Abstract
Background:
Population-based estimates of absolute risk of lung cancer recurrence, and of mortality rates after recurrence, can inform clinical management.
Methods:
We evaluated prognostic factors for recurrences and survival in 2098 lung cancer case patients from the general population of Lombardy, Italy, from 2002 to 2005. We conducted survival analyses and estimated absolute risks separately for stage IA to IIIA surgically treated and stage IIIB to IV non–surgically treated patients.
Results:
Absolute risk of metastases exceeded that of local recurrence in every stage and cell type, highlighting the systemic threat of lung cancer. In stage I, the probability of dying within the first year after diagnosis was 2.7%, but it was 48.3% within first year after recurrence; in stage IV, the probabilities were 57.3% and 80.6%, respectively. Over half the patients died within one year of first metastasis. Although in stages IA to IB about one-third of patients had a recurrence, stage IIA patients had a recurrence risk (61.2%) similar to stage IIB (57.9%) and IIIA (62.8%) patients. Risk of brain metastases in stage IA to IIIA surgically treated non–small cell lung cancer patients increased with increasing tumor grade. Absolute risk of recurrence was virtually identical in adenocarcinoma and squamous cell carcinoma patients.
Conclusions:
This population-based study provides clinically useful estimates of risks of lung cancer recurrence and mortality that are applicable to the general population. These data highlight the need for more effective adjuvant treatments overall and within specific subgroups. The estimated risks of various endpoints are useful for designing clinical trials, whose power depends on absolute numbers of events.
Lung cancer is the leading cause of cancer deaths worldwide (1). The American Cancer Society estimated that 159260 Americans will die of it in 2014 (2). Up to 70% of surgically treated stage IA lung cancer patients are alive five years later, compared with 2% to 13% with stage IV disease (3,4). Surgical resection strongly impacts survival, and it is routinely performed for stage I, II, and selected stage III lung cancer patients. Nevertheless, 30% to 70% of postoperative lung cancer patients develop one or more recurrent lesions, and another 2% to 5% of patients have a second tumor of the lung (5,6). Patients whose lung cancers have spread locally or systemically when first detected, constituting 70% of all patients (7–9), usually receive chemotherapy and/or radiation therapy instead of surgery.
Local extension and metastases are primary causes of death in lung cancer patients. Identifying subsets of patients at high risk for recurrence and mortality following recurrence may lead to interventions that improve survival. Reliable data are needed to help clinicians understand recurrence risk in various patient subgroups, which would require large studies with well-defined methods and follow-up (10). We therefore investigated overall survival, recurrence rates and patterns, and mortality following recurrence in population-based data from the Environment And Genetics in Lung cancer Etiology (EAGLE) study (11).
EAGLE was conducted from 2002 to 2005 in the Lombardy region of Italy, where over nine million people were served by a network of modern hospitals and health services with universal coverage. We estimated incident recurrence rates to various sites by stage and other clinical characteristics and the absolute risk (sometimes called crude risk or cumulative incidence) of recurrences by taking into account competing risks (12). Absolute risk is useful for clinical management, because the chance of having a recurrence is, in fact, reduced by the risk of mortality before recurrence. To our knowledge, this is the first population-based study to examine absolute risk of recurrence in lung cancer. Finally, we quantified the probability of dying following recurrence and the impact of the site of recurrences on subsequent death rates.
Methods
The EAGLE study enrolled 2098 (84.8%) of the 2473 patients with lung cancer diagnosed from April 22nd, 2002, to February 28th, 2005, in the defined catchment area, including 216 municipalities within Lombardy. To verify whether enrolled patients were similar to those who refused to participate, we compared their distributions by age, sex, and area of residence (Supplementary Table 1, available online). Women and patients age 65 years and older refused participation more than men or younger patients. Because we enrolled the large majority of cases, however, our study cases were only slightly enriched in men (78.6%) compared with 77.8% of all case patients and only slightly enriched in those younger than age 65 years (38.8%) compared with 36.6% of all case patients. Thus, the enrolled patients were quite representative of all the case patients.
Institutional review boards at the National Cancer Institute and all local hospitals approved the study, and patients signed an informed consent. Comprehensive epidemiological and behavioral information was collected from each participant as previously described (11).
Detailed information on tumor characteristics, recurrence, treatment, and follow-up data were recovered from patients’ medical records at the end of the follow-up. The lung cancer diagnosis was based on clinical criteria and was by pathology reports from surgery, biopsy, or cytology samples in approximately 95% of cases and on clinical history and imaging for the remaining 5%. To verify the diagnosis, a team composed of two physicians (DC and ACP) and one pathology technician (MR) examined the clinical history, bronchoscopy, and biopsy results; all available imaging; and hospital discharge letters for each case. In addition, the team reviewed surgery and pathology reports for surgical cases and biopsies and/or cytology reports for nonsurgical cases. All information was recorded on optically scannable forms by the physicians and coded accordingly. Histological diagnoses from 10% of cases were reviewed and confirmed by an experienced independent pulmonary pathologist from the National Cancer Institute, National Institutes of Health.
Although staging was originally coded using the 6th TNM Classification of Malignant Tumors (13) in EAGLE, the study team reviewed all clinical data and revised the staging using the 7th edition (3), reflecting the newest International Association for the Study of Lung Cancer (IASLC) recommendations (4). Staging was based on pathological assessment for surgically treated patients and pathological, when available, or clinical assessment for non–surgically treated case patients. For patients initially diagnosed with a stage IV tumor, and thus presenting with a distant metastasis, only new lesions discovered at least three months after the initial diagnosis were counted as disease progression.
After study completion, we identified follow-up visits and hospital admissions by linkage with the region-wide Regional Health Authority database of hospital admissions. Recurrence history was ascertained through December 31, 2010. Detailed clinical information was retrieved for each hospitalization and/or out-patient visit and reviewed by the clinical team that coded local recurrence and metastases.
“Any recurrence” refers to local recurrence or metastases. We defined “local recurrence” as the first new lesion in the ipsilateral lung or in thoracic organs/tissues (including mediastinal or supraclavicular lymph nodes, thoracic wall, trachea, or heart) during patients’ follow-up visits. First postdiagnosis lesions that developed at sites not included in this list were classified as “metastases.” These include case patients with a single distant lesion in the contralateral or bilateral lung(s), liver, brain, or bone(s), as well case patients classified as “other” metastases with lesions in other organs/tissues or multiple lesions identified at the same time, of which at least one was distant. Single recurrences in the lung(s) with unspecified ipsi- or contralateral site and recurrences with unspecified site were also classified as “other” metastases. For lesions of the esophagus, pericardium, pleura, skin/subcutaneous tissue, and muscle, the local or distant nature of the lesion was established on a case-by-case basis (3).
Vital status was ascertained through December 31, 2010, using population databases or postal follow-up to the municipality of last residence in all Italian territory. Death certificates were obtained from local health units of last residence, reviewed (and revised for 62 case patients) by the study team, and coded by trained personnel (MR) according to the International Classification of Diseases, 10th Revision (ICD-10).
To calculate mortality rates after lung cancer diagnosis, we used all 2098 case patients in EAGLE. To characterize risk of recurrence in well-defined subgroups, we studied separately the 760 surgically treated (ST) patients in stages IA to IIIA and the 971 non–surgically treated (NST) patients in stages IIIB to IV. In these recurrence analyses, we excluded 258 non–surgically treated patients from stages IA to IIIA, 61 surgically treated patients from stages IIIB to IV, and 48 non–surgically treated patients in stage X (unknown). Data on these excluded patients are in the Supplementary Material (available online).
Statistical Analysis
We analyzed (1) survival/mortality after lung cancer diagnosis, (2) survival/mortality after first recurrence, and (3) rates of first recurrence after lung cancer diagnosis.
Chi-squared tests were calculated to compare proportions. We used the Kaplan-Meier estimator of survival curves and calculated log-rank tests to compare survival across selected variables. We then fitted Cox proportional hazards model to estimate hazard ratios (HRs) and their 95% confidence intervals (CIs). Follow-up began at date of diagnosis and ended at the earliest of date of death, loss-to-follow-up (ie, the last known therapy or diagnostic follow-up visit), or at five years after first diagnosis. When analyzing the effect of first recurrence on mortality, history of recurrence was treated as a time-dependent covariate (ie, follow-up time for subjects was considered “without previous recurrence” until the day before the date of recurrence diagnosis and “after recurrence” thereafter). We calculated crude mortality rates before and after first recurrence and crude mortality hazard ratios from Cox models with time-dependent recurrence status. One-year mortality risks were then estimated from annual mortality rates using this formula: Risk = 1 – exp(–Rate×1).
Stage-specific nonparametric estimates of the absolute risk of first recurrences were obtained for all recurrences and separately for local and distant recurrences based on Gaynor et al. (14), as implemented in the Stata command stcompet (15). Mortality was always considered to be a competing risk; when estimating the absolute risk of local or distant first recurrence, an earlier distant or local recurrence, as applicable, was considered to be a competing risk. In supplementary analyses, we calculated histology-specific absolute risks of recurrence separately for adenocarcinomas, squamous cell carcinomas, and small cell carcinomas. For all recurrences, we also estimated (1) recurrences rates and their 95% confidence intervals; (2) crude hazard ratios; (3) hazard ratios adjusted for sex, age at diagnosis (<65 and ≥65 years), cigarette smoking (never, former, and current), stage, morphology; and (4) hazard ratios additionally adjusted for therapy (none, surgery, chemotherapy, and radiation therapy).
We evaluated the impact of age at diagnosis, sex, cigarette smoking status, stage, histology, grade, therapy, and weight change before diagnosis on survival by fitting Cox models adjusted for stage and, separately, stratified by stage at lung cancer diagnosis (I, II, II, IV). We analyzed weight loss in the year preceding initial diagnosis as a percentage of the baseline weight one year before the lung cancer diagnosis using the following categories: no change, weight gain, moderate weight loss (1% to 9%), and marked weight loss (10% to 45%).
After fitting Cox models we tested the proportional hazards (PH) assumption on the basis of Schoenfeld residuals (Stata command: estat phtest) for all recurrences combined and mortality. Only the hazard ratios of NST patient mortality comparing stage IV with IIIB and the hazard ratios of overall recurrences by morphology in ST patients and by grade in NST patients varied over time. We tested the homogeneity of hazard ratios of recurrence (occurence, type, and tissue) across stage using Wald tests for the interaction term(s). The analyses were performed with Stata 13 (StataCorp, College Station, TX).
All statistical tests were two-sided, and P values of less than .05 were considered statistically significant.
Results
Table 1 describes clinical characteristics of surgical and nonsurgical patients in stages I to IIIA and IIIB to IV. Almost 80% were male, mean age ranged from 63.7 to 68.8 years, median follow-up ranged from 0.7 to 4.1 years, and the most frequent categories included current smokers, tumors localized to the upper lobes, adenocarcinoma subtype, and stage IV.
Table 1.
Selected characteristics of 2050 lung cancer patients, stages I to IV*, enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005, by stage and treatment
| Stage IA to IIIA | Stage IIIB and IV | |||
|---|---|---|---|---|
| Surgical† | Nonsurgical | Surgical† | Nonsurgical | |
| Patients’ characteristics | No. (%) | No. (%) | No. (%) | No. (%) |
| Total | 760 (100) | 258 (100) | 61 (100) | 971 (100) |
| Sex | ||||
| Male | 590 (77.6) | 206 (79.8) | 49 (80.3) | 761 (78.4) |
| Female | 170 (22.4) | 52 (20.2) | 12 (19.7) | 210 (21.6) |
| Age at diagnosis, y‡ | 66.4 (8.2) | 68.8 (7.8) | 63.7 (8.6) | 65.9 (8.6) |
| Follow-up, y‡ | 3.4/4.1 (1.7–5.0) | 1.8/1.4 (0.8–2.2) | 2.2/1.6 (0.7–3.1) | 1.0/0.7 (0.3–1.2) |
| Cigarette smoking | ||||
| Never | 52 (6.8) | 15 (5.8) | 5 (8.2) | 75 (7.7) |
| Former (quit >6 mo) | 339 (44.6) | 112 (43.4) | 28 (45.9) | 374 (38.5) |
| Current | 363 (47.8) | 129 (50.0) | 28 (45.9) | 501 (51.6) |
| NA | 6 (0.8) | 2 (0.8) | 0 (0.0) | 21 (2.2) |
| LC site (ICD-10) | ||||
| Trachea (C33) | 0 (0.0) | 2 (0.8) | 0 (0.0) | 1 (0.1) |
| Main Bronchus (C34.0) | 26 (3.4) | 39 (15.1) | 0 (0.0) | 132 (13.6) |
| Upper lobe (C34.1) | 433 (57.0) | 140 (54.3) | 36 (50.0) | 484 (49.8) |
| Middle lobe (C34.2) | 36 (4.7) | 12 (4.6) | 4 (6.6) | 39 (4.0) |
| Lower lobe (C34.3) | 247 (32.5) | 61 (23.6) | 16 (26.2) | 199 (20.5) |
| Multiple/NA (C34.8/9) | 18 (2.4) | 4 (1.6) | 5 (8.2) | 116 (11.9) |
| LC side | ||||
| Right | 409 (53.8) | 139 (53.9) | 32 (52.5) | 526 (54.2) |
| Left | 348 (45.8) | 116 (45.0) | 27 (44.3) | 387 (39.9) |
| Other/NA | 3 (0.4) | 3 (1.2) | 2 (3.3) | 58 (6.0) |
| LC morphology | ||||
| NSCLC | 722 (95.0) | 195 (75.6) | 57 (93.4) | 700 (72.1) |
| Adenocarcinoma | 410 (54.0) | 65 (25.2) | 41 (67.2) | 334 (34.4) |
| SCC | 238 (31.3) | 86 (33.3) | 11 (18.0) | 181 (18.6) |
| Large cell carcinoma | 54 (7.1) | 3 (1.2) | 2 (3.3) | 36 (3.7) |
| NSCLC, NOS | 20 (2.6) | 41 (15.9) | 3 (4.9) | 149 (15.4) |
| SCLC | 17 (2.2) | 36 (13.9) | 0 (0.0) | 152 (15.7) |
| Other | 21 (2.8) | 10 (3.9) | 4 (6.6) | 41 (4.2) |
| NA | 0 (0.0) | 17 (6.6) | 0 (0.0) | 78 (8.0) |
| Stage§ | ||||
| IA | 171 (22.5) | 16 (6.2) | - - | - - |
| IB | 185 (24.3) | 20 (7.7) | - - | - - |
| IIA | 129 (17.0) | 8 (3.1) | - - | - - |
| IIB | 95 (12.5) | 22 (8.5) | - - | - - |
| IIIA | 180 (23.7) | 192 (74.4) | - - | - - |
| IIIB | - - | - | 20 (32.8) | 193 (19.9) |
| IV | - - | - | 41 (67.2) | 778 (80.1) |
| Grade | ||||
| I | 46 (6.0) | 4 (1.5) | 2 (3.3) | 17 (1.8) |
| II | 271 (35.7) | 13 (5.0) | 18 (29.5) | 37 (3.8) |
| III | 361 (47.5) | 68 (26.4) | 30 (49.2) | 308 (31.7) |
| NA | 82 (10.8) | 173 (67.1) | 11 (18.0) | 609 (62.7) |
| Therapy | ||||
| None | - - | 39 (15.1) | - - | 158 (16.3) |
| Surgery only | 367 (48.3) | - - | 9 (14.8) | - - |
| CT only | 122 (16.1) | 78 (30.2) | 13 (21.3) | 403 (41.5) |
| RT only | 79 (10.4) | 21 (8.1) | 6 (9.8) | 54 (5.6) |
| CT + RT | 191 (25.1) | 120 (46.5) | 33 (54.1) | 355 (36.6) |
| CT/RT NA | 1 (0.1) | 0 (0.0) | 0 (0.0) | 1 (0.1) |
* Not including 48 patients with stage X (16 untreated, 20 treated with CT only, 12 treated with CT+RT). CT = chemotherapy; ICD-10 = International Classification of Diseases, 10th Revision; LC = lung cancer; NA = not available; NOS = not otherwise specified; NSCLC = non–small cell lung carcinoma; RT = radiation therapy; SCC = squamous cell carcinoma; SCLC = small cell lung carcinoma.
† Including patients treated exclusively with surgery and patients also treated with CT/RT.
‡ Mean (SD) or mean/median (25th–75th percentile).
§ Pathological (when available) or clinical stage according to the TNM Classification of Malignant Tumors, 7th edition.
Survival Prognosis
Five years after the diagnosis of lung cancer, 19.0% of patients were alive. Among the deceased, 95.4% died from lung cancer (Table 2). The following analyses describe the effects of stage (7th TNM Edition), surgical treatment, tumor histology, and weight loss on survival and lung cancer mortality rates.
Table 2.
Vital status within 5 years since diagnosis for 2098 lung cancer patients enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005
| Vital status | No. (%) |
|---|---|
| Total | 2098 (100) |
| Alive | 399 (19.0) |
| Lost to follow-up | 7 (0.3) |
| Dead* | 1692 (80.6) |
| Deaths from lung cancer (ICD-10: C33-C34) | 1614 (95.4†) |
| Deaths from other causes | 78 (4.6†) |
* Not including 67 deaths that occurred more than five years after diagnosis: 48 from lung cancer (ICD-9: C34), 16 from other defined causes, three from unknown cause (ICD-9: R99). ICD-10 = International Classification of Diseases, 10th Revision.
† Denominator: 1692 deaths.
For all 2098 patients (Supplementary Figure 1, available online) and for the subgroups of 760 stage I to IIIA surgically treated (ST) and 971 stage IIIB to IV non–surgically treated (NST) patients (Figure 1), stage-specific Kaplan-Meier survival curves decreased strongly with increasing stage at diagnosis. Among ST patients compared with stage IA (reference with 48 deaths), the stages IB (66 deaths), IIA (70 deaths), IIB (60 deaths), and IIIA (120 deaths) had respective hazard ratios of 1.34 (95% CI = 0.92 to 1.94), 2.37 (95% CI = 1.64 to 3.43), 3.10 (95% CI = 2.12 to 4.53), and 3.29 (95% CI = 2.35 to 4.60). Survival for stage IIA ST patients was very similar to that of stage IB ST patients for the first year after diagnosis but worse thereafter. By contrast, among 971 NST patients, only a modest difference in survival was detected between stage IIIB (182 deaths, reference) and stage IV (753 deaths, HR = 1.46, 95% CI = 1.24 to 1.72) (Figure 1).
Figure 1.
Survival (Kaplan-Meier function) by stage according to the 7th edition of the TNM Classification of Malignant Tumors among 760 surgically treated stages IA to IIIA (left panel) and 971 non-surgically treated stages IIIB and IV (right panel) lung cancer patients enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005.
As expected, the 258 stage I to IIIA NST patients had lower survival (223 deaths, HR = 3.46, 95% CI = 2.91 to 4.10) than the stage I to IIIA surgically treated patients (364 deaths) (Supplementary Figure 2, available online), and outcome was more favorable in the 61 surgically treated stage IIIB to IV patients (47 deaths, HR = 0.40, 95% CI = 0.30 to 0.54) than in the non–surgically treated individuals with the same stage (935 deaths) (Supplementary Figure 2, available online).
We examined the impact of tumor histology on survival (Supplementary Figure 3, available online). Among ST patients, those with adenocarcinoma had a slightly more favorable (log-rank P = .007) five-year survival rate (54.4%, 177 deaths) than subjects with squamous cell carcinoma (45.6%, 125 deaths). No substantial difference by histology was observed in NST patients (2 d.f. log-rank P = .34).
We studied the effect of weight loss in the year before lung cancer diagnosis. Among 702 subjects who lost weight, the mortality hazard ratio was 1.24 (95% CI = 1.09 to 1.41) with 1% to 9% weight loss and 1.36 (95% CI = 1.16 to 1.58) with 10% or higher weight loss compared with the 922 subjects without weight loss (Supplementary Table 2, available online).
Local Recurrences and Metastases
Separately for ST and NST patients, the following analyses present total and site-specific percentages with recurrence, recurrence rates by prognostic factors, and recurrence hazard ratios for prognostic factors.
Tables 3 and 4 present the risks of local recurrences and metastases in stage IA to IIIA ST patients and stage IIIB to IV NST patients, respectively. These tables present the proportions with recurrence during follow-up overall and by site (in percent), the overall recurrence rate per person-year, and site-specific relative risks in ST and NST patients. Men and women had similar recurrence patterns, both for ST and NST.
Table 3.
Proportions with recurrence (%), recurrence rates, and Cox hazard ratios of recurrence among 760 surgically treated, stages IA to IIIA lung cancer patients enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005
| Recurrences by organNo. (%) | Rate(95% CI)* | Hazard ratio of recurrence(95% CI) | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All | Local recurrences | Distantmetastases | All | All | Localrecurrences | Distantmetastases | ||||||||||||
| Patients’ characteristics | No.† (%) | LungIpsi | Thoracic‡ | LungCB | Liver | Brain | Bones | Other§ | LungIpsi | Thoracic‡ | LungCB | Liver | Brain | Bones | Other§ | |||
| Total | 760(100) | 375(49.3) | 54(7.1) | 34(4.5) | 58(7.6) | 19(2.5) | 56(7.4) | 26(3.4) | 128(16.8) | 0.17(0.16 to 0.19) | ||||||||
| Sex | ||||||||||||||||||
| Men | 590(100) | 291(49.3) | 44(7.5) | 28(4.7) | 42(7.1) | 14(2.4) | 40(6.8) | 20(3.4) | 103(17.5) | 0.18(0.16 to 0.20) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| Women | 170(100) | 84(49.4) | 10(5.9) | 6(3.5) | 16(9.4) | 5(2.9) | 16(9.4) | 6(3.5) | 25(14.7) | 0.17(0.13 to 0.21) | 0.97(0.76 to 1.24) | 0.76(0.38 to 1.50) | 0.73(0.30 to 1.77) | 1.27(0.71 to 2.25) | 1.22(0.44 to 3.39) | 1.35(0.76 to 2.42) | 1.03(0.41 to 2.57) | 0.82(0.53 to 1.27) |
| Age, y | ||||||||||||||||||
| <65 | 293(100) | 158(53.9) | 26(8.9) | 17(5.8) | 20(6.8) | 5(1.7) | 24(8.2) | 11(3.8) | 55(18.8) | 0.19(0.16 to 0.22) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| 65+ | 467(100) | 217(46.5) | 28(6.0) | 17(3.6) | 38(8.1) | 14(3.0) | 32(6.9) | 15(3.2) | 73(15.6) | 0.16(0.14 to 0.19) | 0.87(0.71 to 1.07) | 0.68(0.40 to 1.16) | 0.64(0.32 to 1.24) | 1.20(0.70 to 2.07) | 1.77(0.64 to 4.91) | 0.84(0.50 to 1.43) | 0.87(0.40 to 1.89) | 0.84(0.59 to 1.19) |
| Stage|| | ||||||||||||||||||
| IA | 171(100) | 58(33.9) | 13(7.6) | 4(2.3) | 9(5.3) | 4(2.3) | 7(4.1) | 3(1.8) | 18(10.5) | 0.09(0.07 to 0.12) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| IB | 185(100) | 70(37.8) | 10(5.4) | 4(2.2) | 14(7.6) | 4(2.2) | 10(5.4) | 2(1.1) | 26(14.1) | 0.11(0.09 to 0.15) | 1.22(0.86 to 1.73) | 0.79(0.34 to 1.79) | 1.01(0.25 to 4.04) | 1.61(0.70 to 3.72) | 1.01(0.25 to 4.03) | 1.45(0.55 to 3.80) | 0.65(0.11 to 3.93) | 1.46(0.80 to 2.66) |
| IIA | 129(100) | 79(61.2) | 12(9.3) | 7(5.4) | 11(8.5) | 6(4.7) | 12(9.3) | 8(6.2) | 23(17.8) | 0.26(0.21 to 0.32) | 2.63(1.87 to 3.69) | 1.84(0.84 to 4.05) | 3.39(0.99 to 11.6) | 2.55(1.05 to 6.16) | 2.77(0.78 to 9.86) | 3.26(1.28 to 8.29) | 4.45(1.18 to 16.8) | 2.43(1.31 to 4.52) |
| IIB | 95(100) | 55(57.9) | 8(8.4) | 6(6.3) | 6(6.3) | 3(3.2) | 7(7.4) | 3(3.2) | 22(23.2) | 0.25(0.19 to 0.33) | 2.77(1.85 to 3.87) | 1.78(0.74 to 4.31) | 4.35(1.23 to 15.4) | 2.00(0.71 to 5.63) | 2.08(0.46 to 9.31) | 2.78(0.97 to 7.92) | 2.42(0.49 to 12.0) | 3.42(1.83 to 6.38) |
| IIIA | 180(100) | 113(62.8) | 11(6.1) | 13(7.2) | 18(10.0) | 2(1.1) | 20(11.1) | 10(5.6) | 39(21.7) | 0.28(0.23 to 0.33) | 2.81(2.04 to 3.86) | 1.27(0.57 to 2.84) | 4.67(1.52 to 14.3) | 3.14(1.41 to 7.02) | 0.70(0.13 to 3.80) | 4.06(1.71 to 9.64) | 4.13(1.13 to 15.0) | 3.08(1.76 to 5.39) |
| Morphology | ||||||||||||||||||
| ADC | 410(100) | 206(50.2) | 34(8.3) | 15(3.7) | 32(7.8) | 11(2.7) | 35(8.5) | 14(3.4) | 65(15.8) | 0.17(0.15 to 0.19) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| SCC | 238(100) | 108(45.4) | 17(7.1) | 11(4.6) | 14(5.9) | 4(1.7) | 11(4.6) | 7(2.9) | 44(18.5) | 0.17(0.14 to 0.20) | 1.00(0.79 to 1.26) | 0.96(0.54 to 1.72) | 1.42(0.65 to 3.10) | 0.85(0.45 to 1.59) | 0.70(0.22 to 2.19) | 0.59(0.30 to 1.17) | 0.92(0.37 to 2.27) | 1.28(0.88 to 1.88) |
| SCLC | 17(100) | 9(52.9) | 0(0.0) | 0(0.0) | 1(5.9) | 1(5.9) | 5(29.4) | 1(5.9) | 1(5.9) | 0.20(0.10 to 0.38) | 1.19¶(0.61 to 2.32) | -- | -- | 0.86(0.12 to 6.33) | 2.47(0.32 to 19.1) | 3.86(1.51 to 9.87) | 1.84(0.24 to 14.0) | 0.42(0.06 to 3.03) |
| Other/NA | 95(100) | 52(54.7) | 3(3.2) | 8(8.4) | 11(11.6) | 3(3.2) | 5(5.3) | 4(4.2) | 18(19.0) | 0.21(0.16 to 0.28) | 1.33¶(0.98 to 1.81) | 0.46(0.14 to 1.51) | 2.92(1.24 to 6.88) | 1.83(0.92 to 3.63) | 1.47(0.41 to 5.27) | 0.75(0.29 to 1.92) | 1.44(0.47 to 4.38) | 1.46(0.86 to 2.46) |
| Grade | ||||||||||||||||||
| I | 46(100) | 16(34.8) | 6(13.0) | 1(2.2) | 3(6.5) | 1(2.2) | 1(2.2) | 1(2.2) | 3(6.5) | 0.09(0.06 to 0.15) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| II | 271(100) | 135(49.8) | 24(8.9) | 6(2.2) | 21(7.8) | 9(3.3) | 18(6.6) | 11(4.1) | 46(17.0) | 0.17(0.15 to 0.21) | 1.81(1.08 to 3.03) | 0.87(0.36 to 2.14) | 1.28(0.15 to 10.6) | 1.56(0.47 to 5.24) | 1.88(0.24 to 14.9) | 3.84(0.51 to 28.8) | 2.17(0.28 to 16.8) | 3.26(1.01 to 10.5) |
| III | 361(100) | 186(51.5) | 20(5.5) | 22(6.1) | 31(8.6) | 8(2.2) | 30(8.3) | 12(3.3) | 63(17.4) | 0.19(0.17 to 0.22) | 2.01(1.21 to 3.36) | 0.59(0.24 to 1.47) | 3.87(0.52 to 28.7) | 1.88(0.57 to 6.14) | 1.36(0.17 to 10.9) | 5.16(0.70 to 37.9) | 1.87(0.24 to 14.4) | 3.61(1.13 to 11.5) |
| NA | 82(100) | 38(46.3) | 4(4.9) | 5(6.1) | 3(3.7) | 1(1.2) | 7(8.5) | 2(2.4) | 16(19.5) | 0.15(0.11 to 0.21) | 1.65(0.92 to 2.96) | 0.46(0.13 to 1.64) | 3.54(0.41 to 30.3) | 0.70(0.14 to 3.49) | 0.69(0.04 to 11.1) | 4.86(0.60 to 39.5) | 1.31(0.12 to 14.5) | 3.70(1.08 to 12.7) |
| Therapy | ||||||||||||||||||
| S only | 367(100) | 73(19.9) | 11(3.0) | 4(1.1) | 17(4.6) | 6(1.6) | 13(3.5) | 3(0.8) | 19(5.2) | 0.05(0.04 to 0.07) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| S+CT | 122(100) | 85(69.7) | 6(4.9) | 6(4.9) | 19(15.6) | 9(7.4) | 4(3.3) | 1(0.8) | 40(32.8) | 0.28(0.23 to 0.35) | 4.93(3.12 to 7.80) | 2.45(0.90 to 6.66) | 2.45(0.90 to 6.66) | 5.04(2.61 to 9.73) | 5.87(2.08 to 16.6) | 1.27(0.41 to 3.91) | 1.18(0.12 to 11.3) | 8.66(5.00 to 15.0) |
| S+RT | 79(100) | 49(62.0) | 4(5.1) | 2(2.5) | 4(5.1) | 1(1.3) | 10(12.7) | 10(12.7) | 18(22.8) | 0.27(0.21 to 0.36) | 4.76(2.78 to 8.13) | 2.79(0.88 to 8.80) | 2.79(0.88 to 8.80) | 1.83(0.61 to 5.46) | 1.10(0.13 to 9.15) | 5.31(2.32 to 12.2) | 19.0(5.20 to 69.7) | 6.57(3.44 to 12.6) |
| S+CT+RT | 191(100) | 168(88.0) | 33(17.3) | 22(11.5) | 18(9.4) | 3(1.6) | 29(15.2) | 12(6.3) | 51(26.7) | 0.48(0.41 to 0.56) | 7.02(4.60 to 10.7) | 12.2(6.06 to 24.5) | 12.2(6.06 to 24.5) | 4.38(2.22 to 8.62) | 1.58(0.39 to 6.41) | 7.70(3.94 to 15.0) | 10.6(2.95 to 38.2) | 9.27(5.43 to 15.8) |
| NA | 1 | 0 | ||||||||||||||||
* Crude recurrence rates per person-year. ADC = adenocarcinoma; CB = contralateral or bilateral; CI = confidence interval; CT = chemotherapy; Ipsi = ipsilateral; NA = not available; RT = radiation therapy; S = surgery; SCC = squamous cell carcinoma; SCLC = small cell lung carcinoma.
† Number (%) of patients.
‡ Including mediastinal (No. 26) or supraclavicular (No. 3) lymph-nodes, thoracic wall (No. 4), and muscle (No. 1).
§ Including abdomen (No. 2), adrenal gland (No. 9), intestine (No. 2), kidney (No. 3), lung unspecified (No. 46), axillary (No. 1) or cervical (No. 1) lymph-nodes, pancreas (No. 1), pleura (No. 1), skin (No. 1), unspecified sites (No. 6), and multiple sites (No. 55).
|| Pathological (when available) or clinical stage according to the TNM Classification of Malignant Tumors, 7th edition.
¶ Hazard ratios of squamous cell carcinoma vs adenocarcinoma and other histologies vs adenocarcinoma varied over time (test of proportional-hazards P = .007 and P = .004, respectively).
Table 4.
Proportions with recurrence (%), recurrence rates, and Cox hazard ratios of recurrence among 971 non-surgically treated, stages IIIB to IV lung cancer patients enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005
| Patients’ characteristics | Recurrences by organ No. (%) |
Rate (95% CI)* |
Hazard ratio of recurrence (95% CI) |
|||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All | Local recurrences | Distant metastases |
All | All | Local recurrences |
Distant metastases |
||||||||||||
| No.† (%) |
Lung Ipsi |
Thoracic‡ | Lung CB |
Liver | Brain | Bones | Other§ | Lung Ipsi |
Thoracic‡ | Lung CB |
Liver | Brain | Bones | Other§ | ||||
| Total | 971 (100) |
258 (26.6) |
8 (0.8) |
2 (0.2) |
7 (0.7) |
30 (3.1) |
71 (7.3) |
74 (7.6) |
66 (6.8) |
0.32 (0.28 to 0.36) |
||||||||
| Sex | ||||||||||||||||||
| Men | 761 (100) |
192 (25.2) |
6 (0.8) |
1 (0.1) |
4 (0.5) |
20 (2.6) |
56 (7.4) |
54 (7.1) |
51 (6.7) |
0.31 (0.27 to 0.36) |
1.00 - |
1.00 - |
1.00 - |
1.00 - |
1.00 - |
1.00 - |
1.00 - |
1.00 - |
| Women | 210 (100) |
66 (31.4) |
2 (1.0) |
1 (0.5) |
3 (1.4) |
10 (4.8) |
15 (7.1) |
20 (9.5) |
15 (7.1) |
0.33 (0.26 to 0.43) |
1.07 (0.81 to 1.41) |
0.88 (0.18 to 4.35) |
3.29 (0.20 to 52.8) |
2.30 (0.51 to 10.3) |
1.50 (0.70 to 3.21) |
0.82 (0.46 to 1.45) |
1.22 (0.73 to 2.04) |
0.91 (0.51 to 1.61) |
| Age, y | ||||||||||||||||||
| <65 | 396 (100) |
129 (32.6) |
4 (1.0) |
2 (0.5) |
5 (1.3) |
16 (4.0) |
33 (8.3) |
36 (9.1) |
33 (8.3) |
0.40 (0.33 to 0.47) |
1.00 - |
1.00 - |
1.00 - |
1.00 - |
1.00- | 1.00- | 1.00- | 1.00- |
| 65+ | 575(100) | 129(22.4) | 4(0.7) | 0(0.0) | 2(0.3) | 14(2.4) | 38(6.6) | 38(6.6) | 33(5.7) | 0.26(0.22 to 0.31) | 0.67(0.52 to 0.85) | 0.63(0.16 to 2.50) | -- | 0.25(0.05 to 1.31) | 0.57(0.28 to 1.17) | 0.77(0.48 to 1.23) | 0.71(0.45 to 1.13) | 0.66(0.41 to 1.08) |
| Stage|| | ||||||||||||||||||
| IIIB | 193(100) | 69(35.8) | 4(2.1) | 2(1.0) | 3(1.6) | 11(5.7) | 14(7.3) | 16(8.3) | 19(9.8) | 0.34(0.27 to 0.43) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| IV | 778(100) | 189(24.3) | 4(0.5) | 0(0.0) | 4(0.5) | 19(2.4) | 57(7.3) | 58(7.5) | 47(6.0) | 0.31(0.27 to 0.36) | 0.96(0.73 to 1.27) | 0.41(0.10 to 1.63) | -- | 0.51(0.11 to 2.27) | 0.62(0.29 to 1.31) | 1.46(0.81 to 2.62) | 1.21(0.69 to 2.10) | 0.87(0.51 to 1.48) |
| Morphology | ||||||||||||||||||
| ADC | 334(100) | 92(27.5) | 3(0.9) | 0(0.0) | 5(1.5) | 10(3.0) | 21(6.3) | 27(8.1) | 26(7.8) | 0.32(0.26 to 0.39) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| SCC | 181(100) | 38(21.0) | 1(0.6) | 0(0.0) | 1(0.6) | 1(0.6) | 5(2.8) | 21(11.6) | 9(5.0) | 0.21(0.15 to 0.29) | 0.68(0.46 to 0.99) | 0.51(0.05 to 4.88) | -- | 0.32(0.04 to 2.71) | 0.16(0.02 to 1.25) | 0.39(0.15 to 1.03) | 1.31(0.74 to 2.31) | 0.56(0.26 to 1.20) |
| SCLC | 152(100) | 52(34.2) | 3(2.0) | 1(0.7) | 1(0.7) | 8(5.3) | 24(15.8) | 7(4.6) | 8(5.3) | 0.42(0.32 to 0.55) | 1.38(0.98 to 1.93) | 2.91(0.59 to 14.4) | -- | 0.54(0.06 to 4.63) | 2.00(0.79 to 5.08) | 2.85(1.58 to 5.12) | 0.60(0.26 to 1.38) | 0.74(0.33 to 1.64) |
| Other/ NA | 304(100) | 76(25.0) | 1(0.3) | 1(0.3) | 0(0.0) | 11(3.6) | 21(6.9) | 19(6.3) | 23(7.6) | 0.34(0.27 to 0.43) | 1.10(0.81 to 1.50) | 0.55(0.06 to 5.32) | -- | -- | 1.52(0.64 to 3.59) | 1.36(0.74 to 2.50) | 0.87(0.48 to 1.57) | 1.21(0.69 to 2.13) |
| Grade | ||||||||||||||||||
| I | 17(100) | 5(29.4) | 1(5.9) | 0(0.0) | 0(0.0) | 0(0.0) | 1(5.9) | 0(0.0) | 3(17.6) | 0.32(0.13 to 0.77) | 0.94¶(0.38 to 2.32) | 8.05(0.73 to 89.2) | -- | -- | - | 0.64(0.09 to 4.75) | - | 2.36(0.70 to 7.93) |
| II | 37(100) | 12(32.4) | 0(0.0) | 0(0.0) | 3(8.1) | 1(2.7) | 2(5.4) | 5(13.5) | 1(2.7) | 0.30(0.17 to 0.53) | 0.83¶(0.45 to 1.51) | -- | -- | 8.15(1.36 to 49.0) | 0.64(0.08 to 5.06) | 0.46(0.11 to 1.96) | 1.26(0.48 to 3.30) | 0.28(0.04 to 2.11) |
| III | 308(100) | 86(27.9) | 2(0.6) | 2(0.6) | 2(0.6) | 9(2.9) | 25(8.1) | 25(8.1) | 21(6.8) | 0.33(0.27 to 0.41) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| NA | 609(100) | 155(25.5) | 5(0.8) | 0(0.0) | 2(0.3) | 20(3.3) | 43(7.1) | 44(7.2) | 41(6.7) | 0.31(0.26 to 0.36) | 0.90(0.69 to 1.17) | 1.35(0.26 to 7.00) | -- | 0.49(0.07 to 3.52) | 1.15(0.52 to 2.52) | 0.85(0.52 to 1.40) | 0.86(0.52 to 1.40) | 0.98(0.58 to 1.66) |
| Therapy | ||||||||||||||||||
| None | 158(100) | 2(1.3) | 0(0.0) | 0(0.0) | 0(0.0) | 0(0.0) | 1(0.6) | 1(0.6) | 0(0.0) | 0.03(0.01 to 0.11) | 0.11(0.03 to 0.47) | -- | -- | -- | -- | 0.19(0.03 to 1.43) | 0.24(0.03 to 1.76) | -- |
| CT only | 403(100) | 88(21.8) | 2(0.5) | 1(0.2) | 3(0.7) | 9(2.2) | 27(6.7) | 21(5.2) | 25(6.2) | 0.25(0.20 to 0.31) | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- | 1.00- |
| RT only | 54(100) | 10(18.5) | 0(0.0) | 0(0.0) | 0(0.0) | 1(1.9) | 1(1.9) | 4(7.4) | 4(7.4) | 0.35(0.19 to 0.64) | 1.46(0.76 to 2.81) | -- | -- | -- | 1.44(0.18 to 11.4) | 0.49(0.07 to 3.62) | 2.27(0.78 to 6.63) | 2.10(0.73 to 6.06) |
| CT+RT | 355(100) | 157(44.2) | 6(1.7) | 1(0.3) | 4(1.1) | 20(5.6) | 42(11.8) | 48(13.5) | 36(10.1) | 0.43(0.37 to 0.50) | 1.65(1.27 to 2.14) | 2.68(0.54 to 13.3) | 0.89(0.06 to 14.2) | 1.19(0.27 to 5.33) | 2.09(0.95 to 4.59) | 1.42(0.87 to 2.30) | 2.15(1.29 to 3.59) | 1.34(0.81 to 2.24) |
| NA | 1 | 1 | 1 | |||||||||||||||
* Crude recurrence rates per person-year. ADC = adenocarcinoma; CB = contralateral or bilateral; CI = confidence interval; CT = chemotherapy; Ipsi = ipsilateral; NA = not available; RT = radiation therapy; SCC = squamous cell carcinoma; SCLC = small cell lung carcinoma.
† Number (%) of patients.
‡ Including mediastinal lymph-nodes (No. 1) and heart (No 1).
§ Including abdomen (No. 2), adrenal gland (No. 11), eye (No. 1), intestine (No. 1), kidney (No. 1), lung unspecified (No. 1), abdominal (No. 1) or cervical (No. 1) lymph-nodes, perineum (No. 1), pleura (No. 8), skin (No. 2), spleen (No. 1), testis (No. 1), and multiple sites (No. 34).
|| Pathological (when available) or clinical stage according to the TNM Classification of Malignant Tumors, 7th edition.
¶ Hazard ratios of grade II vs grade I varied over time (test of proportional-hazards P = .001).
Among ST patients, 49.3% had local or distant recurrence during the study (Table 3). The percentages of patients with any recurrence increased from 33.9% in stage IA to 62.8% in stage IIIA. Stage IIA patients had a risk of any recurrence (61.2%) similar to stage IIB (57.9%) and much higher than stage IB patients (37.8%), consistent with the survival differences in Figure 1. Overall, risk of metastases (including “other”) exceeded risk of local recurrences for every stage at diagnosis (Table 3).
In ST patients the rate of recurrence, calculated in the absence of competing risks, is 0.17 per person-year, which implies that approximately 1-exp(-17×1) = 16% of patients would have recurrence within one year. Overall recurrence rates were comparable for patients with adenocarcinoma and squamous cell carcinoma and increased with stage. Relative hazards from the Cox model show the impact of stage on risk of recurrences at various sites. There was an increasing frequency of brain metastases with increasing tumor grade in non–small cell lung cancer case patients (Supplementary Table 3, available online). In grade III tumors, 25 of the 30 patients with brain metastases had non–small cell lung cancer (Supplementary Table 3, available online).
Recurrence rates in ST patients were higher in those with any adjuvant treatment and especially in those who had combined chemotherapy and radiation therapy (Table 3), probably reflecting more aggressive therapeutic indications for those with poor prognosis (16) rather than treatment effect. Stage-specific analyses of treatment effects confirm these observations (Supplementary Table 4, available online). The overall hazard ratios of recurrences with mutual adjustment for sex, age at diagnosis, smoking status, stage, morphology, grade, and therapy (Supplementary Table 5, available online) are very similar to the unadjusted (marginal) hazard ratios (Table 3), although the inclusion of stage in the model tends to attenuate associations with grade and histology. Never smokers show a slightly better prognosis than ever smokers (Supplementary Table 5, available online).
Among NST patients (Table 4), the recurrence hazard rate (0.32 per person-year) is nearly twice the rate in ST patients (0.17). Nonetheless, the overall percentage of NST patients with recurrences is 26.6%, about half that for ST patients (49.3%), because of competing risk from mortality in NST patients. Relative risk estimates for any recurrence were lower in squamous cell carcinoma (HR = 0.68, 95% CI = 0.46 to 0.99) than in adenocarcinoma (referent) and, although not statistically significantly, for every site except bones (Table 4). NST patients with small cell tumors had a higher risk of metastases in the brain than in other sites. Grade was not strongly associated with recurrence (progression) rates among NST patients (Table 4; Supplementary Table 6, available online). NST patients had more metastases in each distant organ than in the lungs. This may partly reflect the operational definition of local recurrence, which is different in patients with residual local disease from patients with resected local disease. Having radiation therapy in addition to chemotherapy did not reduce local lung recurrences (HR = 2.68, 95% CI = 0.54 to 13.3) compared with chemotherapy alone in our data (Table 4). The overall hazard ratios of recurrences with mutual adjustment for sex, age at diagnosis, smoking status, stage, morphology, grade, and therapy in NST patients are reported in Supplementary Table 7 (available online). NST patients who received chemotherapy had lower recurrence rates than those who received no therapy or radiation therapy only (Table 4; Supplementary Table 7, available online), possibly because chemotherapy was given selectively to patients with better prognosis.
Absolute Risk of Recurrence
Absolute risk of first local recurrence is the chance that a local recurrence will be observed in a given time interval before death or the occurrence of distant metastasis. Absolute risks of first distant metastasis and all recurrence are defined similarly.
Absolute risks of local recurrence, distant metastasis and all recurrence are shown in Figure 2. Histology-specific absolute risk curves are in Supplementary Figure 4 (available online). The absolute risk of distant first recurrence (metastasis) exceeded that of local first recurrence, regardless of stage and cell type (Figure 2). In fact, the shape of the overall absolute risk curve largely reflects the absolute risk of metastases. Among ST patients, the absolute risk of any recurrence at five years was 0.33 (or 33%), 0.38, 0.61, 0.57, and 0.52 for stages IA, IB, IIA, IIB, and IIIA, respectively. For stages IIIB and IV, the absolute risk of any recurrence at five years was 0.38 and 0.27, respectively, despite the higher recurrence hazard rates for NST patients (compare Tables 3 and 4). The absolute risk of any recurrence increased rapidly in the first two years and more slowly thereafter, regardless of stage. The absolute risk of any recurrence is almost identical for squamous cell and adenocarcinoma (Supplementary Figure 4, available online).
Figure 2.
Absolute risk of any recurrence, local recurrence, and distant metastasis by stage (7th edition of the TNM Classification of Malignant Tumors) among 760 surgically treated patients stages IA to IIIA and 971 non-surgically treated patients stages IIIB and IV enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005.
Mortality Following Recurrence
Death rates per person-year and probabilities of dying in one year increased with stage, both before and after recurrence (Table 5). We evaluated the impact of recurrence on subsequent risk of death by examining death rates before and following recurrence and hazard ratios associated with past recurrence in the Cox model (Table 5). For example, in patients diagnosed with stage I, recurrence raises the death rate from 0.027 per year (2.7% one-year mortality) before recurrence to 0.66 (48.3% one-year mortality) after, with a corresponding hazard ratio of 34.2 (95% CI = 21.9 to 53.3). In stage IV, the probability of dying in one year was 57.3% before recurrence compared with 80.6% after recurrence, but the hazard ratio was only 2.24 (95% CI = 1.89 to 2.65). Local recurrences were associated with lower mortality rates and hazard ratios than distant metastases. Overall, patients who developed a local recurrence had a higher probability of surviving after the secondary lesion than those who developed a distant metastasis.
Table 5.
Lung cancer death rates and Cox hazard ratios before and after first recurrence, by stage (pathological, when available, or clinical stage according to the 7th edition of the TNM Classification of Malignant Tumors), among 2050 lung cancer patients stages I to IV enrolled in the Environment And Genetics in Lung cancer Etiology (EAGLE) study, Italy, 2002–2005
| Recurrence | Stage IA-IB (No. 392) |
Stage IIA-IIB (No. 254) |
Stage IIIA-IIIB (No. 585) |
Stage IV (No. 819) |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| D | Death rate† (95% CI) | HR (95% CI) |
D | Death rate† (95% CI) | HR (95% CI) |
D | Death rate† (95% CI) | HR (95% CI) |
D | Death rate† (95% CI) | HR (95% CI) |
|
| Total | 138 | 156 | 491 | 784 | ||||||||
| No recurrence | 36 | 0.027 (0.020 to 0.038) |
1.00 - |
32 | 0.056 (0.040 to 0.080) |
1.00 - |
244 | 0.27 (0.24 to 0.31) |
1.00 - |
580 | 0.85 (0.78 to 0.92) |
1.00 - |
| 1-year mortality | 2.7% | 5.4% | 23.7% | 57.3% | ||||||||
| Recurrence‡ | ||||||||||||
| Yes | 102 | 0.66 (0.54 to 0.80) |
34.2 (21.9 to 53.3) |
124 | 0.71 (0.59 to 0.84) |
14.3 (9.34 to 21.9) |
247 | 1.16 (1.03 to 1.32) |
4.82 (3.99 to 5.83) |
204 | 1.64 (1.43 to 1.88) |
2.24 (1.89 to 2.65) |
| 1-year mortality | 48.3% | 50.8% | 68.7% | 80.6% | ||||||||
| Type of recurrence‡ | ||||||||||||
| Local§ | 21 | 0.48 (0.31 to 0.74) |
25.5 (14.1 to 46.0) |
23 | 0.40 (0.27 to 0.60) |
8.49 (4.79 to 15.0) |
27 | 0.51 (0.35 to 0.74) |
2.38 (1.58 to 3.58) |
6 | 0.81 (0.37 to 1.81) |
1.55 (0.68 to 3.50) |
| 1-year mortality | 38.1% | 33.0% | 40.0% | 55.5% | ||||||||
| Distant|| | 81 | 0.73 (0.58 to 0.90) |
37.2 (23.6 to 58.5) |
101 | 0.86 (0.70 to 1.04) |
16.5 (10.7 to 25.4) |
220 | 1.38 (1.21 to 1.58) |
5.45 (4.49 to 6.60) |
198 | 1.69 (1.47 to 1.94) |
2.26 (1.91 to 2.68) |
| 1-year mortality | 51.8% | 57.7% | 74.8% | 81.5% | ||||||||
| Tissue of recurrence‡ | ||||||||||||
| Local | ||||||||||||
| Lung Ipsi | 14 | 0.40 (0.24 to 0.68) |
21.0 (10.8 to 40.7) |
14 | 0.39 (0.23 to 0.66) |
7.97 (4.14 to 15.3) |
14 | 0.53 (0.31 to 0.89) |
2.39 (1.39 to 4.13) |
4 | 0.60 (0.22 to 1.59) |
1.21 (0.45 to 3.27) |
| 1-year mortality | 33.0% | 32.3% | 41.1% | 45.1% | ||||||||
| Thoracic | 7 | 0.80 (0.39 to 1.69) |
38.5 (16.5 to 89.8) |
9 | 0.42 (0.22 to 0.80) |
8.63 (3.99 to 18.7) |
13 | 0.48 (0.28 to 0.83) |
2.31 (1.31 to 4.08) |
2 | 3.01 (0.75 to 12.0) |
3.24 (0.80 to 13.1) |
| 1-year mortality | 55.1% | 34.3% | 38.1% | 95.1% | ||||||||
| Distant | ||||||||||||
| Lung CB | 17 | 0.45 (0.28 to 0.73) |
23.8 (12.7 to 44.4) |
13 | 0.46 (0.27 to 0.80) |
9.50 (4.84 to 18.6) |
23 | 0.74 (0.49 to 1.11) |
3.33 (2.16 to 5.15) |
6 | 0.62 (0.28 to 1.39) |
1.00 (0.45 to 2.26) |
| 1-year mortality | 36.2% | 36.9% | 52.3% | 46.2% | ||||||||
| Liver | 10 | 1.66 (0.89 to 3.08) |
85.0 (40.4 to 179) |
11 | 1.99 (1.10 to 3.60) |
34.1 (16.6 to 70.2) |
21 | 2.07 (1.35 to 3.17) |
6.41 (4.09 to 10.0) |
20 | 2.34 (1.51 to 3.63) |
2.91 (1.86 to 4.57) |
| 1-year mortality | 81.0% | 86.3% | 87.4% | 90.4% | ||||||||
| Brain | 15 | 0.94 (0.57 to 1.55) |
43.9 (23.1 to 83.6) |
19 | 0.84 (0.53 to 1.31) |
15.5 (8.54 to 28.2) |
50 | 1.50 (1.14 to 1.98) |
5.79 (4.24 to 7.91) |
62 | 1.84 (1.43 to 2.35) |
2.55 (1.94 to 3.34) |
| 1-year mortality | 60.9% | 56.8% | 77.7% | 84.1% | ||||||||
| Bones | 5 | 2.27 (0.94 to 5.44) |
64.8 (24.4 to 172) |
15 | 0.98 (0.59 to 1.63) |
17.5 (9.34 to 32.7) |
48 | 1.60 (1.21 to 2.12) |
5.89 (4.30 to 8.05) |
60 | 1.59 (1.23 to 2.05) |
1.92 (1.51 to 2.58) |
| 1-year mortality | 89.7% | 62.5% | 79.8% | 79.6% | ||||||||
| Other¶ | 34 | 0.68 (0.49 to 0.96) |
34.6 (20.5 to 58.3) |
43 | 0.93 (0.69 to 1.25) |
17.8 (11.0 to 28.9) |
78 | 1.44 (1.15 to 1.80) |
5.72 (4.39 to 7.45) |
50 | 1.81 (1.37 to 2.39) |
2.47 (1.84 to 3.32) |
| 1-year mortality | 49.3% | 60.5% | 76.3% | 83.6% | ||||||||
* Not including 48 patients with stage X. CB = contralateral or bilateral; CI = confidence interval; D = number of deaths from lung cancer; HR = hazard ratio; Ipsi = ipsilateral.
† Crude death rates per person-year. Rates before recurrence were computed as the number of deaths before recurrence divided by the person-years exposure before recurrence. Rates after recurrence were computed as the number of deaths following recurrence divided by the person-years exposure after recurrence.
‡ Two-sided P value from Wald test of HR homogeneity across stages was statistically significant (P < .001).
§ Including lung ipsilateral and thoracic.
|| Including distant metastases (single or multiple), lung unspecified and unspecified sites.
¶ Including lung unspecified, other organs/tissues, multiple sites, and unspecified sites.
Discussion
We provide a comprehensive description of prognostic indicators for lung cancer recurrences and survival. Importantly, we report absolute risk of recurrence at specific sites, which are clinically relevant, because they account for the reality that the patient is subject to competing risks of mortality. For example, absolute recurrence risks (Figure 2) are low in NST patients, who tend to die quickly without much time to develop recurrences, even though the recurrence hazard rates are high (Table 4). Moreover, while recurrence rates (Table 4) and survival (Supplementary Figure 3, available online) were slightly different between adenocarcinoma and squamous cell carcinoma NST and ST patients, respectively, the absolute risks of recurrence were virtually identical (Supplementary Figure 4, available online). We also emphasized death rates before and after recurrence (Table 5) and factors affecting these death rates. Death rates are more useful than relative risks for advising patients and management strategy. After recurrence, death rates (and one-year probabilities of death) increase with increasing stage, while the Cox hazard ratios comparing mortality after recurrence to before recurrence decrease with stage.
Estimates of risk in a population with universal access to care and where international diagnosis and treatment guidelines are commonly followed are likely to be more relevant to community settings with high-quality care than estimates from highly selected patients seen in clinical trials or referral centers (5,10,17–24). Moreover, our study’s large number of case patients, nearly complete ascertainment, careful staging using up-to-date classification, and thorough follow-up permit detailed prognostic assessments. Only 104 EAGLE patients received biologically targeted treatments. Except for two patients with advanced disease, the patients were given targeted treatments only after failing conventional chemotherapy. Currently, targeted treatment is available for a small proportion of cases, mostly with adenocarcinomas, especially in never smokers. Until effective systemic treatments are available for a large proportion of patients, the results of our study should provide valuable guidance for todays’ clinical practice.
We note an important limitation of our study: We did not have data on performance status, which is a useful prognostic factor. Similar population-based studies in other settings are needed to assess the extent to which our results generalize.
Several findings have clinical implications. First, lung cancer should be regarded as a potentially systemic disease; in fact, the absolute risk of metastasis exceeds that of local recurrence, even in stage I (Figure 2). Moreover, metastases presage very high death rates at any stage (Table 5), indicating the need for systemic treatments to prevent metastases. Second, stage IIA patients have a much higher risk of recurrence (Table 3) and death (Figure 1) than those with stage IB. A similar survival gradient was observed within the data that led to the 7th staging classification system (4,25–27), but we now show that survival mirrors the risk of recurrence. Third, surgically treated non–small cell carcinoma patients have increasing risk of brain metastases with increasing tumor grade (Table 3; Supplementary Table 3, available online). The brain is a frequent site of NSCLC metastases (28,29). However, prophylactic cranial irradiation in NSCLC curtails brain metastases but may not improve overall survival (30) and may affect cognitive functions (31). If our observation is independently confirmed, the inclusion of tumor grade could aid in the therapy decision process for surgical NSCLC cases.
In conclusion, our population-based findings can inform clinical management, provide generalizable information for designing clinical trials whose power depends on absolute numbers of events, and suggest therapeutic needs for specific lung cancer subgroups.
Funding
This work was supported by the Intramural Research Program of the National Institutes of Health, the National Cancer Institute, Division of Cancer Epidemiology and Genetics; the Lombardy Region, Italy (program 14013-2010); and the Italian Ministry of Health and INAIL (PMS/42/06).
Supplementary Material
The study sponsors had no role in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.
We thank Vincenzo Coviello, the author of Stata command stcompet, for his technical advices on calculation of absolute risks. We thank the EAGLE participants and all collaborators who made this study possible. More details on EAGLE can be found on the EAGLE website http://eagle.cancer.gov/.
References
- 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69–90. [DOI] [PubMed] [Google Scholar]
- 2. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin. 2013;63(1):11–30. [DOI] [PubMed] [Google Scholar]
- 3. UICC International Union Against Cancer. Lung. In: Sobin LH, Gospodarowicz MK, Wittekind Ch, editors. TNM classification of malignant tumours. 7th edition Chichester: Wiley-Blackwell; 2009;138–146. [Google Scholar]
- 4. Goldstraw P, Crowley J, Chansky K, et al. The IASLC lung cancer staging project: proposals for the revision of the TNM stage groupings in the forthcoming (seventh) edition of the TNM Classification of malignant tumours. J Thorac Oncol. 2007;2(8):706–714. [DOI] [PubMed] [Google Scholar]
- 5. Sugimura H, Nichols FC, Yang P, et al. Survival after recurrent nonsmall-cell lung cancer after complete pulmonary resection. Ann Thorac Surg. 2007;83(2):409–417. Discussion in 417–418. [DOI] [PubMed] [Google Scholar]
- 6. Kelsey CR, Fornili M, Ambrogi F, et al. Metastasis dynamics for non-small-cell lung cancer: effect of patient and tumor-related factors. Clin Lung Cancer. 2013;14(4):425–432. [DOI] [PubMed] [Google Scholar]
- 7. Molina JR, Yang P, Cassivi SD, Schild SE, Adjei AA. Non-small cell lung cancer: epidemiology, risk factors, treatment, and survivorship. Mayo Clin Proc. 2008;83(5):584–594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Provencio M, Isla D, Sánchez A, Cantos B. Inoperable stage III non-small cell lung cancer: Current treatment and role of vinorelbine. J Thorac Dis. 2011;3(3):197–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lehto R, Therrien B. Death concerns among individuals newly diagnosed with lung cancer. Death Stud. 2010;34(10):931–946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Fedor D, Johnson WR, Singhal S. Local recurrence following lung cancer surgery: incidence, risk factors, and outcomes. Surg Oncol. 2013;22(3):156–161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Landi MT, Consonni D, Rotunno M, et al. Environment And Genetics in Lung cancer Etiology (EAGLE) study: an integrative population-based case-control study of lung cancer. BMC Public Health. 2008;8:203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Gail MH, Benichou J, eds. Encyclopedia of epidemiological methods. Chichester: John Wiley & Sons; 2000;1–17. [Google Scholar]
- 13. Sobin L, Wittekind Ch, eds. TNM classification of malignant tumours, Sixth Edition New York: Wiley-Liss; 2002;99–103. [Google Scholar]
- 14. Gaynor JJ, Feuer E, Tan C, et al. On the use of cause-specific failure and conditional failure probabilities: Examples from clinical oncology data. J Am Stat Assoc. 1993;88:400–409. [Google Scholar]
- 15. Coviello V, Boggess M. Cumulative incidence estimation in the presence of competing risks. Stata J. 2004;4(2):103–112. [Google Scholar]
- 16. Signorello LB, McLaughlin JK, Lipworth L, Friis S, Sørensen HT, Blot WJ. Confounding by indication in epidemiologic studies of commonly used analgesics. Am J Ther. 2002;9(3):199–205. [DOI] [PubMed] [Google Scholar]
- 17. Taylor MD, Nagji AS, Bhamidipati CM, et al. Tumor recurrence after complete resection for non-small cell lung cancer. Ann Thorac Surg. 2012;93(6):1813–1821. [DOI] [PubMed] [Google Scholar]
- 18. Saynak M, Veeramachaneni NK, Hubbs JL, et al. Local failure after complete resection of N0-1 non-small cell lung cancer. Lung Cancer. 2011;71(2):156–165. [DOI] [PubMed] [Google Scholar]
- 19. Kelsey CR, Marks LB, Hollis D, et al. Local recurrence after surgery for early stage lung cancer. Cancer. 2009;115(22):5218–5227. [DOI] [PubMed] [Google Scholar]
- 20. Hung JJ, Hsu WH, Hsieh CC, et al. Post-recurrence survival in completely resected stage I non-small cell lung cancer with local recurrence. Thorax. 2009;64(3):192–196. [DOI] [PubMed] [Google Scholar]
- 21. Nakagawa T, Okumura N, Ohata K, Igai H, Matsuoka T, Kameyama K. Postrecurrence survival in patients with stage I non-small cell lung cancer. Eur J Cardiothorac Surg. 2008;34(3):499–504. [DOI] [PubMed] [Google Scholar]
- 22. Jang KM, Lee KS, Shim YM, et al. The rates and CT patterns of locoregional recurrence after resection surgery of lung cancer: correlation with histopathology and tumor staging. J Thorac Imaging. 2003;18(4):225–230. [DOI] [PubMed] [Google Scholar]
- 23. Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg. 1995;109(1):120–129. [DOI] [PubMed] [Google Scholar]
- 24. Baldini EH, DeCamp MM Jr, Katz MS, et al. Patterns of recurrence and outcome for patients with clinical stage II nonsmall-cell lung cancer. Am J Clin Oncol. 1999;22(1):8–14. [DOI] [PubMed] [Google Scholar]
- 25. Groome PA, Bolejack V, Crowley JJ, et al. The IASLC lung cancer staging project: validation of the proposals for revision of the T, N, and M descriptors and consequent stage groupings in the forthcoming (seventh) edition of the TNM classification of malignant tumours. J Thorac Oncol. 2007;2(8):694–705. [DOI] [PubMed] [Google Scholar]
- 26. Rena O, Massera F, Robustellini M, et al. Use of the proposals of the international association for the study of lung cancer in the forthcoming edition of lung cancer staging system to predict long-term prognosis of operated patients. Cancer J. 2010;16(2):176–181. [DOI] [PubMed] [Google Scholar]
- 27. Kameyama K, Takahashi M, Ohata K, et al. Evaluation of the new TNM staging system proposed by the International Association for the Study of Lung Cancer at a single institution. J Thorac Cardiovasc Surg. 2009;137(5):1180–1184. [DOI] [PubMed] [Google Scholar]
- 28. Carolan H, Sun AY, Bezjak A, et al. Does the incidence and outcome of brain metastases in locally advanced non-small cell lung cancer justify prophylactic cranial irradiation or early detection? Lung Cancer. 2005;49(1):109–115. [DOI] [PubMed] [Google Scholar]
- 29. Mamon HJ, Yeap BY, Jänne PA, et al. High risk of brain metastases in surgically staged IIIA non-small-cell lung cancer patients treated with surgery, chemotherapy, and radiation. J Clin Oncol. 2005;23(7):1530–1537. [DOI] [PubMed] [Google Scholar]
- 30. Gore EM, Bae K, Wong SJ, et al. Phase III comparison of prophylactic cranial irradiation versus observation in patients with locally advanced non-small-cell lung cancer: primary analysis of radiation therapy oncology group study RTOG 0214. J Clin Oncol. 2011;29(3):272–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Sun A, Bae K, Gore EM, et al. Phase III trial of prophylactic cranial irradiation compared with observation in patients with locally advanced non-small-cell lung cancer: neurocognitive and quality-of-life analysis. J Clin Oncol. 2011;29(3):279–286. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.