Abstract
Background: Tyrosine kinase inhibitors (TKIs) directed against epidermal growth factor receptor (EGFR) are important in the treatment of non-small cell lung cancer (NSCLC), especially those harboring EGFR mutations. But little is known regarding the clinical value of serum tumor marker levels measured prior to treatment. Patients and Methods: We retrospectively reviewed 95 patients with advanced NSCLC treated with EGFR-TKIs, and inspected the relationship between serum tumor marker levels and clinical outcome. Results: Forty-three patients with an elevated serum level of cytokeratin 19 fragment (CYFRA 21-1) had shorter progression-free (PFS) and overall (OS) survival than 52 patients with normal serum CYFRA 21-1 levels (99 vs. 123.5 days p=0.011; and 385 vs. 607 days, respectively, p=0.001). Regardless of EGFR mutation status, patients had shorter progression-free survival when serum CYFRA 21-1 was elevated. Conclusion: Serum CYFRA 21-1 level may be a predictive factor for patients with NSCLC treated with EGFR-TKIs, regardless of EGFR mutation status.
Tyrosine kinase inhibitors (TKIs) directed against epidermal growth factor receptor (EGFR) are very effective against non-small cell lung cancer (NSCLC) harboring EGFR mutations (1-6). Erlotinib, one of the first-generation EGFR-TKIs, is a treatment for NSCLC without EGFR mutation (7). Beyond EGFR mutation however, there are no clear predictive clinical markers for the therapeutic effect of EGFR-TKIs.
Meta-analysis has demonstrated that an elevated serum level of cytokeratin 19 fragment (CYFRA 21-1) may be associated with poor prognosis in patients with NSCLC (8). CYFRA 21-1 is usually used as a tumor marker in lung cancer, especially in squamous cell tumors (9). It has also been reported that an increase in pretreatment serum level of CYFRA 21-1 and carcinoembryonic antigen (CEA) are associated with poor outcome in patients with NSCLC treated with erlotinib (10). There is little information, however, about the predictive value of serum tumor markers for patients with advanced NSCLC treated with EGFR-TKIs. The present study aimed to examine the predictive and prognostic value of pretreatment serum CYFRA 21-1 and CEA levels in patients with advanced NSCLC treated with first- and second-generation TKIs (gefitinib, erlotinib and afatinib).
Patients and Methods
We retrospectively screened cases of advanced NSCLC treated with EGFR-TKI therapy at Nagoya City University Hospital (Japan) between June 2006 and August 2016. All patients had stage IIIB or IV disease. Patients received gefitinib (250 mg/day), erlotinib (150 mg/day), or afatinib (40 mg/day) until detection of progressive disease (PD) or intolerable toxicity. Dose interruption or reduction was permitted in cases of treatment-related toxicity. Our Institutional Ethics Committee approved the protocol of this study (No. 1301), and all medical data were anonymized.
Serum CYFRA 21-1 and CEA levels were measured prior to initial chemotherapy. The serum CEA level was measured by using a commercial electrochemiluminescence immunoassay on the HISCL-5000 system (Sysmex, Hyogo, Japan). Serum CYFRA 21-1 level was measured using a commercial electrochemiluminescence immunoassay on a LUMIPULSE Presto II system (Fujirebio Inc., Tokyo, Japan). Measurements were performed at our hospital, using cut-off values of 3.5 ng/ml for CYFRA 21-1, and 5.0 ng/ml for CEA, which were the upper limit of normal values at our hospital.
Statistical analysis. Progression-free survival (PFS) was defined as the time from the first day of TKI treatment until the date of disease progression, death, or last follow-up. Overall survival (OS) was defined as the time from the first day of TKI treatment until the date of death or last follow-up. PFS and OS were analyzed using the Kaplan–Meier method and compared using the log-rank test, with p<0.05 considered statistically significant. Multivariate analysis using a Cox proportional hazards model was performed to identify associations between clinical characteristics and survival, using a probability of p=0.10 as the threshold for adding a covariable to or removing it from the model, with p<0.05 again considered significant. All statistical analyses were performed with EZR (Saitama Medical center, Jichi Medical University, Saitama, Japan), which is a graphical use interface for R (the R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of R commander designed to add statistical functions frequently used in biostatistics (11).
Results
The clinicopathological characteristics of the 95 patients are shown in Table I. All patients had a performance status of 0 or 1. Fifty-four patients carried an EGFR mutation (29 had exon 19 deletion, 23 had exon 21 L858R, and two had exon 21 L861Q), 34 patients had wild-type EGFR, and in seven patients the EGFR mutation status was not evaluated. Forty-three patients had an elevated serum level of CYFRA 21-1 (>3.5 ng/ml) prior to the start of EGFR-TKI treatment, and 69 patients had an elevated serum level of CEA (>5.0 ng/ml). Both PFS and OS were significantly shorter in the group with an elevated level of CYFRA 21-1 than those with normal serum levels (99 vs. 123.5 days, p=0.011; and 385 vs. 607 days, p=0.001, respectively; Figure 1A and B). In contrast, there was no significant difference in PFS or OS between those who had elevated or normal CEA level (124 vs. 97 days, p=0.757; and 542 vs. 357 days, p=0.059, respectively; Figure 1C and D). In univariate analysis, CYFRA 21-1 level and EGFR mutation were significant factors for PFS, while CYFRA 21-1 level, EGFR mutation, and tumor pathology were significant factors for OS (Table II). In multivariate analysis, CYFRA 21-1 and EGFR mutation were significant factors for both PFS [hazard ratio (HR)=2.17, p<0.001; HR=2.92, p=0.001, respectively; Table III] and OS (HR=2.29, p=0.001; and HR=4.72, p<0.001, respectively; Table IV).
In patients with EGFR mutation, PFS and OS were both found to be significantly shorter in those with elevated CYFRA 21-1 than those with normal levels (168 vs. 244.5 days, p=0.032; and 471 vs. 1023 days, p<0.001, respectively; Figure 2A and B). On the other hand, in patients with wild-type/unknown EGFR status, there was again a significantly shorter PFS but only a trend for shorter OS in those with elevated CYFRA 21-1 compared to those with normal levels (30 vs. 53.5 days, p=0.027; and 106 vs. 252.5 days, p=0.115, respectively; Figure 2C and D). There was no association between CEA and length of PFS or OS in EGFR mutation-positive patients (193 vs. 362.5 days, p=0.297; and 767 vs. 1314 days, p=0.951, respectively), nor in those with wild-type/unknown EGFR status (48 vs. 45.5 days, p=0.66; and 194 vs. 285 days, p=0.827, respectively).
Discussion
In this retrospective study, EGFR-TKI-treated patients with an elevated serum level of CYFRA 21-1 had shorter PFS and OS times than those with normal levels. In contrast, an elevated serum level of CEA was not found to influence the duration of PFS or OS. In multivariate analysis, CYFRA 21-1 and EGFR mutation were significant factors for both PFS and OS. It is known that EGFR-TKIs for patients with advanced NSCLC harboring EGFR mutation improve PFS compared with chemotherapy (1-6). In our subgroup analysis, an elevated level of serum CYFRA 21-1 was associated with shorter PFS and OS in EGFR-mutated patients, and shorter PFS in those with with wild-type/unknown EGFR status. Thus, CYFRA 21-1 was a predictive factor not only in EGFR mutation-positive patients treated with EGFR-TKIs, but also in those with wild-type/unknown EGFR status.
A previous meta-analysis reported that an elevated serum level of CYFRA 21-1 is associated with shorter OS in patients with advanced NSCLC (8), an observation that our findings are consistent with. It has been shown that CYFRA 21-1 is a predictive marker of PFS in patients with NSCLC with EGFR mutation who were treated with EGFR-TKIs (12). An elevated serum level of CYFRA 21-1 has also been associated with poor PFS in patients with NSCLC treated with EGFR-TKIs (10). However in this study, because EGFR status was not considered, it was difficult to determine how much effect CYFRA 21-1 had on PFS and OS in patients with NSCLC with wild-type/unknown EGFR status. Our study showed that elevation of serum CYFRA 21-1 was also a negative predictive factor for PFS in patients with NSCLC with wild-type EGFR or unknown EGFR status. Since elevated CYFRA 21-1 has been associated with the presence of mediastinal lymph node metastases and poor performance status (9), this marker may reflect tumor volume. In contrast, serum CEA had no association with PFS or OS in our study. Not only tumor volume but also tumor characteristics may have an important role in determining patient survival.
CYFRA 21-1 is a fragment of the cytokeratin 19 protein that forms part of the cytoskeleton of epithelial cells (9). Activated proteases in malignant epithelial cells increase cytokeratin degradation, with the consequent release of cytokeratin fragments, particularly fragment 19, into the blood (14). Cytokeratin 19 is thought to be expressed more widely in cases of poorly differentiated squamous cell cancer (15), where its serum levels have been shown to provide significantly greater sensitivity and specificity for the purpose of diagnosis than other histological markers (9). Due to the histological heterogeneity of NSCLC, it is often difficult to distinguish squamous from non-squamous cell cancer using small biopsies or cytological specimens (16). Pemetrexed, which has more efficacy in patients with non-squamous NSCLC, in combination with a platinum-derivative therapy, has been associated with shorter PFS and OS, and worse disease control when the serum level of CYFRA 21-1 is elevated (17). Given this observation, higher levels of CYFRA 21-1 may be indicative of a squamous cell cancer component, explaining such differences in survival with this treatment approach.
It is important to highlight that this study has several limitations. Firstly, this was a single-institute retrospective study, so it is possible that selection bias may have affected our findings. Secondly, the molecular basis for how cytokeratin 19 plays a role in the growth and invasion of squamous cells in lung cancer is not yet understood.
In conclusion, elevated serum CYFRA 21-1 was associated with shorter PFS and OS of patients with NSCLC treated with EGFR-TKI. Importantly however, CYFRA 21-1 was associated with shorter PFS even in patients with wild-type or unknown EGFR status, who were poor responders to this therapy. Thus, the serum level of CYFRA 21-1 may be useful to help determine the appropriate use of EGFR-TKI therapy in all patients with NSCLC, regardless of EGFR mutation status. Further research into the molecular basis for this observation, along with prospective clinical studies, are now warranted to verify these findings.
- Received July 12, 2017.
- Revision received August 7, 2017.
- Accepted August 8, 2017.
- Copyright© 2017, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved