might be possible to identify lesions that respond less to immunotherapy. FDG-PET/CT, in combination with NLR and PLR, might identify candidate lesions amenable to stereotactic radiation. We note that histopathological analysis is not always available for several patients, either for local lesions, which are not easily biopsied due to their anatomical location, or for distant metastases. Furthermore, primary tumor and nodal metastases show poor to fair agreement when comparing biopsy material and resection tissue [34]. Predictive biomarkers, such as the composite positive score developed by Roach et al. in 2017 [35], have the potential to surmount some of these challenges and have shown promise in NSCLC [34,36,37]. The idea of using HPs in combination with PET/CT-MPs in patients with oligometastatic disease as a clinically useful prognostic tool for therapy response of a metastasis that is difficult to access through biopsy holds potential, not only in NSCLC but also in HNSCC.
More is known about the relationship between systemic inflammatory parameters and outcomes. In multiple studies, retrospective analyses and meta-analyses have shown that baseline circulatory NLR is a strong predictor of survival outcomes after radio(chemo)therapy for patients with HNSCC [38][39][40][41][42]. Cancer is a systemic disease whose course and prognosis is influenced by inflammatory reactions in the micromillieu, as a hallmark of cancer by tumor invasion and metastasis [12]. Two decades ago, Balkwill et al. reported that inflammatory cells and cytokines in the tumor could contribute to tumor growth, progression, and immunosuppression [43]. Since then, several HPs, such as platelets, neutrophils, lymphocytes, NLR, PLR, and MPV, have been studied in various malignant tumors for their effects on tumor pathology [44][45][46][47][48]. In patients with HNSCC, numerous analyses have been performed concerning PET/CT-MPs regarding tumor control and prognosis, reporting MTV defined from pretreatment FDG-PET scans as the strongest predictor of patient outcome after radio(chemo)therapy [49]. The effect of pretreatment SUVmax and SUVmean on patient outcome seems to be a less important factor compared with MTV [49].
There is no clear answer on the relationship between systemic inflammatory parameters and outcome. A possible explanation is that distant (micro)metastases are already manifest at initial diagnosis, which cannot yet be detected by any imaging modality [50]. For clinical routine, it might be interesting to compute a score comprising PET/CT-MPs and HPs for the prediction of already existing but subclinical distant metastases, in order to adapt individual therapy concepts.
In addition to its retrospective design, this hypothesisgenerating study has a number of limitations. Using our small cohort, we did not observe a correlation between NLR and PET/CT-MP. Studies with larger or pre-defined cohort sizes are required to further test these potential relationships, because of a substantial type II error risk due to our under-powered sample size. At the time patients in this study were examined, HPV status was not routinely acquired, and we are unable to assess the influence of HPV status on the parameters measured in this study. Furthermore, the effect of some PET/CT-MPs, in particular MTV, could be influenced by the tumor lesion's size, similar to the influence of tumor volume on the TNM stage [51,52]. For example, the relationship with PLR could be a result of the platelets' contribution to tumor growth [53]. Moreover, we have only recorded PET/CT-MPs for the primary lesions, and other FDGavid metastatic lesions should be considered by future studies, ideally of prospective design. Finally, although our historical patient collective (2007-2010) afforded a follow-up for clinical outcomes, the scanner type used does not represent the state of the art, where digital and long axial field-of-view PET/CT systems can show improved lesion quantification [54].

Conclusions
In patients with HNSCC treated with definitive radio(chemo)therapy or oncologic surgery followed by adjuvant RT, baseline PLR is correlated with MTV and TLG but not with SUV. NLR did not correlate with any PET/CT-MPs analyzed in our study. Confirmatory studies and further investigation concerning immunohistochemistry in HNSCC are needed, and a potential   interaction between tumor microenvironment, host inflammation, and FDG-PET/CT measures warrants more detailed research, maybe yielding information on tumor behavior and decision-making in radioimmunotherapeutic or treatment deescalating options.