Relationship between serum tumor markers and Anaplastic Lymphoma Kinase mutations in stage IV lung adenocarcinoma in Hubei province, Central China

Abstract Objective The aim of this study was to explore the predictive value of carcinoembryonic antigen (CEA), squamous cell carcinoma antigen (SCCAg), and neuron‐specific enolase (NSE) in the prediction of anaplastic lymphoma kinase (ALK) mutations in advance stage non‐small cell lung cancer (NSCLC). Subjects and Methods A total of 482 cases with untreated lung adenocarcinoma were retrospectively reviewed. Finally, 72 patients with stage IV were enrolled because of intact data of the detection of ALK rearrangement and serum tumor markers, as well they have not received any previous anticancer therapy. We used the one‐way ANOVA analysis, correlation analysis, and multiple logistic regression analysis to evaluate the relationship between the level of serum tumor markers and ALK mutations. Results Fifteen cases with ALK mutations and 57 cases without mutations were identified. The result of the one‐way ANOVA analysis showed only CEA was significantly associated with ALK mutations (95% CI:39.05‐148.88; P = .001). The area under the ROC curve (AUC) of CEA was 0.705 (95%CI:0.567‐0.843; P = .015). However, no significant association was observed between CEA and ALK mutations though the result of correlation analysis (P = .069) and multivariate logistic regression analysis (OR = 0.988, 95% CI: 0.972‐1.003, P = .111). Conclusions In our study, we performed on the patients with stage IV lung adenocarcinoma in our region and found preoperative serum levels of SCCAg, CYRF21‐1, and NSE not suitable for the detection of ALK mutation. Although we observed a significant association between CEA and ALK mutations; however, it was not strong enough to distinguish ALK status for the patients in our region.


| INTRODUC TI ON
Anaplastic Lymphoma Kinase (ALK) rearrangements are responsible for approximately 3%-5% of non-small cell lung cancer (NSCLC). 1 Patients harboring ALK gene mutations benefit from ALK tyrosine kinase inhibitors (TKIs) treatment, which is a research hotspot concerning about targeted therapy for lung cancer in recent years. 2 Currently, FDA-approved ALK-TKIs include the following: (a) crizotinib, the first generation of ALK-TKI; (b) the second generation of ALK-TKI, including ceritinib, alectinib, and brigatinib; and (c) the third generation of ALK-TKI, such as Lorlatinib. 3 These developments have prolonged the survival time of patients with advanced NSCLC, which led to the detection of ALK rearrangement as the standard of care for advanced NSCLC patients.
At present, there are three detection methods for ALK rearrangement, including real-time polymerase chain reaction (RT-PCR), immunohistochemistry (IHC), and fluorescence in situ hybridization(FISH). 4 However, lacking of suitable sample for genetic testing in some cases restricts the detection of ALK rearrangement, which would limit the reasonable application of ALK-TKI. Therefore, to analyze the characteristic findings of advanced NSCLC patients with ALK positive will help to distinguish ALK mutations.
Our previous work was to investigate the predictive power of 18F-FDG PET/CT for predicting positive ALK expression, which demonstrated that ALK-positive patients tended to have a high the maximal standard uptake value (SUVmax) of lymph node. 5 However, PET/CT examination is somewhat expensive and has still not become prevalent. Therefore, looking for an economical and convenient method to distinguish ALK mutations is promising. Blood examination as routine examination is the most widely used approach and its price is not high. So, we suggest that whether or not blood examination could be a valuable method for predicting the ALK rearrangement in NSCLC.
Carcinoembryonic antigen (CEA), squamous cell carcinoma antigen (SCCAg), CYFRA21-1, and neuron-specific enolase (NSE) served as the serum tumor markers have been widely used for the routine examination of NSCLC patients. 6 In this study, we investigated the predictive value of tumor markers (CEA, SCCAg, CYFRA21-1, and NSE) in the prediction of ALK rearrangements. There was a little similar study, which we would mention in discussion.

| Study design
We retrospectively reviewed 482 cases with untreated lung ad-

| ALK mutation analysis/
Mutational analysis of ALK was used Ventana IHC assay that described in our previous work. 5 In brief, 4 μm-thick formalin-fixed,

| Statistical analyses
The data of normal distribution were analyzed by independent samples t test, while the data of questionably normal distribution were analyzed by Mann-Whiney U test. Point-biserial correlation coefficient was calculated to test the relativity of serum tumor markers and the ALK gene mutation. The value of CEA in the prediction of mutation in the ALK oncogene was evaluated with receiver operating characteristic curve (ROC) curve. Multiple logistic regression analysis was used to assess the predictive power of preoperative serum tumor markers levels and missing data were supplemented by the method of the Expectation-Maximization (EM) Algorithm. All analyses were performed by SPSS, v.22.0 (IBM), and P < .05 was considered statistically significant.

| RE SULTS
A total of 482 cases with untreated lung adenocarcinoma were retrospectively reviewed.  We then used correlation analysis and multivariate logistic regression analysis to further evaluate the relationship between serum tumor markers and ALK mutations. As described above, all patients in the study had detected the level of CEA without missing value.

| D ISCUSS I ON
ALK gene rearrangements often mean that patients may be benefit from ALK-TKIs treatment, which is glad tidings, especially for F I G U R E 1 Selection process for patients in stage IV untreated lung adenocarcinoma with the detection of ALK mutations. Abbreviations: ALK, anaplastic lymphoma kinase; NSCLC, nonsmall cell lung cancer those with stage IV lung adenocarcinoma. So, the detection of ALK rearrangements play an important role in the targeted therapy.
Considering the limitation of the obtainment of specimen, a lot of research were conducted to search a new method or index for predicting ALK mutations. Our previous work was to investigate the predictive power of parameters of PET/CT for the prediction of ALK mutations and found that a high-level of SUVmax of lymph node related to ALK mutations significantly. 5 However, the limitation of PET/ CT includes low popularity and its high price. Therefore, we proposed research hypothesis for serum tumor markers, which is economical and convenient for every patient in the prediction of ALK mutations.
We have investigated the relationship between serum tumor markers (CEA, SCCAg, CYFRA21-1, and NSE) and ALK mutations in the patients in our region. No significant correlation was observed between serum tumor markers and ALK mutations except CEA (P = .015, AUC: 0.705); however, its predictive value was not a very strong.
CEA was one of clinical commonly used tumor markers and has been regarded as an assistant diagnostic index for NSCLC, especially for adenocarcinoma. 7 Therefore, CEA has been studied to help predict the clinical features of the patients with NSCLC, including diagnosis and prognosis. 8,9 The study showed by Wen-  F I G U R E 2 ROC curve for the predictive powers of CEA in the prediction of ALK mutations. Abbreviations: CEA, carcinoembryonic antigen; ROC, receiver operating characteristic curve associated with ALK mutations in patients with lung adenocarcinoma. 10 The similar result also observed in the study of Zeng Wang et al, in which they identified 20 patients with ALK mutations from 422 cases with lung adenocarcinoma in the stage I ~ III and found that CEA could predict independently ALK mutations. 11 Recently, Jie Liu et al performed on 701 patients with advanced lung adenocarcinoma(stage Ⅲb and Ⅳ) and observed that CEA (OR = 0.481, 95% CI: 0.275-0.842, P = .010) were independent predictors of ALK mutations. 12 Similarly, in our current study, a significant difference between CEA and ALK mutations was overserved (95% CI:39.05-148.88; P = .001). However, the predictive power of CEA was not a very strong, although its AUC was 0.705 (95% CI: 0.567-0.843, P = .015), P value of correlation analysis and multivariate logistic regression analysis was .111 and .069, respectively. The reasons for the different results may be due to different research objects. In our study, we mainly study patients in the stage IV, whose lifesaving straw usually was targeted therapy. However, we shouldn't neglect the limitation of our study. First, although we mainly studied the population of our region, our study is a single institutional study and the sample size was relatively small due to our strict inclusion criteria. Second, we were unable to control all factors which were associated with the level of tumor makers or ALK mutations, which may be due to introduced bias of retrospective design.
In summary, we studied on 72 patients in stage IV lung adenocarcinoma in Hubei province and investigated the predictive value of serum tumor markers in the prediction of ALK mutations, which we cannot predict through the level of tumor markers (SCCAg, CYRF21-1, and NSE). Only the level of CEA was significantly associated with ALK mutations; however, It was not strong, which needs a large number of samples to extend these findings.

ACK N OWLED G M ENTS
Qi Tan wants to thank the kindness, patience, and care from Dr Bingxin Guo in the past years. You like flowers, love you love to have no place to hide.