Treatment of Small Cell Lung Cancer in Academic and Community Settings: Factors Associated with Receiving Standard Therapy and Survival

Abstract

Purpose

Patients diagnosed with small cell lung cancer (SCLC) historically have poor prognosis. Clinical trials have demonstrated improved survival among patients receiving standard platinum/etoposide-based chemotherapy. While treatment patterns and outcomes have been evaluated for patients with SCLC in clinical trials, population-based practice patterns are not well known.

Methods

The National Cancer Institute’s Patterns of Care study was used to evaluate patient and provider factors associated with standard treatment, clinical trial enrollment, and 12-month relative hazard of death.

Results

Among 931 patients diagnosed in 2007 with SCLC in academic and community settings, 72.2% of patients with limited-stage (LS) disease received chemoradiation and 42.2% of patients with extensive-stage (ES) disease received chemotherapy only; the expected treatment scenarios by stage. Less than 1% of patients enrolled in clinical trials and 2.1% LS and 3.4% ES patients refused any type of treatment. Patients ≥80 years at diagnosis and those with pneumonia/lung collapse were less likely to receive chemoradiation for LS disease. Patients treated in hospitals with residency programs were more likely and patients ≥80 were less likely to receive chemotherapy for ES disease. Finally, female patients with LS disease, black patients with ES disease and all patients who received chemotherapy compared to receiving radiation alone or no therapy experienced significantly lower mortality.

Discussion

Despite the demonstrated lower mortality, a relatively large proportion of patients with SCLC are not treated with a standard treatment regimen. Future studies should evaluate efforts to promote use of appropriate treatment regimens and encourage clinical trial participation.

Small cell lung cancer (SCLC) is the fifth leading cause of cancer mortality in the United States and accounts for approximately 15% of the more than 200,000 lung cancers diagnosed each year.[1,2] Patients diagnosed with SCLC have a very poor prognosis, with median survival of 16–24 months for limited-stage (LS) disease and 10–12 months for extensive-stage (ES) disease.[3,4] Compared with other types of lung cancer, the majority of patients with SCLC present with widely disseminated disease at diagnosis.[4] Although chemotherapy and radiation can favorably impact the natural course of the disease, overall improvements in outcomes for SCLC over the past 25 years have been limited.[3–5] In fact, the inclusion of prophylactic cranial irradiation (PCI) for both LS and ES patients with very good response to initial therapy could be considered the most recent significant advance in the management of this disease.[6,7]

Phase III clinical trials have demonstrated improved survival in patients treated with a combination of etoposide with cisplatinum or carboplatin.[5] Current practice guidelines recommend that patients presenting with limited stage (LS) disease receive a combination of chemotherapy and radiation, while those with extensive stage (ES) disease receive chemotherapy only.[5,8,9] Patients with ES who have a very good response to initial systemic treatment may be offered additional thoracic radiation to the primary mass to further improve disease control or to sites of metastatic disease for palliation purposes.[5,10] However, there is no definitive evidence of added survival benefit with this approach and, for the purpose of this analysis, we will consider chemotherapy only as the standard approach for treatment of ES. No consistent benefit of alternative regimens has been demonstrated over this standard.[10–16] Improvements in the survival of patients with SCLC have been achieved through well-designed clinical trials that build upon the best available therapies,[11,12] so patient entry for such trials is highly desirable. Treatments and outcomes for patients with SCLC have been evaluated in the clinical trial and comprehensive cancer center setting,[17] but population-based practice patterns in the U.S. community and academic settings are not well described.

The purpose of our study was to evaluate patterns of treatment for patients newly diagnosed with SCLC in the U.S. community and academic settings and to evaluate the use of chemotherapy combinations, thoracic radiation, and clinical trial participation on outcomes. Using the National Cancer Institute’s Patterns of Care Study, we present a population-based analysis of patient, provider and tumor characteristics associated with receipt of recommended therapy by stage at diagnosis for SCLC. We evaluated these factors for associations with 12-month relative hazard of death.

MATERIALS AND METHODS

Data and Sampling Methodology

We identified SCLC patients registered in the Surveillance, Epidemiology and End Results (SEER) program and diagnosed in 2007. SEER is a population-based set of cancer registries that collects information on incident cancer diagnoses and survival for approximately 28% of the US population.[18] The SEER program routinely collects information on cancer stage, initial therapy, patient demographics and follow-up for vital status. However, because SEER data collection is primarily hospital based, therapy administered in an outpatient setting can be under reported. To obtain more complete treatment information, the annual National Cancer Institute’s Patterns of Care (POC) study reviews medical records and queries physicians to obtain information on cancer therapies for selected cancers.[19]

SEER patients diagnosed with SCLC were stratified by age, race/ethnicity and registry and a random sample was selected from each stratum for the POC study. Non-Hispanic blacks, Hispanics, Asian/Pacific Islanders and American Indians/Alaskan Natives were oversampled to obtain more stable estimates. Institutional Review Board approval was received as required by the registries. Each patient’s doctor was contacted to verify specific treatments administered, including agents and dates of initial therapy and provide the name (s) of other physicians who might have treated the patient and these additional physicians were also contacted. Hospital medical records were reviewed to identify patient comorbidities, confirm tumor characteristics and treatments. Treatment was verified by physician report or hospital records in 98.8% of patients. Finally, information about hospital characteristics was recorded.

Patients

We included patients diagnosed with histologically confirmed SCLC or mixed histologies. Patients with lung sarcoma were ineligible. Patients were also ineligible if they had a history of cancer (other than non-melanoma skin cancer); were diagnosed simultaneously with a second primary cancer; or were diagnosed at autopsy or on death certificates only. Patients undergoing surgical resection (n=27) as a component of their initial treatment were excluded as information regarding what led to the choice of surgery (i.e. suspected non-small cell histology) was unavailable in our study and would make medical interpretation of our results difficult.[20,21] After exclusions, our final sample included 931 patients with SCLC.

Staging and Therapy

Patients were grouped according to the standard classification of LS or ES disease at presentation according to physician’s report or hospital record verification of the stage at diagnosis.[5] We evaluated initial therapeutic treatment received and 12-month mortality for newly diagnosed patients with SCLC. Initial therapy was defined as treatment that was administered or planned before progression or recurrence of the disease, as verified by the physician. Treatment for SCLC was classified as: chemotherapy alone, thoracic radiation therapy alone to the primary site, chemotherapy plus thoracic radiation, and no/unknown treatment. Chemotherapy regimens were further classified as etoposide with cisplatin or carboplatin or ”other” chemotherapy. We considered US standard chemotherapy the receipt of the doublet etoposide plus either cisplatin or carboplatin. Patients receiving “other” chemotherapy within 2 months of diagnosis, even if administered in addition to the standard doublet were categorized as “other” chemotherapy because a three drug regimen was not considered standard of care. Although the combination of platinum plus irinotecan is considered by some as an acceptable initial standard treatment, the confirmatory phase 3 study in the US population (SWOG S0124) did not confirm the superiority of that doublet against the US standard of cisplatin/etoposide,[22] as seen in the Japanese studies. Activity was similar with both treatment combinations, but irinotecan led to a worse toxicity profile for the US population. The conclusion arising from SWOG S0124 was that platinum/etoposide remains as the reference treatment standard for the US population. Therefore, for the purpose of this analysis, we considered the platinum/etoposide combination as the standard regimen for ES SCLC.

The codes for radiation site were to the 1) primary site (i.e., thoracic) and/or 2) to an unspecified non-primary site at the time of initial treatment. Radiation to the non-primary site was prior to progression or recurrence; not for palliative therapy. We used this data to investigate the administration of CNS radiation for LS patients. By definition, LS patients do not have metastatic disease, so, although the coding does not specifically refer to radiation to a CNS site, radiation to a non-primary site was attributed to CNS radiation (e.g. PCI). Because by LS definition there were no metastatic sites, it is highly unlikely that this radiation would have been given to any non-primary site other than brain. Therefore, even with the noted limitation, it is reasonable to infer that radiation was delivered to the brain as PCI. Time to death for 12-month mortality calculations was defined as the time, in months, from a patient’s diagnosis until death.

Analysis

We used chi-square analyses to assess the unadjusted association between patient, provider and treatment characteristics and stage at diagnosis. Logistic regression evaluated the association between patient and provider characteristics and treatment for SCLC. Models were stratified by extent of disease at presentation. For patients with LS, we evaluated factors associated with 1) etoposide with cisplatin or carboplatin and thoracic radiation and 2) any chemotherapy with thoracic radiation. For patients with ES, we evaluated factors associated with 1) etoposide with cisplatin or carboplatin as well as 2) any chemotherapy. Finally, Cox proportional hazards regression was used to evaluate factors associated with 12-month relative hazard of death. Each model was adjusted for type of treatment, age at diagnosis, race/ethnicity, sex, insurance, Charlson co-morbidity score, pneumonia/lung collapse at presentation, hospital ownership[23–25] and the presence of a residency program in the treating facility. In separate models we defined chemotherapy as etopside plus cisplatin or carboplatin or as “other” chemotherapy. These models were not substantially different likely due to the relatively small number of patients receiving “other” agents, therefore we present the hazards models with “any” chemotherapy. All analyses were performed using SUDAAN (Research Triangle Institute, Research Triangle Park, NC) statistical software. All percentages are weighted to the SEER population from which the data were obtained using the inverse of the sampling fraction. The numbers are unweighted. All tests of statistical significance were assessed using the Wald-type F-statistics with a p-value <0.05 considered statistically significant.

RESULTS

In this population-based study of 931 patients with SCLC treated in the community and academic settings, 71.4% of patients were diagnosed with ES disease while 28.6% presented with LS disease (Table 1), which is similar to the distribution of SCLC cases in the U.S. as a whole. Unadjusted analyses identified no significant differences in between LS and ES at presentation by age, race/ethnicity, sex, insurance type, smoking status, or Charlson co-morbidity score. However, patients with ES disease were more likely to be treated in not-for-profit hospitals.

Table 1.

Demographic, Hospital and Treatment Characteristics of Patients with Pathologically Confirmed Small Cell Lung Cancer, by Extent of Disease (N=931)

Demographics	Limited Stage Disease	Extensive Stage Disease	p-Value
	(N=288) N, Wt %	(N=643) N, Wt %
Age at diagnosis			0.13
<60	59 (24.1)	156 (25.5)
60–69	85 (36.6)	201 (32.3)
70–79	106 (32.5)	209 (30.5)
≥80	38 (6.8)	77 (11.6)
Sex			0.2
Male	146 (45.1)	337 (52.4)
Female	142 (54.9)	306 (47.6)
Race/Ethnicity			0.34
White	121 (81.2)	295 (84.1)
Black	74 (8.5)	174 (8.0)
Hispanic	49 (5.3)	102 (4.5)
Other/Unknown	44 (5.0)	72 (3.4)
Insurance			0.35
Private	141 (57.3)	347 (59.8)
Any Medicaid	80 (17.8)	152 (18.3)
Medicare only	45 (16.7)	103 (13.4)
IHS/Tricare/VA/Military	12 (6.0)	14 (3.1)
Unknown/No Insurance	10 (2.2)	27 (5.4)
Smoking Status			0.56
Current	124 (50.2)	319 (53.3)
Previous	129 (43.0)	259 (38.1)
Never/Unknown	35 (6.8)	65 (8.6)
Charlson Comorbidity Score			0.32
0	115 (40.8)	257 (33.8)
1	109 (37.4)	261 (46.2)
≥2	64 (21.8)	125 (20.0)
Pneumonia/Lung Collapse			0.94
Yes	110 (42.4)	264 (42.8)
No/Unknown	178 (57.6)	379 (57.2)
Treating Hospital Characteristics
Bed Size			0.37
<199	68 (26.1)	124 (23.5)
200–299	64 (19.9)	168 (28.4)
300–399	72 (21.8)	132 (18.8)
400+	77 (29.3)	202 (25.5)
Unknown/Outpatient	7 (2.9)	17 (3.8)
Approved Residency Training Program			1.0
Yes	130 (41.4)	322 (41.4)
No/Unknown	158 (58.6)	321 (58.6)
Hospital Ownership			0.03
For Profit	27 (10.7)	62 (8.9)
Not-for-Profit	195 (62.7)	445 (76.1)
Government/Other	66 (26.7)	136 (15.0)
Therapy
Treatment			<0.0001
Etoposide with Cis or Carbo-platin	27 (12.0)	215 (30.2)
Other Chemotherapy*	8 (2.5)	76 (9.8)
Thoracic Radiation Only	10 (3.2)	17 (2.1)
Etoposide with Cis or Carbo-platin & Thoracic Radiation	159 (59.5)	113 (18.8)
Other Chemotherapy *& Thoracic Radiation	33 (12.7)	32 (6.3)
Refused Treatment	13 (2.1)	33 (3.4)
No/Unknown Treatment	38 (8.0)	157 (29.4)
Clinical Trial Participant			0.16
Registered	2 (0.2)	9 (0.6)
Pet Scan Received			0.004
Yes	111 (43.0)	171 (26.4)
No/Unknown	177 (57.0)	472 (73.6)

Wt: Weighted to the population from which the sample was drawn

^*Note: Other chemotherapy regimens were defined failure to receive a combination of Cisplatin or Carboplatin with Etoposide

Type of Treatments Received

Limited-Stage Disease

Of patients with LS disease 72.2% received chemoradiation as the initial treatment (Table 1), with 2.1% indicating refusal of any type of treatment (n=13). Further, only 59.5% of all LS patients received etoposide with cisplatin or carboplatin and radiation. This decreased with age from 60% to 29% for those age <60 and 80+, respectively (data not shown). Few patients (N=2) participated in clinical trials. PET scans were received by 43% of LS patients as part of their diagnosis and treatment planning.

After adjusting for patient and provider characteristics, we found that age at diagnosis and Charlson score were associated with receipt of etoposide with cisplatin or carboplatin plus radiation. Age, race/ethnicity, pneumonia/lung collapse were associated with receiving any chemotherapy plus radiation (Table 2). Specifically, we found that older patients (80+) were less likely to receive etoposide with cisplatin or carboplatin and thoracic radiation (OR: 0.15 (95% CI 0.04, 0.64)) or any chemoradiation (OR: 0.06 (95% CI 0.01, 0.29)). Patients with a Charlson score of 2+ were more likely to receive etoposide with cisplatin or carboplatin and radiation (OR: 2.81 (95% CI 1.08, 7.44)). Patients of other/unknown race/ethnicity (OR: 0.29 (95% CI 0.09, 0.94)) or with pneumonia/lung collapse (OR: 0.32 (95% CI 0.15, 0.72)) were less likely to receive any chemotherapy and radiation. Thirty-five percent of patients (n=84) determined to be LS received radiation to a non-primary site as part of their initial therapy.

Table 2.

Demographic and Provider Characteristics Associated with Treatment by Stage of Disease, Multivariate Logistic Regression, Odds Ratios (95% Confidence Intervals)

	Limited Stage Disease		Extensive Stage Disease
	Receipt of Etoposide with Cis or Carbo-platin only + Radiation (Yes vs. No)	Receipt of Any Chemotherapy+ Radiation (Yes vs. No)	Receipt of Etoposide with Cis or Carbo-platin Only (Yes vs. No)	Receipt of Any Chemotherapy (Yes vs. No)
	N=288	N=288	N=643	N=643
Age at diagnosis
<60	Ref	Ref	Ref	Ref
60–69	0.88 (0.33, 2.36)	0.96 (0.26, 3.51)	2.14 (1.01, 4.53)	1.74 (0.88, 3.45)
70–79	0.53 (0.18, 1.56)	0.52 (0.15, 1.81)	1.75 (0.79, 3.88)	0.92 (0.46, 1.85)
80+	0.16 (0.04, 0.64)	0.06 (0.01, 0.29)	0.63 (0.25, 1.57)	0.43 (0.20, 0.94)
Sex
Male	Ref	Ref	Ref	Ref
Female	0.63 (0.27, 1.44)	0.43 (0.16, 1.17)	1.03 (0.61, 1.74)	1.19 (0.75, 1.90)
Race/Ethnicity
White	Ref	Ref	Ref	Ref
Black	1.06 (0.43, 2.59)	0.91 (0.30, 2.75)	1.06 (0.63, 1.79)	1.20 (0.73, 1.98)
Hispanic	0.43 (0.16, 1.14)	0.37 (0.11, 1.20)	1.62 (0.91, 2.90)	1.31 (0.74, 2.33)
Other/Unknown	0.47 (0.14, 1.55)	0.29 (0.09, 0.94)	0.99 (0.49, 2.02)	1.51 (0.85, 2.67)
Insurance
Private	Ref	Ref	Ref	Ref
Any Medicaid	1.28 (0.39, 4.17)	0.67 (0.18, 2.56)	1.14 (0.58, 2.23)	0.82 (0.42, 1.60)
Medicare only	1.26 (0.46, 3.42)	0.56 (0.20, 1.54)	0.90 (0.40, 2.02)	0.92 (0.46, 1.87)
IHS/Tricare/VA/Military	2.55 (0.43, 15.18)	0.75 (0.13, 4.40)	1.56 (0.42, 5.71)	0.81 (0.23, 2.87)
Unknown/No Insurance	0.21 (0.02, 1.85)	0.15 (0.01, 1.79)	1.35 (0.34, 5.41)	1.66 (0.45, 6.17 )
Charlson Comorbidity Score
0	Ref	Ref	Ref	Ref
1	2.02 (0.84, 4.90)	1.16 (0.44, 3.08)	1.12 (0.61, 2.04)	0.94 (0.52, 1.69)
2+	2.81 (1.06, 7.44)	1.98 (0.68, 5.74)	0.72 (0.33, 1.60)	0.73 (0.36, 1.48)
Pneumonia/Lung Collapse
No/Unknown	Ref	Ref	Ref	Ref
Yes	0.54 (0.25, 1.19)	0.32 (0.15, 0.72)	1.04 (0.59, 1.83)	1.18 (0.71, 1.97)
Residency Program
No/Unknown	Ref	Ref	Ref	Ref
Yes	0.84 (0.40, 1.74)	0.48(0.22, 1.01)	1.64 (0.92, 2.94)	2.01 (1.19, 3.41)
Hospital Ownership
Not-for-Profit	Ref	Ref	Ref	Ref
For Profit	0.51 (0.12, 2.22)	0.22 (0.05, 1.00)	1.24 (0.45, 3.38)	1.78 (0.72, 4.37)
Government/Unknown	0.48 (0.18, 1.28)	0.39 (0.14, 1.08)	0.86 (0.48, 1.54)	0.80 (0.44, 1.45)

Extensive-Stage Disease

In patients with ES disease at presentation, 30.2% received etoposide with cisplatin or carboplatin and 9.8% received ”other” chemotherapy as the primary treatment with no radiation given (Table 1). A further 18.8% received etoposide with cisplatin or carboplatin and 6.3% ”other” chemotherapy with thoracic radiation. However, in the community and academic settings 32.8% of patients with ES disease received no chemotherapy or radiation, including 3.4% (n=33) who refused any type of treatment; 18% of those <60 years and 74% of those 80+ years received no treatment (data not shown). Less than 1% (n=9) of patients were entered onto a clinical trial. A PET scan was part of the diagnosis and treatment planning for 26.4% of patients initially diagnosed as ES.

After adjusting for patient and provider characteristics, we found that only age was associated with receipt of etoposide plus cisplatin or carboplatin with no radiation and age and the presence of an approved residency program at the treatment site was associated with receipt of any chemotherapy (Table 2). Patients age 60–69 were more likely to receive etoposide plus cisplatin or carboplatin (OR: 2.14 (95% CI 1.01, 4.53)). Patient age ≥80 were less likely to receive any chemotherapy (OR: 0.43 (95% CI 0.20, 0.94)). Patients treated in hospitals with residency programs were more than twice as likely to receive chemotherapy compared to facilities without a residency program (OR: 2.01 (95% CI 1.19, 3.41)).

12-month Mortality

Limited-Stage Disease

More than 43% of patients with LS disease in our study died within 12-months following diagnosis. After taking patient and hospital factors into account, we found that the type of treatment patients received was highly associated with 12-month relative hazard of death (Table 3). Compared to patients treated with chemoradiation, patients treated with radiation alone (HR: 11.66 (95% CI 5.09, 26.67)) and those receiving no treatment (HR:16.12 (95% CI 7.89, 32.95)) experienced significantly poorer survival. Race/ethnicity was significantly associated with mortality. Overall, patients of black and other/unknown race/ethnicity experienced significantly lower relative hazard of death. Finally, patients treated in hospitals with an approved residency program experienced significantly higher relative mortality compared to facilities without training programs (HR: 1.84 (95% CI 1.00, 3.36)).

Table 3.

Patient and Hospital Factors Associated with 12 Month Relative Hazard of Death-Cox Proportional Hazard Models

	Limited-Stage Disease	Extensive-Stage Disease
	Hazard Ratio (95% CI) N=288	Hazard Ratio (95% CI) N=643
Age at diagnosis
<60	Ref	Ref
60–69	1.30 (0.44, 3.79)	1.31 (0.91, 1.88)
70–79	2.16 (0.91, 5.14)	1.66 (1.12, 2.45)
80+	1.81 (0.66, 4.97)	1.32 (0.84, 2.07)
Sex
Male	Ref	Ref
Female	0.96 (0.54, 1.73)	0.67 (0.51, 0.90)
Race/Ethnicity
White	Ref	Ref
Black	0.41 (0.17, 0.94)	1.16 (0.90, 1.49)
Hispanic	0.59 (0.24, 1.46)	0.73 (0.52, 1.02)
Other/Unknown	0.33 (0.14, 0.75)	0.79 (0.52, 1.20)
Insurance
Private	Ref	Ref
Any Medicaid	1.10 (0.42, 2.85)	0.91 (0.63, 1.33)
Medicare only	0.81 (0.41, 1.60)	1.48 (1.06, 2.06)
IHS/Tricare/VA/Military	1.80 (0.49, 6.61)	0.98 (0.42, 2.30)
Unknown/No Insurance	0.71 (0.20, 2.51)	1.28 (0.68, 2.41)
Charlson Comorbidity Score
0	0.48 (0.24, 0.96)	1.03 (0.76, 1.41)
1	Ref	Ref
2+	0.89 (0.47, 1.70)	1.15 (0.74, 1.78)
Pneumonia/Lung Collapse
No/Unknown	Ref	Ref
Yes	1.41 (0.73, 2.70)	1.50 (1.12, 2.01)
Treating Hospital Characteristics
Residency Program
No/Unknown	Ref	Ref
Yes	1.84 (1.00, 3.36)	0.76 (0.57, 1.00)
Hospital Ownership
Not-for-profit	Ref	Ref
For Profit	0.78 (0.30, 2.00)	0.87 (0.53, 1.44)
Government/Other	0.79 (0.36, 1.72)	0.94 (0.67, 1.32)
Treatment
No or unknown treatment	16.12 (7.89, 32.95)	5.45 (3.71, 8.01)
RT only	11.66 (5.09, 26.67)	3.60 (2.33, 5.58)
Any Chemo only	1.79 (0.59, 5.49)	Ref
Any Chemo & RT	Ref	0.80 (0.55, 1.16)
Overall-Unadjusted 12-Month Mortality	43.3% (N=129)	72.5% (N=469)

Extensive-Stage Disease

Of patients with ES disease treated in the community and academic settings 72% were deceased 12-months following diagnosis. We found treatment received was highly associated with 12-month relative hazard of death after adjusting for patient and hospital factors (Table 3). As with patients diagnosed with LS disease, those treated with radiation therapy only (HR: 3.60 (95% CI 2.33, 5.58)) or received no treatment (HR: 5.45 (95% CI 3.71, 8.01)) experienced significantly poorer survival compared to patients treated with any chemotherapy. Additionally, female patients had significantly lower relative hazard of death compared to males (HR: 0.67 (95% CI 0.51, 0.90)). Finally, patients age 70–79 (HR: 1.66 (95% CI 1.12, 2.45)), patients with Medicare only (HR: 1.48 95% CI 1.05, 2.06)) and patients with pneumonia/lung collapse experienced higher 12-month relative hazard of death (HR: 1.50 (95% CI 1.12, 2.01)).

DISCUSSION

Among all patients in this study, we found approximately 29% presented with LS disease, while over 71% were diagnosed ES disease; these study prevalences are similar to previous national estimates.[17] However, few studies have evaluated how these patients with SCLC are treated in community and academic settings.

We found that 59.5% of patients with LS disease received chemoradiation containing the etoposide plus cisplatin or carboplatin doublet. For ES disease 30.2% received the cisplatin or carboplatin with etoposide regimen only as the initial treatment. Patients who received chemoradiation for LS or chemotherapy for ES experienced significantly lower mortality compared to those who received no therapy or radiation only. Elderly patients (≥80 at diagnosis), patients of other/unknown race/ethnicity, and those with pneumonia/lung collapse were less likely to receive any chemotherapy plus radiation for LS disease. Patients treated in hospitals with residency programs were more likely to receive some chemotherapeutic agents for ES disease. These findings provide further insight into specific segments of the U.S. population with SCLC who might benefit from efforts to improve receipt of standard therapy.

Overall, our findings indicate more extensive use of etoposide with cisplatin or carboplatin treatment for SCLC compared to studies of practice patterns in large-hospital and comprehensive cancer centers more than 2 decades ago.[17] In a study examining treatment patterns of patients in the 1988–91 National Cancer Database, Fry et al. found that 33% of patients with AJCC stage I and II disease were treated with chemotherapy and radiation while 39% of AJCC stage III and IV were treated with chemotherapy only.[17] Although use of cisplatin or carboplatin with etoposide was still somewhat low during this period, use of this therapy is more widespread than previous studies report. Cisplatin or carboplatin and etoposide were given as a doublet to 71.5% of LS and 49% of ES patients in our study. We did not define the use of a platin/irinotecan combination as a US standard therapy in our study. During the study period its use was not frequent. Only 23 (4.9%) patients with ES received irinotecan in a doublet with a platinum agent as part of their initial therapy. Although the use of irinotecan may have increased after 2007, definitive results of SWOG 0124 were not enough to change the standard of care at least in the US.[22]

This study allows us to identify those individuals who are at higher risk for not receiving recommended treatment. In a 2002 study of SCLC patients treated in the British Columbia Cancer Agency, increasing age and comorbidity were both associated with lower use of combined chemoradiation.[26] Our study expands this treatment profile to evaluate treatment of patients with both LS and ES disease in the U.S. community and academic setting. Among patients with LS disease, older patients were less likely to receive etoposide with cisplatin or carboplatin plus radiation for their disease, but those with a Charlson score of 2+ were more likely to receive this therapy. This may be a result of physicians providing a more complete recording of comorbidities for patients slated to receive more extensive therapy. In addition, radiation plus any chemotherapy was prescribed less frequently for patients 80 or older, other/unknown race/ethnicity and for patients with pneumonia or lung collapse. We also found that 34% of patients with LS disease received radiation to a non-primary site as part of their initial therapy. Because of limitations on how the data were collected, we are unable to determine the exact radiation site. It is reasonable to speculate that the majority was directed to the use of PCI after initial good control of their disease as per the established standard at the time of the study period.[7]

For patients with ES disease, however, only age ≥80 and the presence of a residency program where they were treated was associated with receiving any chemotherapy. These findings may be due to a combination of both patient and structural factors. Specifically, facilities with residency programs were also more likely to be large, non-profit facilities that typically have greater access to specialized resources for dedicated cancer care. Combined, these results indicate that a patient’s risk profile and treating hospital characteristics may play a significant role in treatment planning for SCLC. Age as the only factor should not bias against receiving standard treatment when other variables are not increasing patient risk. Future studies should evaluate the most effective mechanisms for, at a minimum, promoting receipt of standard treatment regimens while encouraging entry onto clinical trials.

Our study also adds to the growing body of literature examining factors that influence survival after SCLC diagnosis. Overall, our study confirms the poor prognosis of patients with SCLC in both community and academic settings, with survival rates that are comparable to other single institution and comprehensive cancer center settings.[17,26–29] We found that LS patients treated in hospitals with residency programs experienced significantly higher mortality compared to those treated in a hospital with no approved residency program. Although underlying reasons are not clear, this finding might reflect characteristics of patients choosing specialized centers, where residents are likely present (i.e., lower functional status, more serious comorbidities, and perhaps more disseminated disease, even within limited stage).

Interestingly, this was not true for patients with ES. However, this is not a randomized controlled trial and the number of patients receiving “other” agents was relatively small. This may account for the lack of significant difference between the use of etoposide and ciplatin or carboplatin and other agents. Future studies might examine the impact of other factors such as patient frailty that might be associated with survival in this population. Importantly, female gender, other and unknown race/ethnicity and the absence of pneumonia/lung collapse are important factors for improved survival with ES disease. The specific factors leading to improved outcome in female patients needs to be further investigated. Other/unknown race in our study included Asian Americans, who historically have experienced better overall cancer survival relative to other racial/ethnic groups.[30] Improved outcome among patients without pneumonia/lung collapse may reflect a worse functional status and the inability of patients with pneumonia/lung collapse to receive standard therapy. Further, we found that patients with Medicare coverage only and those aged 70–79 had poorer survival. Medical care is costly and patients with Medicare alone have significant out of pocket costs. Patients may choose not to receive all the care that is indicated due to these costs. Interestingly, patients age 70–79 had a higher hazard of death, which may reflect differences in functional status compared to those less than age 60 at diagnosis. Finally, those who did not receive treatment may have had such extensive disease that they were not candidates for therapy, and as a result experienced higher mortality.

Our study provides information regarding the use of PET scans for diagnosis and treatment planning in the community and academic setting. Our finding that 43% of patients with LS disease and 26.4% of those with ES disease received a PET scan during the course of diagnosis and treatment is similar to other studies. In a 2010 study, Dinan et al. found that 35–53% of Medicare beneficiaries with incident lung cancer including all histologies diagnosed between 1999–2006 received PET scans during their initial treatment.[31] The implications for the use of PET scans in the initial staging strategy of LS disease are evident since a more accurate determination of stage will affect prognosis and treatment. The appropriate integration and eventual value of this technology in ES SCLC remains to be well established.

Finally, an important concern is the very low enrollment on clinical trials in this population, with participation rates less than 1% among patients with both LS and ES disease. Most efforts for improving survival for both LS and ES SCLC have focused on the use of novel therapeutic approaches in the clinical trial setting. The lack of significant progress in new systemic therapies for SCLC in the past two decades is reflective of the little progress in trials of therapy.[16] Identifying new therapies based on improved knowledge in the biology and molecular traits of SCLC will be crucial for improving the low-survival rates. Further, access to clinical trials can be a real concern for newly diagnosed patients as they create their treatment plan with their oncologist. Only 53 first-line trials were open between January and December 2007 according to clinicaltrials.gov, which could create logistical challenges for enrollment such as long distances to enrolling facilities, travel and other out-of-pocket costs.

Although our study provides insight into the practice patterns and outcomes of both LS and ES SCLC patients treated in the US population, we acknowledge several data-related limitations. First, the available data do not include measures of frailty, degree of weight loss or performance status in the patients and these may have influenced eligibility for or receipt of chemotherapy and/or radiation for SCLC. However, we adjusted for comorbidities in the Charlson score, which has been shown to be associated with 1- and 10-year mortality in hospitalized patients. Additionally, the data do not include information on the timing of pneumonia/lung collapse in patients. Our study data do not include information on all factors, including patient or physician preferences in determining treatment planning that may influence treatment or outcomes. Future studies should prospectively plan to assess how these factors may impact treatment and outcome. Finally, as mentioned earlier, because of limitations in data collection, including information on the specific type of remission, we cannot provide definitive information regarding the use of PCI.

Acknowledging these limitations, our study is one of the first to evaluate treatment practices in a population-based sample of patients with SCLC in both community and academic settings and to assess survival outcomes. Specifically, we provide further insight into specific segments of the U.S. population with SCLC who may benefit from efforts to improve receipt of standard therapy. As a result, future efforts to improve receipt of standard therapy might focus on patient-related factors such as older age for LS disease, while examining hospital or system-level factors for improving receipt of standard therapy in ES disease.

In conclusion, in our study of patients with SCLC treated in the community and academic setting, we found that still a somewhat large proportion of patients are not treated with a standard treatment regimen due to a mixture of patient and system-related factors. Efforts to promote and facilitate clinical trial enrollment to test novel therapies should be explored further.