To estimate the risk for second primary cancers in survivors of small-cell lung cancer, we evaluated all patients with this malignancy who remained disease free for 2 years after the initiation of therapy at the National Cancer Institute. The potential change in the risk for a second cancer during follow-up was studied to determine whether the risk increases or decreases with time. In addition, the influence of smoking cessation on the development of subsequent primary smoking-related cancers is undefined [13-20]. Smoking history in our patient population was therefore carefully assessed, and the effect of smoking cessation was evaluated.

Methods

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Patients

From April 1973 to December 1989, 540 consecutive patients with histologically confirmed, previously untreated small-cell lung cancer were treated in National Cancer Institute intramural therapeutic trials. All had primary small-cell carcinoma of the lung with evaluable tumor lesions; none were treated in an adjuvant setting after surgical resection. The majority of patients were assigned a performance status, metastatic sites identified, and they were classified as having either limited- or extensive-stage disease as previously defined [21, 22]. The patients began treatment with combination chemotherapy with or without radiotherapy using various regimens that have been reported previously [11, 21-29]. The current status of all patients was determined.

The locations of the initial small-cell lung cancers were identified by reviewing reports of chest roentgenograms, thoracic computed tomography, fiberoptic bronchoscopies, mediastinoscopies, and thoracotomies in all patients who were free of cancer for 2 or more years after the initiation of chemotherapy. The sites of subsequent intrathoracic cancer were localized by reviewing reports of chest roentgenograms, thoracic computed tomography, fiberoptic bronchoscopies, thoracotomies, and autopsies. Chest roentgenograms and computed tomographic scans from patients with second primary non-small-cell lung cancers were compared with pretreatment radiotherapy chest roentgenograms showing treatment ports. All initial and subsequent pathologic material was reviewed by one of two pathologists from our institution. The diagnosis of small-cell lung cancer was made and confirmed using the previously defined histologic criteria [30, 31]. Smoking history in all 55 2-year cancer-free survivors was repeatedly determined by a combination of patient follow-up visits, chart and protocol database review, and contact of either patients alive at the time of manuscript preparation or relatives of deceased patients. Smoking cessation was defined as completely stopping smoking by the end of treatment, which was usually of 6 months duration.

Definitions

The upper aerodigestive tract includes the epithelial regions of the head and neck, lung, and esophagus [13, 32-34]. Smoking-related cancers include cancers of the lung, head and neck, esophagus, bladder, pancreas, kidney, and possibly the stomach [35]. A second primary non-small-cell lung cancer after an initial small-cell lung cancer occurs when 1) tumor histologic findings show non-small-cell lung cancer without small-cell elements; 2) no evidence suggests local or distant recurrence of small-cell lung cancer; and 3) the second cancer is identified more than 2 years after diagnosis of the original small-cell lung cancer [3, 36, 37].

Statistical Analysis

Top Methods Statistical Analysis Results Discussion Author & Article Info References

For estimation of the expected values of second cancer development, the period of risk began 2 years after diagnosis of the small-cell lung cancer and ended with the date of death, date of last follow-up, or date of diagnosis of a second cancer, whichever occurred first. Person-years of observation were accumulated using the computer program of Monson [38]. Age, sex, and time-specific rates for mortality obtained from the National Mortality Statistics and for cancer incidence obtained from the Surveillance, Epidemiology, and End Results (SEER) program were applied to the appropriate person-years of observation, had this population experienced the same rates prevailing in the national or SEER populations, respectively. Statistical methods for risk estimation were based on the assumption that the observed number of second cancers followed a Poisson distribution [39]. Test of significance and confidence intervals (CIs) for the relative risk (observed and expected) were calculated by using exact Poisson probabilities. To determine the absolute risk, or excess cases of cancer per 1000 persons per year, the number of expected cases was subtracted from the number observed; the difference was divided by the number of person-years of observation and multiplied by 10⁴. Trends and homogeneity were tested as described by Breslow and colleagues [39].

Results

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Patient characteristics at time of initiation of treatment for both the entire patient population and for the 2-year cancer-free survivors are summarized in Table 1. Of the 540 patients, 55 (10%) were alive and free of cancer 2 or more years after the initiation of treatment protocols Figure 1, corresponding to 22% of 204 patients with limited-stage and 3% of 336 patients with extensive-stage disease. Eighty-six of the patients did not have a performance status assigned or metastatic sites identified prospectively. Compared with the entire patient population, the 2-year cancer-free survivors had a higher percentage of women, of patients with limited-stage disease, and of fully ambulatory patients (Eastern Cooperative Oncology Group performance status, 0-1). These favorable prognostic factors are thoroughly described in the literature [40-43]. The other 485 patients died or had relapsed with small-cell lung cancer within 2 years. The median follow-up from initiation of therapy was 6.1 years (range, 2.1 to 15.1 years) for the 55 patients. Ten patients have remained free of cancer since initial treatment. Five other patients remain alive; three developed second primary cancers that were successfully treated; and two relapsed with small-cell lung cancer, one achieving a prolonged second complete response (>3 years) after recurrence and the other currently receiving salvage chemotherapy. Of the 55 patients, 40 have died: 16 from recurrent small-cell lung cancer, 13 from second primary cancers, 1 from a suspected second primary lung cancer (lung mass without histologic confirmation), and 10 from other causes.

View larger version (19K): [in this window] [in a new window]   Figure 1. Flow diagram of the outcome of 540 patients with small-cell lung cancer treated at the National Cancer Institute from 1973 to 1989. One of the 10 patients who died of causes other than cancer had an incidental esophageal cancer discovered at postmortem examination and is not included in the risk analysis of second cancers after successful treatment of small-cell lung cancer. The 13 patients with non-small-cell lung cancer included 1 who had a lung mass that was not confirmed by biopsy and 1 with a sarcoma of the lung.

Eighteen of the 55 patients developed recurrent small-cell lung cancer 2.1 to 12.2 years (median, 3.2 years) after beginning chemotherapy (see Figure 1). The histopathologic features of 4 of the 18 (22%) recurrent small-cell lung cancer tumors included subpopulations of cells with prominent nucleoli but were still in the histologic spectrum of small-cell lung cancer [30].

Eighteen of the 55 2-year cancer-free survivors developed one or more second primary cancers 3.5 to 13.3 years (median, 7.5 years) after beginning therapy for small-cell lung cancer. One of the 18 patients with a second primary cancer had invasive squamous cell carcinoma of the lip, which metastasized to the regional lymph nodes. Another patient died of a central nervous system disorder but had an incidental esophageal cancer discovered at postmortem examination. These cancers were not included in the calculations of the relative risk because incidence data are not available from SEER on nonmelanoma skin cancer and on incidental cancers discovered at postmortem examination (Table 2). The sarcoma was included with other lung cancers because the SEER incidence rates are based on the classification of cancers by site of origin, not by histologic findings.

The risk for development of a second cancer more than 2 years after the diagnosis of small-cell lung cancer increased by approximately four times, mostly due to the 16-fold increase in second non-small-cell lung cancers (nine squamous; one adenocarcinoma, one large cell carcinoma, and one sarcoma; and one suspected second primary cancer without histologic confirmation) (see Table 2). A patient who developed an adenocarcinoma of the lung 12 years after the initial diagnosis of small-cell lung cancer had an initial biopsy specimen with predominant small-cell lung cancer histologic features and an adenocarcinoma component. The patient who developed a squamous cell cancer of the head and neck 8.9 years after the initiation of treatment for small-cell lung cancer had mixed small-cell/large-cell histologic features. The other 53 patients had only small-cell lung cancer in their biopsy specimens at presentation. Seven cancers developed in the lung contralateral to the original small-cell lung cancer, two in a different lobe of the ipsilateral lung, three in the same lobe, and one site could not be accurately determined because of the presence of massive bilateral pleural effusions. The relative risk for a second primary non-small-cell lung cancer compared with that in the general population increased significantly over time from 6 times at 2 to 4 years to 36 times after 10 years (P = 0.02) (Table 3). The risk for other smoking-related cancers appeared to increase, although the number of events was small and the statistical evidence that these risks were truly elevated was modest.

Two of the 18 patients with second primary cancers had a third primary cancer. One patient who developed a second primary non-small-cell lung cancer was successfully treated for stage IB ovarian papillary cystadenocarcinoma 14 months after diagnosis of small-cell lung cancer (see Table 2). The patient with erythroleukemia was found to have a non-small-cell lung cancer at postmortem examination. The cause of death has been reported as acute nonlymphocytic leukemia, and only the erythroleukemia has been included in the risk analysis of second primary tumors.

The smoking histories were obtained directly from 15 patients (27%), from relatives for 10 (18%), and from the patients' records for 30 (55%). Twelve patients continued to smoke more than 6 months after initiating treatment for small-cell lung cancer, and 43 did not. Of the 12 who continued to smoke, 8 developed subsequent second primary cancers (7 of 8 were smoking-related cancers). In those patients who stopped smoking within 6 months of initiating treatment for small cell lung cancer, the risk for a subsequent second primary lung cancer was increased 11 times compared with that in the general population. Those who continued to smoke showed a 32-fold increase (Table 4). The 12 patients who continued to smoke had a threefold (CI, 1.0 to 9.1) increased risk for a second primary lung cancer compared directly with the 43 who stopped smoking (P = 0.02). Similarly, the risk for development of any second primary smoking-related cancer compared with that in the general population was significantly higher in those who continued to smoke compared with those who stopped smoking at diagnosis (relative risk of 21; CI, 8.8 to 40 compared with a relative risk of 5.3; CI, 2.1 to 11, respectively). In both groups, the risk for a second lung cancer increased over time, to 100 times after 10 years among those who continued to smoke (P = 0.05) and 25 times after 10 years among those who stopped smoking (P = 0.08).

The cumulative smoking history did not appear to affect the risk for second lung cancers or any smoking-related cancer (Table 5). The relative risk of a second lung cancer in those with a smoking history of less than the median 55-pack years was similar to the relative risk in those with a smoking history of more than 55-pack years (relative risks of 17 and 14, respectively). The relative risk for a second primary smoking-related cancer in those with a smoking history of less than 55-pack years was 9.3 compared with 8.1 in those with a smoking history of more than 55-pack years. Similarly, smoking history and thoracic irradiation appeared to have no additive effect on the risk for a second primary lung cancer (Table 5). Patients who received thoracic irradiation and had a greater than 55-pack year smoking history had a relative risk for a second lung cancer of 21 compared with a relative risk of 22 in patients who did not receive thoracic irradiation and had a smoking history of less than 55-pack years.

Of the 43 2-year cancer-free survivors who were treated with thoracic irradiation, 10 developed second primary lung cancers (Table 6). Seven arose within the radiation port, two at the margin of the port, and one site could not be accurately determined because of bilateral pleural effusions. Among irradiated patients, the risk for a second primary lung cancer increased 18 times. Five of the 10 second primary cancers in this group occurred within the first 7 years (relative risk, 12), and the remaining 5 occurred after 7 years (relative risk, 26.3). Among the 12 patients who did not receive thoracic irradiation, 3 developed second primary lung cancers. The risk for second primary lung cancer increased 12 times, and all 3 occurred after 7 years (relative risk, 25). No significant difference in risk for developing second primary non-small-cell lung cancer was noted between those who received thoracic irradiation and those who did not (P = 0.34).

Overall mortality among 2-year cancer-free survivors increased six times compared with that in the general population (Table 7). The 16-fold increased mortality from malignancies included deaths from second primary cancers and from the recurrent small-cell lung cancers. Ten patients died of causes other than cancer. One patient who died of a central nervous system disorder had an incidental esophageal cancer discovered at postmortem examination; therefore, nine died without evidence of cancer. The mortality from central nervous system disease (cerebrovascular disease in two patients and degenerative disease in two patients) was also significantly elevated (P = 0.009). Three of these four patients had received cranial irradiation. The risk for death from pneumonia was approximately 20 times higher than the risk in the general population of the same age on the basis of three cases. Although three patients died suddenly of presumed cardiovascular causes, 2.5 cardiovascular deaths would be expected in a population of that age and sex.

Discussion

Top Methods Statistical Analysis Results Discussion Author & Article Info References

Of 540 patients with small-cell lung cancer, 55 (10%) survived cancer-free for 2 or more years and have been followed for a median of 6.1 years (range, 2.0 to >15.1 years) from the initiation of treatment for small-cell lung cancer. Pathologic review of the biopsy specimens from 55 patients showed that none of the patients had pulmonary neuroendocrine carcinoma (atypical carcinoid or well-differentiated neuroendocrine carcinoma), cancers that have been commonly reported after pathologic review of other long-term survivors of small-cell lung cancer [44-46]. Two of 55 patients (4%) presented with a mixed histology, an incidence which is similar to the estimated incidence (1% to 2%) of small-cell lung cancer with non-small-cell lung cancer components [31].

Heyne and colleagues [7] recently reported similar 2-year survival data in which 51 of 446 patients (11%) with small-cell lung cancer survived free of cancer for 2 or more years. The long duration of follow-up allowed an assessment of risk for development of second primary cancer for 2 to 10 or more years. Both these studies showed that most second primary tumors were lung cancers and that the incidence of second primary tumors increased during the 10 years of observation. In contrast to these two studies, Sagman and colleagues [10] reported that most second primary tumors in patients treated for small-cell lung cancer were solid tumors other than lung cancer [10]. The authors did not identify the outcome of all patients, however, and the duration of follow-up was unclear. Their calculated relative risk for second tumors in their sample was 3.7, similar to the relative risk of 4.4 observed in our patients. In contrast, their relative risk for second lung cancers was 6.8 compared with a relative risk of 16 in our study. This discrepancy may be caused by a shorter duration of follow-up because both our study and that by Heyne and associates [7] showed that the risk for second primary tumors increases with time.

The high incidence of second primary aerodigestive cancers can probably be attributed to a common etiology, that is, exposure to cigarette smoke. The 16-fold elevated relative risk for developing non-small-cell lung cancer observed in this study (see Table 2) is probably a conservative estimate, because the expected value of 0.83 included small-cell lung cancer (20% of all lung cancers), and all of the observed cases were non-small-cell lung cancer. Nearly all reports of long-term survivors of small-cell lung cancer quote a much lower incidence of second primary cancer than that seen in our sample [1, 5, 6, 9, 10]. However, the median follow-up of patients in this analysis and that by Heyne and coworkers [7] is longer than those of other studies, a factor that may explain the higher relative risks and cumulative incidence rates. Our experience parallels reports of the development of second malignancies in head and neck cancer and in non-small-cell lung cancer. The estimated risk for development of a metachronous second primary cancer in patients surviving head and neck cancer has been reported to be as high as 23% at 8 years from diagnosis of first cancer [47-51]. In patients with non-small-cell lung cancer who survive for more than 2 years from initial diagnosis, the incidence of second primary lung cancer progressively rises to as high as 32% at 15 years [36, 37, 52-55].

The increased risk of 146 second primary non-small-cell lung cancers per 1000 person-years in our group of long-term survivors of small-cell lung cancer is more than 20 times greater than the published incidence rate of initial primary lung cancers of 5.5 per 1000 person-years in male smokers over 45 years of age who participated in a lung cancer screening program [56]. This result suggests that patients with a previous primary lung cancer have an inherent, markedly increased susceptibility to further primary lung cancers compared with smoking men in whom lung cancer has not yet occurred. This predilection to the development of second primary cancers is further supported by the tendency toward the development of the second cancer in the same primary organ as the original cancer, both in patients with head and neck cancer and in those with small-cell lung cancer [7, 13-15, 20, 47, 50, 51, 57]. Exposure of susceptible mucosa to cigarette smoke has been proposed to cause diffuse mucosal diathesis or "field cancerization" and may explain the occurrence of multiple primary epithelial cancers in the aerodigestive tract regions of the head and neck, lung, and esophagus [13, 32-34]. In our patients, we only observed an increased risk for second cancers in sites that are directly in contact with cigarette smoke (that is, the lungs, head and neck, and esophagus) and not in other smoking-related cancer sites (bladder, pancreas, and kidney).

Smoking history after diagnosis of small-cell lung cancer influenced the risk and time to development of second aerodigestive cancer. Patients who stopped smoking reduced their relative risk for a second primary lung cancer to 25 at 10 years, compared to 100 for the patients who continued to smoke. This fourfold reduction in the risk for lung cancer at 10 years is similar in magnitude to the reduction in relative risk seen in other adult smokers 10 years after discontinuing smoking [58]. This finding is consistent with some earlier reports stating that cessation of smoking after diagnosis of head and neck cancer reduced the risk for second primary cancers of the head and neck [13, 15]. Further, of the 12 patients who continued to smoke, 8 developed second primary cancers (7 aerodigestive and 1 gastric), and none of the 10 patients who remain alive and continuously free of cancer continued to smoke. Among those patients who developed a second primary smoking-related cancer, no tendency was seen toward higher cigarette intake. This observation is contrary to published literature documenting that risk for a second lung cancer rises with greater numbers of cigarettes smoked [14, 58, 59]. A shortcoming of our study was the absence of serum and urinary cotinine levels in our patients to confirm their smoking history. Cotinine is the major metabolite of nicotine, and levels in the serum and urine are elevated in patients who are exposed to cigarette smoke by active or passive smoking; discrepancies between cotinine levels and reported smoking histories have been identified [60-62].

The high incidence of second primary lung cancers in our series of patients could also be related to the use of combined-modality therapy (chemotherapy and thoracic irradiation) in 80% of our patients. The increased risk for lung cancer is well described in patients treated for Hodgkin disease, and it has been suggested that thoracic irradiation (with and without chemotherapy) plays a role in increasing lung cancer risk [63-68]. Most patients who have subsequently developed lung cancer after treatment for Hodgkin disease also smoked, and the influence of exposure to cigarette smoke in this population has not been analyzed. In our series, no significant increase in risk for second primary lung cancer was seen in patients who were treated with thoracic irradiation compared with those who did not receive thoracic irradiation; however, only limited data exist over 10 years, when radiation-induced lung cancer is expected to occur. The data were too sparse to simultaneously evaluate both radiation and smoking. From our analysis, thoracic irradiation had no significant detectable effect on the development of second primary aerodigestive tumors. Because of the small number of second primary cancers, the relatively large number of chemotherapeutic agents used, and the diversity of regimens over time, it is difficult to evaluate adequately the possible role of chemotherapy in the cause of second primary cancers.

Only one case of acute nonlymphocytic leukemia occurred in our series (relative risk of 69; CI, 0.9 to 384). The patient received alkylating agents and radiation but no etoposide. Although 40% of the 540 patients were treated with etoposide, none has yet developed acute leukemia. This finding contrasts with recent reports of a high risk for etoposide-related acute nonlymphocytic leukemia after treatment for non-small-cell lung cancer and germ cell tumors [69, 70]. The actuarial risk for the development of treatment-related nonlymphocytic leukemia has been reported to be as high as 25% at 3 years [71], but this finding is not consistent with our experience, nor with that of the published literature, in which only 24 cases have been recorded [10, 72-74] in a cancer with an incidence of 30 000 new cases in the United States in 1992 alone [75].

In summary, it is clear from our data and from the published literature that up to 6% of patients with small-cell lung cancer can be cured of their disease with current therapeutic approaches [1, 3, 5, 7-9, 12]. Thirty-one of our patients remained alive and free of cancer at 5 years from the initiation of therapy for small-cell lung cancer. Of these, 25 patients subsequently remained free of recurrent small-cell lung cancer. Despite the decreasing incidence of recurrent small-cell lung cancer with time, the longevity of long-term survivors of small-cell lung cancer continues to be compromised by the increasing incidence of second primary smoking-related cancers, as described in patients with early stage non- small-cell lung cancer treated with surgical resection [76]. Continued cigarette smoking after the initiation of treatment for small-cell lung cancer is an important factor identified in our patients both in the development of second primary aerodigestive cancers and in the time to development of these malignancies. Type of treatment had no discernible effect on the development of second primary tumors, but the data were sparse in the relevant time intervals. The number of patients involved in this study did not permit a multivariate analysis of covariables, but a multi-institutional analysis of patients with small-cell lung cancer who survive for 2 or more years without recurrent cancer is under way to independently evaluate the risk for second tumors with time, the effect of smoking cessation, and the effect of therapeutic radiation.

In view of these findings, it appears that application of successful preventive interventions could have a large potential benefit in all patients with small-cell lung cancer who remain free of cancer for at least 2 years. The simplest, and probably the most important, intervention should be to actively encourage patients to stop smoking. Daily treatment with isotretinoin was effective in preventing second primary cancers of the head and neck in patients who have been previously treated for squamous-cell carcinoma of the head and neck in one study [20]. The use of retinol palmitate in a randomized trial involving 307 patients with early-stage non-small-cell lung cancer treated with complete surgical resection showed a trend toward a lower incidence of smoking-related cancers in the treatment group (10 smoking-related cancers compared with 16 in the control group) [77]. Because lung cancer is a smoking related malignancy of the aerodigestive tract, similar to head and neck cancer, retinoids may prevent the occurrence of second primary smoking-related cancers in patients with lung cancer with sufficiently prolonged survival to be at risk for this complication. Plans for a cooperative group trial to address this issue are currently being developed.

The opinions contained in this report are those of the authors and do not necessarily reflect those of the Department of the Navy or the Department of Defense.