ABSTRACT: The majority of individuals diagnosed with lung cancer in the United States are 70 years of age and older. Defining appropriate therapy for older patients with non–small-cell lung cancer (NSCLC) is becoming a major focus of clinical research. In this article, we review the available data on clinical predictors of risk and benefit for elderly NSCLC patients receiving treatment via a variety of modalities, including surgery, radiotherapy, combined radiotherapy and chemotherapy, and chemotherapy alone. The data demonstrate that subgroups of elderly patients benefit from appropriately selected treatment. Participation of older patients in clinical trials designed to assess efficacy, toxicity, and quality-of-life outcomes for recently developed treatment modalities in this population is critical.
Age-Adjusted Incidence Rates for Lung Cancer
Age Distribution of Lung Cancer Cases
Lung cancer is the leading cause of cancer death in the United States among both women and men. In 2002, an estimated 169,400 people were diagnosed with lung cancer, and 154,900 died of the disease.[ 1] Advancing age is associated with an increased incidence of lung cancer (Figure 1).[2] Approximately 51% of patients diagnosed with lung cancer are 70 years of age and older, and 16% are ≥ 80 years old (Figure 2).[3]
The treatment of older persons (age ≥ 70) with lung cancer deserves careful attention because this subgroup represents a substantial and growing segment of the population, and most importantly, because clinical decision- making is fairly complex in these patients. Aging involves physiologic changes that may affect drug pharmacology and treatment tolerance, including decreased renal function, bone marrow reserve, and lean body mass.[4] In addition, the older population comprises those with a wide range of physiologic reserve, including the most fit (who typically participate in clinical trials), patients with significant comorbid disease or poor performance status, and the very old. A key question for oncologists is how to assess the characteristics of an older patient in order to predict the therapeutic ratio for that patient and make appropriate recommendations.
Fortunately, the evaluation of appropriate therapies for older patients with lung cancer has indeed become a major focus of research.[5] The results of several clinical trials specifically designed to evaluate treatment modalities in older patients have recently been published. This article will review the available data pertaining to clinical predictors of risk and benefit for specific treatments in older patients with non-small-cell lung cancer (NSCLC).
Surgery
Surgical resection is the treatment of choice for stage I/II NSCLC because it offers the best chance for long-term cure. Overall 5-year survival rates range from 60% to 80% for patients with stage I disease and 25% to 50% for patients with stage II disease.[6] Age alone should not exclude a patient from consideration for surgical therapy, as many older patients live long enough to enjoy the survival benefit conferred by surgical resection. The average life expectancy for US men and women 80 years of age is 6.7 and 8.8 years, respectively, and for men and women 85 years of age, 5.3 and 6.7 years.[7] The critical questions to be addressed are whether the overall survival of older patients undergoing surgical resection for NSCLC is similar to that of younger patients, and whether surgical risks for older patients are too high to achieve a survival benefit.
Long-Term Outcomes
• Older vs Younger Patients—The results of two large retrospective analyses suggest that long-term outcomes for appropriately selected older patients undergoing surgical resection are comparable to those of younger patients. Van Rens et al conducted a retrospective study analyzing survival among 2,361 patients who underwent surgical resection for stage I, II, or IIIA primary NSCLC.[8] A total of 1,115 patients (47%) were over age 65 at the time of surgery, with the age range extending to 85. In comparing the overall survival of patients over 65 to that of patients younger than 65, no difference was observed at up to 4 years of followup.At 5 years, the overall survival of older patients was slightly inferior to that of younger patients (38% vs 44%, P < .0001).
Outcomes for Patients 65 ≥ Years Old Treated With Surgery
Ishida et al conducted a retrospective analysis of patients with NSCLC who underwent treatment with surgical resection at a university hospital in Japan between 1974 and 1989.[9] Of 662 cases analyzed, 185 (28%) were age 70 and older (Table 1). Outcomes were compared for elderly patients (≥ 70) and younger patients (< 70). No significant differences in overall 5-year survival were found, and the 5-year overall survival of the older group compared favorably to that of the younger group (67% vs 68% for stage I disease and 40% vs 17% for stage II disease). The inferior survival of younger patients with stage II disease was not explained by the authors.[9]
• Case Series—Kamiyoshihara et al published a series on 37 patients aged ≥ 70 years with resectable stage I-IIIA lung cancer. These investigators observed a 35% 5-year survival rate.[10]
With five small case series, researchers reported on the outcomes of patients ≥ 80 years old who underwent lung resection for NSCLC (Table 1). Three series involving 103 patients with stage I, II, or IIIA disease showed acceptable 5-year survival rates ranging from 30% to 55%.[11-13] Two studies focused on patients with stage I NSCLC. The largest series included 54 patients aged ≥ 80 years and demonstrated a 57% 5-year survival rate.[14] The smaller study involved 18 patients and showed a 43% 5-year survival rate.[15] Shirakusa and colleagues also performed a subset analysis on data from patients with stage I disease and reported an impressive 5-year survival rate of 79% in 18 patients.[11]
Morbidity and Mortality
Although appropriately selected older patients can enjoy long-term benefits from surgical resection of stage I and II NSCLC, evidence suggests that this population also faces increased short-term morbidity and mortality associated with surgical therapy. Ginsberg et al analyzed 30- day surgical mortality rates for resections performed between 1979 and 1981 at centers participating in the Lung Cancer Study Group. Among enrolled patients, 416 were age 70 to 79, and 37 were age 80 or older.
By univariate analysis, the postoperative mortality rate increased from 1.3% among patients aged < 60 years to 4.1% among patients aged 60 to 69, 7% among those aged 70 to 79, and 8.1% among those aged 80 and older (P < .01). Compared to patients aged 50 to 59 years, complication rates were 2.5 times greater for patients aged 70 to 79 and 2 times greater for those aged 80 and older. Common causes of surgical mortality included pneumonia, respiratory failure, bronchopleural fistula, empyema, and myocardial infarction.[16]
Although Ginsberg's results are frequently used to demonstrate an association between age and increased postoperative morbidity and mortality, their applicability to contemporary clinical decision-making is limited. These results describe outcomes for cases performed 20 years ago and do not reflect advances in surgical technique or postoperative care, which may influence outcomes. Additionally, because only univariate analysis was performed, Ginsberg's results do not help us determine whether the more powerful predictor of surgical risk is advancing age or the comborbid diseases commonly associated with aging.
Romano and Mark analyzed inhospital mortality for 12,437 lung cancer resections performed in California hospitals between 1983 and 1986.[17] Compared to patients less than age 60, the adjusted odds ratio (OR) for postoperative death was 2.3 for patients aged 60 to 69, 3.6 for those aged 70 to 79, and 5.4 for those > 79. Other significant predictors of postoperative death included chronic heart disease (adjusted OR = 1.8), extended resection (OR = 1.8), diabetes (OR = 1.5), male gender (OR = 1.5), chronic obstructive pulmonary disease (OR = 1.4), and volume of lung resections performed at the hospital (OR = 0.6).[17] This study may have underestimated the contribution of comorbid disease to in-hospital mortality because only comorbid diseases recorded as secondary diagnoses at hospital discharge were noted.
• Other Variables—In a more recent study, several variables were found to be more powerful independent predictors of postoperative morbidity and mortality than age. The National Veterans Affairs Surgical Quality Improvement Program conducted a prospective multicenter study involving 194,319 major pulmonary resections performed between 1991 and 1995. In multivariate models, age was a significant predictor of postoperative morbidity and mortality. Compared to 50-year-old patients, the postoperative mortality risk was 1.33 for 60-year-old patients, 1.66 for 70-year-old patients, and 1.99 for 80-year-old patients. Other independent prognostic variables included do-not-resucitate status (OR = 4.3), pneumonectomy (OR = 3.0), impaired sensorium (OR = 2.7), hypoalbuminemia (OR = 1.7), and dyspnea (OR = 1.4).[18]
In another recent prospective single- institution study in 500 patients, Bernard et al found that age was associated with increased risk of postoperative mortality only in the univariate analysis.[19] In a logistic model accounting for indication for surgery, type of procedure performed, pulmonary function, comorbidity indices, and preoperative chemotherapy, age was not an independent predictor of postoperative mortality.[ 19] Both these results and those of the Veterans Affairs Surgical Quality Improvement Program suggest that adequate physiologic reserve (as measured by pulmonary function, comorbidity indices, and nutrition) is a more useful criterion for appropriate selection of patients than age alone.
To our knowledge, no validated models have been developed to predict the risk of surgical morbidity and mortality among older patients undergoing lung resection. In two of the series cited in Table 1, the authors specified that they used standard criteria to assess cardiopulmonary function when selecting patients for surgery or chose the type of resection based on prediction of postoperative pulmonary function.[14,9] In other series, the method of patient selection was not specified, but the prevalence of ischemic heart disease was low.
In Kamiyoshihara's series, 5% of patients aged 70 years and older had a history of ischemic heart disease.[ 10] In the series published by Osaki et al, 12% of patients had ischemic heart disease (angina or history of myocardial infarction), and in Hanagiri's series, 17% of patients had a history of angina and 6%, a history of myocardial infarction.[13,15] The prevalence of arrhythmias in these three series was also fairly low (19%, 30%, and 22%).[10,13,15]
• Choice of Procedure—In addition to the factors discussed above, surgical morbidity and mortality are dependent on the procedure performed. In all of the case series cited, the use of pneumonectomy was avoided or limited, because the procedure is associated with a high rate of postoperative complications and mortality. Additionally, Dyszkiewicz et al compared pneumonectomy to lobectomy among 90 patients older than age 70.[20] Postoperative complications developed in 79% of patients undergoing pneumonectomy vs 58% of patients undergoing lobectomy.[ 20] De Perrot compared postoperative mortality rates for pneumonectomy in 208 patients aged ≥ 70 years to those in 416 patients aged < 60. The mortality rate among the older patients was 13.7%, compared with 6.5% among the younger patients.[21]
Landreneau et al evaluated wedge resection vs lobectomy in 219 patients with T1, N0, M0 NSCLC and found that the mean hospital stay decreased by 3 to 4 days in the wedge resection population, even though this group of patients was older and had poorer lung function.[22] Additionally, wedge resection via video-assisted thoracic surgery was associated with decreased postoperative pain and better postoperative pulmonary function. This may be especially beneficial to older patients, for whom postoperative pulmonary rehabilitation may be slower and the risk of postoperative delirium, higher.
Five-year survival was inferior for patients undergoing wedge resection (58% for open wedge, 65% for video-assisted wedge, and 70% for lobectomy, P = .02), but these differences in survival were due to an increased death rate from other causes. At 5 years, the non-cancerrelated death rate was 38% for patients who had undergone wedge resection vs 18% for patients who had a lobectomy.[22]
Conclusions Regarding Surgery
Overall, the evidence suggests that 5-year survival for patients ≥ 70 years old who undergo surgical resection is similar to that of patients < 70 years old. In addition, 5-year survival for appropriately selected octogenarians is quite favorable. Older patients should not, therefore, be excluded from consideration for surgical therapy on the basis of age alone. Rather, older individuals should be risk-stratified so that appropriate recommendations for therapy are made.
Unfortunately, no validated method has been developed to enable physicians to take into account the multiple factors that influence postoperative risk and accurately predict that risk for older patients undergoing lung resection. In most studies, age was an independent risk factor for postoperative morbidity and mortality. The presence of comorbid disease, including cardiopulmonary disease, diabetes, poor nutrition, and impaired sensorium, further increased surgical risk. Postoperative risk was also influenced by the extent of the surgical procedure and the volume of lung resections performed at a given institution.
Recommendations
Surgical therapy should be offered to older patients with stage I or II NSCLC who have minimal significant comorbid disease. Surgery should be performed by experienced thoracic surgeons at high-volume institutions with care to avoid pneumonectomy. Further research is needed to define methods of optimizing postoperative care and rehabilitation for older patients who undergo curative lung resection.
