Background
Non-small cell lung cancer (NSCLC) is a heterogeneous group of carcinomas with different biological behaviors and prognoses. Histological classification and staging are critical in constituting a treatment strategy and predicting prognosis for NSCLC [
1]. [
18 F]-2-fluoro-deoxy-D-glucose (FDG)-positron emission tomography (PET) is a metabolic imaging technique which has become an essential tool for staging of NSCLC patients. The integration of PET with computed tomography (PET/CT) provides an accurate anatomic localization and improved staging especially for mediastinal lymph nodes and occult distant metastases [
2,
3]. It also provides prognostic information, monitors response to therapy and can be used to follow up patients after treatment [
4]. The rationale for using FDG-PET in oncology is its ability to measure increased glucose metabolism of tumor cells. Elevated FDG uptake suggests that the lesions or tissues harbor tumor cells. The maximum standardized uptake value (SUVmax) greater than 2.5 is often used as a cut off value for malignancy. However it has been shown that there is a significant number of false positivity (due to inflammatory diseases) and false negativity (due to low-grade malignancies) in the evaluation of primary tumor [
5]. The major reasons of false negative and false positive lymph nodes are microscopic metastasis beyond the spatial resolution of PET/CT and lymph node involvement by underlying inflammatory processes such as immune reaction due to the presence of lung tumor, obstructive pneumonia, anthracosis or granulomatous inflammation [
6‐
8].
It is often assumed that FDG uptake is primarily within the malignant tumor cells, and SUVmax is a well known measure indicating the aggressiveness of tumor [
9,
10]. But other cellular components of such as normal parenchymal cells, atypical cells, inflammatory cells, fibroblasts, or hematopoietic progenitor cells may also uptake FDG. To the best of our knowledge, there is only one study investigating the correlation between SUVmax and specific cellular components of the tumor conducted in patients with resected stage 1 NSCLC. In that study the authors found that the cellular composition of the tumor was highly variable and there wasn’t any correlation between a specific tumor cellular component and FDG activity [
11].
SUVmax of the primary tumor is a risk factor for occult mediastinal metastasis in clinical stage 1 NSCLC patients [
12‐
14]. The likelihood of lymph node metastasis increases with the increase of tumor SUVmax [
6,
15]. SUVmax of lymph node is also important for predicting metastasis, but false positivity is an important challenge [
6,
7]. For this reason some authors claimed to use a higher SUVmax value instead of the traditional value of 2.5 in order to increase the accuracy for presence of metastasis [
7].
In this study we aimed to investigate the correlation of tumor SUVmax value with largest tumor diameter, tumor histology, differentiation, number of mitosis, degree of stromal inflammation and necrosis, and to determine whether a T/LN SUVmax ratio can predict the presence of metastasis in mediastinal or hilar lymph nodes in NSCLC patients.
Discussion
The present study revealed that SUVmax of the primary tumor is positively correlated with the largest tumor diameter, number of mitosis, and pathological stage of the disease. Mean SUVmax and number of mitosis were higher in SCC compared to AC. N stages were advanced in SCC. Reactive hyperplasia and anthracosis were the major etiologies for positive lymph nodes on PET/CT. A T/LN SUVmax ratio of 5 or lower was suggestive for a malignant lymph node with high sensitivity (92.8%) but low specificity (47%).
FDG activity in other words SUVmax measured on PET/CT is a semiquantitative value that indicates the degree of glucose uptake in a lesion. The exact mechanism of FDG uptake remains unknown. Most lung cancers accumulate FDG preferentially that SUVmax may vary widely [
11,
18]. SUVmax of a tumor is a product of several fundamental factors including glucose metabolism and the type/number of cells present in the tumor. Higher SUVmax may be a result of either higher mitotic activity of tumor cells or higher number of inflammatory cells (lymphocytes, macrophages) that competitively uptake FDG-glucose. Alternatively, there may be a large number of tumor cells with low metabolic activity or a low number of tumor cells with high metabolic activity [
11]. In the present study, we found a positive correlation between tumor SUVmax and number of mitosis supporting the role of higher mitotic activity in the mechanism of FDG uptake. There was no correlation between tumor SUVmax and the degree of stromal inflammation.
The tumor size and the presence of necrosis are other factors that affect the SUVmax of a tumor. Previous studies demonstrated the positive correlation between tumor diameter and SUVmax [
19‐
22]. The increase in the tumor diameter was also correlated with more glucose transporter-1 (Glut-1) expression on the surface of tumor cells, leading to increased FDG uptake [
23]. In this study, the largest tumor diameter was positively correlated with SUVmax in AC, SCC and the whole study group. As expected, larger tumors were more necrotic. Contrary to the conventional wisdom that tumor necrosis dilutes standart uptake value, more necrosis was not associated with lower SUVmax. There was no correlation between SUVmax and the degree of necrosis. This condition can be explained by the fact that SUVmax is calculated from the highest FDG uptake regions of the tumor.
Consistent with the previous studies [
6,
19,
21,
22,
24], this study revealed that SCC had higher SUVmax values compared to AC. This can be explained in two ways. First, SCCs are rapidly growing tumors with shorter doubling times, thus leading to higher levels of glucose metabolism [
18]. Second, the expression of Glut-1 is higher in SCCs. The localization of Glut-1 is also important in the uptake process of FDG. While Glut-1 is mainly located in cytoplasm of tumor cells in ACs, it is mainly located on cell membranes of tumor cells in SCCs. The cell membrane localization is more important in FDG uptake process [
25,
26].
Recently, a study investigating Glut-1 expression and FDG uptake in histological subtypes of pulmonary AC documented that solid predominant ACs had a significantly higher Glut-1 expressions and SUVmax than those with other predominant histology [
27]. In the present study, despite the fact that there are limited number of cases in AC subgroups, we compared solid predominant ACs (n = 9) with acinary predominant ACs (n = 16 cases) and found that the largest tumor diameters, number of mitosis, and the degree of necrosis were higher in solid ACs. Mean tumor SUVmax was also higher in solid predominant ACs but the difference did not reach statistical significance. All these findings supported the aggressive behavior of solid predominant ACs.
The relationship between tumor SUVmax and differentiation is conflicting in the literature. Some studies showed that tumor SUVmax was higher in poorly differentiated NSCLCs [
22,
28]. Another study found no correlation between SUVmax and the degree of differentiation in SCCs [
19]. In the present study, there was no relation between tumor SUVmax and differentiation in SCCs. But the degree of stromal inflammation was higher in poorly differentiated tumors.
Higher SUVmax of the primary tumor is found to be a strong predictor of lymphtic vessel invasion and lymph node metastasis in studies consisting resected early stage NSCLC patients [
21,
29]. In a study conducted on a small group of resected lung cancer patients, the authors investigated the correlation of SUVmax with the likelihood of lymph node metastasis and reported that, when the SUVmax of the primary tumor was greater than 12, the probability of lymph node metastasis was high, reaching 70%, irrespective of the degree of FDG activity in the lymph node stations [
6]. In the present study, tumor SUVmax was correlated with pathological T stage, disease stage, but not with N stage. Mean tumor SUVmax was not different in patients with and without metastatic lymph nodes (Table
6). Contrary to the current literature reporting AC histology as a risk factor for occult N2 lymph node metastasis [
12,
13], in this study N stages were advanced in SCCs that have higher SUVmax compared to ACs.
PET/CT is now regarded as the most accurate imaging modality in N-staging of lung cancer. However there are a significant number of false positivity and false negativity. The major reason for false negativity is microscopic metastasis beyond the spatial resolution of PET/CT. The major reasons for false positivity are lymph node involvement by underlying inflammatory disease (tuberculosis, histoplasmosis), lymphadenopathies secondary to obstructive pneumonia, immune reaction due to the presence of lung tumor, antracosis, and silicotic nodules [
6,
8]. The major cause of false positivity may vary from region to region. In a study from Alabama, histoplasmosis infection was the most common cause of false positivity [
7]. Silicosis has been found to be a cause of false positivity in a study from Germany [
30]. In the present study, there were antracosis in 40% of the PET/CT positive lymph nodes probably due to the intensive indoor air pollution or biomass exposure. Granulomatous inflammation (4%) and silicosis (3%) was low.
In the present study, distinctively from the literature, we considered a lymph node station as positive if there was a FDG uptake higher than the surrounding mediastinal tissue, with regard to the idea that a tumor with low FDG activity might have a metastatic lymph node with low FDG activity. Thus we set a low threshold (any FDG uptake higher than the surrounding mediastinal blood pool) in order to avoid false negative lymph nodes and accept a higher incidence of false positives. All of the metastatic lymph nodes had a SUVmax higher than 2.5. The sensitivity (63.6%) and specificity (72.4%) of PET/CT was lower than the current literature due to the lower threshold we choose for PET/CT positivity.
In the present study we hypothesized that a T/LN SUVmax ratio could be a predictor of lymph node metastasis. This hypothesis was constituted based on the clinical observation that a cut off value of 2.5 for the prediction of metastasis is too low in countries where inflamatory reactions are more prevelant. Besides in clinical practice we realized that some tumors with low SUVmax values have metastatic lymph nodes after resection despite lower SUVmax values than 2.5 on PET/CT obtained before surgery. The present study favors this hypothesis. As demonstrated in Table
6, while mean SUVmax values were not different between metastatic and non metastatic lymph nodes, T/LN ratios were significantly lower in metastatic ones. T/LN SUVmax ratios were similar in anthracotic and reactive lymph nodes. A T/LN SUVmax ratio of 5 or lower was suggestive for a malignant lymph node with high sensitivity but low specificity. Since this study is performed in a population where the prevalance of inflammatory disease is high, T/LN SUVmax ratios may show variations in different patient populations.
In the literature
Cerfolio et al also determined a ratio of LN/T SUVmax as a universal predictor of lymph node metastasis in order to eliminate the variation of SUVmax among different PET scanners, and documented that when the ratio is 0.56 or greater, there is a 94% chance that the node is malignant [
31].
There are some limitations of the present study. First it is a retrospective study with limited number of patients. The study population consists of patients who underwent surgery only that may cause a selection bias. Since all cases were early lung cancers who are candidates for curative surgery, the number of metastatic lymph nodes was low. Prospective studies including more lymph node stations are needed to examine the validation of a T/LN SUVmax ratio for prediction of metastasis.
Competing interests
All authors disclose that there is not any actual or potential conflict of interest including any financial, personal or other relationships with other people or organizations that could inappropriately influence (bias) their work.
Authors’ contributions
DK designed the study, collected the data, interpret the results and draft the article. FD carried out the pathological evaluation. HB conceived of the study, and participated in its design and coordination and helped to draft the manuscript. OO carried out the interpretation of PET/CT imagings. ET carried out the interpretation of PET/CT imagings. BB participated in the design of the study and performed the statistical analysis. KA participated in the design of the study and coordination. EY participated in the design of the study and coordination. All authors read and approved the final manuscript.