Outcomes of Major Lung Resection After Induction Therapy for Non-Small Cell Lung Cancer in Elderly Patients

Abstract

Background

This study analyzes the impact of age on perioperative outcomes and long-term survival of patients undergoing surgery after induction chemotherapy for non-small cell lung cancer.

Methods

Short- and long-term outcomes of patients with non-small cell lung cancer who were at least 70 years and received induction chemotherapy followed by major lung resection (lobectomy or pneumonectomy) from 1996 to 2012 were assessed using multivariable logistic regression, Kaplan-Meier, and Cox proportional hazard analysis. The outcomes of these elderly patients were compared with those of patients younger than 70 years who underwent the same treatment from 1996 to 2012.

Results

Of the 317 patients who met the study criteria, 53 patients were at least 70 years. The median age was 74 years (range, 70 to 82 years) in the elderly group, and induction chemoradiation was used in 24 (45%) patients. Thirty-day mortality was similar between the younger (n = 12) and elderly (n = 3) patients (5% versus 6%; p = 0.52). There were no significant differences in the incidence of postoperative complications between younger and elderly patients (49% versus 57%; p = 0.30). Patients younger than 70 years had a median overall survival (30 months; 95% confidence interval [CI], 24 to 43) and a 5-year survival (39%; 95% CI, 33 to 45) that was not significantly different from patients at least 70 years (median overall survival, 30 months; 95% CI, 18 to 68; and 5-year overall survival, 36%; 95% CI, 21 to 51). However, there was a trend toward worse survival in the elderly group after multivariable adjustment (hazard ratio, 1.43; 95% CI, 0.97 to 2.12; p = 0.071).

Conclusions

Major lung resection after induction chemotherapy can be performed with acceptable short-and long-term results in appropriately selected patients at least 70 years, with outcomes that are comparable to those of younger patients.

Age has been shown to be an important risk factor for morbidity and mortality after lung resection for non-small cell lung cancer (NSCLC) [1–6]. Probably because of increased risk, curative lung cancer surgery is offered less often to older patients [7, 8]. Given that induction therapy may also potentially increase the risk of pulmonary resection [9], selecting elderly patients who are likely to tolerate and also receive long-term benefit from pulmonary resection after induction therapy can be a difficult clinical decision.

Overall, specific outcomes for elderly patients with NSCLC after major lung resection after induction therapy have not been extensively evaluated. Only one single-center study by Marquez-Medina and colleagues [10] has reported long-term survival outcomes among elderly patients who underwent induction therapy followed by surgery. Quantitative data to support difficult treatment decisions about when to offer surgery after induction therapy for elderly patients are needed, as nearly half of locally advanced NSCLC cases occur in patients 70 years of age or older [8], and a subset of these patients with favorable prognostic factors may experience extended survival by undergoing surgery.

The purpose of this study was to investigate morbidity and survival after major lung resection following induction therapy for NSCLC in patients 70 years or older. Our objective was to test the hypothesis that appropriately selected elderly patients undergoing induction therapy followed by surgery have similar short-term and long-term outcomes when compared with younger patients.

Material and Methods

Study Design

A retrospective analysis was performed of patients who underwent NSCLC resection at the Duke University Medical Center from 1996 to 2012. The study received institutional review board approval with individual patient consent being waived. All patients who received preoperative chemotherapy with or without radiation therapy and had a lobectomy or pneumonectomy within 1 year of preoperative chemotherapy were selected for inclusion. Patients were selected based on the use of induction chemotherapy and not by specific cancer stage, as the preoperative choice of chemotherapy alone or chemotherapy with concurrent radiation was dependent on physician preference and availability of induction therapy protocols throughout the course of the study.

Baseline and outcome variables included demographics, comorbidities, pulmonary function, preoperative clinical stage, pathologic stage, chest tube duration, length of hospitalization, postoperative bleeding requiring reoperation, postoperative bleeding requiring blood transfusion, pneumonia, prolonged air leak, respiratory failure, other major complications, and overall and recurrence-free survival. Preoperative staging was determined using information before resection, including computed tomographic scan, positron emission tomographic scan, brain imaging with computed tomographic scan or magnetic resonance imaging, bronchoscopy, mediastinoscopy, thoracoscopic staging, and needle biopsy. Clinical staging data were completed using the American Joint Committee on Cancer 7th edition TNM classifications [11]. The postoperative pathologic stage was determined from the final pathology report and review by a senior surgeon.

Statistical Analysis

Baseline characteristics and outcomes were compared between patients younger than 70 years and patients 70 years and older. Univariable analysis was done using the Pearson χ² test or Fisher’s exact test when applicable for categorical variables and the Student’s t test or Wilcoxon rank-sum test when applicable for continuous variables. Multivariable logistic regression was used to identify factors associated with overall morbidity, which was defined as the occurrence of at least one postoperative event. Variables in the model were chosen a priori based on suspected clinical relevance, and included age group (younger than 70 years and 70 years or older) as the primary potential predictor of interest, in addition to sex, pathologic M status, pathologic T status, pathologic N status, chronic obstructive pulmonary disease, diabetes, renal insufficiency, coronary artery disease, neoadjuvant radiation, operative year, tumor location (right or left lung), and type of surgical resection (lobectomy or pneumonectomy).

Overall and recurrence-free survivals were evaluated using the Kaplan-Meier method, and compared between the younger and older patients using the log-rank test. Overall survival was determined from the time of surgery to death from any cause, with patients censored at the time of last follow-up at Duke University Medical Center. A Cox proportional hazards model was then used to further evaluate the impact of age on overall survival, adjusting for variables chosen based on clinical relevance, which included age group (younger than 70 years and 70 years and older) as well as sex, pathologic M status, pathologic T status, pathologic N status, chronic obstructive pulmonary disease, diabetes, renal insufficiency, coronary artery disease, neoadjuvant radiation, operative year, tumor location (right or left lung), and type of surgical resection (lobectomy or pneumonectomy). As age was our primary predictor of interest, the model was run with age as a continuous variable, with age as a binary variable (younger than 75 years and 75 years and older), and again with age categorized by the following strata: younger than 70 years, 70 to 74 years, and 75 years or older.

Statistical significance for all tests was set at a probability value of less than 0.05. All statistical analyses were performed using STATA version 13.0 (StataCorp LP, College Station, TX) and R version 3.1 (Vienna, Austria).

Results

Of the 317 patients who met study inclusion criteria, 53 patients were 70 years or older (range, 70 to 82 years). Table 1 shows the baseline characteristics of the patients in each group. Compared with the younger age group (range, 33 to 69 years), the elderly patients had a higher prevalence of hypertension and coronary artery disease. Preoperative radiation was used more often in the younger cohort. Otherwise there were no other significant differences in baseline characteristics between the two groups. The number of elderly patients who underwent surgery did not significantly increase with time (Supplemental Fig 1). A figure displaying the type of preoperative treatment (induction chemotherapy or induction chemoradiation) by operative year can be found in Supplemental Figure 2.

Table 1.

Patient Characteristics

Characteristic	<70 Years (n = 264)	≥70 Years (n = 53)	p Value
Sex (n, %)			0.41
Male	116 (44)	20 (38)
Female	148 (56)	33 (62)
Age (y)			<0.001
Mean ± SD	56.9 ± 8.3	74.4 ± 3.1
Median (IQR)	58 (51–64)	74 (72–76)
Ethnicity (n, %)			0.16
White	221 (84)	50 (94)
Black	40 (15)	3 (6)
Native American	3 (1)	0 (0)
FEV1 (percent predicted)			0.84
Mean ± SD	71.3 ± 17.8	73.0 ± 20.8
Median (IQR)	73 (53–83)	73 (57–85)
DLCO (percent predicted)			0.59
Mean ± SD	70.9 ± 18.5	72.5 ± 18.1
Median (IQR)	70 (58–83)	74.5 (58–85)
BMI at operation (kg/m2)			0.38
Mean ± SD	26.3 ± 5.1	27.1 ± 4.9
Median (IQR)	25.5 (22.6–29.7)	26 (23.2–29.1)
History of diabetes (n, %)	34 (13)	12 (23)	0.07
Renal insufficiency (n, %)	5 (2)	2 (4)	0.33
Hypertension (n, %)	88 (33)	28 (53)	0.01
COPD (n, %)	57 (22)	17 (32)	0.10
Peripheral vascular disease (n, %)	12 (5)	3 (6)	0.72
Prior thoracic surgery (n, %)	67 (25)	18 (34)	0.20
Congestive heart failure (n, %)	5 (2)	1 (2)	0.95
Cerebrovascular disease (n, %)	10 (4)	3 (57)	0.53
Coronary artery disease (n, %)	28 (11)	11 (21)	0.04
Zubrod score (n, %)			0.90
0	83 (31)	19 (36)
1	168 (64)	33 (62)
2	6 (2)	1 (2)
3	1 (0.5)	0 (0)
4	1 (0.5)	0 (0)
N/A	5 (2)	0 (0)
Induction radiation (n, %)	168 (64)	24 (45)	0.01
Type of resection (n, %)			0.28
Lobectomy	224 (85)	48 (91)
Pneumonectomy	40 (15)	5 (9)

BMI = body mass index; COPD = chronic obstructive pulmonary disease; DLCO = diffusing capacity of the lung for carbon monoxide; FEV₁ = forced expiratory volume in 1 second; IQR = interquartile rank; N/A = not available; SD = standard deviation.

Table 2 shows the clinical and pathologic staging details for the patients in each age group. The majority of patients overall were classified as clinical stage IIIA. Of the patients whose tumors were noted to be pathologic T0, 88% of them were complete responders (T0 N0 M0). There was no significant difference between the age groups in type of surgery performed (p = 0.28) or location of pneumonectomy (p = 0.83), but a significant difference was detected between the groups in terms of location of lobectomy (p = 0.004) and the surgical approach (video-assisted thoracoscopic surgery versus open surgery) for lobectomy cases (p = 0.013; Table 3).

Table 2.

Clinical Staging and Pathologic Results

Characteristic	<70 Years (n = 264)	≥70 Years (n = 53)	p Value
Clinical T status			0.44
1a	23 (9)	6 (11)
1b	22 (8)	9 (17)
2a	58 (22)	11 (21)
2b	41 (16)	7 (13)
3	78 (30)	13 (25)
4	17 (6)	2 (4)
N/A	25 (9)	5 (9)
Clinical N status			0.56
0	96 (36)	19 (36)
1	19 (7)	2 (4)
2	116 (44)	26 (49)
3	5 (2)	0 (0)
N/A	28 (11)	6 (11)
Clinical M status			0.25
0	193 (73)	43 (81)
1	14 (5)	1 (2)
N/A	57 (22)	9 (17)
Clinical stage before induction therapy (n, %)			0.35
IA	5 (2)	0 (0)
IB	13 (5)	4 (8)
IIA	20 (8)	5 (9)
IIB	53 (20)	7 (13)
IIIA	130 (49)	31 (58)
IIIB	15 (6)	0 (0)
IV	16 (6)	2 (4)
N/A	12 (5)	4 (8)
Pathologic T status (n, %)			0.87
T0	64 (24)	13 (25)
T1a	44 (17)	8 (15)
T1b	17 (6)	6 (11)
T2a	59 (22)	12 (23)
T2b	26 (10)	3 (6)
T3	47 (18)	10 (19)
T4	7 (3)	1 (2)
Pathologic N status (n, %)			0.99
N0	172 (65)	35 (66)
N1	42 (16)	8 (15)
N2	50 (19)	10 (19)
Postinduction therapy pathologic stage (n, %)			0.45
Complete remission (ypT0 N0 M0)	59 (22)	10 (19)
Stage IA	30 (11)	6 (11)
Stage IB	27 (10)	8 (15)
Stage IIA	31 (12)	8 (15)
Stage IIB	39 (15)	4 (8)
Stage IIIA	58 (22)	16 (30)
Stage IIIB	2 (1)	0 (0)
Stage IV	18 (7)	1 (2)
Postinduction therapy tumor sizea (cm)			0.90
Mean ± SD	3.0 ± 2.7	3.0 ± 2.5
Median (IQR)	2.9 (0.01–4.5)	2.5 (0.9–4.2)

^aData only available for 313 of 317 patients (<70 years, n = 260; ≥70 years, n = 53).

IQR = interquartile rank; N/A = not available; SD = standard deviation.

Table 3.

Type and Location of Surgery by Age Group

Characteristic	<70 Years	≥70 Years	p Value
Type of surgery	n = 264	n = 53	0.28
Lobectomy (n, %)	224 (85)	48 (91)
Pneumonectomy (n, %)	40 (15)	5 (9)
Lobectomy surgical approach	n = 224	n = 48	0.013
Open (n, %)	174 (78)	29 (60)
VATS (n, %)	50 (22)	19 (40)
Lobectomy location	n = 224	n = 48	0.004
Right upper (n, %)	125 (56)	19 (40)
Right middle (n, %)	15 (7)	6 (12)
Right lower (n, %)	10 (4)	9 (19)
Left upper (n, %)	55 (25)	11 (23)
Left lower (n, %)	19 (8)	3 (6)
Pneumonectomy location	n = 40	n = 5	0.83
Right (n, %)	18 (45)	2 (40)
Left (n, %)	22 (55)	3 (60)

VATS =video-assisted thoracoscopic surgery.

Perioperative Outcomes

Patients 70 years or older had increased incidence of perioperative respiratory failure (p = 0.03) (Table 4). There were no significant differences in 30-day mortality, 90-day mortality, atrial fibrillation, prolonged air leak, or other major complications between the two groups. Further, there was no significant difference between the two groups in terms of chest tube duration or length of hospitalization. In multivariable analysis, older age (70 years or older) was not significantly associated with increased occurrence of perioperative complications (odds ratio, 1.41; 95% confidence interval [CI], 0.74 to 2.68; p = 0.30; Table 5).

Table 4.

Perioperative Outcomes

Characteristic	<70 Years (n = 264)	≥70 Years (n = 53)	p Value
30-day mortality (n, %)	12 (5)	3 (6)	0.52
90-day mortality (n, %)	20 (8)	6 (11)	0.37
Length of hospitalization (days), median (IQR)	5 (4–7)	5 (4–7)	0.36
Chest tube duration (days), median (IQR)	3 (2–5)	3 (2–5)	0.89
Overall complications (n, %)	129 (49)	30 (57)	0.30
Postoperative bleeding requiring blood transfusion (n, %)	32 (12)	8 (15)	0.51
Postoperative bleeding requiring reoperation (n, %)	6 (2)	0 (0)	0.59
Pneumonia (n, %)	23 (9)	4 (8)	0.78
Atrial fibrillation (n, %)	54 (20)	17 (32)	0.06
Prolonged air leaks (n, %)	32 (12)	7 (13)	0.83
Respiratory failure (n, %)	4 (2)	4 (8)	0.03

IQR = interquartile rank.

Table 5.

Predictors of Perioperative Complications

Characteristic (n = 307)	Odds Ratio	p Value	95 % CI
Age 70 y or older	1.41	0.30	0.74–2.68
Sex (female)	1.14	0.60	0.70–1.87
Pathologic M status	0.31	0.042	0.10–0.96
Pathologic T status
0	ref	ref	ref
1a	1.36	0.43	0.64–2.92
1b	2.16	0.14	0.78–5.96
2a	1.65	0.19	0.78–3.48
2b	2.75	0.044	1.03–7.34
3	1.45	0.33	0.68–3.07
4	2.54	0.30	0.43–14.8
Pathologic N status
0	ref	ref	ref
1	0.66	0.27	0.31–1.38
2	0.55	0.079	0.29–1.07
COPD	1.68	0.078	0.94–2.98
Diabetes	0.90	0.77	0.45–1.80
Renal insufficiency	6.87	0.11	0.66–71.58
Coronary artery disease	0.87	0.72	0.42–1.82
Induction chemoradiation (ref = induction chemotherapy)	1.00	1.00	0.57–1.75
Operative year	0.96	0.168	0.90–1.02
Tumor location in right lung (ref = left)	1.19	0.51	0.72–1.97
Pneumonectomy (ref = lobectomy)	3.33	0.003	1.50–7.41

CI = confidence interval; COPD = chronic obstructive pulmonary disease.

Survival Analysis: Overall Survival

In univariable analysis, there was no significant difference (p = 0.16) in overall survival observed between the younger (<70 years) and older (≥70 years) age groups in the short-term (1-year survival, 76%; 95% CI, 70 to 81 versus 67%; 95% CI, 53 to 78), mid-term (3-year survival, 47%; 95% CI, 41 to 53 versus 43%; 95% CI, 28 to 57), and long-term (5-year survival, 39%; 95% CI, 33 to 45 versus 36%; 95% CI, 21 to 51; Fig 1A). The younger age group had a median survival of 30.1 months (95% CI, 23.6 to 43.2), and the older group had a median survival of 30.1 months (95% CI, 17.8 to 67.7). After multivariable adjustment, there was a trend toward worse overall survival for the 70 or older age group (hazard ratio [HR], 1.43; 95% CI, 0.97 to 2.12; p = 0.071; Table 6).

Fig 1.

(A) The Kaplan-Meier overall survival estimates of the 70 years or older group (dashed line) versus the younger than 70 years group (solid line) of patients undergoing induction therapy followed by major lung resection is shown. (B) The Kaplan-Meier recurrence-free survival estimates of the 70 years or older group (dashed line) versus the younger than 70 years group (solid line) of patients undergoing induction therapy followed by major lung resection is shown. Tick marks represent censored subjects.

Table 6.

Cox Proportional Hazards Analysis of Overall Survival

Characteristic (n = 307)	Hazard Ratio	p Value	95 % CI
Age 70 or older	1.43	0.071	0.97–2.12
Sex (female)	1.03	0.83	0.736–1.41
Pathologic M status	1.27	0.42	0.71–2.26
Pathologic T status
0	ref	ref	ref
1a	1.06	0.80	0.67–1.70
1b	1.58	0.14	0.86–2.91
2a	1.49	0.12	0.91–2.31
2b	2.61	0.001	1.47–4.63
3	1.91	0.008	1.18–3.08
4	2.57	0.060	0.96–6.84
Pathologic N status
0	ref	ref	ref
1	1.22	0.37	0.79–1.87
2	1.55	0.021	1.07–2.26
COPD	1.64	0.006	1.16–2.33
Diabetes	1.05	0.83	0.71–1.55
Renal insufficiency	2.25	0.096	0.87–5.85
Coronary artery disease	0.91	0.685	0.59–1.42
Induction chemoradiation (ref = induction chemotherapy)	1.46	0.020	1.06–2.10
Operative year	0.94	0.001	0.90–0.97
Tumor location in right lung (ref = left)	1.23	0.18	0.91–1.65
Pneumonectomy (ref = lobectomy)	1.19	0.39	0.80–1.77

CI = confidence interval; COPD = chronic obstructive pulmonary disease.

In a multivariable analysis comparing the overall survival of patients 75 years and older (n = 21) with patients younger than 75 (n = 296), we found no significant differences in survival between the two groups (HR, 1.74; 95% CI, 0.90 to 3.36; p = 0.099). Additionally, we performed a multivariable analysis in which we compared the survival of patients younger than 70 years (n = 264), patients 70 to 74 years (n = 32), and patients 75 years and older (n = 21). Compared with patients younger than 70 years, patients in the 70 to 74 years group (HR, 1.32; 95% CI, 0.83 to 2.09; p = 0.24) and patients in the 75 years and older group (HR, 1.83; 95% CI, 0.92 to 3.64; p = 0.084) did not have significantly worse survival (Fig 2A). Finally, in a multivariable analysis in which age was modeled as a continuous variable, increasing age was found to be associated with worse overall survival (HR, 1.02; 95% CI, 1.00 to 1.04; p = 0.024).

Fig 2.

(A) The Kaplan-Meier overall survival estimates of the younger than 70 years group (solid line), 70 to 74 years group (dash dot line), and the 75 years or older group (dashed line) of patients undergoing induction therapy followed by major lung resection is shown. (B) The Kaplan-Meier recurrence-free survival estimates of the younger than 70 years group (solid line), 70 to 74 years group (dash dot line), and the 75 years or older group (dashed line) of patients undergoing induction therapy followed by major lung resection is shown. Tick marks represent censored subjects.

Survival Analysis: Recurrence-Free Survival

In univariable analysis, there was no significant difference (p = 0.99) in recurrence-free survival observed between the younger (<70 years) and older (≥70) age groups in the short-term (1-year survival, 53%; 95% CI, 46 to 59 versus 47%; 95% CI, 32 to 61), mid-term (3-year survival, 28%; 95% CI, 22 to 35 versus 34%; 95% CI, 20 to 48), and long-term (5-year survival, 22%; 95% CI, 16 to 29 versus 27%; 95% CI, 14 to 42; Fig 1B). The younger age group had a median recurrence-free survival of 12.7 months (95% CI, 10.8 to 15.0), and the older group had a median survival of 11.1 months (95% CI, 5.9 to 20.8). After multivariable adjustment, there remained no significant difference in recurrence-free survival observed between the younger and older age groups (HR, 0.99; 95% CI, 0.66 to 1.47; p = 0.94).

In a multivariable analysis comparing the recurrence-free survival of patients 75 years and older (n = 21) with that of patients younger than 75 years (n = 296), we found no significant differences in survival between the two groups (HR, 0.86; 95% CI, 0.41 to 1.81; p = 0.69). Additionally, we performed a multivariable analysis in which we compared the recurrence-free survival of patients younger than 70 years (n = 264), patients 70 to 74 years (n = 32), and patients 75 years and older (n = 21). Compared with patients younger than 70 years, patients in the 70 to 74 years group (HR, 1.03; 95% CI, 0.66 to 1.61; p = 0.89) and patients in the 75 years and older group (HR, 0.86; 95% CI, 0.41 to 1.82; p = 0.70) did not have significantly worse recurrence-free survival (Fig 2B). Finally, in a multivariable analysis in which age was modeled as a continuous variable, age was not found to be an important predictor of overall survival (HR, 0.99; 95% CI, 0.98 to 1.01; p = 0.30).

Comment

This study compared the short-term and long-term outcomes of elderly patients (70 years or older) versus those of younger patients (<70 years) who underwent lobectomy or pneumonectomy after induction therapy for NSCLC. In both univariable and multivariable analyses, increased age was not significantly associated with increased 30-day and 90-day mortality. Although elderly patients were more likely to have postoperative respiratory failure, the percentage of patients with respiratory failure was low and there were no other significant differences in complications between the two groups. In unadjusted analysis, there was no significant difference in overall survival between the age groups, but after multivariable adjustment, there was a trend toward worse survival in the elderly age group. These results support the use of major lung resection after induction therapy in carefully selected elderly patients.

To date, only two studies have detailed the impact of age on outcomes of patients with NSCLC who underwent major lung resection after induction therapy. Marquez-Medina and colleagues [10] conducted a study comparing the outcomes of elderly (age ≥70 years; n = 44) and younger (age <70 years; n = 64) patients treated with induction therapy followed by definitive treatment (radiotherapy or surgery). Of the 50 patients who underwent surgery after induction therapy, there was no difference in response rate, operability, and overall and disease-free survival between the two age groups in univariable analysis. Our study findings are largely consistent with those reported by Marquez-Medina and colleagues [10], although we additionally report a multivariable analysis that showed a trend toward worsened survival in the elderly age group. A key difference between the two studies is that Marquez-Medina and colleagues included sublobar resection in the surgery group whereas the present study included only major lung resections (lobectomy and pneumonectomy).

Rivera and colleagues [9] conducted a matched-analysis of 81 elderly (age 75 years or older) patients with NSCLC treated with neoadjuvant chemotherapy followed by resection (pneumonectomy, bilobectomy, lobectomy, sublobar resection, or explorative thoracotomy) compared with patients younger than 75 years undergoing the same treatment. In contrast to findings in the current study, the authors reported significantly greater number of complications (22.2% versus 14.8%; p = 0.03) and longer length of hospital stay (14.9 versus 12.0 days; p < 0.001) for elderly patients when compared with their younger counterparts. The discrepancies between the present study and that reported by Rivera and associates [9] may be in part related to differences in the definition of elderly age between the two studies.

It should be noted that the postoperative transfusion rates of 12% to 15% observed in this study are somewhat higher than typically seen in series of lobectomy patients. This observation may be explained by clinicians having a lower threshold for transfusion in this series of postinduction therapy patients, who often come to surgery relatively anemic after completing the induction treatment and may be thought to have a somewhat impaired ability to quickly regenerate blood counts owing to both nutritional and posttherapy factors.

An important limitation of this study is that only patients who had surgery were included. All patients included in the study passed the important step of having been judged appropriate candidates for surgery after a thorough evaluation. Therefore, the study findings are not necessarily generalizable to all elderly patients with locally advanced but potentially resectable lung cancer but only those who are deemed acceptable resection candidates after a surgical evaluation. In general, the criteria used to determine whether patients were acceptable resection candidates in this study consisted of: 1) using pulmonary function tests as well as the patient’s level of activity and functional status to judge whether the patient would tolerate the extent of resection required to remove all disease, 2) an evaluation of the patient’s comorbid conditions to consider their risk of morbidity with induction therapy as well as general anesthesia and surgery, and 3) a subjective evaluation of whether the patient had other comorbid conditions that were believed likely to lead to death before their lung cancer diagnosis.

There are other limitations to the study. First, this is a single-institution, retrospective study, and although efforts were made through multivariable adjustment to reduce the potential effects of selection bias, unobserved confounding could still be present. Further, the study population is not entirely homogeneous; the study included all patients who had undergone induction therapy followed by lobectomy or pneumonectomy, and although the majority of patients received induction therapy for stage IIIA disease, patients in the study underwent induction therapy for a wide range of reasons including for superior sulcus tumors and tumors invading the chest wall. In addition, these findings are from a single institution that performs a high volume of lobectomies and pneumonectomies and may be limited in its generalizability. Recognizing that the evolution in induction therapy for NSCLC during the study period may have affected our outcomes, we did try to account for differences in treatment eras by including the operative year in multivariable analysis. Also, we were not able to compare disease-specific survival between the two groups. Finally, the follow-up time for patients was relatively short, which may have limited the long-term findings of the study; longer follow-up time would help strengthen our findings.

In conclusion, major lung resection after induction chemotherapy can be performed with acceptable short-and long-term results in appropriately selected patients 70 years and older, with outcomes that are comparable to those of younger patients. Although it is important to consider age when considering multimodality therapy for NSCLC, there is no age threshold beyond which surgery should be excluded from patients who are potential candidates for multimodality therapy. Future research should focus on identifying which comorbidities and characteristics are most important in the elderly population to optimize both perioperative outcomes and long-term survival.

DISCUSSION

DR JOSHUA R. SONETT (New York, NY): Excellent paper, and it is important for everybody out there to know that we can get older patients through lung resections. Since your conclusion is that older patients need not be excluded, why do we put any age on it? Why is it 70? To me, 70 is not old anymore for a lot of our patients, and there are 40-year-olds who look horrible and 70-year-olds who look really good. In that respect for your respiratory failure ones, did you tease apart to see if they had other risk factors beyond just age 70 and just consider getting enough groups together to triage this to see if there is actually ever a breakpoint in age versus physiologic statues of patient?

DR YANG: Thank you, Dr Sonett. With regard to your first question, our study is limited by the sample size. We chose an age cutoff that would allow for comparisons between two groups, and 70 is a commonly selected age in the literature, but to your point, using 70 years as a cutoff is somewhat arbitrary.

To describe the nonlinear effect of age on overall survival, we did do a restricted cubic spline analysis where we found that the risk of death increases with age, but there was no particular cutoff where the risk of death increased exponentially. There is no particular age point that should be used to exclude patients.

In terms of your other question to more specifically get to what kind of elderly patient could benefit from surgery, we did try to look into that in more detail but found that our sample size was too small to build, test, and validate a robust and discriminative risk model. We did examine whether incorporating additional clinical variables could improve the discriminative power of the existing T, N, and M model. We did not observe any significant increases in the concordance index or improvements in the calibration curves to justify the addition of other clinical variables.

DR SONETT: Excellent. Maybe you can rework your same study with the STS [Society of Thoracic Surgeons] database and you will have more robust numbers to augment your series.

DR YANG: Thank you.

DR DANIEL RAYMOND (Cleveland, OH): I enjoyed your talk. It seems to me what you are getting at is the whole notion of physiologic age versus chronologic age and something that probably all of us have discussions with our patients about every day in the clinic. Truly to address this what it seems like you need to do is go after kind of a frailty assessment. Is that something that your group is working on? Do you think that is an important component of your evaluation of lung cancer patients?

DR YANG: Thank you for your question. Doctor Tong, I believe, is working on a frailty assessment, and it is certainly an important aspect of care for patients. For this particular study we were not able to evaluate frailty in detail, but that is a great recommendation for a future study.

DR JOSEPH I. MILLER (Atlanta, GA): At last week’s Focus on Thoracic Surgery conference in Boston they had an identical presentation, and I congratulate you. There was a panel of two oncologists, a radiation therapist, and a surgeon, and it addressed surgery after induction therapy. All agreed that chemotherapy should be the only preop[erative] induction given, then surgery, then adjuvant radiation postop[eratively]. But it showed no difference in control groups between induction chemotherapy and just surgery for stage I and IIA lung cancer. Just a comment.

DR YANG: Thank you. I look forward to reading that manuscript when it comes out.

DR TRAVES CRABTREE (St. Louis MO): Doctor Yang, why is there a difference between the age groups and the number of patients that had a VATS [video-assisted thoracoscopic surgery] approach? It was only 19% for the less than 70 and 38% for those older than 70. Was that related to radiation therapy or tumor factors? Can you comment on that?

DR YANG: It is probably related to radiation use. In a multivariable model adjusting for radiation use, older patients were no longer associated with a significantly increased likelihood of undergoing a VATS approach when compared to younger patients.

DR BETTY C. TONG (Durham, NC): I just want to add to Jeff’s talk. As you know, it is standard at our institution to at least consider VATS as first line no matter what sort of preop[erative] therapy they have. I think the difference between those who had thoracotomy versus VATS is going to be due to the inclusion of radiation; at least 4 out of 5 of us would strongly consider a muscle flap if given preoperative radiation just from the get-go.