Timing of Chemotherapy and Radiotherapy

Two strategies of concurrent radiation therapy and chemotherapy have been piloted. McCracken and colleagues reported results from a Phase II trial of the SWOG in which two courses of cisplatin, etoposide, and vincristine were given with concurrent radiation therapy using one fraction of 1.8 Gy per day, 5 days per week, to a total dose of 45 Gy.⁶³⁷ Additional chemotherapy with vincristine, methotrexate, and etoposide, alternating with doxorubicin and cyclophosphamide for 12 weeks, followed the concurrent therapy. They evaluated 154 patients, and with a minimum period of observation of 3 years, the 2-year survival rate was 42%, and the 4-year rate was 30%. When updated, this study showed a 5-year survival rate of 26%.⁶³⁸ Turrisi and Johnson and colleagues reported a small series of patients treated with concurrent cisplatin and etoposide with accelerated fractionation: 1.5 Gy twice daily, 5 days per week, was given for 3 weeks to a total dose of 45 Gy.^639,640 Two-year survival rates were 57% and 65% for the Turrisi and Johnson and colleagues studies, respectively. The updated 4-year survival rate from Turrisi was 36%.^639,640

The two studies were compared at the University of Texas M.D. Anderson Cancer Center between 1987 and 1989.⁶⁴¹ The COPE (cyclophosphamide 750 mg/m² IV day 1, vincristine mg/m² IV day 8, cisplatin 20 mg/m² days 1–3, etoposide 100 mg/m²) trial attempted to blend the alternating approach similar with the approach employing accelerated treatment, in regard to both dose per day and overall time to complete the radiation therapy.

After 3 cycles of chemotherapy with COPE, patients underwent surgical resection. Those patients without a complete response were offered postoperative chemotherapy and interdigitating radiotherapy. Radiotherapy was given at .5 Gy bid for one week, repeated twice at 3 week intervals to a total of 45 Gy over 9 weeks; COPE was given between each cycle of radiotherapy and for an additional two more cycles. The other study between 1990 and 1992 was part of a national cooperative study to give chemotherapy (cisplatin 60 mg/m² IV day 1, and etoposide 120 mg /m² IV days 1–3 for four cycles, every 3 weeks) with concurrent radiotherapy 45 Gy in 5 weeks (1.8 Gy daily) or in 3 weeks (1.5 Gy bid ) starting on day 1 of chemotherapy.

The earlier study (follow-up ranging from 7 to 50 mo, with a median of 17.5 mo) enrolled 28 patients with limited-stage SCLC; 20 patients achieved complete response and 8 patients achieved partial response with 79%, 39%, and 30% survival rates at 1, 2, and 3 years, respectively.

The latter study (follow-up ranging from 1 to 39 mo, with a median of 21 months) enrolled 33 patients; 29 patients achieved complete response and 4 patients achieved partial response, with 93%, 70%, and 53% survival rates at 1, 2, and 3 years, respectively. Rates of locoregional recurrence and distant metastasis were 44% and 48% in the interdigitating group and 30% and 33% in the concurrent group, respectively.

The major toxicity in both protocols was granulocytopenia. Grade 3-4 esophagitis was significantly greater in the concurrent group (5 grade 3 and 2 grade 4), than in the interdigitating group (2 grade 3, 0 grade 4). The data suggest that concurrent chemotherapy plus radiation therapy appears to be more effective than induction chemotherapy followed by interdigitating chest radiotherapy and chemotherapy for both local and distant disease control as well as 2- and 3-year overall survival.

The NCI of Canada Clinical Trial Group studied early versus late thoracic radiation therapy in a randomized comparative fashion. Their trial enrolled 308 patients and showed that the early thoracic irradiation group had significantly better survival and fewer brain metastases compared with the late thoracic irradiation group.⁶⁴²

Fractionation

The question of standard once-a-day fractionation (1.8 Gy per fraction) versus accelerated fractionation (1.5 Gy bid) with a total dose of 45 Gy has been answered by a cooperative randomized trial.^643,644 Radiotherapy was given with concurrent chemotherapy consisting of cisplatin 60 mg/m² IV on day 1 and etoposide 120 mg/m² IV on days 1–3 in four cycles. Four hundred and nineteen patients were enrolled in this randomized trial between May 1989 and July 1992, with 381 eligible and evaluable patients. Twice-daily treatment beginning with the first cycle of chemotherapy significantly improved survival compared with concurrent once-daily radiotherapy (p = .04 by the log-rank test). After a median follow-up of almost 8 years, the median survival was 19 months for the once-daily group and 23 months for the twice-daily group. The survival rates for patients receiving once-daily radiotherapy were 41% at 2 years and 16% at 5 years. For patients receiving twice-daily radiotherapy, the survival rates were 47% at 2 years and 26% at 5 years. Grade 3 esophagitis was significantly more frequent with twice-daily thoracic radiotherapy, occurring in 26% of patients compared with 11% in the once-daily group (p < .001). Despite the major myelosuppression noted in approximately 90% of patients in both groups, there was only one death from myelotoxicity. There were 11 treatment deaths, 6 in the twice-daily treatment group and 5 in the once-daily group. Overall, four cycles of cisplatin plus etoposide and a course of concurrent radiotherapy, 45 Gy given either once or twice daily beginning with cycle 1 of chemotherapy, resulted in 2- and 5-year survival rates of 44% and 23%, respectively, which are twice as good as they were a decade ago. However, sequencing and timing of chemotherapy and thoracic radiotherapy are still controversial. The NCI of Canada clinical trial group studied early versus late thoracic radiation therapy in a randomized comparative fashion. Their trial enrolled 308 patients and showed that the early thoracic irradiation group had a significantly better survival and fewer brain metastases compared with the late thoracic irradiation group.⁶⁴²

The question of standard once-a-day fractionation (1.8 Gy per fraction) versus accelerated fractionation (1.5 Gy bid) with a total dose of 45 Gy was investigated in a cooperative randomized trial.⁶⁴⁴ Radiotherapy was given with concurrent chemotherapy consisting of cisplatin 60 mg/m² IV day 1 and etoposide 120 mg/m² IV days 1–3 for 4 cycles. Four hundred and nineteen patients were enrolled in this randomized trial between 1988 and 1992, with 383 eligible and evaluable.

This intergroup study (ECOG 3588/RTOG 8815) randomly assigned 417 patients with limited SCLC to receive a total of 45 Gy thoracic radiation therapy (TRT), given either once daily over a period of 5 weeks (arm 1) or twice-daily over a 3-week period (accelerated hyperfractionation, arm 2). The hyperfractionated and accelerated TRT group (arm 2) significantly improved the 5-year survival by 26% compared to 16% among patients treated by daily fractionated prolonged TRT (arm 1) (p = .04). The median survivals were 19 months in arm 1 and 23 months in arm 2. The 2-year survival rates were 26% in arm 1 and 46% in arm 2, although acute grade 3 esophagitis was significantly more frequent in arm 2 (27%) compared with 11% in arm 1 (p < .001).⁶⁴⁴

Total Tolerable Dose

The radiation dose to the thorax is another controversial area.⁶⁴⁵ The National Cancer Institute of Canada developed an important study to show dose-response to the thorax.⁶⁴⁶ They have shown a clear dose response with increased thoracic progression-free survival by giving 37.5 Gy in 15 fractions in 3 weeks compared to 25 Gy in 10 fractions in 2 weeks as a consolidation after completion of cisplatin-etoposide and cyclophosphamide-doxorubicin-vincristin alternating or sequential chemotherapy. Arriagada and coworkers published a report of 173 patients with limited SCLC treated in 3 consecutive trials at the Institute of Gustave-Roussy, France.⁶⁴⁷ The total dose of thoracic radiotherapy increased from 45Gy (15-15-15), 55Gy (20-20-15) to 65 Gy (20-20-20), which was given by split courses interdigitating with chemotherapy. Their 3-year local control rates were 66%, 70%, and 70%, respectively, and 5-year survival rates were 16 % 16%, and 20%, respectively. There was a 10 % rate of lethal toxicity without significant difference depending on the doses. They concluded there were no significant differences in the local control and survival with treatment between 45 Gy and 65 Gy when effective chemotherapy was given in this interdigitating manner.⁶⁴⁷

Choi and colleagues published a Phase I study to determine the mean tolerated dose (MTD) of radiation in daily and twice daily fractionation with concurrent chemotherapy.⁶⁴⁸ MTD for hyperfractionated radiotherapy was reached at 45Gy/30fractions/19 days. However, MTD for daily fractionation was not reached at 66Gy/33fractions/45 days. Therefore, patients were accrued to to the treatment arm consisting of 70 Gy/35 fractions/47 days, which might be too long to treat rapidly proliferating tumors such as SCLC. The tumor response rates varied from 78% to 100% without significant difference among the different dose levels. Grade 3 or more esophagitis and granulocytopenia were more marked among the patients who were treated by hyperfractionated and accelerated fractionation.

A Phase I study of RTOG 97-12 was initiated to improve locoregional control by escalating the total dose of radiation therapy, using daily fractionation to larger fields and then boost field and bid radiation therapy toward the end of the thoracic radiation therapy, with concurrent etoposide and cisplatin. In this way accelerated radiation therapy could be given toward the end, when residual tumors are thought to be more resistant and proliferating rapidly. The boost field was much smaller than the original AP/PA field to reduce toxicity to the esophagus. The previously discussed intergroup trial⁶⁴⁴ demonstrated increased toxicity to the espohagus (11% versus 26%) when one-a-day fractionation (1.8 Gy per day) versus accelerated fractionation (1.5 Gy bid) was used. Therefore, if we can start daily fractionation to the larger field and boost toward the end with the smaller field to the residual tumor, dose escalation can be achieved without extending the duration of the thoracic radiation therapy, which would be detrimental because of rapid proliferating tumor cells.

The Phase I study of RTOG 97-12 was activated in February 1998. Arms 1, 2, 3, 4, and 5 have been opened and closed during the dose escalation accrual component. A 6th arm was activated on September 21, 2001, and was closed to accrual January 11, 2002. This study has accrued 62 patients across these 6 treatment arms. Arm 1 was considered too toxic, with two observed grade 3 esophagitis toxicities within the first five patients. The dose of chemotherapy was modified, and arm 5 was activated. Two grade 3 esophagitis cases were observed in the first, ninth, and tenth patients in arm 5, so the total dose of radiation therapy was escalated in arm 2. Two toxicities were observed in arm 2, and the dose was escalated. One grade 3 esophagitis was observed in the first 10 patients in arm 3, so the total dose of radiation therapy was escalated in arm 4. One additional grade 3 esophagitis occurred in arm 3, and it occurred in the 11th patient. Two grade 3 esophagitis cases occurred in arm 4, so the total dose of radiation therapy was escalated in arm 6. In arm 6, there were four patients who developed grade 3 nonhematologic toxicities in the first 10 patients; therefore, a MTD of 61.2 Gy (arm 4) was chosen. We are proposing a phase II study for patients with limited SCLC and good performance status to receive the same chemotherapy as RTOG 97-12 with concurrent thoracic radiotherapy of a total dose of 61.2 Gy in 3 weeks.

Cranial Irradiation in SCLC

Brain metastases are detected in approximately 10% of SCLC patients at initial presentation, and the likelihood for development increases with lengthening survival. In the absence of specific therapy to the central nervous system (CNS), the actuarial cumulative probability reached 80% at 28 months of follow-up when the cases of CNS metastasis found at autopsy were included and 58% at 24 months when only clinical brain metastases were counted.^286,649 At postmortem examination, much higher incidences were reported, ranging up to 65%.^650,651 With the hope that the CNS would be the only site of residual disease, prophylactic cranial irradiation (PCI) was incorporated as an initial part of treatment for SCLC in the early 1970s.⁶⁵²

Since the early 1980s, PCI has been recommended for complete responders on the basis of the findings from a retrospective study from the NCI.^653,654 Between 1977 and 1995, 11 prospective randomized trials of PCI were published. Although all but two trials reported a significant reduction in the incidence of brain metastasis, none demonstrated an improvement in survival with PCI.^655,656 Four trials have addressed the issue of PCI in complete responders.^657–660 All four studies failed to demonstrate any survival benefit of PCI even, among the complete responders, despite significant reduction in the frequency of brain metastases. This lack of survival benefit was attributed to the fact that most patients were destined to have their disease progress at other sites, which is the predominant factor influencing survival. In a most definitive large randomized study of 300 SCLC patients who achieved complete responses following various chemotherapy regimens, PCI (24 Gy given in 8 fractions over 12 days) significantly reduced the rate of brain metastasis as the isolated first site of relapse (45% vs 19% at 2 years, p < .01, and overall brain relapse rate at 2 years 67% vs 40%; p < .01).⁶⁵⁹ The 2-year survival rate of 29% in the PCI group was not significantly different from 21.5% for the control group who did not receive PCI (p = .14). Although these trials failed to demonstrate any survival benefit of PCI, a recent meta-analysis concluded that PCI improves both overall survival and disease-free survival among SCLC patients in complete remission.⁶⁶⁴ This analysis was based on individual data from 987 patients who were enrolled in seven randomized trials to evaluate the role of PCI in complete remission. The relative risk of death in the treatment group, compared with the control group, was 0.84 (95% confidence interval, 0.73–0.97; p = .01), which corresponds to a 5.4% increase in the rate of survival at 3 years (15.3% in the control group vs 20.7% in the PCI group). PCI has not been routinely applied; even to patients who achieved complete response (CR) to the thoracic radiotherapy and chemotherapy because of concern of the neurological toxicities, which were claimed to be due to the PCI and lack of overall survival benefit by application of PCI.⁶⁶¹ However, when more fractionated radiotherapy without concurrent PCI was given to CR patients within optimal timing, there was reduction of intracranial metastasis without severe neurotoxicities.⁶⁶² Furthermore, when patients with limited SCLC were evaluated before and after PCI by an expert neuro-pschycologist applying detailed neuropsychological testing, 83% (25 of 30 patients) were found to have minor cognitive dysfunction prior to PCI, and no significant differences were found in the comparison of pre- and postneuropsychological testing.⁶⁶³

With regard to the impact on overall survival made by PCI, for CR patients, a recent report based on meta-analysis of seven randomized clinical trials has demonstrated that PCI improved overall survival by 5.4%, from 15.3% to 20.7% in 3 years (p = .01), disease-free survival by 8.8%, from 13.5% to 22.3% (p = .001), and cranial recurrence was reduced by 25.3% from 58.6% to 33.3% (p < .001) without significant short-term neurotoxicities from PCI.⁶⁶⁴ There was a trend in the reduction of brain metastasis in the subset of PCI patients who received at least 36 Gy in 12 fractions compared to lower doses of PCI.

In summary, the management of patients with limited-stage SCLC has improved median, long-term, and disease-free survival rates in the past 25 years, with some increase in toxicities, although some of the toxicities could be related to the disease itself and others were never seen because the patients did not live long enough. Patient selection with consideration of tumor histology, clinical stage, patient's performance status, and response to chemotherapy; as well as prompt supportive care with antibiotics, granulocyte-macrophage colony-stimulating factor (GM-CSF), antiemetics, hydration, correction of electrolytes and aggressive nutritional support, have all contributed to this improvement.

In the future, a multidisciplinary approach—to prevent this ominous disease, to detect early stage with better imaging or tumor markers, to select patients for appropriate treatment, to institute more efficacious systemic treatment with early application of thoracic radiation and PCI with normal tissue protector or more confined radiotherapy ports, and to provide appropriate supportive care—is warranted by the application of knowledge from basic and clinical researches.

Combination Chemotherapy

Only a few prospective randomized trials were conducted in the 1970s to directly compare combination regimens with single-agent regimens. Cyclophosphamide-based combinations yielded better response rates and survival rates than did the cyclophosphamide-alone regimen.^682,683 Simultaneous administration of four agents yielded a better response rate than did the sequential single-agent administration of the same agents but without any survival advantage.⁶⁸⁴ The number of drugs in combination regimens had been a subject of study and extensively reviewed.^685,686 In general, randomized studies demonstrate predictable advantages of combination programs. On the basis of a large number of combination chemotherapy trials, which reported much higher response rates and better survival outcome compared with the results from single-agent trials, combination chemotherapy has been accepted as the standard of treatment.^536,626 Some exceptions exist; elderly or medically compromised patients may be better treated with single agents.

Chemotherapy is the basis for management of SCLC, and commonly used chemotherapy regimens are shown in Table 92-20. During the 1980s, cyclophosphamide-based regimens predominated. The CAV combination was considered to have much activity, high response rates, and was effective in improving median and long-term survival.^536,687 This regimen was considered in the 1980s to be the standard regimen against which newer approaches were compared.

Table 92-20

Commonly Used Chemotherapy Regimens for Small-Cell Lung Cancer.

With the discovery of etoposide as an active single agent in SCLC, many randomized trials evaluated whether adding etoposide to or substituting it for one of the components of CAV would be beneficial. Controlled trials compared CAV plus etoposide (CAVE) with CAV alone.^666,688 In an equitoxic comparison of standard-dose CAV plus etoposide versus higher-dose CAV for the initial two cycles of therapy followed by five additional cycles of CAV, there were no differences in response rates or survival rates.⁶⁸⁹ When the CAV doses in the CAVE arm were kept identical to the CAV arm, multiple studies reported significantly greater response rates with CAVE, but survival was not significantly improved.^666,688,690 With etoposide substituting for one of the drugs in CAV (either doxorubicin or vincristine), there was no evident superiority.⁶⁹¹ CEV (cyclophosphamide, etoposide, vincristine) was compared with the same regimen without etoposide, and the three-drug regimen significantly improved the response survival rates in extensive-stage SCLC patients.⁶⁹² Similarly, with etoposide substituted for methotrexate at two different time points in a 4-week, four-drug regimen of cyclophosphamide, methotrexate, lomustine, and vincristine, there was improved survival when etoposide was given earlier, but not when given later, in the chemotherapy cycle.⁶⁹³ Therefore, the superiority of the etoposide-containing regimens has not been unequivocally proven, compared with the regimens without etoposide, in patients with SCLC.

The most commonly used regimen in the US is the combination of cisplatin and etoposide (PE).^536,693 Preclinical studies showed therapeutic synergy in murine leukemia.⁶⁹⁴ Both cisplatin and etoposide have limited single-agent activity in previously treated SCLC patients. Nevertheless, when used together, the PE regimen has generated objective tumor responses in more than 50% of patients with previously treated SCLC.^695,696 In chemotherapy-naïve patients, the PE regimen yields survival results comparable to those obtained with cyclophosphamide and/or doxorubicin-based regimens.^697–699 At least two randomized trials compared the CAV and PE regimens as sole initial chemotherapy.^691,700 Patients treated with PE had improved response rates, and overall survival was similar to that in those given CAV. This lack of survival improvement may be related to the fact that PE given as salvage therapy after CAV yielded response rates twice as high as CAV given as salvage treatment after PE in these trials. Moreover, two additional cycles of PE given after six cycles of regular CAV in patients with limited-disease SCLC resulted in a 7-month improvement in median survival in a randomized trial reported by Einhorn and colleagues.⁷⁰¹ PE is associated with less myelosuppression and infection than CAV, providing room for the addition of a third agent. The utility of this regimen is perhaps most apparent in patients with limited-stage disease because the PE regimen is better suited to the simultaneous administration of radiotherapy than the doxorubicin-containing regimens.

Noda and colleagues recently published an important randomized Phase III trial.⁷⁰² In extensive-stage patients, cisplatin and irinotecan was compared with cisplatin and etoposide. Two hundred fifty-four patients were entered, and there was a statistically significant difference in overall survival favoring patients treated with the irinotecan arm (12.8 vs 9.4 mo, p = .002). The cisplatin and irinotecan combination was also associated with superior 1-year survival (54% vs 37%) and 2 year (20% vs 5%) survival rates. The experimental arm in this study consisted of cisplatin and irinotecan given at 4-week cycles, with cisplatin 60 mg/m² on day 1 and irinotecan 60 mg/m² on days 1, 8, and 15. These data are provocative and are prompting further study of irinotecan as a component of first line therapy.

Second-Line Therapy for SCLC

The prognosis is poor for patients who receive second-line therapy after relapse.⁶⁸⁷ Response is influenced by the time to progression after cessation of first-line therapy. Patients who relapse less than 3 months after first-line therapy are commonly termed refractory and have response rates that are lower than for those patients who relapse more than 3 months after therapy and are usually termed sensitive. There is no standard regimen after first-line failure. CAV has been used in second-line therapy with response rates of 13% and 28% and median response durations of 26 and 24 weeks, respectively, in two studies.^703,704 In a study comparing PE and CAV in patients who had received the alternate regimen as first-line therapy, the response rate was higher with PE, but overall survival was similar. Early studies with oral etoposide reported response rates as high as 45%.⁷⁰⁵

A number of newer agents have been shown to be active in SCLC (Table 92-21). Topotecan is the most extensively studied, and a recent randomized trial demonstrated that topotecan (1.5 mg/m² as a 30-minute IV infusion daily for 5 days every 21 days) is at least as effective as CAV for SCLC patients who had relapsed at least 60 days after completion of first-line therapy.⁷⁰⁶ The response rate was 24.3% in 107 topotecan-treated patients compared with 18.3% in 104 patients treated with CAV (p = .285). There was no difference in median time to progression (13.3 vs 12.3 wk, p = .552) or median survival (25.0 vs 24.7 wk, p = .795) between the two groups. However, there was more symptom improvement in the topotecan group than in the CAV group for four of eight symptoms evaluated, including dyspnea, anorexia, hoarseness, and fatigue, as well as interference with daily activity (p ≤ .043). Grade 4 neutropenia occurred in 38% of topotecan courses versus 51% of CAV courses (p < .001). Grade 4 thrombocytopenia and grade 3/4 anemia occurred more frequently with topotecan, occurring in 10% and 18% of topotecan courses versus 1% and 7% of CAV courses, respectively (p < .001 for both). These results warrant further study of topotecan in combination with other active agents as a first-line therapy. Irinotecan also has demonstrated activity for chemotherapy-naïve patients as well as in the second-line setting.⁷⁰⁷

Table 92-21

Objective Response of Small-Cell Lung Cancer to Single Agents.

Dose Intensity

In preclinical studies, dose escalation has been an important strategy for overcoming drug resistance; thereby achieving higher complete remission rates and increasing curability.⁷⁰⁸ Several randomized trials have studied this concept. Cohen and colleagues administered twofold higher doses of the cyclophosphamide, methotrexate, and lomustine (CMC) regimen for an initial 6-week induction.⁷⁰⁹ Although there was significant improvement in response rates and median survival with higher drug doses, the doses of the “standard” CMC regimen used in comparison might be considered suboptimal. A much larger controlled trial using the CMC regimen with two different cyclophosphamide doses (ie, 700 vs 1,500 mg/m²) with the same dose of methotrexate (15 mg/m²) and lomustine (70 mg/m²) during the first 6 weeks demonstrated significant but modest increases in response rates and survival, particularly in patients with limited-stage disease.⁷¹⁰

However, more recent attempts to build on these early observations have not been successful. In two randomized studies of the CAV regimen, the doses of cyclophosphamide (CTX) were increased from 20% to 56% and the doses of doxorubicin were increased from 18% to 75% during the first 9 to 12 weeks of chemotherapy.^685,686 Although the complete response rate was superior with the high-dose regimen in one trial with extensive-stage SCLC, there was no difference in either the overall response rates or median survival rates in either study.⁶⁸⁵ Life-threatening leukopenia and serious infections requiring hospitalization were more common with high-dose CAV.^685,686 Also, there was no advantage of high-dose methotrexate with levamisole rescue compared with standard-dose methotrexate when added to the CAV regimen.⁷¹¹ More recently, Ihde and colleagues compared standard-dose and high-dose regimens involving administration of 67% higher doses of PE over the first 6 weeks of therapy.⁷¹² Despite 68% higher actual doses delivered and 46% higher dose-rate intensity, there was no benefit in terms of complete response rates (23% vs 22%, p = .44) or survival (median 10.7 vs 11.4 mo, p = .68). There was increased myelosuppression with high-dose PE.

A recent French trial randomized limited stage SCLC patients to receive a 33% higher dose of cyclophosphamide and a 25% higher dose of cisplatin in the first of six cycles of treatment with cyclophosphamide, cisplatin, doxorubicin, and etoposide.⁷¹³ Patients given higher doses of the two drugs survived significantly longer than those who received regular doses. It is possible that benefits of higher doses are of greater magnitude in limited-stage SCLC patients, but most controlled trials have failed to establish any benefit from higher-than-standard doses of commonly used combination therapy regimens in SCLC.

To obtain increased tumor cell kill, several investigators used even higher doses of chemotherapy, sometimes with autologous bone marrow transplantation support and often in a protected environment.^714–720 Although higher complete response rates have been reported, survival outcome is similar to results attained with less intensive regular programs.

Reasoning that only the patients with smaller tumor burden would benefit from the intensive treatment approach, several investigators have evaluated high-dose chemotherapy with autologous bone marrow transplantation as a late intensification.^{694,721–724} Some patients who were not in complete response after standard therapy responded further to achieve a complete response. However, the fraction of disease-free long-term survivors has been small, and the majority of such patients were those who had achieved excellent responses even before entering the late-intensification program. Overall, there have been no convincing data supporting the benefit of late-intensification therapy with autologous bone marrow transplantation.⁵³⁶ A recent meta-analysis of the dose-rate intensity of 60 published studies, which accounts for dose and interval of chemotherapy administration, could not show consistently better therapeutic results with dose-intensification programs of CAV, CAVE, or PE chemotherapy.⁷²⁵ There are currently no data justifying chemotherapy given as a routine initial treatment in doses that cause more than moderately severe myelotoxicity.

Elias and colleagues reported a 2- to 5-year disease-free survival rate of 57% in a small highly selected group of patients who received late-intensification therapy after achieving complete response or near complete response with the regular induction regimen.⁷²⁶ These patients were also given radiation to the chest (50 to 60 Gy) and the brain. Until a more definitive study confirms the survival benefit of this approach, the use of very-high-dose chemotherapy with autologous bone marrow transplantation support is recommended only on research protocols. The availability of hematopoietic growth factors has renewed the interest in investigating the dose escalation issue in patients with limited-stage SCLC.^727,728 However, given the lack of data supporting the survival advantage of intense chemotherapy plus marrow transplantation programs, the true benefit of dose escalation that can be achieved with growth factor support remains to be determined.

Another approach to increasing dose intensity is to shorten the intervals between cycles of chemotherapy.^729–731 Phase II pilot trials of weekly intensive chemotherapy reported high response rates and survival outcome comparable with those reported with conventional 3-week chemotherapy regimens. However, randomized trials have not established the benefit of weekly intensive chemotherapy over the standard combination regimen.^732,733 Because of the associated hematologic toxicity, which is similar in most weekly intensive chemotherapy trials, the maximum achievable dose intensity was only 74% of that attempted in one study and 77% in the other randomized study.^732,733 When granulocyte colony-stimulating factor (GCSF) was added to the weekly regimen, neutropenia could be mitigated, but only a 10% increase in intensity was achieved before other toxicities became dose limiting.⁷³⁴

Alternating Chemotherapy/Maintenance Therapy

Goldie and Coldman have provided a mathematical model of the spontaneous development of drug-resistant clones at a mutation rate proportional to the number of actively dividing tumor cells.⁶⁰⁴ This model predicts that multiple active agents should be given at full doses to maximize the chance of eradication of the entire tumor cell population; but because of overlapping myelosuppressive toxicity, all possible drugs cannot be given simultaneously. An alternative approach is to alternate the administration of two equally effective, non-cross-resistant combination regimens. This strategy has been tested in a series of trials.^735–739

Most of the recently completed trials compared CAV with CAV alternating with PE. In an early Canadian trial, the alternating schedule of CAV/PE improved the overall survival in patients with extensive-stage SCLC compared with six cycles of CAV (9.6 vs 8.0 mo, p = .03).⁷³⁷ However, the survival difference did not reach statistical significance when patients with only locoregional disease were excluded from the analysis. Also, because PE alone was not tested in this trial, the possibility could not be eliminated that the superiority of the alternating regimen was due to the better treatment regimen rather than the alternating treatment strategy. Two subsequent studies compared treatment with CAV and PE alone and alternating the two regimens.^691,700 The Japanese trial showed significant improvement in the overall survival of patients with the CAV/PE alternating regimen (median survival 11.8 mo) compared with the CAV or PE induction group (median survival 9.9 mo).⁷⁰⁰ However, this survival benefit was seen only in patients with limited-stage SCLC; the median survival of the CAV/PE arm was 16.8 months, significantly superior to that in the CAV arm (12.4 mo, p = .014) or the PE arm (11.7 mo, p = .023). For the patients with extensive-stage SCLC, there was no difference in overall survival among the three treatment arms. Similar results were obtained for extensive-stage SCLC by the Southeastern Cancer Study Group (SECSG) investigators and others.^691,739

In another large randomized trial of 557 patients with extensive stage SCLC, ECOG investigators reported a modest but statistically significant improvement in overall survival after alternating induction chemotherapy with CAV/HEC, compared with induction chemo-therapy with CAV (10.7 vs 9.9 mo, p = .002).⁷⁴⁰ These investigators cautioned that this survival benefit must be weighed against the higher life-threatening toxicities associated with this alternating induction regimen and did not exclude the possibility that the observed survival prolongation may simply be an effect of using a new active agent, such as etoposide, early in the treatment course.

For patients with limited-stage SCLC, SWOG investigators conducted a large randomized trial comparing alternating CAV/PE regimen with induction with CAVE.⁷³⁸ There was no difference in median survival between the two arms (16.5 vs 15.1 mo, p = .58). These data suggest that the simple addition of etoposide to the CAV regimen could account for the improved median survival of the CAV/PE arm rather than the treatment strategy. Canadian investigators failed to show the advantage of the alternating CAV/PE strategy in patients with limited-stage SCLC when it was compared with sequential treatment with CAV (3 times) followed by PE (3 times).⁷⁴¹ German investigators also failed to show the advantage of the alternating chemotherapy treatment strategy over the sequential treatment using ifosfamide/etoposide followed by CAV.⁷⁴²

Nevertheless, these results do not invalidate the Goldie-Coldman hypothesis because none of these trials used regimens that were truly non-cross-resistant. In addition, many patients who failed to respond to CAV proceeded to the PE regimen. This may have blurred any potential survival advantage of the alternating treatment strategy and should be taken into consideration when designing a similar trial with truly non-cross-resistant regimens.

Randomized clinical trials have addressed the value of maintenance chemotherapy.^{701,740,743–748} These trials fail to show that the prolonged administration of the same regimen used as an initial treatment improves overall survival, although the duration of initial tumor response may be prolonged. Two large British studies compared 6 versus 12 cycles and 4 versus 8 cycles of chemotherapy, and an EORTC study compared 5 versus 12 cycles of chemotherapy.^743,744 No significant difference in survival is seen, if chemotherapy is given at relapse in patients randomized to short-term chemotherapy.⁷⁴⁴ In patients with limited-stage SCLC, results are quite controversial.⁷⁰¹ Overall, available data indicate that there is no proven benefit to maintenance chemotherapy. It is recommended that six cycles of initial chemotherapy is adequate, provided that additional chemotherapy is to be given at relapse. However, it does not completely eliminate the possibility that sequential or alternating applications of more effective regimens would improve the outcome, as exemplified by the SECSG study in which two additional cycles of PE chemotherapy following six cycles of CAV improved the median survival by almost 7 months among the responding patients with limited-stage SCLC.⁷⁰¹

Chemotherapy in Special Clinical Situations

For patients with SVC syndrome, acute airway obstruction, and CNS metastases, radiotherapy is a powerful local treatment modality and is commonly used. But, as more experience has accumulated with chemotherapy, it is apparent that chemotherapy provides rapid palliation of SVC syndrome in SCLC patients.^740,749,750 Therefore, chemotherapy alone is appropriate at least for the initial treatment of SVC syndrome unless chest irradiation is indicated otherwise. Similarly, combination chemotherapy alone is also appropriate treatment for airway obstruction at initial presentation.

For the brain metastases, systemic chemotherapy was believed to play a limited role until recently because of the widely held view of the brain as a pharmacologic sanctuary.⁷⁴⁰ In the 1980s, several groups observed in prospective trials that systemic chemotherapy alone, without brain irradiation, could induce objective regression of the brain lesions.^751–754 These results clearly indicate that the blood-brain barrier is not the major obstacle for the successful treatment of brain metastases and, in part, explain why patients who present with brain metastases at initial diagnosis of SCLC fare better than those who develop brain metastases later during the disease course.^755–757 As has been the case with SVC syndrome in SCLC, first-line therapy with cranial irradiation may not be necessary, even though this has not been addressed in a randomized trial setting. However, for patients with intracranial relapse, the available data indicate that the efficacy of second-line chemotherapy is generally much less than in the previously untreated patients. Conversely, patients with recurrent brain metastases after previous cranial irradiation should not be denied a trial of systemic chemotherapy solely because of the presence of intracranial lesions.