Clinical Characteristics and Risk of Second Primary Lung Cancer After Cervix Cancer: A Population-Based Study

Background Lung cancer as a second primary malignancy is increasingly common, but the clinical characteristics of second primary non-small cell lung cancer after cervix cancer (CC-NSCLC) in comparison with first primary non-small cell lung cancer (NSCLC1) is unknown. Methods The Surveillance, Epidemiology and EndResults (SEER) cancer registry between 1998 and 2010 was used to conduct a large population-based cohort analysis. Demographic and clinical characteristics as well as prognostic data were systematically analyzed. We further compared overall survival (OS) in the two cohorts. Risk factors of secondary primary lung cancer in cervical cancer patients were also analyzed. Results 557 (3.52%) had developed second primary lung cancer after cervix cancer and 451 were eligible for inclusion in the final analyses. In comparison to NSCLC1, patients with CC-NSCLC had a higher rate of squamous cell carcinoma (SCC) (36.59% vs. 19.07%, p<0.01). Median OS was longer for CC-NSCLC than for NSCLC1 before propensity score matching (PSM) (16 vs. 13months) but there was no significant difference after PSM. High-risk factors in cervical cancer to developing CC-NSCLC include: 50-79years old, black race (OR 1.417; 95%CI 1.095-1.834; p<0.05)and history of radiotherapy (OR 1.392; 95%CI 1.053-1.841; p<0.05). Conclusion 50-79years old, black race and history of radiotherapy were independent risk factors of second primary lung cancer in cervical cancer patient. CC-NSCLC patients had distinctive clinical characteristics and a better prognosis compared with NSCLC1 patients.


Introduction
Cervical cancer is the fourth most common cancer worldwide, with an estimated 310 000 annual deaths globally [1][2][3] . However, improvements in early detection and cancer treatment have led to long survival among cervical cancer patients. Subsequently, the possibility for patients to develop a subsequent primary cancer becomes a more important consideration 4 , with a 17% higher rate of cancer than general population in female. What's more, cervix cancer survivors had more than double the rate of lung In particular, lung cancer as a second primary malignancy is increasingly common. Indeed, lung cancer is the leading cause of cancer incidence and mortality around the world, with 2.1 million new lung cancer cases and 1.8 million deaths predicted in 2018, representing 18.4% cancer deaths 1 . Among women, lung cancer constitutes one of the 3 most commonly diagnosed cancers besides breast and colorectal cancers 2 .
Notably, lung cancer incidence rates are now higher among young women than among young men in non-Hispanic whites and Hispanics Americans 6 . As regards both cervical cancer and lung cancer, approximately 10% of cervical cancer survivors have developed a second malignancy, in which lung cancer accounts for one of the largest numbers 7,8 . However, risk factors of secondary primary lung cancer in cervical cancer patients are not known. Similarly, differences between CC-NSCLC and NSCLC on clinical characteristics and survival have not been studied.
Consequently, there is a crucial need to characterize the lung cancer disease in this specific subgroup as regards both their high incidence rates. Thereby, our study aims to focus on clinical differences between CC-NSCLC and NSCLC1 as well as risk factors for secondary primary lung cancer in patients with cervical cancer.

Ethical Statement
The study was approved by the Research Ethics Committee of Daping Hospital. Data obtained from SEER database did not require informed patient consent because cancer is a reportable disease in the United States.

Population
We identified cervical cancer cases and NSCLC cases from the SEER program of the National Cancer Institute (http://seer.cancer.gov/). The cohort was composed of adult patients who were pathologically confirmed with cervical cancer or NSCLC from the SEER database from 1998 to 2010. This SEER program released a 18 population-based cancer registries of incidence rate and survival rate in United States, covering about 28% of the general population. Exclusion criteria were: confirmed by autopsy, unknown age of diagnosis, unknown marriage status, undetermined grade of disease, unknown stage of disease, unknown pathological type. A total of 173272 NSCLC1 patients, 15809 cervical cancer patients and 451 CC-NSCLC patients were eligible for inclusion in the final analyses. Domestic status was recorded as follows: never married as "unmarried", married as married or unmarried but having domestic partner; separated, divorced and widowed status were classified as "other". Except for squamous cell neoplasm and adenocarcinoma, other histology types were recorded as "other", including NSCLC not otherwise specified (NSCLC-NOS).

Statistical analysis
Categorical measurements were described as count and percentage, while continuous measurements were presented as mean( median) and range. The chi-square was used to compare the categorical measurements while the t test was used for continuous ones. Survival data were measured from the lung cancer date of diagnosis to the date of all-cause death or the last follow-up. Cumulative survival curves were generated by the Kaplan-Meier method. Differences in survival were compared using the log-rank (Mantel-Cox) tests. PSM was used to balance the difference from baseline characteristics between CC-NSCLC and NSCLC1 groups. According to 1 to 3 matches, 449 CC-NSCLC patients were matched successfully. After PSM, there were 449 cases in the CC-NSCLC group and 1347 cases in the NSCLC1 group, and there were no significant differences in histology, age at lung cancer diagnosis, race, year of lung cancer diagnosis, stage of lung cancer, marital status, radiotherapy records, chemotherapy records, surgery records and grade between the two groups. Logistic multiple regression analysis was performed to identify independent risk factors and odds ratios (OR) of second primary lung cancer in cervical cancer patient. All p values were two-sided, with p < 0.05 considered statistically significant. The incidence of second primary lung cancer was compared to the literature. All the analyses were done using SPSS statistical software, version 23 (IBM Corp, Armonk, NY).

Patients Characteristics
A total of 15809 cervical cancer patients and 173272 NSCLC1 patients between 1998 and 2010 were involved. 557 patients (3.52%) were diagnosed with second primary non-small cell lung cancer after cervix cancer (CC-NSCLC) and 451 patients with complete information were eligible for inclusion in the final analyses. The demographic and clinic-pathologic features of NSCLC1 and CC-NSCLC patients are listed in Table 1.
There are significant differences in histology, age at diagnosis, race, year at diagnosis, marital and cause of death between CC-NSCLC and NSCLC1. No significant was detected in stage, radiotherapy records, chemotherapy records, surgery records and grade. The mean time to NSCLC diagnosis was 57 months after cervical cancer, with a range of 12-192 months. The mean age of cervical cancer diagnosis was 58.2 years whereas the mean age at CC-NSCLC diagnosis was 62.9 years. The majority of CC-NSCLC were adenocarcinomas (38.36%) while 36.59% were SCC and 25.1% were other. Of the 173272 NSCLC1 patients in the database, a vast majority was adenocarcinomas (49.05%), and 19.07% of patients were SCC while 31.89% of patients were other. The proportion of SCC in CC-NSCLC patients was apparently higher than that in NSCLC1 patients (36.59% vs. 19.07%). The difference in pathologic type distribution between these two cohorts is significant (p<0.01).

Clinical features in CC-NSCLC patients
The impact of demographic characteristics and clinical features of cervical cancer on pathological types and clinical stages of lung cancer in CC-NSCLC patients are listed in Table 2. Latency, stage, histology, radiotherapy records, chemotherapy records and grade of cervical cancer were associated with pathological types of lung cancer, rather than race, age at cervical cancer diagnosis, year of cervical cancer diagnosis, marital status, surgery records. There was no significant correlation between clinical factors of cervical cancer and stages of lung cancer in CC-NSCLC patients.
The impact of demographic characteristics and clinical features of cervical cancer on the cause of death in CC-NSCLC patients are listed in Table 3. Latency, age at diagnosis, stage, histology, marital status, radiotherapy records, chemotherapy records, surgery records and grade were associated with the causes of death, rather than race and year of cervical cancer diagnosis. Patients with a latency≤1year were more likely to die of cervical cancer, and those with a latency＞5years were more likely to survive. Married patients with young age, regional stage, treated by radiation or chemotherapy, died more often from cervical cancer. Patients with cervical adenocarcinoma, well or moderately differentiated in terms of histological grade and treated by surgery were more likely to survive. Lung cancer was the most common cause of death(44.8%).

Risk factors of secondary primary lung cancer in cervical cancer patients
High-risk factors of developing secondary primary lung cancer in cervical cancer patients include: age between 50 and 79years old, black race and history of radiotherapy. All significant independent factors from logistic multiple regression analysis to develop second primary lung cancer are shown in Table 4.

Survival Analysis
The median OS was 16months (range,

Discussion
Over the past three decades, advances in early detection and treatment of cervix cancer have resulted in significant survival improvement among cervical cancer patients. Such survival after a cervical cancer diagnosis are higher than they have ever been, due to improvements in cancer therapy and current emerging issues concern long-term events in survivors, with notably the occurrence of second cancer 9  decreasing the risk of cancer recurrence and improving survival, but such treatments are also associated to an increased risk of second malignancies after exposure, especially, in long-term smokers [18][19][20][21] . However, no significant high standardized incidence ratios were observed among the radiation group in a large population-based study using SEER data.
The authors explained that a half of the patients were none/unknown status of radiotherapy which could explained such results 17 . In our population-based study using SEER data, significant high incidence ratios were observed in radiation group. With the improvement of radiotherapy technology, the application of Intensity-modulated radiation therapy(IMRT) and 3dimensional conformal radiation therapy(3D-CRT) has become more and more common 22,23 . Our data cannot assess whether IMRT and 3D-CRT will increase the risk of second primary tumor. No significant SIR was detected in surgical treatment group.
Finally, our study has several limitations including notably the lack of biological data including PD-L1 status [24][25][26] . We failed to collect this crucial data as biomarker analysis has not been universalized in clinical practice until very recent years. Consequently, further studies are warranted for targetable driver mutations and PD-L1 in CC-NSCL.
Furthermore, our study did not include the interactions of all possible risk factors on cervical cancer patients. Furthermore, the main limitation of the SEER data, like any retrospective study of treatment effects, is the lack of randomness in treatment regimens, which leads to confounding factors, and result may be biased and interpreted with caution despite the use of PSM to remedy this defect.

Conflicts:
The authors declare no potential conflicts of interest.