Overexpression of Glut1 in Lymphoid Follicles Correlates with False-Positive ¹⁸F-FDG PET Results in Lung Cancer Staging

Subjects

Sixty-two patients with NSCLC (48 men, 14 women; age range, 39–77 y; mean, 62.5 y) who had surgery were enrolled in this study. Before being enrolled, all patients provided written informed consent for PET scanning and to have their tissues studied experimentally. The patients underwent preoperative ¹⁸F-FDG PET as a part of the staging work-up, and mediastinal lymph node sampling was performed at thoracotomy or mediastinoscopy.

CT examination was performed during preoperative work-up in all patients. The maximal diameter of enlarged lymph nodes, as determined from CT scans, was taken as an index of the size. The histologic diagnoses were 28 squamous cell carcinomas, 21 adenocarcinomas, 7 large cell carcinomas, 4 adenosquamous carcinomas, and 2 sarcomatoid carcinomas. One-hundred eleven mediastinal lymph nodes from 62 patients were selected for this study. Since the specific aim of this study was not to evaluate the sensitivity or specificity of PET, we selected all false results but only a portion of true results to serve as a control. The lymph nodes consisted of 4 groups: 41 true-positive (TP), 31 true-negative (TN), 27 FP, and 12 false-negative (FN) (Table 1).

¹⁸F-FDG PET

¹⁸F-FDG PET was done 1–2 wk before surgery. All patients fasted for at least 6 h before PET imaging. Six patients had diabetes mellitus. In these patients, the dosage of insulin or oral hypoglycemic agent was reduced before scanning, and their blood glucose levels before scanning were <150 mg/dL. PET was performed using an Advance PET scanner (General Electric Medical Systems). Emission scans were obtained from the orbitomeatal line of the head to the thigh for 6 min per frame, 45 min after the intravenous injection of 370 MBq ¹⁸F-FDG. Transmission scans were performed for 3 min per frame using the ⁶⁸Ge ring source and were used for the subsequent attenuation correction of the emission scans. Tomographic images were reconstructed with attenuation correction by an ordered-subsets expectation maximization iterative algorithm with a Hanning filter (cutoff frequency, 8.0 mm) and displayed as coronal, sagittal, and transaxial slices on a workstation. The standardized uptake value (SUV) was defined as the radioactivity concentration inside the region of interest divided by the injected activity per kilogram of body weight. Two nuclear physicians who were unaware of the CT and histologic results reviewed together and interpreted the PET images by consensus. Regional lymph nodes were considered positive for malignancy if focal prominent ¹⁸F-FDG uptake was found in 2 or more consecutive transaxial slices. Lymph nodes with an SUV of 3 or more were considered positive in case of failure of consensus as described (18). Each lymph node region visualized on PET was matched to a surgical specimen by the surgeon with expertise in mediastinal lymph node dissection.

Histologic Examinations of Lymph Nodes

Three serial sections of the lymph nodes were examined, and immunohistochemical staining for cytokeratin was performed to rule out micrometastasis in all negative lymph nodes. For histologically proven negative lymph nodes, Glut1 expression and the grade of lymphoid follicular hyperplasia were examined. Two independent pathologists graded lymphoid follicular hyperplasia by reviewing the same histologic sections without knowledge of the ¹⁸F-FDG PET results. Based on the area occupied by follicles in the lymph node, the grading of lymphoid follicular hyperplasia was assigned as follows: grade 1, <25%; grade 2, 25%–49%; grade 3, 50%–74%; and grade 4, 75%–100%.

For histologically positive lymph nodes, the proportion of tumor cells and Glut1 expression were determined. The proportion of tumor cells was determined as the relative tumor-occupying area in the cut surface of the involved lymph nodes.

Glut1 Immunostaining

One-hundred eleven lymph nodes from 62 patients were submitted for immunohistochemical study. Formalin-fixed, paraffin-embedded tissue sections were immunostained with rabbit anti-Glut1 polyclonal antibody (1:1,000; Chemicon). Four-micron sections were deparaffinized in xylene and rehydrated in a graded ethanol series. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide in methanol. The sections were then washed with phosphate-buffered saline (PBS, pH 7.6), and microwaved in 10 mmol/L citrate buffer (pH 6.0) for 10 min to expose the antigens. After washing with PBS, the sections were incubated with bovine serum albumin for 10 min at room temperature to suppress nonspecific protein binding. Treated tissue sections were incubated with anti-Glut1 antibody at 37°C for 60 min and then rinsed with PBS. Sections were then incubated with biotinylated secondary antibody and streptavidin-biotin-peroxidase complexes (LSAB kit; DAKO). After washing, 3,3′-diaminobenzidine was used as a chromogen. Finally, the sections were counterstained with hematoxylin, dehydrated, and mounted. Adjacent sections incubated with rabbit IgG were used as negative controls. Red blood cells present in each section served as positive controls for Glut1. Glut1 expression was considered positive only if distinct membrane staining was present. Because the staining pattern is heterogeneous in the same tissue, we developed our own scoring method that includes staining intensity as well as the proportion of positive cells. Staining intensity was categorized as follows: 1, weak; 2, weak to medium; 3, medium; and 4, intense. The proportion of positive cells was graded on a semiquantitative scale as 1, 1%–9%; 2, 10%–24%; 3, 25%–49%; 4, 50%–74%; and 5, 75%–100%. Then the 2 values are multiplied and, finally, the Glut1 score is determined as follows: 1, ≤6; 2, ≤10; 3, ≤15; and 4, ≤20.

Statistical Analysis

The data are expressed as mean ± SD. Probability values of <0.05 indicated a statistically significant difference. Statistical analysis of size and tumor proportion between false and true results by PET was performed by the χ² test. Statistical analysis of the grade of follicular hyperplasia and Glut1 score between false and true results by PET was tested by the Mann–Whitney nonparametric 2-sample test (SPSS version 10; SPSS Inc.).

Expression of Glut1

All metastatic tumors stained positively for anti-Glut1 antibody, and the staining patterns were heterogeneous by histologic type. The staining intensity and percentage of positive cells were considerably greater in the cell membranes of squamous cell carcinomas than that of adenocarcinomas or other NSCLCs (Fig. 2). Adenocarcinomas showed the most pronounced heterogeneity, with intracytoplasmic granular staining as well as weak membranous staining. All lymphoid follicular cells showed strong positivity along the cell membranes (Fig. 3).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Glut1 immunostaining in metastatic non-small cell carcinomas. (A) Squamous cell carcinoma with strong membranous Glut1 immunostaining (STAIN, ×200). (B) Adenocarcinoma with focal cytoplasmic Glut1 immunostaining (immunostain, ×200).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Glut1 immunostaining in lymphoid follicles. Note intense linear membranous staining in lymphoid follicular cells and erythrocytes as an internal positive control (immunostain, ×200).

Histologically Proven Negative Lymph Nodes

Glut1 expression was diffuse and uniform in the lymphoid follicles but was negative or faintly stained in the extrafollicular lymphoid cells. The grade of lymphoid follicular hyperplasia was 2.7 ± 1.1 in FP nodes and 1.6 ± 0.7 in TN nodes (P < 0.001) (Table 2). The Glut1 score was 3.0 ± 0.8 and 1.5 ± 0.9 in FP and TN nodes, respectively (P < 0.001). Maximal lymph node diameters estimated from CT scans were 11.0 ± 2.9 and 9.7 ± 2.8 mm for FP and TN nodes, respectively (P = 0.071).

Histologically Proven Positive Lymph Nodes

The Glut1 score of the metastatic tumor cells was 3.1 ± 0.9 and 1.8 ± 0.7 in TP and FN nodes, respectively (P < 0.001) (Table 3). The Glut1 score of the metastatic adenocarcinoma was 2.8 ± 0.6 and 1.5 ± 1.1 in TP and FN nodes, respectively (P = 0.015). The proportion of tumor cells in lymph nodes was 67.9% ± 35.0% (range, 5%–100%) and 52.5% ± 31.2% (range, 10%–90%) in TP and FN nodes, respectively (P = 0.095). The maximal lymph node diameters of TP and FN nodes were 13.3 ± 5.0 and 11.2 ± 2.7 mm, respectively (P = 0.584).