Patients with solid tumors, who were planned to receive chemotherapy, were asked to participate in the pharmacogenomic study at Seoul National University Hospital (IRB Registration No. H-0610-020-186). After obtaining informed consent, patients donated blood for research purposes, and were registered in the database. From this database, we selected patients with stage IIIB or IV NSCLC who were treated with platinum-based combination chemotherapy as a first-line treatment between September 2005 and September 2008. Patients with histologically confirmed NSCLC were eligible for this study. Platinum-based combination chemotherapy was defined as cisplatin or carboplatin, plus one of the following: taxane, gemcitabine, vinorelbine, and etoposide. Patients who received concurrent chemoradiation as a first-line treatment or who had any history of prior chemotherapy, including adjuvant therapy, were excluded.

The main objective of this study was to evaluate the association between VEGF-A, VEGF₁₆₅b, IL-1β, and TGF-β1 serum levels and treatment outcomes of patients. In addition, the correlation of serum cytokine levels with leukocyte or platelet counts was evaluated. The prognostic impact of leukocyte or platelet counts was also studied.

Data regarding patient demographics, pathologic classification, treatment response, progression-free survival (PFS), and overall survival (OS) were obtained via a medical record review. We also collected clinical data regarding leukocyte and platelet counts at the time of serum sampling, along with the history of infection requiring antibiotics treatment within 1 month prior to chemotherapy. Patients who had smoked more than 100 cigarettes during the lifetime were defined as smokers. Tumor histology and subtypes were classified by the World Health Organization (WHO) criteria [19]. The treatment response was evaluated using a computed tomography scan by the Response Evaluation Criteria in Solid Tumors (RECIST) ver. 1.0 [20]. PFS was calculated from the initiation of chemotherapy to documented disease progression or death from any cause. OS was calculated from the start of chemotherapy to death from any cause.

Serum samples acquired before chemotherapy were used for this study. The blood samples were collected and allowed to clot before centrifugation. Serum was removed and stored in liquid nitrogen until used. To measure cytokine levels from the baseline serum samples, enzyme-linked immunosorbent assay (ELISA) was performed, using commercially available kits, following the manufacturer's instructions for the following markers: VEGF-A (R&D Systems, Minneapolis, MN), VEGF₁₆₅b (R&D Systems), IL-1β (R&D Systems), and TGF-β1 (R&D Systems). To measure serum levels of latent complexes of TGF-β1, the activation procedure by acidification was performed before ELISA as the manufacturer's instructions.

The capture antibody was diluted to the working concentration in phosphate buffered saline without carrier protein. A 96-well microplate was coated with the diluted capture antibody. The plate was sealed and incubated overnight at room temperature. After incubation, each well was aspirated and washed with wash buffer three times. Each well was then blocked with block buffer and incubated at room temperature for 1 hour. Each well was aspirated and washed again with a wash buffer three times. When the plate was prepared, the samples or standards in an appropriate diluent were added to each well. The plate was covered with an adhesive strip, which was then incubated at room temperature according to the manufacturer's instructions. After incubation, each well was washed with a wash buffer three times. Then, the detection antibody was added to each well. The plate was covered with a new adhesive strip and incubated at room temperature. After incubation, each well was washed again with a wash buffer three times. The working dilution of streptavidin-HRP was added to each well. The plate was covered and incubated again at room temperature, avoiding any direct light. After incubation, each well was washed again with a wash buffer three times. A substrate solution was added to each well. Then, the plate was incubated again at room temperature, avoiding any direct light. After incubation, a stop solution was added to each well. The optical density of each well was determined using a microplate reader (Multiskan Ascent, MTX Lab Systems, Inc., Vienna, VA), which was set to 450 nm with wavelength correction at 540 nm. Using measured optical density values, the concentration of each well was calculated. The samples were assayed in duplicate.

The median PFS and OS were calculated using the Kaplan-Meier method. The univariate and multivariate analyses of categorical variables was performed using logistic regression. The univariate and multivariate analyses of risk factors for survival data were performed using the Cox proportional hazard model. Variables with a p-value in the univariate analysis of less than 0.05 were selected for the multivariate analysis. Mann-Whitney U test was used to compare two different groups, consisting of non-parametric continuous variables. To correlate between serum cytokine levels and leukocyte counts, the Pearson correlation was used. Two-sided p-values below 0.05 were considered statistically significant. SPSS ver. 17.0 (SPSS Inc., Chicago, IL) software was used for all the statistical analyses.

Signed informed consents for chemotherapy and blood sample collection for the pharmacogenomic study were obtained from all patients before treatment. The protocol of this study was reviewed and approved by the institutional review board of Seoul National University Hospital (IRB Registration No. H-0906-071-284) and conducted in accordance with the precepts established by Helsinki Declaration.

A total of 140 patients were enrolled in this study. The median age of patients was 64 years (range, 25 to 86 years). One hundred four patients (74.3%) were male. Ninety patients (64.3%) were previous or current smokers. Tumor histology was as follows: adenocarcinoma in 73 patients (52.1%), squamous cell carcinoma in 29 patients (20.7%), large cell neuroendocrine carcinoma in 4 patients (2.9%), and not otherwise specified cases in 34 patients (24.3%). Twentyfour patients (17.1%) were initially diagnosed with stage IIIB disease. Among the other 116 patients (82.9%) with initial stage IV, 109 patients initially had metastatic disease and 7 patients had recurrent disease after curative treatment. Eastern Cooperative Oncology Group performance status were grade 0 in 11 patients (7.9%), 1 in 107 patients (76.4%), and 2 in 22 patients (15.7%). Chemotherapy regimens were as follows: taxane plus platinum in 98 patients (70.0%), gemcitabine plus platinum in 41 patients (29.3%), and etoposide plus platinum in 1 patient (0.7%).

During a median follow-up of 20.5 months (range, 1.0 to 40.7 months), 125 patients (89.3%) experienced disease progression and 82 patients (58.6%) died. The treatment response could be evaluated in 132 patients. The response rate of platinum-based combination chemotherapy was 47.0%. The median PFS was 3.9 months (95% confidence interval [CI], 3.0 to 4.7 months). The median OS was 13.0 months (95% CI, 9.7 to 16.3 months).

The median VEGF-A level was 660 pg/mL (range, 49 to 3,722 pg/mL). VEGF₁₆₅b levels were detectable (>15.6 pg/mL) by ELISA in 17 samples (12.1%). The median TGF-β1 level was 11,180 pg/mL (range, 511 to 41,210 pg/mL). IL-1β levels were detectable (>1.56 pg/mL) by ELISA in 19 samples (13.6%). The associations among serum cytokine levels were not statistically significant (data not shown).

The results of the univariate analysis are shown in Table 1. Smoking history was a significant factor for OS (p=0.020). Histology was a significant factor for the treatment response (p=0.038) and OS (p=0.013). The performance status was a significant factor for PFS (p<0.001) and OS (p=0.001). The association between cytokine levels and treatment responses was not significant. Patients with high VEGF-A levels (≥1,000 pg/mL, n=41) had shorter OS (median, 8.2 months [95% CI, 4.6 to 11.8 months] vs. 13.8 months [95% CI, 8.6 to 19.0 months]; p=0.035). VEGF₁₆₅b was not associated with clinical outcomes of patients. High TGF-β1 levels (≥10,000 pg/mL, n=77) were associated with shorter PFS (median, 3.1 months [95% CI, 2.4 to 3.8 months] vs. 4.9 months [95% CI, 4.0-5.8 months]; p=0.025). High IL-1β levels (≥3.00 pg/mL, n=10) were associated with shorter PFS (median, 3.5 months [95% CI, 1.9 to 5.1 months] vs. 4.2 months [95% CI, 3.2 to 5.3 months]; p=0.046) and OS (median, 7.1 months [95% CI, 4.5 to 9.7 months] vs. 13.5 months [95% CI, 9.9 to 17.1 months]; p=0.030).

The results of the multivariate analysis are shown in Table 2. After adjustment for performance status, which was a significant variable in the univariate analysis, high TGF-β1 and IL-1β levels tended to be associated with a high risk of disease progression with a marginal significance (p=0.075 and p=0.070, respectively). VEGF-A (Fig. 1A) and IL-1β (Fig. 1B) were significant prognostic factors (p=0.0497 and p=0.047, respectively) after adjustment for smoking history, histology, and performance status.

Serum VEGF-A levels were correlated with leukocyte counts (r=0.422, p<0.001) (Fig. 2A) and platelet counts (r=0.462, p<0.001; data not shown). Serum TGF-β1 levels were also correlated with leukocyte counts (r=0.197, p=0.020) (Fig. 2B) and platelet counts (r=0.359, p<0.001; data not shown). However, serum VEGF₁₆₅b and IL-1β levels were associated with neither leukocyte counts (p=0.071 and p=0.378, respectively) nor platelet counts (p=0.495 and p=0.816, respectively).

The median leukocyte and platelet counts at the time of treatment were 7,505/mm³ (range, 3,940 to 55,560/mm³) and 278,000/mm³ (range, 68,000 to 692,000/mm³), respectively. The effects of the baseline leukocyte and platelet counts on OS are shown in Table 3. High leukocyte counts (≥9,000/mm³, n=39) and high platelet counts (≥400,000/mm³, n=20) were significant risk factors for OS in the univariate analysis. After adjustment for smoking history, histology, and performance status, the baseline leukocyte count was an unfavorable prognostic indicator (p<0.001) (Fig. 3).

Nineteen patients (13.6%) required antibiotics treatment for pneumonia within 1 month before the initiation of chemotherapy. No other infectious diseases developed. All patients had recovered from pneumonia, and antibiotics were stopped at the time of treatment. No significant association was found between the history of pneumonia and baseline leukocyte counts (p=0.937). The baseline platelet counts were also not significantly different according to the history of pneumonia (p=0.775).