Abstract
Objective We described the patterns of chemotherapy use and outcomes in patients diagnosed with stage III or IV non-small cell lung cancer (NSCLC) in England.
Method In this retrospective population-based study, we included 20,716 (62% stage IV) patients with NSCLC diagnosed in 2014-17 treated with chemotherapy. We used the Systemic Anti-Cancer Treatment (SACT) dataset to describe changes in treatment plan and estimated 30 and 90-day mortality rates and median, 6- and 12-month overall survival (OS) using Kaplan Meier estimator for patients aged <75 and ≥ 75 by stage. Using flexible hazard regression models we assessed the impact of age, stage, treatment intent and performance status on survival.
Results Patients aged ≥ 75 years were less likely to receive 2 or more regimens, more likely to have their treatment modified because of comorbidities and their doses reduced compared to younger patients. However, early mortality rates and overall survival were similar across ages, apart from the oldest patients with stage III disease.
Conclusion This observational study demonstrates that age is associated with treatment choices in an older population with advanced NSCLC in England. Despite reflecting a pre-immunotherapy period, given the median age of NSCLC patients and increasingly older population, these results suggest older patients (>75yrs) may need additional support to access more intense treatments that may influence overall survival.
Introduction
Non-small cell lung cancer (NSCLC) represents about 85% of new lung cancer diagnoses, affecting about 20,500 people in England annually1. Over half of patients with primary NSCLC are diagnosed at an advanced stage1. Net survival is low; in 2010-2014 at 1 year from diagnosis it was estimated at 44% (all stages and ages combined) in England and only 22% at three years.1.
Before the routine use of immunotherapy from 2017, cytotoxic chemotherapy was the main treatment for advanced NSCLC management for patients with good performance status.2 As a result of the underrepresentation of older adults in randomized clinical trials2, physicians do not have clear evidence on the benefit-risk balance in this population. Older adults commonly have poorer health status and fitness, more concomitant conditions, age-related physiological changes, and support needs compared to their younger counterparts3. Information on outcomes after chemotherapy in older patients with advanced NSCLC and the impact of chemotherapy on mortality at population level is scarce.
The Systemic Anti-Cancer Treatment (SACT) dataset includes data on systemic therapy in all patients diagnosed with cancer in England since 2014 4.
Using the SACT dataset, we described the patterns of chemotherapy use, and associated outcomes in patients diagnosed with stage III or IV NSCLC from 2014 to 2017 in England in relation to age.
Methods
In this retrospective observational population-based study, we included all patients diagnosed with stage III and IV NSCLC (ICD-10 codes: C34; morphology codes: see supplemental material) between January 1, 2014, and December 31, 2017, aged 18 or over, recorded in the SACT dataset as having received at least one dose of chemotherapy, regardless of treatment intent. We restricted our analysis to the use of cytotoxic chemotherapy for lung cancer (see supplemental Table 1). In case of multiple NSCLC diagnoses in a patient, we kept the first diagnosis in the analysis. All patients were followed for vital status up to December 31, 2018. All cycles of chemotherapy received after this date were excluded from analysis. We further excluded cycles which were recorded as “trial”, receiving “not chemo”, “not matched” or “missing treatment group”, and duplicate records (Figure 1). We retrieved characteristics of patients (age, sex, ethnicity) and tumours (morphology, stage at diagnosis) as well as vital status and date of death, if applicable, from the National Cancer Registration and Analysis Service (NCRAS) data linked to SACT data. From the SACT dataset, we analysed regimen data that included start date, treatment intent, treatment given, and whether the regimen was adapted based on comorbidities, whether doses were modified or if there were changes in treatment interval. Treatment intent was recoded into three categories: curative (including neo-adjuvant and adjuvant treatment), palliative, disease modification. A cancer clinician (KNF) checked for inconsistencies between chemotherapy regimen and treatment intent. In case of inconsistencies and missing data, intent was recoded based on review of the regimens used in the context of NSCLC management (supplemental Table 2). If a chemotherapy regimen may be used in both palliative and curative intent, we kept the original coding. If the original coding was “disease modification”, we kept as is.
Flow chart
We also analysed the following outcome data4:
Dose reduction = This identifies any regimens modified by reducing the dose of any anti-cancer drug administered at any point in the regimen after it’s commencement.
Time delay = This identifies any regimens modified by extending the time between administration dates at any point in the regimen after its commencement.
Regimen stopped early = This identifies any regimens modified by reducing the administration days below the number originally planned.
For each patient, we kept information regarding the first regimen they received. We also created a variable to indicate whether patients who received a first regimen with curative intent received chemotherapy with palliative intent subsequently.
We also used the start date of cycle of chemotherapy4 and we calculated the number of cycles each patient received.
In our analysis, we kept the performance status recorded at the start of the first chemotherapy cycle and recoded it into 0,1, 2+ and missing. This categorisation was based on NICE guidelines that, at the time, recommended palliative chemotherapy only for patients with a performance score of 0 or 16.
Statistical analysis
All analyses were stratified by stage. Percentages and medians with their interquartile range (IQR) were used to describe categorical and continuous variables (age, number of cycles), respectively.
We estimated mortality rates within 30 and 90 days of the first chemotherapy administration and we used the Kaplan Meier estimator to estimate median, 6- and 12-month overall survival (OS) from the start of the first chemotherapy cycle.
To better describe how age at diagnosis influenced OS, we derived OS from the estimation of individual mortality hazard using a flexible hazard regression model. We modelled the hazard function as the exponential of B-spline of degree 3 with a knot located at the median of the distribution of survival times in patients who died. We included age as a B-spline of degree 3 with a knot at the median of the distribution of age in the whole sample. We first ran models including treatment intent of the first regimen as a factor, and then performance status as a factor.
We performed statistical analyses using R statistical software (version 3.4.0; R Development Core Team, 2017) and we used the R ‘mexhaz’ package to model hazard and estimate survival. 7
Ethics
North of Scotland Research Ethics Service gave ethical approval for this work (REC/19/NS/0057).
Results
Out of 43 505 lung cancer patients diagnosed between 1st January 2014 and 31 December 2017, we retained 7 884 stage III and 12 832 stage IV NSCLC (44% females, respectively). Figure 1 shows the flow chart.
Age at diagnosis was similarly distributed across stages (Figure 2), with comparable median age (68 [IQR: 62-73] in stage III and 67 [IQR: 60-73] years old in stage IV). Overall, 19.5% and 17.8% patients with stage III and IV NSCLC, respectively, were 75 years or older.
Distribution of age at diagnosis based on stage at diagnosis
Table 1 presents characteristics of patients under and over 75 years old by stage at diagnosis. There were fewer females than males in older than younger patients in both stages (∼38% vs 45%). Ethnicity was equally distributed across age groups in both stages (>90% White). Older patients had poorer performance scores than their younger peers regardless of stage. In both stages, older patients were more likely to receive curative treatment than their younger peers (28.6% vs 21.6%, respectively, in stage III patients and 17.1% vs 12.8%, respectively, in stage IV patients). The median number of cycles received was 4 in all groups except for older patients with stage IV disease (median number = 3).
Characteristics of patients diagnosed with non-small cell lung cancer in England in 2014-2017 treated with chemotherapy by stage and age group
Changes in chemotherapy treatment plans by stage at diagnosis and age group
Changes in treatment plan
Older patients were more likely to have their treatment adjusted because of the presence of comorbidities (24% vs 18%) and to have reduced doses of chemotherapy than younger patients (36% vs 27%) at both stages of disease.
30% of stage III NSCLC patients had delayed treatment regardless of age. However, treatment delay was more frequently recorded in older stage IV patients compared to younger patients (32% vs 28%). ∼18-19% patients had their treatment stopped earlier than planned regardless of age and stage.
Older NSCLC patients were less likely to have ≥2 regimens reported than younger ones, regardless of stage (42-44% vs 49%). Among patients who received ≥ 1 regimen, the percentage of those for whom a change in treatment plan was recorded was not noticeably different between age groups or stages. Older patients were more likely to have reduced doses of chemotherapy than younger patients (34% in patients under 75 years old in both stages vs 43% and 40% in older patients with stage III and IV, respectively).
Early mortality and overall survival (Table 3)
30-day and 90-day mortality rates were ∼3% and 8-10%, respectively in stage III patients, and 5-6% and 18-20%, respectively, in stage IV patients with no age-related differences within stage.
30- and 90-day mortality rates, 6- and 12-month and median overall survival from the start of the 1st cycle of chemotherapy in non-small cell lung cancer patients by stage at diagnosis and age group
In respect to overall survival, older patients with stage III disease had a lower 6- and 12-month overall survival by 3.4%-points and 6.6%-points, respectively, and a shortened median survival by 3 months compared to patients below 75. However, we did not observe age-related differences in 6-or 12-month OS (60% and 33%, respectively) or median survival (∼8 months in both age groups) in patients with stage IV NSCLC.
Figure 3 shows overall survival over the initial year after the start of the first chemotherapy cycle by age, stage and treatment intent (Figure 3 A and B) and performance status (Figure 3 C). Regardless of treatment intent, overall survival decreased over the first year and as age increased in stage III NSCLC, but there was no difference in overall survival across ages in stage IV patients. There was also no age-related difference in 6- and 12-month overall survival (Figure 3 B) except in stage III patients receiving palliative treatment in whom survival gradually decreased from the age of 65.
Overall survival (OS) over the initial year of the initiation of chemotherapy at ages 50, 60, 70, and 80; (B) 6, and 12-month OS by age at diagnosis and (C) 6-month OS by age and performance status in patients diagnosed with stage III and IV non-small cell lung cancer in 2014-2017 and treated with chemotherapy (Note. Solid lines represent overall survival estimates and dashed lines represent 95% confidence intervals)
Finally, 6-month overall survival did not differ between patients with a performance score of 0 or 1 but was poorer in those with score ≥2 in both stage III and IV (Figure 3 C). Overall survival estimates in patients with missing performance status were in between those with score of 0/1 and those with score ≥2.
Discussion
This population-based study is the first to describe the use of chemotherapy and associated outcomes in a cohort of NSCLC patients treated with chemotherapy in England in relation to age in the era before the introduction of immunotherapy. We showed chemotherapy treatments were more likely to be modified in older patients than their younger peers, and older patients were less likely to have more than one line of treatment. However, age-related differences in early mortality and overall survival were marginal, notably in stage IV NSCLC, suggesting proper selection of patients for treatment.
There are a few observational studies that describe the use of chemotherapy in NSCLC patients covering more or less the same pre-immunotherapy period. However, most of them were single institution studies and did not stratify by treatment intent or age, making comparisons difficult 8–12. Moreover, some estimated survival from diagnosis precluding comparison with our findings8. A study conducted in Leeds in England reported survival estimates stratified by histology and stage and showed an improvement in survival in stage IIIA and III-B-IV NSCLC, mainly in non-squamous cell carcinoma between 2007-2012 and 2013-201710. However, results were not provided by age categories or by type of systemic therapy given. A US study reported a median survival of 9.7 months from the initiation of anticancer therapy in stage IV NSCLC patients (median age = 67; 45% females) treated with chemotherapy or immunotherapy (∼20%), higher than in our study, but they did not report estimates by age9.
Although there were more frequent modifications of treatment regimen in older patients, they had similar or somewhat better outcomes than younger patients. This finding suggests the selection of patients to receive chemotherapy was appropriate regardless of age. However, data were not available for the population who did not receive chemotherapy so precluding any direction comparison. It would seem likely, however, that older patients who received chemotherapy were fitter than those who did not receive. This hypothesis should be confirmed in future studies.
The main strength of the present study is the inclusion of all patients diagnosed with advanced NSCLC between 2014 and 2017 treated with chemotherapy in England, representing ∼22% of all NSCLC patients diagnosed during the study period (based on an estimate of 82,000 NSCLC patients over the entire period1). Other strengths include good completeness of stage at diagnosis (about 2% missing stages), and long follow-up of patients. Our study has, however, limitations. Data (e.g., cycle number or regimen number) may be missing or incorrectly inputted4 and characteristics such as comorbidity, and geriatric conditions (e.g., frailty) for older patients were incomplete precluding in-depth description of those patients who received chemotherapy.
Conclusion
This national population-based study provides a snapshot of the chemotherapy use and its associated outcomes in terms of early mortality and overall survival in advanced NSCLC patients under and over 75 years old in England. Our results showed the absence of age-related difference in early mortality and survival. Our findings serve as baseline to monitor chemotherapy outcomes in both age groups as new immunotherapy for NSCLC was made available since 2017.
Data Availability
De-personalised study data may be made available on request to accredited researchers who submit a proposal that is approved by the PHE Office for Data Release.
Funding information
SP was supported by the European Union’s Horizon 2020 Research and Innovation Programme, Belgium under the Marie Sk∤odowska-Curie grant agreement No 842817. DD is supported by Cancer Research UK (grant no C8225/A21133).
Funders had no involvement in the present study.
