Usefulness of monitoring the circulating Krebs von den Lungen-6 levels to predict the clinical outcome of patients with advanced nonsmall cell lung cancer treated with epidermal growth factor receptor tyrosine kinase inhibitors

Authors

  • Nobuhisa Ishikawa,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Noboru Hattori,

    Corresponding author
    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    • Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8551, Japan
    Search for more papers by this author
    • Fax: +81-82-255-7360

  • Akihoto Yokoyama,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    2. Department of Hematology and Respiratory Medicine, Kochi Medical School, Kochi University, Kochi, Japan
    Search for more papers by this author
  • Sonosuke Tanaka,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Ryohei Nishino,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Koji Yoshioka,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Shinichiro Ohshimo,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Kazunori Fujitaka,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Hiroshi Ohnishi,

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Hironobu Hamada,

    1. Department of Integrated Medicine and Informatics, Ehime University Graduate School of Medicine, Ehime, Japan
    Search for more papers by this author
  • Koji Arihiro,

    1. Department of Anatomical Pathology, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author
  • Nobuoki Kohno

    1. Department of Molecular and Internal Medicine, Graduate School of Biomedical Sciences, Hiroshima University, Hiroshima, Japan
    Search for more papers by this author

Abstract

Krebs von den Lungen-6 (KL-6) is a high molecular weight glycoprotein classified in the category of human MUC1 mucin. KL-6 has been reported to serve as a sensitive marker for interstitial pneumonia; however, recent studies have suggested that it can also be used as a tumor marker as its origin shows. To further elucidate the clinicopathological significance of circulating KL-6 in lung cancer, we monitored the circulating KL-6 levels in advanced nonsmall cell lung cancer (NSCLC) patients and analyzed the association between these levels and the clinical outcome of EGFR-TKI treatment. The pretreatment levels of circulating KL-6 were found to be significantly higher in progressive disease (PD) patients than disease-controlled (partial response (PR) and stable disease (SD)) patients. Multivariate analyses revealed the circulating KL-6 level to be an independent prognostic factor for overall survival as well as progression-free survival. In addition to these observations, we found that changes in circulating KL-6 levels at 2 weeks after the start of EGFR-TKI treatment from the baseline could quite precisely discriminate PD cases from PR or SD patients and the clinical outcome of EGFR-TKI in NSCLC patients. These results indicate that the monitoring of circulating KL-6 levels in NSCLC patients is effective for both selecting patients to be treated with EGFR-TKI and predicting the clinical outcome of EGFR-TKI. In addition, the findings suggest that the circulating KL-6 level could be used as a clinically relevant biomarker in patients with NSCLC, particularly those who are candidates for EGFR-TKI treatment. © 2008 Wiley-Liss, Inc.

Lung cancer is the leading cause of cancer deaths worldwide, and nonsmall cell lung cancer (NSCLC) accounts for nearly 80% of those cases.1 The epidermal growth factor receptor (EGFR) is expressed at high levels in the majority of NSCLC cases and becomes a major target for molecular targeted therapies.2, 3 Gefitinib (ZD1839, Iressa) is an orally active, selective EGFR tyrosine kinase inhibitor (EGFR-TKI) that blocks the signal transduction pathways that have been implicated in the proliferation and survival of cancer cells.4 EGFR-TKI has been shown to improve the symptoms and quality-of-life in a subset of patients with advanced NSCLC, including women, nonsmokers, patients with lung adenocarcinoma (ADC) histology and patients of Asian origin, who did not respond to platinum-based chemotherapy.5–9 In spite of these favorable effects, treatment with EGFR-TKI was found to cause serious side effects, such as EGFR-TKI-induced interstitial lung disease, in about 4% of patients in Japan.10 Considering these observations, the establishment of a method to predict the clinical outcome of EGFR-TKI in NSCLC patients would be highly useful.

Recently, somatic mutations in the tyrosine kinase domain of EGFR have been identified in NSCLC cells and the responsiveness to EGFR-TKI has been shown to correlate with the presence of EGFR mutations.11, 12 Furthermore, recent studies have demonstrated that the presence of EGFR mutations predict a prolonged survival in NSCLC patients treated with EGFR-TKI.13–15 Therefore, the detection of EGFR mutations seems to be a promising method to predict the responsiveness and survival benefit of EGFR-TKI treatment; however, the analysis of EGFR mutations requires a decent size of tissue specimen obtained by surgery or biopsy, which is not routinely performed for advanced NSCLC and may cause various complications. Furthermore, previous reports and our observations have demonstrated that no single factor examined so far, including somatic mutations of EGFR, has been able to precisely determine the clinical benefit of EGFR-TKI. Therefore, practical clinical tests, such as serological markers, that can predict the clinical outcome of EGFR-TKI in NSCLC patients are urgently required.

Krebs von den Lungen-6 (KL-6) is a high-molecular-weight glycoprotein classified as “Cluster 9 (MUC1)” of lung tumor and differentiation antigens according to the findings of immunohistochemical and flow cytometry studies.16, 17 Although cluster 9 antibodies can react with carbohydrate chains not expressed in MUC1 glycoforms, a previous study from our laboratory clearly demonstrated KL-6 to be a submolecule of MUC1 mucin.18 KL-6 has been reported to serve as a sensitive serum marker for interstitial pneumonia19, 20 and is now clinically used to detect the presence of interstitial pneumonia in Japan; however, recent studies have suggested that it can also be used as a tumor marker as its origin shows.21 KL-6 expression is increased in breast, lung, pancreas, ovary, colon and hepatocellular carcinomas,22–28 thus suggesting that this aberration of KL-6 expression is a common property of adenocarcinomas. Furthermore, elevated circulating KL-6 levels are frequently observed in patients with NSCLC, pancreatic cancer and breast cancer.21, 22, 29 Although the prognostic value of KL-6 or its physiological function has not yet been clarified, the relationship between MUC1 mucin, and the growth and survival of cancer cells, has been clearly indicated.30 In addition, MUC1 has been reported as a binding partner and a substrate of EGFR,31, 32 and its expression can potentiate EGFR-dependent signal transduction31 and also inhibit the degradation of ligand-activated EGFR.32 On the basis of these observations, we hypothesized that the KL-6 expression may therefore affect the growth and survival of cancer cells and, thus, circulating KL-6 levels may be able to provide information about the clinical outcome of NSCLC patients, particularly treated with EGFR-TKI. To prove this hypothesis, we measured the serum KL-6 levels in advanced NSCLC patients treated with EGFR-TKI prior to, 2 weeks after, and 4weeks after the start of EGFR-TKI treatment, and analyzed these levels along with the clinical outcome of EGFR-TKI.

Material and methods

Patients

Between 2002 and 2006, 74 consecutive advanced NSCLC patients who had previously failed chemotherapy were treated with EGFR-TKI, gefitinib (250 mg/day) at Hiroshima University Hospital (Hiroshima, Japan) and Ehime University Hospital (Ehime, Japan). Patients with locally advanced (stage IIIB), metastasized (stage IV) and postsurgically relapsed NSCLCs who were resistant to one or more regimens of conventional chemotherapy were included in this retrospective study. The disease staging in all 74 cases was supported by a computed tomography (CT) scan of the chest and abdomen, bone scintigraphy and magnetic resonance imaging (MRI) of the head. The inclusion criteria were almost the same as described in our previous reports.33 In brief, they were as follows: (i) age of 20 years or older; (ii) performance status (PS) of 0, 1, 2 or 3; (iii) no significant abnormalities in liver or kidney functions; and (iv) absence of active interstitial lung diseases. EGFR-TKI was orally administered and the treatment was continued until the patient was dropped from the study due to progression of disease, intolerable toxicity or withdrawal of consent. We assessed the objective tumor response as complete response (CR), partial response (PR), stable disease (SD) or progressive disease (PD) in accordance with the Response Evaluation Criteria in Solid Tumors for patients with measurable disease.34 In addition to the serum samples, we obtained formalin-fixed NSCLC tissue samples from 33 patients (24 adenocarcinomas, 6 squamous cell carcinomas and 3 adenosquamous cell carcinomas) and who were treated with EGFR-TKI for their recurrent and chemoresistant diseases. Thirteen samples were obtained by a surgical resection and 20 cases were by lung biopsies from the patients. The characteristics of these 33 patients are shown in Table III. The use of all clinical materials obtained with written informed consent was approved by the Institutional Research Ethics Committees.

Electrochemiluminescence immunoassay to determine the circulating KL-6 level

Serum specimens were obtained within 1 week prior to, 2 weeks after, and 4 weeks after the start of EGFR-TKI administration and stored at −80°C. Serum KL-6 levels were measured by sandwich-type electrochemiluminescence immunoassay (ECLIA) using a Picolumi 8220 Analyzer (Sanko Junyaku, Tokyo, Japan), as previously described.35 In brief, the KL-6 solution was incubated with anti-KL-6 antibody-coated magnetic beads and the beads were then separated using a magnetic rack. Ruthenium-labeled anti-KL-6 antibody was added to the beads as a second antibody following the wash with PBS. The reaction mixture was placed into the electrode, and the photon emitted from the Ruthenium was measured with a photomultiplier. The cut-off level of circulating KL-6 was set at 500 U/ml based on the levels of healthy individuals as reported previously.20

Measurement of the serum carcinoembryonic antigen and cytokeratin 19 fragments (CYFRA 21-1) levels

The serum carcinoembryonic antigen (CEA) level was measured within 1 week prior to the start of EGFR-TKI by using a commercial electrochemiluminescence immunoassay on the ARCHITECT i2000SR system (Abbott Diagnostics, Tokyo, Japan). The serum CYFRA 21-1 level in a limited number (65) of cases was measured by using a commercial electrochemiluminescence immunoassay on the ElecSys 2010 system (Roche Diagnostics Corp Indianapolis, IN). The selected cut-off value for CEA and CYFRA21-1 was 5.8 and 2.8 ng/ml, respectively.

EGFR mutation

EGFR mutations in the region of exons 18 to 21 of EGFR, which has been reported to be a hotspot for mutations (Refs.11 and12; from p-loop to activation loop, codon position 709–870), were assessed in the surgically resected tumor tissues obtained from the 33 NSCLC patients who were eventually treated with EGFR-TKI. We used the peptide nucleic acid-locked nucleic acid polymerase chain reaction (PNA-LNA PCR) clamp test for EGFR mutations that can detect G719S, G719C, L858R, L861Q and 7 different exon 19 deletions.36, 37

Statistical analysis

Statistical analyses were performed using the StatView statistical software program (SaS, Cary, NC) in order to compare patient characteristics with responses to therapy. Associations between clinicopathological variables including serum KL-6 levels, and the response to EGFR-TKI were analyzed by Fisher's exact tests or Mann-Whitney test. In order to test differences among the variables evaluated prior to, 2 weeks after, and 4 weeks after the start of EGFR-TKI, the Wilcoxon test was used. The serum KL-6 levels in the 3 tumor response patterns (PR, SD, and PD) were compared using the Kruskal Wallis test. The progression-free survival (PFS) was defined as the interval from the date of blood sampling to the date of documented disease progression, death from any cause, or the last follow -up. The overall survival (OS) was defined as the interval from the date of blood sampling to the date of death from any cause, or the last follow-up. Survival curve and 95% confidence intervals (CIs) were analyzed by the Kaplan-Meier method, and differences between the 2 groups were compared with the log-rank test. Risk factors associated with the prognosis of the patients were evaluated using Cox's proportional-hazard regression model with a step-down procedure. Only variables shown to be statistically significant in the univariate analysis were evaluated in a multivariate analysis. The criterion for removing a variable was the likelihood ratio statistic, which was based on the maximum partial likelihood estimate (default p value of 0.05 for removal from the model).

Results

Clinical characteristics of the patients

Among the 74 patients recruited in our study, 4 patients (5.4%) developed ILD, which was likely induced by EGFR-TKI treatment, and 1 patient (1.4%) died from ILD. These 4 patients were excluded from the analysis. The clinical characteristics of the 70 patients who fulfilled the selection criteria and were treated with 250 mg of EGFR-TKI, gefitinib per day are summarized in Table IA.

Table IA. Association between the NSCLC Patient Characteristics and the Response to EGFR-TKI Therapy (N = 70)
VariablesTotal (n = 70)PR (n = 25)SD (n = 9)PD (n = 36)p-value (PR vs. SD + PD)p-value (PR + SD vs. PD)
  • ADC, adenocarcinoma; SCC, squamous cell carcinoma; ASC, adenosquamous cell carcinoma.

  • 1p < 0.05 (Fisher's exact test or Mann-Whitney test).

  • *

    Adenocarcinoma versus other histology (squamous cell carcinoma and adenosquamous cell carcinoma).

  • **

    IIIB versus others.

  • ***

    Never versus others.

Age (years)
 <603294190.31710.241
 ≥603816517  
Sex
 Male428529<0.0011<0.0011
 Female281747  
Histologic type
 ADC58227290.517*0.754
 SCC9315  
 ASC3012  
Disease stage
 IIIB81340.244**>0.999**
 IV4517325  
 Recurrence after surgery17737  
Performance status
 <255229240.2260.0191
 ≥2153012  
Smoking history
 Never301848<0.0011***<0.0011***
 Former215313  
 Current192215  
No. of prior chemotherapy regimens
 <245197190.1930.0481
 ≥2256217  
Serum KL-6 level (U/ml)
 Medium 403298648.50.1800.0391
 Range 186–4,201200–9,475155–8,300  
Serum CEA level (ng/ml)
 Medium 24.611.022.60.3940.996
 Range 1.2–736.41.0–408.20.8–1040.4  
Serum CYFRA 21-1 level
 Medium 6.13.55.10.9670.477
 Range 0.9–58.40.9–9.91.5–267.5  

The response to EGFR-TKI was assessable in all patients included in our study. Twenty-five patients were classified as PR and none as CR; 9 patients were classified as SD and 36 as PD. The tumor-response rate (CR + PR/CR + PR + SD + PD) for EGFR-TKI treatment was 35.7%, and the disease control rate (CR + PR + SD/CR + PR + SD + PD) was 48.6%. The median follow-up was 250 days (range, 32–1,138 days). Table IA also shows associations between the clinicopathological factors and responses to EGFR-TKI therapy among the 70 patients. In our study, gender (female versus male; p < 0.001 by Fisher's exact test) and smoking status (never smoker versus smoker; p < 0.001) were significantly associated with the responsiveness to EGFR-TKI. Gender (female versus male; p < 0.001), smoking status (never smoker versus smoker; p < 0.001) and number of prior chemotherapy regimens (<2 versus ≥2; p = 0.0481) were significantly associated with the disease control rate (PR and SD) achieved by EGFR-TKI treatment. Although circulating levels of KL-6 were not significantly higher in responder (PR) patients than nonresponder (SD and PD) patients (p = 0.180, Mann-Whitney test), circulating levels of KL-6 were significantly higher in PD patients than disease-controlled (PR and SD) patients (p = 0.039, Mann-Whitney test).

Association of circulating KL-6 levels at baseline with clinicopathological variables

A previous report from our laboratory revealed that the circulating KL-6 level in the normal subjects was less than 500 U/ml.20 We thus set this value as the cut-off level for circulating KL-6 and, on the basis of this value, we divided the studied patients into those with high (≥500 U/ml, high KL-6 group) and those with normal (<500 U/ml, normal KL-6 group) serum KL-6 levels. The responses to EGFR-TKI therapy in each group are summarized in Table IB. Serum KL-6 levels were found not to be associated with any of the clinicopathological factors in our study.

Table IB. Association between the Characteristics of NSCLC Patients Who were Treated with EGFR-TKI and circulating KL-6 levels (N = 70)
VariablesTotal (n = 70)KL-6 ≥ 500 (n = 35)KL-6 < 500 (n = 35)p-value
  • ADC, adenocarcinoma; SCC, squamous cell carcinoma; ASC, adenosquamous cell carcinoma.

  • 1p < 0.05 (Fisher's exact test or Mann-Whitney test).

  • *

    Adenocarcinoma versus other histology (squamous cell carcinoma and adenosquamouscell carcinoma).

  • **

    IIIB versus others.

  • ***

    Never versus others.

Age
 <60321616>0.999
 ≥60381919 
Sex
 Male4222200.808
 Female281315 
Histologic type
 ADC582929>0.999*
 SCC945 
 ASC321 
Disease stage
 IIIB844>0.999**
 IV452520 
 Recurrence after surgery17611 
Performance Status
 <25526290.561
 ≥21596 
Smoking history
 Never301515>0.999***
 Former21813 
 Current19127 
No. of prior chemotherapy regimens
 <2452322>0.999
 ≥2251213 
Serum CEA level
 Normal1688>0.999
 High542727 
Serum CYFRA21-1 level
 Normal17980.779
 High482226 

Association of circulating KL-6 levels with survival

As shown in Figures 1a and 1b, the OS and PFS in the patients of the high KL-6 group were significantly poorer than those of the patients in the normal KL-6 group (p < 0.001 and p = 0.007, respectively, by log-rank test). The median OS in the high and normal KL-6 groups was 180 and 574 days, respectively. The median PFS in the high and normal KL-6 groups was 47 and 219 days, respectively. On the other hand, the pretreatment serum level of CEA was no found to correlate with OS nor PFS in NSCLC patients treated with EGFR-TKI (p = 0.159 and 0.199, respectively, by the log-rank test; Figs. 1c and 1d). In addition to CEA, we also analyzed serum CYFRA 21-1 levels in a limited number (65) of cases because CYFRA 21-1 was shown to be significantly more sensitive than CEA for predicting clinical outcome in NSCLC patients.38 However, we again found that the pretreatment serum CYFRA 21-1 level did not correlate with PFS nor OS in NSCLC patients treated with EGFR-TKI (data not shown). Next, to determine the prognostic importance of clinical characteristics and serum KL-6 level in NSCLC patients treated with EGFR-TKI, we performed a Cox proportional hazard regression analysis on the parameters listed in Table II. Univariate analyses revealed that a high serum KL-6 level (odds ratio, 3.336; 95% CI, 1.744–6.380; p < 0.001) and performance status (odds ratio, 7.289; 95% CI, 3.463–15.346; p < 0.001) were significant prognostic factors for the OS, and a high serum KL-6 level (odds ratio, 2.081; 95% CI, 1.205–3.594; p = 0.009), sex (odds ratio, 2.661; 95% CI, 1.506–4.704; p = 0.001) and smoking history (odds ratio, 2.059; 95% CI, 1.198–3.538; p = 0.009) were significant prognostic factors for the PFS (Tables IIA and IIB). In addition, a multivariate analysis demonstrated that a high serum KL-6 level (odds ratio, 3.285; 95% CI, 1.714–6.295; p = 0.003) and performance status (odds ratio, 7.163; 95% CI, 3.395–15.113; p < 0.001) were significant independent prognostic factors for the OS in NSCLC patients treated with EGFR-TKI. On the other hand, another multivariate analysis revealed that only a high serum KL-6 level (odds ratio, 2.028; 95% CI, 1.148–3.584; p = 0.015) was an independent prognostic factor for the PFS in such patients.

Figure 1.

The overall survival and progression-free survival of the 70 patients with advanced NSCLC treated with EGFR-TKI in relation to the serum KL-6 (a and b) and CEA (c and d) levels. Differences between the 2 groups were evaluated using the log-rank test.

Table IIA. Cox's Proportional Hazards Model Analysis of the Overall Survival in Patients with NSCLC
VariablesHazards ratio95% CIUnfavorable/favorablep-value
  • ADC, adenocarcinoma.

  • 1

    p < 0.05 (Cox's proportional-hazard regression model with a step-down procedure).

Univariate analysis
 Age (years)1.2910.710–2.35160≥/<600.403
 Sex1.6800.913–3.289Male/Female0.095
 Histology1.6500.812–3.444Others/ADC0.166
 Stage1.0710.476–2.410Others/IIIB0.869
 Performance status7.2893.463–15.3462≥/<2<0.0011
 Smoking history1.5390.844–2.807Smoker/Never0.160
 Serum KL-6 level3.3361.744–6.380High/Normal<0.0011
 Serum CEA level1.8460.824–4.136High/Normal0.403
 Serum CYFRA21-1 level1.9730.912–4.271High / Normal0.084
Multivariate analysis
 Performance status7.1633.395–15.1132≥/<2<0.0011
 Serum KL-6 level3.2851.714–6.295High/Normal0.0031
Table IIB. Cox's Proportional Hazards Model Analysis of the Progression-Free Survival in Patients with NSCLC
VariablesHazards ratio95% CIUnfavorable/favorablep-value
  • ADC, adenocarcinoma.

  • 1

    p < 0.05 (Cox's proportional-hazard regression model with a step-down procedure).

Univariate analysis
 Age (years)1.0690.632–1.80860≥/<600.805
 Sex2.6611.506–4.704Male/Female0.0011
 Histology1.5630.811–3.012Others/ADC0.183
 Stage1.3910.654–2.958Others/IIIB0.391
 Performance status1.8290.965–3.4652≥/<20.064
 Smoking history2.0591.198–3.538Smoker/Never0.0091
 Serum KL-6 level2.0811.205–3.594High/Normal0.0091
 Serum CEA level1.5130.799–2.866High/Normal0.204
 Serum CYFRA21-1 level1.2440.651–2.375High/Normal0.509
Multivariate analysis
 Sex2.0110.959–4.218Male/Female0.065
 Smoking history1.4130.692–2.886Smoker/Never0.342
 Serum KL-6 level2.0281.148–3.584High/Normal0.0151

Impact of circulating KL-6 monitoring on response to EGFR-TKI and survival rate

At 2 and 4 weeks after the start of EGFR-TKI, we measured the serum KL-6 levels and compared them with the pretreatment level in each studied patient. The serum KL-6 levels in the PR patients decreased significantly from 864 ± 1,168 U/ml to 643 ± 755 U/ml (Wilcoxon test; p = 0.008) at 2 weeks and 512 ± 622 U/ml (p < 0.001) at 4 weeks after the start of EGFR-TKI treatment, respectively. On the other hand, the serum KL-6 levels in the PD patients tended to increase from 648 ± 1,412 U/ml to 1,001 ± 1,983 U/ml (p = 0.006) at 2 weeks and 1,244 ± 2,182 U/ml (p = 0.001) at 4 weeks after the start of EGFR-TKI treatment, respectively. Furthermore, the serum KL-6 levels in the SD patients did not change significantly. Next, we calculated the ratios of the serum KL-6 levels obtained 2 and 4 weeks after the start of EGFR-TKI treatment to the pretreatment level, and plotted them according to the 3 groups defined by the treatment outcome, namely PR, SD and PD. As shown in Figures 2a and 2b, the difference in the ratios of serum KL-6 levels at 2 and 4 weeks after the start of EGFR-TKI treatment among the PR, SD and PD groups was found to be statistically significant (Kruskal Wallis test; p < 0.001 and 0.001, respectively). We then determined an optimal cut-off index for the ratio of serum KL-6 level to predict prognosis, using receiver-operating characteristic (ROC) curves according to the death at median follow-up (250 days). The optimal cut-off for the ratio of serum KL-6 level was determined to be 1.2. The OS rate of the patients with higher ratios of serum KL-6 level at 2 and 4 weeks than the cut-off was significantly poorer than those with a lower ratio (p = 0.002 and 0.004, respectively, by log-rank test; Figs. 3a and 3b).

Figure 2.

The ratios of the serum KL-6 levels obtained 2 weeks (a) and 4 weeks (b) after the start of EGFR-TKI to the pretreatment serum KL-6 level. The response to EGFR-TKI was evaluated in the 70 patients with advanced NSCLC. Twenty-five, 9, and 36 patients were categorized into partial response (PR), stable disease (SD), and progressive disease (PD), respectively. Differences among the 3 groups were evaluated with the Kruskal Wallis test.

Figure 3.

The overall survival of the 70 patients with advanced NSCLC treated with EGFR-TKI in comparison to the ratios of the serum KL-6 levels obtained 2 weeks (a) and 4 weeks (b) after the start of EGFR-TKI to the pretreatment serum KL-6 level. After we set the cut-off value for this ratio as described in the Results section, the studied patients were divided into those with high (≥1.2, high KL-6 ratio group) and those with low (<1.2, normal KL-6 ratio group) serum KL-6 ratios. Differences between the 2 groups were evaluated using the log-rank test.

Association of EGFR mutation with the circulating KL-6 levels and survival rate

In 33 of the studied patients, we could carry out nucleic acid-locked nucleic acid polymerase chain reaction clamp-based tests for EGFR mutations, which can detect G719S, G719C, L858R, L861Q and 7 different exon 19 deletions. The detailed characteristics of these patients are listed in Table III. Among these 33 patients, 18 patients had EGFR mutations in their tumors. Although the number of patients included in the study was not sufficient to carry out a valid statistical analysis, no significant association between the serum KL-6 levels and EGFR mutations was found. We also compared the OS and PFS between the patients with and without EGFR mutations. The OS (median, 395 versus 205 days; p = 0.2868) were not significantly longer in the patients with EGFR mutations than in the patients with wild-type EGFR, whereas the PFS (median, 254 versus 39 days; p = 0.0157) were significantly longer.

Table III. Clinicopathological Characteristics and the EGFR Mutation Status of 33 NSCLC Patients
CaseAgeGenderSmoking statusHistologic typeECOG PSKL-6 (Before)KL-6 ratio (2 weeks)KL-6 ratio (4 weeks)EGFR mutationResponse to EGFR-TKIOS (days)
  1. OS, overall survival; ADC, adenocarcinoma; SCC, squamous cell carcinoma; ASC, adenosquamous cell carcinoma; PR, partial response; SD, stable disease; PD, progressive disease, ND, no mutation was detected.

139FemaleNeverADC13580.5640.514DeletionPR>1024
253MaleCurrentASC19,4750.5710.526NDSD168
330MaleCurrentADC33640.6840.937DeletionPD162
464FemaleNeverADC08940.7020.405DeletionPR>225
566FemaleNeverADC13310.7490.716DeletionPR391
648MaleCurrentADC14250.7530.706DeletionPR>946
756FemaleNeverADC07030.7970.535DeletionPR395
862FemaleNeverADC07410.8460.874DeletionPR>194
953MaleFormerADC11,4130.8980.910NDPD460
1064MaleFormerSCC13480.8990.899NDPR791
1167FemaleNeverSCC02650.9250.860DeletionPR338
1257MaleFormerADC12220.9501.432L858RPD86
1350MaleFormerASC01550.9550.994NDPD>458
1452MaleFormerSCC03360.9581.063NDSD240
1559MaleFormerADC11,0920.9630.961L858RPD404
1646MaleCurrentADC02,4360.9681.514L858RPD128
1768MaleNeverSCC35560.9800.716L858RPR278
1858MaleCurrentSCC31,2340.9851.210DeletionPD106
1940MaleCurrentADC33651.0051.118NDPD51
2069FemaleNeverADC13001.0430.823L858RPR288
2148FemaleNeverADC12281.0531.057NDSD1052
2262MaleFormerASC21,8771.0590.910NDPD130
2364FemaleNeverADC18211.0631.557NDPD204
2464MaleFormerADC02541.0980.945L858RPR>447
2563MaleCurrentADC04481.1141.188L858RPD451
2669MaleFormerADC29841.1311.013NDPD48
2769FemaleNeverADC12091.1630.957DeletionPR595
2880MaleNeverSCC21,0591.2231.131NDPD49
2967MaleFormerADC33211.2961.249L858RPR>242
3078MaleFormerADC12791.3301.337NDPD574
3169FemaleNeverADC01,1751.4562.511NDPD540
3254MaleCurrentADC17461.6331.511NDPD149
3350FemaleNeverADC25432.092 NDPD34

Discussion

In the present study, we showed that the pretreatment level of circulating KL-6 is an independent prognostic factor in NSCLC patients treated with EGFR-TKI. Furthermore, we have demonstrated that changes in the circulating KL-6 level at 2 weeks after the start of EGFR-TKI treatment from the baseline can precisely predict the responsiveness and clinical outcome of EGFR-TKI.

EGFR-TKI is a key drug of molecular targeting therapy for treating patients with NSCLC, and the presence of mutations in EGFR has been identified as a powerful predictor of its response to NSCLC.11, 12 However, analyses of EGFR mutations are time-consuming and the methods to determine EGFR mutations have not yet been standardized. Variations in the results between investigators and/or institutions are frequently experienced. Although simple and sensitive methods to detect EGFR mutations, such as the peptide nucleic acid-locked nucleic acid (PNA-LNA) PCR clamp assay and Scorpions Amplified Refractory Mutation System (ARMS)36, 37, 39 have been established, an analysis of mutations still requires tissue samples containing a sufficient number of cancer cells. Highly sensitive PCR methods to detect EGFR mutations in the serum genomic DNA and a mass spectrometry analysis of serum using the predictive algorithm based on matrix-assisted laser desorption ionization have been reported to be new approaches to select the subgroups of NSCLC patients suitable for treatment with EGFR-TKI, however, these methods require special equipment and often lack sufficient reproducibility.39, 40 In the present study, we demonstrate that a high pretreatment level of circulating KL-6 is an independent prognostic factor for the OS and PFS in NSCLC patients treated with EGFR-TKI, while neither CEA nor CYFRA 21-1 are significant prognostic factors. More importantly, the greater impact of monitoring circulating KL-6 levels in NSCLC patients who receive EGFR-TKI is its ability to specifically discriminate PD cases from PR or SD patients, in contrast to the analysis of EGFR mutations which mainly selects patients with PR to EGFR-TKI but has no significant power to discriminate PD or SD patients. Because measuring the circulating KL-6 level seems to be more reproducible, rapid and inexpensive compared with the highly sensitive PCR method and the mass spectrometry analysis, these findings suggest that the assessment of the circulating KL-6 level may therefore be an extremely useful diagnostic modality in clinical practice to predict the clinical outcome of EGFR-TKI treatment in NSCLC patients.

In the present study, the importance of measuring the circulating KL-6 levels in NSCLC patients who receive anticancer therapy other than EGFR-TKI is not determined yet. A previous study from our laboratory has demonstrated that circulating levels of KL-6 did not correlate with the OS in 30 nonresectable NSCLC patients who did not receive EGFR-TKI.41 In our unpublished data analyzing 114 stage IIIB and IV NSCLC patients who were diagnosed at Hiroshima University between 2002 and 2006 and received anticancer therapy other than EGFR-TKI, high levels of pretreatment circulating KL-6 were observed in 45 out of 114 patients but were likely not correlated with the OS in this cohort (p = 0.867, by log-rank test; data not shown). Based on these observations, a high circulating KL-6 level is not a prognostic factor at least in nonresectable NSCLC patients who would not be treated with EGFR-TKI. We thus would be able to conclude that an elevated level of circulating KL-6 is a prognostic factor particularly in NSCLC patients treated with EGFR-TKI. Unfortunately, however, we could not obtain convincing data to explain why it can predict the prognosis in such NSCLS patients in the present study.

A question regarding whether the origin of KL-6 in the sera of NSCLC patients is associated with tumors may be raised. In our unpublished data comparing the circulating KL-6 levels in the preoperative sera with those in the postoperative sera in the NSCLC patients, the KL-6 levels were found to significantly decrease after a surgical resection of the primary tumors. Furthermore, the immunohistochemical staining of these resected tumors using anti-KL-6 antibody has demonstrated the variable levels of KL-6 expression in the primary tumors. These results strongly suggest the origin of KL-6 in the sera of NSCLC patients to be tumor related and the circulating KL-6 level can thus be used as a biomarker for NSCLC (an article regarding this point is now in preparation).

Besides its ability to predict survival and help identify PD cases in NSCLC patients treated with gefitinib, the monitoring of the circulating KL-6 levels may also be useful to detect patients who develop life-threatening ILD. Among the recruited patients in our study, 4 patients developed gefitinib-related ILD. Among these 4 patients, 1 patient died from ILD while the others survived. We recognized that, at the onset of ILD, the serum KL-6 level of the patient who died from ILD had significantly increased from the baseline, whereas those of the survivors did not increase (data not shown). A previous study from our laboratory reported that the serum KL-6 levels increased only in the more severe forms of drug-induced pneumonitis, such as those defined as diffuse alveolar damage (DAD) pattern or chronic interstitial pneumonia (CIP) pattern on high resolution computed tomography (HRCT) images.42 It should be noted that the DAD and CIP patterns of drug-induced pneumonitis are occasionally refractory to treatment and are also related to high morbidity and mortality. Furthermore, the serum KL-6 levels were herein shown to decrease in accordance with the improvement of the diseases in patients with DAD or CIP pattern ILD. Taken together, the monitoring of circulating KL-6 levels in NSCLC patients being treated with gefitinib would contribute to identifying the patients who would not respond to gefitinib treatment and may also predict the occurrence of life-threatening gefitinib-induced ILD. Because there is no way to distinguish whether an elevation of the circulating KL-6 level is due to the growth of the tumor or gefitinib-induced ILD at present, HRCT scans should therefore be performed as soon as possible when this phenomenon is observed during gefitinib treatment.

The significance of MUC1 in the growth, survival, invasiveness and metastatic potential of cancer cells is indisputable,30, 43 and several lines of evidence suggest that the overexpression of MUC1 in tumor is associated with a shortened survival in patients with surgically resected NSCLC.44, 45 In addition, the suppression of MUC1 in lung and breast cancer cells has also been reported to show increased sensitivity to genotoxic anticancer agents in vitro and in vivo.46 Regarding the relationship between MUC1 and EGFR, MUC1 has been reported as a binding partner and a substrate of EGFR, and its expression can potentiate EGFR-dependent signal transduction,31 while also enhancing EGFR stability by inhibiting its down-regulation upon EGFR stimulation.32 In addition to MUC1, recent studies have shown that the expression level of E-cadherin in NSCLC cell lines can determine the sensitivity of the cells to gefitinib and erlotinib,47, 48 suggesting the involvement of E-cadherin in the signal transduction through EGFR. Previous studies from our laboratory have demonstrated that E-cadherin can be functionally suppressed by the overexpression of MUC149 and anti-KL-6 mAb induces capping of MUC1 and facilitate E-cadherin-mediated cell–cell interaction.50 The precise role of circulating KL-6 in the development and progression of NSCLC has not yet been elucidated; however, these observations suggest that KL-6/MUC1 affects EGFR signaling directly by binding with EGFR or indirectly through its interaction with E-cadherin, thus modulating the clinical outcome of EGFR-TKI treatment.

Because of its severe adverse effects in a subset of patients, the clinical use of EGFR-TKI is avoided in several countries. However, EGFR is still an important target molecule in NSCLC and the mission of the therapies targeting EGFR would be carried out by gefitinib and may be succeeded by its analogs, including erlotinib. Therefore, we believe that the monitoring of circulating KL-6 levels will provide valuable information in making the decision to treat NSCLC patients with gefitinib and, possibly, other drugs targeting the tyrosine kinase of EGFR. Although promising results were obtained, we are aware that our study has a number of limitations. First, the number of patients included in the study was not sufficient to carry out a valid statistical analysis. A prospective study monitoring the circulating KL-6 levels in larger number of NSCLC patients receiving EGFR-TKI treatment is needed to confirm its usefulness. Second, the studied patients were only Japanese. Considering ethnic differences in the efficacy of EGFR-TKI treatment and/or the occurrence of adverse side effects induced by EGFR-TKI, we should carefully interpret the results when this monitoring system is applied to non-Japanese patients.

In conclusion, our results indicate that the monitoring of circulating KL-6 levels in NSCLC patients is effective for both selecting patients to be treated with EGFR-TKI and predicting the clinical outcome of EGFR-TKI. In addition, these findings suggest that the circulating KL-6 level could be used as a clinically relevant biomarker in patients with NSCLC, particularly those considered candidates for EGFR-TKI treatment.

Ancillary