CYFRA 21-1 predicts the efficacy of nivolumab in patients with advanced lung adenocarcinoma

Abstract

CYFRA 21-1 is a prognostic marker for non–small cell lung cancer. The serum CYFRA 21-1 level is also known as an adjunct for the diagnosis of lung squamous cell carcinoma. This study aimed to examine whether CYFRA 21-1 has predictive implications for nivolumab therapy in patients with advanced lung adenocarcinoma. Of the 79 patients who were treated with nivolumab therapy at the Shizuoka Cancer Center between December 2015 and September 2016, we retrospectively reviewed the data of 50 patients. The patient characteristics were as follows: age <70/≥70 years: 43 (86%)/7; male/female: 31 (62.0%)/19; Eastern Cooperative Oncology Group performance status 0–1/2: 43 (86%)/7; smoking status: no/yes: 18 (36%)/32; epidermal growth factor receptor mutation status negative/positive: 36 (72%)/14; CYFRA 21-1 ≥2.2/<2.2 ng/mL: 28 (56%)/22; carcinoembryonic antigen ≥5/<5 ng/mL: 29 (58%)/21; and number of prior regimens 2–3/≥4: 16 (32%)/34. With a median follow-up of 263.5 (range, 64–352) days, the median progression-free survival was 70 days. The clinical variables investigated using univariate analysis were as follows: age (p = 0.423), carcinoembryonic antigen (p = 0.888), epidermal growth factor receptor mutation status (p = 0.105), performance status (p = 0.968), sex (p = 0.210), number of prior regimens (p = 0.146), CYFRA 21-1 (p = 0.026), and smoking status (p = 0.041). A multivariate analysis identified a serum CYFRA 21-1 level ≥2.2 ng/mL as an independent predictor of a favorable outcome (hazard ratio, 0.44; 95% confidence interval, 0.23–0.85; p = 0.015; median progression-free survival, 155 vs 51.5 days). In conclusion, CYFRA 21-1 might be an independent predictor of outcome for patients with advanced lung adenocarcinoma treated with nivolumab.

Introduction

Lung cancer remains the leading cause of cancer death globally. Nivolumab, a fully human IgG4 programmed death 1 (PD-1) immune-checkpoint-inhibitor antibody, has been shown to confer a survival benefit over docetaxel for patients with previously treated squamous or non-squamous non–small cell lung cancer (NSCLC).^1,2 In Japan, nivolumab is approved for use in patients with either squamous or non-squamous NSCLC regardless of their PD-L1 expression level.

CYFRA 21-1 is a prognostic marker for NSCLC, ³ and lung squamous cell carcinomas (SCCs) show elevated values compared with other histological types. A low CYFRA 21-1 level has also been reported as a predictor of the response to epidermal growth factor receptor (EGFR)-tyrosine kinase inhibitor (TKI) treatment in EGFR-mutated NSCLC patients ⁴ and is regarded as a potential independent prognostic factor, along with the EGFR mutation status. ⁵ According to recent phase III studies, a squamous histology predicts the efficacy of nivolumab therapy regardless of the PD-L1 expression level, ² while the efficacy of nivolumab therapy for a non-squamous histology was predicted to depend on the PD-L1 expression level. ¹ Some studies have suggested possible predictors, such as the nonsynonymous mutation burden, ⁶ and the presence of CD8+ tumor-infiltrating lymphocytes. ⁷ However, these biomarkers are not entirely reliable, and their investigation is labor-intensive and impractical for daily clinical practice.

The CYFRA 21-1 level reflects an increasing tendency toward differentiation into SCC. ⁵ Lung SCCs have a higher mutation burden than lung adenocarcinomas. ⁸ Therefore, we hypothesized that a high serum CYFRA 21-1 level is correlated with a greater efficacy of nivolumab therapy in lung adenocarcinoma patients. In this study, we investigated the predictive implications of the CYFRA 21-1 level in advanced lung adenocarcinoma patients treated with nivolumab therapy.

Materials and methods

Study design

We retrospectively reviewed data from 79 advanced lung adenocarcinoma patients treated with nivolumab at the Shizuoka Cancer Center between December 2015 and September 2016.

Patients

The inclusion criteria were as follows: histologically confirmed adenocarcinoma; diagnosis of stage IV including postoperative recurrence; Eastern Cooperative Oncology Group performance status (ECOG PS) of 0–2; treated with at least one previous chemotherapy regimen; and adequate organ functions. The pathological diagnoses were performed according to the World Health Organization (WHO) classification criteria. ⁹ The study was conducted with the approval of the Shizuoka Cancer Center Institutional Review Board (28-J150-28-1-2).

Progression-free survival (PFS) was defined as the period from the initiation of nivolumab treatment until tumor progression or death. Tumor progression was assessed using the Response Evaluation Criteria in Solid Tumors (RECIST; version 1.1). ¹⁰ Data for patients who had not progressed or who were still living were censored at the time of the last tumor assessment or the last confirmed contact.

The serum CYFRA 21-1 and serum carcinoembryonic antigen (CEA) concentrations were measured at the time of the diagnosis of stage IV disease, including postoperative recurrences, prior to the initial chemotherapy. The serum CYFRA 21-1 concentration was measured using a Lumipulse Presto kit (FUJIREBIO Inc., Tokyo, Japan), which is based on a CLEIA (chemiluminescent enzyme immunoassay) method, while the serum CEA concentrations were measured using an ARCHITECT kit (Abbott Japan, Tokyo, Japan). For the serum CYFRA 21-1 level as measured using an enzyme immunoassay (EIA), the reported upper limit of normal for the diagnosis of NSCLC and the upper limit of the percentiles for healthy individuals are 3.5 and 2.8 ng/mL, respectively. ¹¹ In contrast, for serum CYFRA 21-1 measurements using the CLEIA method, the reported upper limit of the percentiles for healthy individuals is 1.6 ng/mL, ¹² which is lower than that set for measurements using the EIA method. Therefore, for our study, we set the cutoff value for CYFRA 21-1 at 2.2 ng/mL based on the mean value for healthy subjects + 3 standard deviation (SD) according to the protocol used to set the cutoff value in previous studies,^13,14 which was lower than that set for measurements using the EIA method. The cutoff value for serum CEA was set at 5.0 ng/mL, which is the upper limit of normal. ⁵ Adverse events were evaluated according to the Common Terminology Criteria for Adverse Events, v4.0.

Statistical analysis

Categorical variables were compared using the Fisher exact test. Survival was estimated using the Kaplan–Meier method. A log-rank test was performed to evaluate the significance of the differences in the survival periods among the groups. The univariate analysis was performed using a log-rank test to evaluate the significance of the differences in the survival periods among the groups. All the variables analyzed in the univariate analysis were included in a multivariate analysis using the Cox proportional hazard model (forward stepwise method) to identify any associations between the clinical variables and survival. The statistical analysis was conducted using EZR, version 1.32. ¹⁵ A p value of <0.05 was considered statistically significant.

Results

Patient selection and characteristics

A flow diagram of the patients included in the analysis is shown in Figure 1. The clinicopathological characteristics of the patients are presented in Table 1: median age (range), 64.5 (39–76) years; male/female, 31/19; PS 0–1/2, 43/7 (one patient with PS2 received nivolumab as a third-line treatment, five patients received nivolumab as a fourth-line treatment, and one patient received nivolumab as a seventh-line treatment); never-smokers/smokers, 18/32; median CEA (range), 7.8 (0.5–2148.2) ng/mL; median CYFRA 21-1 (range), 2.2 (1–25) ng/mL; EGFR mutation negative/positive, 36/14 (8 patients with an exon 19 deletion and 6 patients with L858R in exon 21); anaplastic lymphoma kinase (ALK) fusion gene negative/positive, 49/1; and treatment line 2/3/4 or more, 8/8/34. A total of 34 patients received nivolumab as a fourth or later line of therapy (eight patients received nivolumab as a fifth-line treatment, three patients received nivolumab as a sixth-line treatment, three patients received nivolumab as a seventh-line treatment, two patients received nivolumab as an eighth-line treatment, and one patient received nivolumab as a ninth-line treatment).

Figure 1.

Flow diagram of the patients included in the analysis.

Table 1.

Patients’ characteristics.

Characteristics		%
Gender
Male	31	62
Female	19	38
Median age (range), years	64.5 (39–76)
Performance status (ECOG)
0–1	43	86
2	7	14
Smoking status
Past/current smoker	32	64
Never smoker	18	36
EGFR mutation
Positive	14	28
Negative	36	72
ALK fusion gene
Positive	1	2
Negative	49	98
Median CYFRA21-1 level (range), ng/mL	2.2 (1–25)
Median CEA level (range), ng/mL	7.8 (0.5–2148.2)
Treatment line
2	8	16
3	8	16
≥4	34	68

ECOG: Eastern Cooperative Oncology Group; EGFR: epidermal growth factor receptor; ALK: anaplastic lymphoma kinase; CEA: carcinoembryonic antigen.

Treatment outcome and adverse effects

A median of four cycles (range, 1–23) of nivolumab were administered. The median follow-up period for all the patients was 263.5 (range, 64–352) days. The median PFS was 70 (range, 56–155) days. The overall response rate (ORS) was 14.0% (seven patients). The disease control rate was 48.0% (24 patients). In total, 37 of the 50 patients (74%) exhibited disease progression. Of the 13 patients who were censored, 8 patients had stable disease and 5 patients had a partial response. The clinical variables identified using univariate analysis as being associated with a significantly prolonged PFS included a serum CYFRA 21-1 level ≥2.2 ng/mL (hazard ratio (HR), 0.47; 95% confidence interval (CI), 0.25–0.91; p = 0.026) and a never smoking status (HR, 1.99; 95% CI, 1.03–3.86; p = 0.042). A multivariate analysis identified a serum CYFRA 21-1 level ≥2.2 ng/mL as an independent predictor of a favorable prognosis (HR, 0.44; 95% CI, 0.23–0.85; p = 0.015). The median PFS in the group with a CYFRA 21-1 level ≥2.2 ng/mL was 155 days (95% CI, 65–275), while that in the group with a CYFRA 21-1 level <2.2 ng/mL was 51.5 days (95% CI, 36–70), respectively (Table 2, Figure 2).

Table 2.

Univariate and multivariate analyses for progression survival.

Variables	No.	Univariate analysis			Multivariate analysis
		HR	95% CI	p	HR	95% CI	p
Age (years)
<70	43	0.70	0.29–1.69	0.423
≥70	7
Sex
Male	31	0.66	0.34–1.27	0.210
Female	19
PS
0/1	43	0.98	0.34–2.80	0.968
2	7
Smoking status
Never	18	1.99	1.03–3.86	0.042
Former/current	32
EGFR mutation
Negative	36	0.55	0.27–1.13	0.105
Positive	14
Treatment line
2/3	16	0.58	0.28–1.21	0.146	0.52	0.25–1.09	0.0847
≥4	34
CEA (ng/mL)
<5	21	1.05	0.54–2.03	0.888
≥5	29
CYFRA (ng/mL)
<2.2	22
≥2.2	28	0.47	0.25–0.91	0.026	0.44	0.23–0.85	0.0147

HR: hazard ratio; CI: confidence interval; PS: performance status; EGFR: epidermal growth factor receptor; CEA: carcinoembryonic antigen.

Figure 2.

Kaplan–Meier curves for progression-free survival (PFS) stratified according to the CYFRA 21-1 level.

The frequencies of the immunological adverse effects are presented in Table 3. Some type of immunological adverse effect (any grade) occurred in a total of 20 patients; 9 of these 20 patients had an elevated serum CYFRA 21-1 level (≥2.2 ng/mL). A significant difference in immunological adverse effects was not observed between the group with a CYFRA 21-1 level ≥2.2 ng/mL and the group with a CYFRA 21-1 level <2.2 ng/mL (p = 0.25, Fisher exact test).

Table 3.

The frequencies of immunological adverse effects.

Adverse event	Any grade	Grade 1	Grade 2	Grade 3
	Number (%)
Diarrhea	5 (10)	4 (8)	1 (2)
Hypothyroidism	2 a (4)		2 (4)
Hyperthyroidism	4 (8)	3 (6)	1 (2)
Adrenal insufficiency	1 (2)		1 (2)
Hyperglycemia	2 a (4)			2 (4)
AST/ALT increased	5 a (10)	5 (10)
Skin rash	3 a (6)	2 (4)	1 (2)

AST: aspartate aminotransferase; ALT: alanine aminotransferase.

a

There were some overlapping.

Discussion

Identifying biomarkers that can predict the efficacy of a given therapy, enabling the selection of appropriate candidates for nivolumab therapy, is a crucial topic of ongoing research. In this study, we demonstrated that the serum CYFRA 21-1 level was an independent predictor of the efficacy of nivolumab therapy in patients with advanced lung adenocarcinoma. A CYFRA 21-1 level lower than the cutoff value predicts a favorable outcome among patients with EGFR-mutated NSCLC who are treated with an EGFR-TKI and among patients with lung adenocarcinoma.^4,5,16 In this study, however, a CYFRA 21-1 level higher than the cutoff value has been reported to predict a favorable outcome among patients with lung adenocarcinoma who were treated with nivolumab. In this study, the median PFS for patients with a serum CYFRA 21-1 level of ≥2.2 ng/mL was 5.2 months. This result was similar to the median PFS of patients in a recent phase III trial who exhibited >1%, >5%, and >10% PD-L1 expression positivity. ¹ An increase in the serum CYFRA 21-1 level may also represent a greater degree of degradation and the release of intracellular cytokeratin 19 (CK19) into the serum, with an increasing tendency toward differentiation into SCC. Thus, a relationship may exist between the serum CYFRA 21-1 levels and the degree of tumor differentiation toward squamous epithelium, ⁵ which may be associated with the smoking status and a higher nonsynonymous mutation burden. However, this study showed that the smoking status was not an independent predictor of a response to nivolumab.

A higher nonsynonymous mutation burden in tumors has been reported to be associated with the clinical efficacy of anti PD-1 therapy. ⁶ The prevalence of somatic mutations is highly variable among and within cancer classes. Cancers related to chronic mutagenic exposure exhibited the highest prevalence of somatic mutations. For example, lung cancer is related to tobacco smoking, and malignant melanoma is related to exposure to ultraviolet light. The frequency of somatic mutation in SCC is higher than that in adenocarcinoma.^8,17 The average mutation frequency in smokers is more than 10-fold higher than that in never-smokers. ¹⁸ In the present trial, an association between smoking and a significantly better PFS was also identified in a univariate analysis. A subgroup analysis of a phase III trial showed a survival benefit for current or former smokers among patients treated with nivolumab. ¹ Meanwhile, the EGFR mutation status was not a predictor of the efficacy of nivolumab therapy in the present trial. Also, the EGFR mutation status was not significantly correlated with the different PFS outcomes between two treatment arms in a subgroup analysis of a phase III trial. ¹ A previous study reported that patients with murine double minute 2 (MDM2) family amplifications appear to be at risk for accelerated progression after immunotherapy. ¹⁹ In this study, 41 of the 50 patients were diagnosed based on tissue specimens obtained from tumor biopsies. Because these were small specimens, they were inadequate for whole exome sequencing (WES). Recently, we reported clinical factors associated with the mutation burden using WES and surgically resected specimens of NSCLC. ²⁰ In this same reported cohort, we plan to analyze the molecular mechanism and profile using WES to determine the underlying relationship between the CYFRA 21-1 value and overall survival in patients with surgically resected lung adenocarcinoma

This study had several limitations. First, the study was performed retrospectively with small sample size, and the selection of patients with diverse characteristics resulted in an inherent selection bias in our analysis. Seven patients with PS2 received nivolumab therapy. These patients received nivolumab therapy after at least one regimen, as recommended by the guidelines of the Japan Lung Cancer Society. Furthermore, nivolumab is a recommended regimen for adenocarcinoma patients with a PS of 0–2 who have shown progression after initial therapy according to the National Comprehensive Cancer Network Clinical Practice Guidelines, Version 4, 2017. The diversity in patient characteristics in the present trial might explain why the ORR was lower than the results of previous reports. Second, the serum CYFRA 21-1 levels were not measured immediately before the initiation of nivolumab, but at the time of the diagnosis of stage IV disease. The CYFRA 21-1 level was evaluated in two patients at the end of nivolumab therapy. The CYFRA 21-1 level increased from 2.8 to 11.0 ng/mL in one patient and from 1.0 to 7.2 ng/mL in the other patient. The PFS periods for each patient were 166 and 21 days, respectively. Serial CYFRA 21-1 monitoring needs to be addressed in a future study. Third, tumor markers related to adenocarcinoma or SCC, such as carbohydrate antigen 125 (CA125) or SCC antigen, were not measured in our daily clinical practice. Consequently, we only considered two markers, CEA and CYFRA 21-1, in this study. Other related markers need to be evaluated and compared with CYFRA 21-1 in a future study. Fourth, we did not measure PD-L1 expression. The role of PD-L1 expression as a predictive marker is controversial in the broader oncology community, since many PD-L1 positive tumors do not respond and some PD-L1-negative tumors do respond. ²¹ Furthermore, dynamic changes in PD-L1 expression can reportedly be problematic for evaluations performed at a single time point. ²² In addition, PD-L1 expression was not yet measured during the administration of nivolumab therapy in daily clinical practice during the study period. The potential application of serum CYFRA 21-1 levels might include their use as an indicator for the appropriate use of nivolumab in lung adenocarcinoma patients who exhibit a negative or unmeasurable level of PD-L1 expression.

Conclusion

The serum CYFRA 21-1 level might be an independent predictor of the efficacy of nivolumab therapy among patients with advanced lung adenocarcinoma. A serum CYFRA 21-1 level ≥2.2 ng/mL was regarded as a predictor of the efficacy of nivolumab therapy in lung adenocarcinoma patients. In daily clinical practice, evaluating the serum CYFRA 21-1 level (≥2.2 ng/mL) might be clinically relevant in adenocarcinoma patients treated with nivolumab therapy. This issue should be evaluated and validated in a prospective study.