Prospective multicentre analysis of the therapeutic approach and prognostic factors determining overall survival in elderly patients with non-small-cell lung carcinoma treated with curative intent.

Objective: To analyse patterns of treatment with curative intent commonly used in elderly patients with locally advanced non-small-cell lung carcinoma (NSCLC) and predictive factors of overall survival in routine clinical practice.
Methods: This multicentre prospective study included consecutive patients aged ≥65 years old diagnosed with NSCLC between February 2014 and January 2018. Inclusion criteria: age ≥65 years, stage IIIA/IIIB NSCLC. Treatment decisions were taken by a multidisciplinary committee. Kaplan-Meier curves and log-rank test were used to identify which clinical/treatment-associated variables, or pre-treatment quality of life (QOL) considering EORTC QLQ-C30 (and LC13 module) were predictive of overall survival.
Results: A total of 139 patients were recruited. Median follow-up was 9.9 months (1.18-57.36 months) with a median survival of 14 months (range 11-17 months). In the group>75-year-old patients, the committee recommended chemotherapy and sequential radiotherapy (55.6%) or radiotherapy alone (22.2%), rather than surgery (3.7%) or concomitant radiochemotherapy (16.5%). However, in 65- to 75-year-old patients, surgery and concomitant radiochemotherapy were recommended in half of cases (p=0.003). Regarding multivariate analysis, the risk of death was higher in patients with pre-existing heart disease (p=0.002), low score for physical functioning (p=0.0001), symptoms of dysphagia (p=0,01), chest pain (p=0.001), and those not undergoing surgical treatment (p=0.024). Conclusions: Patients >75 years received more conservative treatments. Surgery improved survival and should be carefully considered, regardless of patient age. Comorbidities and poor baseline QOL are predictive of shorter survival. Advances in knowledge: Measuring these parameters before treatment may help us to define a population of frail patients with a poorer prognosis to facilitate decision making in clinical practice.

Introduction

More than two-thirds of lung cancer cases currently diagnosed are in people over 65 years of age. Indeed, the mean age at diagnosis is 71 years old,[1] most patients being frail patients with comorbidities that may limit their prognosis and tolerance of treatment.[2-4] Therefore, it is becoming increasingly important to establish which management approach is most effective in elderly patients with locally advanced lung cancer. A meta-analysis[5] demonstrated the superiority of concomitant chemotherapy and radiotherapy over sequential chemotherapy and radiotherapy in patients with unresectable stage III lung cancer, with 2- and 5 year survival rates of 36% and 15%, respectively, with concomitant treatment and 30% and 11% with sequential treatment. Nonetheless, there are other treatment options for patients with a poorer general condition, including sequential chemotherapy and radiotherapy[6] or radiation therapy alone.[7-9]

Several studies indicated that the proportion of patients who receive active treatment for lung cancer decreases with advancing age.[10,11] Furthermore, in clinical trials, the evidence for the use of different treatment regimens is generally gathered from fit younger patients. Notably, in the meta-analysis of Auperin et al,[5] most patients included had a good performance status (0–1) and <20% were aged ≥70 years. It is therefore difficult to extrapolate the findings to all patients with non-small-cell lung cancer (NSCLC) who are elderly and have comorbidities.[12-14] Although there have been small studies in elderly patients,[5,15] there is no solid evidence regarding tolerance or the importance of patient clinical characteristics to guide us in deciding which is the best treatment option in this population.

Due to ageing is a vague concept, several tools have been designed to predict toxicity and identify which patients would be good candidates to undergo radical treatment or adapted therapy[16]

In this study, we sought to assess patterns of commonly used treatment modalities with curative intent in elderly patients with locally advanced NSCLC and clinical factors predictive of overall survival in the context of daily clinical practice. By combining these clinical findings, we would be able to identify the best treatment for each patient.

Methods and materials

Patient population

This multicentre prospective observational study included all consecutive patients aged≥65 years old diagnosed with NSCLC between February 2014 and January 2018. The study was approved by the Ethics Committees of the participating hospitals (xx) and was conducted in accordance with the principles of the Declaration of Helsinki. All patients who participated in the study gave written informed consent prior to inclusion.

Inclusion criteria: age≥65 years, a histological diagnosis of NSCLC, locally advanced disease (stages IIIA or IIIB according to the seventh edition of the American Joint Committee on Cancer Staging TNM classification),[17] receiving radiotherapy with radical intent, with a total prescribed dose of ≥50 Gy (undergoing previous surgery) or ≥60 Gy without a history of surgery, and with or without chemotherapy (concomitant/sequential).

Exclusion criteria: previous radiotherapy, recurrence, or previous history of cancer.

Assessment, treatment and follow-up of patients

Patients were assessed at their first visit through obtaining a clinical history and performing a physical examination. All treatment decisions for these patients were taken by a multidisciplinary committee in each of the participating centres. The treatment options planned were classified as follows: surgery and postoperative radiotherapy (sequentially after postoperative chemotherapy), concomitant radiochemotherapy, sequential chemotherapy and radiotherapy, or radiation therapy alone. Data were collected on the following patient characteristics, categorised as indicated (in parentheses): age (65–75 years vs >75 years old); the Karnofsky Performance Scale (KPS) (<70 vs≥70); smoking habit (into three categories,[18] smoker, ex-smoker and non-smoker; and also into three categories by smoking history[19] :≤30 pack-years vs 31–75 pack-years, vs >75 pack-years); baseline haemoglobin levels ([20] (<12 vs≥12 g dl−1); pretreatment weight loss (yes vs no); alcohol abuse (yes vs no); chronic obstructive pulmonary disease (yes vs no)[21] ; pretreatment thromboembolic event (yes vs no), heart disease (yes vs no), diabetes mellitus (yes vs no), type of treatment received (surgery vs concomitant treatment vs sequential treatment vs radiotherapy alone), stage (IIIa vs IIIb), radiation dose (≤60 Gy vs >60 Gy), radiotherapy technique (3D conformal radiotherapy vs volumetric-modulated arc therapy [VMAT]/intensity modulated radiotherapy [IMRT]).

As it has been suggested that pretreatment quality of life (QOL) has prognostic value,[22,23] in this study, QOL questionnaire (consisting of the EORTC QOL-C30 and lung cancer module QLQ-LC13) was administered to all patients at baseline. The aim was to assess the effect on survival of patients’ subjective assessment of their own baseline status before treatment.[24] The EORTC QOL-C30 evaluates QOL in relation to physical, emotional and social factors, considering general level of functioning in oncology patients. The questionnaire is divided into five functional scales (physical functioning, activities of daily living, emotional functioning, cognitive functioning and social functioning), three symptom scales (fatigue, pain, and nausea and vomiting), one global health status domain, and finally six independent items (dyspnoea, insomnia, anorexia, constipation, diarrhoea and economic impact).[25]

The QLQ-LC-13 includes[26] measures of the symptoms associated with lung cancer (cough, haemoptysis, dyspnoea and pain) and the adverse effects of conventional chemotherapy and radiotherapy (hair loss, neuropathy, sore mouth and dysphagia).

High scores in the symptom scales indicate the presence of symptoms associated with cancer that negatively affect the quality of life. On the other hand, High scores on the global health and functional status scales indicate a better QOL.

For this study, we categorised each of the functional and symptom scores from the questionnaires (EORTC QOL-c30 and module LC13) by the pretreatment score (0–100) into the following categories[27] :≤33.3 vs 33.3-66.6 vs>66.6 points.

Treatment

Regarding radiotherapy, immobilisation and treatment planning were performed with patients in the supine position. A vacuum-locked cradle was used for patient immobilisation when deemed necessary. In all patients, a contrast computed tomography (CT) scan was performed with a 0.5 cm thickness, from the atlas bone (C1) to the second lumbar vertebra, approximately, to include the entire neck and the lungs.

Radiation was administered with 3D conformal radiotherapy or VMAT/IMRT using radiological imaging to delineate the gross target volume of the primary tumour (GTV-P) and/or macroscopic lymph node involvement (GTV-N). Any regions of the tumour visible by endoscopy but not seen in the CT images were also included in the GTV-P. The GTV was extended by 6 to 8 mm around the primary tumour and selected lymph nodes to obtain the clinical target volume (CTV), which was, in turn, extended 10 mm laterally and vertically to obtain the planning target volume (PTV). The radical radiotherapy was conventionally fractionated and, in some cases, was preceded by induction or concomitant chemotherapy (doublet therapy with cisplatin or carboplatin) at the discretion of the medical oncologist. Surgery was considered in patients with operable tumours. Thereafter, postoperative radiotherapy was performed in patients found after surgery to have pN2 disease, sequentially after chemotherapy.[28-30]

In designing the treatment, the aim was to use to the minimum dose possible in neighbouring organs at risk: healthy lung tissue, heart, oesophagus, and spinal cord, following the QUANTEC guidelines.[31]

Follow-up

After treatment, check-ups were performed first at 1 month after the radiotherapy and then every 3 months (including a CT scan of the neck and chest every 3–6 months) by each of the specialists who participated in their treatment (thoracic surgeons, medical and radiation oncologists). Any acute (up to 3 months after treatment) and chronic (from then until after the end of the radiotherapy) toxicity was recorded, using the Common Terminology Criteria for Adverse Events ( vs 4.0.).

Statistical analysis

Continuous variables were expressed as medians and range and categorical variables as frequencies and percentages. To compare categorical variables Chi-square test was used or Fisher exact test when expected frequency less than five.

The primary outcome was the overall survival of the population, analysed using Kaplan-Meier curves. To calculate survival, the time interval considered was from the end of radiotherapy to the date of death (all-cause) or the last follow-up.

Analysis was performed to assess the influence of clinical characteristics (age, sex, TNM stage, KPS score, history of heart disease and diabetes, pretreatment weight loss, diagnosed chronic obstructive pulmonary disease, baseline haemoglobin levels, smoking and drinking habits, history of thromboembolism, pretreatment QOL considering EORTC QLQ-C30 and LC-13 scores) and treatment (modality, radiotherapy technique, and radiation dose) on patient survival.

Kaplan-Meier curves and log-rank test was used to identify which clinical or treatment-associated variables were predictive of overall patient survival. Subsequently, variables with a p < 0.2 in the univariate analysis were included in the Cox multivariate regression analysis (using a non-automatic stepwise procedure), to assess whether they were statistically significant independent predictors (p value < 0.05). Cox proportional hazards analysis was performed to calculate hazard ratios (HRs) and confidence intervals (CIs).

The analysis was performed using the IBM SPSS (version 23.0).

Results

We recruited a total of 139 consecutive patients between February 2014 and January 2017, with a median age of 71 years old (65-88), of whom 123 (88.5%) were males and 16 (11.5%) women. Clinical and treatment characteristics of the population included in the study are described in Table 1. In addition, we described characteristics of our study population according to the age (≤75 vs>75 years). See Supplementary Table 6.

Table 1.

Clinical and treatment characteristics of the population included in the study

Characteristics	Patients, n = 139
Age (median and range)	71 years old (65-88)
Karnofsky Performance Scale score≥70<70	135 (97.1%)4 (2.9%)
Sex: Male/Female, n (%)	123 (88.5%)/16 (11.5%)
Histological diagnosis, n (%)
Adenocarcinoma	44 (31.7%)
Giant cell carcinoma	5 (3.6%)
Epidermoid/squamous cell carcinoma	90 (64.7%)
Comorbidities, n (%)
Chronic obstructive pulmonary disease	Yes 70 (50.4%)
Diabetes mellitus	Yes 46 (33.1%)
History of heart disease Arrhythmia Hypertensive heart disease Heart failure Ischaemic heart disease Others	50 (36%)13 (9.4%)3 (2.2%)2 (1.4%)16 (11.5%)16 (11.5%)
History of thromboembolic event (yes), n (%)	16 (11.5%)
Smoking habits, n (%)
Smoker	52 (37.4%)
Ex-smoker	80 (57.6%)
Non-smoker	7 (5%)
Pack/years	67 (0–162)
Alcohol abuse, n (%)
No	86 (61.9%)
Yes	53 (38.1%)
Weight loss, n (%)
No	91 (65.5%)
Yes	48 (34.5%)
a) Baseline haemoglobin	11.6 gr/dl (range: 6.8–16.4)
Stage, n (%)
III A	72 (51.8%)
III B	67 (48.2%)
Previous surgery, n (%)
No	115 (82.7%)
Yes	24 (17.3%)
Radiotherapy technique, n (%)
3D	117 (84.2%)
Intensity-modulated radiation therapy	1 (0.7%)
Volumetric modulated arc therapy	21 (15.1%)
Radiotherapy dose received (median; Gy)	66 Gy (50–66 Gy)

Data on baseline haemoglobin was not available for four patients

Based on treatment modality, we classified all 139 patients into one of four groups: 24 patients received surgery and postoperative radiotherapy (17.3%), 38 concomitant radiochemotherapy (27.3%), 67 sequential chemotherapy and radiotherapy (48.2%), and 10 radiotherapy alone (7.2%). We then broke these treatment groups down as a function of age (65–75 vs >75 years old; Table 2): in the >75-year-old patients, the multidisciplinary committee mostly recommended chemotherapy and sequential radiotherapy (n = 15, 55.6%) or radiotherapy alone (n = 6, 22.2%), rather than surgery (n = 1, 3.7%) or concomitant radiochemotherapy and (n = 5, 16.5%). In contrast, in 65- to 75-year-old patients, surgery and concomitant radiochemotherapy and were recommended by the committee in approximately half of cases. The differences between these groups were significant (p = 0.003) (Table 2).

Table 2.

Distribution of treatment modality by age

Treatment provided	Age	Age	Total
	65–75 years	>75 years
Surgery+postoperative radiotherapy (chemotherapy)	23 (20.5%)	1 (3.7%)	24 (17.3%)
Concomitant radiotherapy and chemotherapy	33 (29.5%)	5 (18.5%)	38 (27.3%)
Sequential radiotherapy and chemotherapy	52 (46.4%)	15 (55.6%)	67 (48.2%)
Radiotherapy alone	4 (3.6%)	6 (22.2%)	10 (7.2%)
Total	112 (100%)	27 (100%)	139 (100%)

It should be noted that the multidisciplinary committee recommended surgical treatment based on multiple clinical parameters such as performance status, clinical staging, and the presence of comorbidities such as history of heart disease or pulmonary function. Patients undergoing surgery had less chronic obstructive pulmonary disease (p = 0.04) and lower T stage (p = 0.01).

In addition, (although not statistically significant) patients with previous history of heart disease underwent surgery less frequently (29.2% vs  70.8%).These data are fully described in Supplementary Table 7.

The median radiation dose was 66 Gy (50-66). The median follow-up was 9.9 months (1.18–57.36 months). The median survival was 14 months (range 11–17 months), and the overall survival rates at 6, 12 and 24 months were 82.7%, 60.9 and 32.3%, respectively (Figure 1).

Figure 1.

Overall Survival

Analysing factors with a potential influence on overall survival, the following variables were found to be significant in the Kaplan Meier analysis (Table 3): pack-year history (p = 0.049); heart disease (p = 0.0001); thromboembolic events (p = 0.012); physical, role, cognitive and social functioning (p = 0.0001, p = 0.0001, p = 0.0001 and p = 0.003 respectively); fatigue (p = 0.017); pain (p = 0.029); loss of appetite (p = 0.001); dyspnoea (p = 0.001); dysphagia (p = 0.003); pain in chest (p = 0.0001); and previous surgery (p = 0.044) (Figure 2).

Table 3.

Univariate analysis

			95% confidence interval
Variables	n*	Median survival	Lower limit	Upper limit	P value
Patient age
65–75 years	112	14.09	11.09	17.09	0.275
>75 years	27	14.42	8.42	20.42
SexFemaleMale		12.214.4	011.6	26.327.2	0.989
Karnofsky Performance Scale score
<70	4	4.27	0	21.27	0.396
≥70	135	14.42	11.28	17.56
Smoking habits
Smoker	52	20.04	12.75	27.32	0.270
Ex-smoker	80	13.20	11.18	15.23
Non-smoker	7	33.84	0	0
Pack-years
≤30	16	33.84	2.89	64.78	0.049
31–75	62	13.76	11.54	15.98
>75	61	16.92	8.59	25.24
Haemoglobin (g/dl)
<12	78	14.09	11.56	16.62	0.947
≥12	57	16.06	10.91	21.21
Weight loss
No	91	12.61	8.15	17.07	0.178
Yes	48	20.46	12.56	28.37
Alcohol abuse
No	86	13.76	10.17	17.35	0.492
Yes	53	14.42	10.36	18.48
Chronic obstructive pulmonary disease
No	69	16.92	9.59	24.25	0.050
Yes	70	13.07	10.37	15.77
Diabetes Mellitus
No	93	13.20	10.69	15.71	0.928
Yes	46	16.06	14.31	17.81
Heart disease
No	89	23.49	13.02	33.95	0.0001
Yes	50	9.98	6.42	13.55
Thromboembolic event
No	123	16.62	11.03	22.21	0.012
Yes	16	12.35	7.27	17.43
Treatment modality
Surgery (yes)	24	33.84	0.69	66.99	0.07
Concomitant chemotherapy	38	14.09	7.79	20.39
Sequential chemotherapy	67	12.61	8.52	16.70
Radiotherapy alone	10	13.37	4.92	21.81
Physical functioning, C30
≤33.3	5	4.27	0.81	7.72	0.0001
33.3–66.6	23	9.56	1.76	17.35
>66.6	99	20.07	15.05	25.09
Fatigue, C30
≤33.3	56	21.06	14.22	27.89	0.017
33.3–66.6	48	12.61	11.35	13.88
>66.6	25	8.90	0	19.70
Nausea and vomiting, C30
≤33.3	120	14.88	11.62	18.14	0.111
33.3–66.6	6	14.09	1.28	26.90
>66.6	3	5.48	0	11.53
Pain, C30
≤33.3	91	16.92	10.70	23.13	0.029
33.3–66.6	22	9.26	4.45	14.07
>66.6	13	9.56	2.14	16.97
Dyspnoea, C30
≤33.3	69	14.42	10.61	18.23	0.407
33.3–66.6	42	15.40	6.44	24.37
>66.6	18	9.56	0	22.01
Loss of appetite, C30
≤33.3	78	20.99	13.38	28.60	0.001
33.3–66.6	27	9.26	1.18	17.34
>66.6	25	10.64	6.31	14.97
Constipation, C30
≤33.3	69	20.46	15.29	25.64	0.055
33.3–66.6	39	13.20	10.74	15.67
>66.6	22	10.64	4.82	16.46
Diarrhoea, C30
≤33.3	110	13.37	10.28	16.46	0.450
33.3–66.6	17	34.82	13.38	56.26
>66.6	3	27.72	0	0
Financial impact, C30
≤33.3	98	16.06	10.40	21.72	0.143
33.3–66.6	22	12.22	4.25	20.18
>66.6	10	0	0	0
Dyspnoea, LC-13
≤33.3	104	16.92	11.15	22.68	0.001
33.3–66.6	18	8.90	5.41	12.39
>66.6	4	4.56	0	13.66
Cough, LC-13
≤33.3	31	13.37	3.81	22.93	0.300
33.3–66.6	67	16.06	12.10	20.02
>66.6	30	12.22	4.53	19.91
Haemoptysis, LC-13
≤33.3	108	15.40	8.25	22.55	0.103
33.3–66.6	16	13.37	6.16	20.57
>66.6	5	12.32	0	29.53
Sore mouth, LC-13
≤33.3	112	14.42	10.92	17.92	0.689
33.3–66.6	14	9.75	0	0
>66.6	3	16.06	0	39.04
Dysphagia, LC-13
≤33.3	106	16.06	10.28	21.84	0.003
33.3–66.6	17	12.38	6.49	18.28
>66.6	6	3.02	0	8.76
Peripheral neuropathy, LC-13
≤33.3	80	13.37	10.07	16.66	0.601
33.3–66.6	31	14.42	9.22	19.61
>66.6	17	27.07	2.31	51.82
Hair loss, LC-13
≤33.3	83	13.07	9.80	16.34	0.084
33.3–66.6	19	13.20	10.06	16.34
>66.6	26	34.82	0	70.93
Pain in chest, LC-13
≤33.3	90	16.06	9.77	22.36	0.0001
33.3–66.6	27	16.92	9.40	24.43
>66.6	12	4.30	2.03	6.57
Pain in arm or shoulder, LC-13
≤33.3	94	14.42	7.29	21.54	0.447
33.3–66.6	25	14.88	8.66	21.10
>66.6	9	10.64	7.47	13.81
Pain in other parts, LC-13
≤33.3	54	14.09	11.28	16.90	0.390
33.3–66.6	20	24.54	15.58	33.50
>66.6	16	13.37	0	28.33
Total dose (Gy)
≤60	36	14.88	12.02	17.74	0.720
>60	103	14.42	9.91	18.93
Previous surgery
No	115	14.09	10.85	17.33	0.044
Yes	24	33.84	0.69	66.99
Technique
Others	22	7.32	3.01	11.64	0.166
3D	117	14.88	11.77	17.99
Stage
IIIa	67	12.32	8.77	15.87	0.083
Others	72	16.92	10.80	23.03

Figure 2.

Overall survival regarding treatment modality

No differences in overall survival by treatment modality reached significance (p = 0.073), although there was a clinical trend towards higher survival in patients who underwent surgery (see Figure 1). Given this, to assess the role of surgery in the study population, the treatment modalities were grouped into two categories (surgery vs other treatment modalities) for the multivariate analysis (Table 3).

According to the multivariate analysis, the risk of death was higher in patients with pre-existing heart disease, a low score for physical functioning, or symptoms of dysphagia and/or chest pain, as well as those who did not receive surgical treatment. These variables are considered significant independent predictors. The results of multivariate analysis are shown in full in Table 4.

Table 4.

Multivariate analysis

Variable	p-value	HR	95,0% CI
			Inferior	Superior
Cardiopathy (present)	,002	2,206	1,334	3,647
Physical Functioning (PF) (>66.66)	,000
PF (33.3–66.6)	,000	1.546	2,901	25,155
PF (<33.3)	,170	8.546	,829	2,881
Dysphagia (ref<33.3)	,038
Dysphagia (33.3–66.6)	,571	1,213	,622	2,365
Dysphagia (>66.66)	,010	4,276	1,406	13,006
Pain in chest (ref<33.3)	,001
Pain in chest (33.3–66.6)	,657	,863	,450	1,656
Pain in chest (>66.66)	,001	4,332	1,883	9,965
Surgery (patients not undergoing surgery)	,024	2,236	1,111	4,501

Data on acute and chronic toxicity are summarised in Table 5. We performed subanalysis to assess whether acute and/or chronic toxicity (oesophagitis, pneumonitis, heart toxicity) experienced by the patients was influenced by age (65–75 vs >75 years), dose (≤60 vs>60 Gy), treatment modality, or baseline KPS (<70 vs≥70). By treatment modality, Grade three acute oesophageal toxicity was observed in 2 patients out of 16 treated with surgery and postoperative radiotherapy (after adjuvant chemotherapy), 5/16 treated with concomitant radiochemotherapy, and 1/16 treated with sequential radiotherapy and chemotherapy, while there were no cases in the group given radiotherapy alone, the rate of Grade three oesophagitis being significantly higher among the patients treated with concomitant chemotherapy (p = 0.022).

Table 5.

Acute and Chronic toxicity

Toxicity	Patients, n = 139
Acute esophagitis n (%)
Yes	69 (49.6)
No	70 (50.4)
Acute esophagitis; grade, n (%)
Grade I	13 (9.4)
Grade II	48 (34.5)
Grade III	8 (5.8)
Chronic esophagitis* n (%)
Yes	18 (12.9)
No	118 (84.9)
Chronic esophagitis; grade, n (%)
Grade I	4 (2.9)
Grade II	12 (8.6)
Grade III	1 (0.7)
Grade IV	0
Grade V	1 (0.7)
Acute pneumonitis; n (%)
Yes	62 (44.6)
No	77 (55.4)
Acute pneumonitis; grade, n (%)
Grade I	12 (8.6)
Grade II	33 (23.7)
Grade III	17 (12.2)
Chronic pneumonitis; n (%)
Yes	62 (44.6)
No	75 (54)
Chronic pneumonitis; grade, n (%)
Grade I	10 (7.2)
Grade II	27 (19.4)
Grade III	21 (15.1)
Grade IV	0
Grade V	4 (2.9)
Chronic cardiac toxicity
Yes	3 (2.2)
No	136 (97.8)
Type of cardiopathy due to chronic toxicity; n (%)
Heart failure	1 (0.7)
Ischaemic heart disease	2 (1.4)

In addition, we have analysed the influence of treatment technique (3D conventional RT vs IMRT/VMAT) in toxicity (including oesophagitis, pneumonitis, heart toxicity). We did not find statistically significant differences. Full data are described in Supplementary Table 8. We did not find any other significant associations between acute or chronic oesophageal, lung or heart toxicity and the aforementioned variables.

Discussion

The aim of this study was to assess survival and the patterns of treatment among unselected elderly patients with locally advanced NSCLC. The median survival was 14 months (11–17 months), while the 1- and 2 year overall survival rates were 60.9 and 32.3% respectively. These data are similar to those of other studies that have analysed the survival in elderly patients,[32,33] and even to those found in younger patients with locally advanced NSCLC.[34]

In our study, 27.3% (38/139) of patients received concomitant radiochemotherapy compared to 48.2% (67/139) who received sequential radiotherapy and chemotherapy (48.2%), in line with other specific studies in elderly patients.[12,32,35]

Regarding the treatment modality used, as was expected, the least common approaches were surgery and concomitant radiochemotherapy, especially among the oldest patients (>75 years old). In this latter group, the most common treatment modalities were sequential radiotherapy and chemotherapy or radiotherapy alone. Such a trend towards more interventional approaches involving surgery and concomitant radiotherapy and chemotherapy in patients aged 65 to 75 years, while more conservative treatments (sequential radiotherapy and chemotherapy or radiotherapy alone) are indicated in elderly patients (>75 years old), has been described previously by other authors.[33]

In our study, we have not observed significant differences in survival as a function whether patients received concomitant or sequential chemotherapy, in line with the recent study by Driessen et al..[32] The survival rate was even similar in patients who received radiotherapy alone and those who received sequential chemotherapy. This finding contrasts with the results of a meta-analysis[36] which indicated higher survival rates in patients given concomitant radiotherapy and chemotherapy than those given sequential radiotherapy and chemotherapy. Although this might be due to the lack of statistical power in our study, it could also be explained by the trials studied only having included young and fit elderly patients, who are not representative of the patients treated in daily clinical practice.[5,36]

On the other hand, according to our results, it seems important that, regardless of patient age, a clinical committee carefully selects candidates for surgery,[33] as our multivariate analysis indicates that surgical treatment may influence survival.

Performance status is a well known factor influencing survival in patients with lung cancer. Indeed, in our sample, patients with KPS <70 had a notable tendency towards a shorter survival (median 4.2 vs 14.4 months) than those with KPS≥70, although these differences were not statistically significant (Table 3). In our opinion this is probably due to the low number of patients (4 out of 139) included in the study with KPS<70. Moreover, considering that only patients undergoing treatment with radical intent were included in this study, it is reasonable in our consideration, including a majority of patients presenting good performance status.

On the other hand, according to our results, general clinical condition as assessed by measuring baseline QOL (in particular, physical functioning, dysphagia and chest pain) was a significant predictor of survival. This is consistent with the results of various other studies.[23,37,38] We believe that this is important since QOL parameters can easily be assessed in daily clinical practice before treatment using the EORTC questionnaires (QLQc-30 and LC-13) to facilitate decision making and inform patients about their prognosis. We believe that this is particularly important since our study has produced similar results to those of previous studies in lung cancer,[23,37,38] but with a focus on a specific sample of patients (elderly patients with locally advanced NSCLC) who are often clinically frail and regarding whom decision making may be a challenge. This finding supports the view that QOL data should be collected in daily clinical practice.

Additionally, it was found that a history of specific comorbidities such as heart disease, which is relatively common in this type of patients, was a significant independent predictor of survival. Grose et al[39] also noted the importance of comorbidities as an independent prognostic factor in early and advanced stages of lung carcinoma. Therefore, the level of comorbidity should be taken into account to stratify patients and interpret the results of clinical trials, especially in elderly patients.[40]

Regarding toxicity, a direct relationship was not found with age (65–75 vs >75 years old), but was found with the treatment modality, especially in patients who received concomitant chemotherapy, having this group a higher rate of acute oesophagitis. Therefore, when making a treatment recommendation in these patients, we should consider the risk-benefit ratio,[41] preferences of the patient regarding survival and treatment tolerance, given that in this unselected population of elderly patients the administration of concomitant chemotherapy did not significantly improve survival.[32]

Regarding smoking, our univariate analysis revealed an association between a higher level of smoking (packs/year) and shorter survival, as was reported by other authors.[19] Nonetheless, this result was not statistically significant in the multivariate analysis. We did not find associations between survival and anaemia, or other clinical parameters related to the treatment such as the technique or radiation dose (Table 3).

Notably, we did not find differences in survival between the oldest patients in the cohort (>75 years old) and those aged 65 to 75 years, age by itself not being found to be determinant in patient survival.

While age itself did not prove prognostic on the multivariate analyses, surgery instead, was a significant factor. This result should be interpreted with caution since only one patient >75 years underwent surgery. Indeed, in the group >75 years of age, only 16 patients could be truly evaluated for surgery (without considering cases with stage IIIB where surgery is clearly not indicated), regarding PS and comorbidities (see Supplementary Table 6). It is, therefore, a small number of patients that could have influenced our results. Future research should focus on predictive patient characteristics to distinguish patients within the heterogeneous older population who can benefit from curative-intent treatment.

After an analysis of overall survival, in line with other studies in younger patients, we affirm that there is no reason to rule out combined treatment for patients based on their age alone.[34,36,42]

However, we should recognise that this study has some limitations. It should be noted that this study was not designed to explore the benefit of a specific treatment approach as it not a randomised trial.

Indeed, the clinical decision may be difficult in elderly patients with lung cancer (usually fragile population) in the absence of high quality data. Considering the risks of surgery and toxicity of chemo-radiotherapy are often increased in the elderly compared with younger patients, patients in this study were therefore, closely scrutinized. Our management recommendations were generally similar to those of general guidelines for the NSCLC population. Careful evaluation was performed to ensure that treatment was guided by patient characteristics, stage, and not by age. All the treatment decisions were based on patient performance status, tumour resectability including T and N stage (see Supplementary Table 7), pulmonary function and presence of comorbidities. The best radical treatment approach was indicated for each patient in a multidisciplinary board. Whenever possible surgery was indicated (±postop-radiotherapy, regarding TNM stage, resectability and comorbidities), followed by concomitant radiochemotherapy or radiotherapy alone.

We also acknowledge that we did not perform a comprehensive geriatric assessment, that might be necessary to provide the best suitable treatment for each patient, and therefore is increasingly been incorporated in oncologic care demonstrating that it can alter treatment decisions.[16]

According to the multivariate analysis, the risk of death was higher in patients with pre-existing heart disease, a low score for physical functioning, or symptoms of dysphagia and/or chest pain, as well as those who did not undergo surgical treatment. We recognize several reasons that justify our findings. First, it should be noted that cardiopathy is a frequent comorbidity in elderly patients with lung cancer and one of the major causes of death in the general population. Second, physical functioning is evaluating the patient fitness, considering that a better performance status is generally associated with better survival. Finally, chest pain and dysphagia are symptoms probably related to locoregionally advanced disease in this population and therefore associated with worse prognosis.

Future research on the use of geriatric evaluation in elderly lung cancer patients should be powered to understand how it could potentially contribute to optimal decision making.[32,43] On the other hand, it can be difficult to conduct research on elderly patients, given the slow recruitment and strict selection criteria for inclusion of patients in trials.[44] In our study, we prospectively assessed an unselected elderly population, an approach which may provide useful insights for daily routine clinical practice.

Conclusions

Patients over 75 years of age tend to receive more conservative treatments, involving less surgery and less concomitant radiochemotherapy. The surgical modality improved survival and therefore, this treatment modality should be carefully considered on case-by-case basis, regardless of patient age. A history of comorbidities and poor baseline QOL according to the EORTC QLQc30 and LC-13 (low physical functioning, marked dysphagia, and chest pain) are predictive of shorter survival. Therefore, measuring these parameters before treatment may help us to define a population of frail patients with a poorer prognosis to facilitate decision making in clinical practice. Prospective studies in this crucial and understudied area are needed.