International Journal of Radiation Oncology * Biology * Physics
Volume 80, Issue 4 , Pages 969-977, 15 July 2011

Summary of Study Outcomes 

Symptom relief 

The goal of chest reirradiation for the majority of patients reported in these studies was palliation. That said, the description of symptoms relieved as well as the efficacy of palliation varied among the studies. Overall, hemoptysis appeared to be controlled in 33% to 100% cases, but if the study by Green was excluded due to the small number of events (1 of 3 responders), efficacy improved significantly (83%–100%). Cough was controlled in 50% to 77% cases, chest pain in 40% to 80% cases, and dyspnea in 35% to 100% cases; like hemoptysis, the efficacy of dyspnea relief was influenced by the small number of events reported in some studies. When global (i.e., any) symptom relief was considered, rates ranged between 70% and 80% of patients reporting some benefit of treatment, with the exception of lower rates for the studies by Green and Melbye  (48% relief rate) and Jackson and Ball.

Survival 

MSTs from reirradiation ranged from 5 to 14 months, while 1- and 2-year survival rates ranged 8.7 to 59%, respectively, and 58 to 27 months, respectively. Three studies that provided the longest MST and highest survival rates were those of Okamoto Wu , and Tada , all of which used reirradiation doses with median values of 50 to 51 Gy. This is in contrast to studies using lower reirradiation doses, where MSTs of 4.9 to 5.6 months were reported. In the series reported by Tada et al., the MST for 14 patients who received the prescribed dose was 10.5 months. Similarly, in the study by Okamoto  for the subset of patients treated with curative intent and higher RT doses, the MST was 15 months, and 2-year survival was 51%. The difference in median survival results suggests a possible dose response effect for EBRT. On the other hand, the study by Jackson and Ball did not report any survival differences between patients retreated with 20 Gy versus those treated with 30 Gy, but this may be because both dose levels were relatively low.

Toxicity 

Reirradiation was associated with relatively low rates of toxicity, with 5% complication rates observed in the earliest studies The most frequent event reported was esophagitis (20%); whereas radiation pneumonitis was reported in only 3% of cases; and RT myelopathy and rib fracture were rare events. Studies covering pre-1990 therapy were particularly limited by the absence of formal scoring systems to routinely document side effects. More contemporary studies employed validated scoring systems. Recent studies suggest low rates of grade 3 esophagitis (4%–6%), and grade 3 pneumonitis was noted in 5% to 21% cases. The higher incidence of grades 3 and 4 esophagitis (21%) reported by Okamoto et al.  may reflect the relatively high total and median cumulative doses used (150 Gy and 110 Gy, respectively). This increase in toxicity with total dose would also seem to be supported by the finding of a 35% rate of grade 2 pneumonitis, rates of which are much lower in the other studies. Toxicity data from the study by Wu et aare particularly important as they were collected prospectively Their results do recall those of Okamoto et al., with 26% of cases having grade 2 to 3 late lung fibrosis, of which only two (9%) were grade 3. Finally, rare but fatal (grade 5) events have been reported. Although in the study of Gressen et al. a single patient died due to possible RT pneumonitis, the authors did not exclude the possibility of disease progression as the cause of death.

Prognostic factors 

Few studies provided statistical analyses that sought prognostic predictors of outcomes. Age was not shown in any study to influence survival, with no significant cutoffs noted (e.g., no differences were noted for a cutoff at 70 years . Neither tumor size nor its location was shown to influence outcome . Patient PS  and treatment response  may positively influence outcomes. In the study by Tada et al. , survival was 12.6, 7.1, and 1.1 months for patients with PS 0 to 1, PS 2, and PS 3, respectively. Time interval between the end of the first RT and reirradiation appears to have a significant impact on survival. Studies by Tada et al.  and Cetingoz et al. showed that longer intervals may be advantageous for better survival. In addition, Kramer et al.  observed that none of the patients with an interval of <10 months survived for >5 months, while all patients surviving >1 year had a tumor-free interval of >1 year. This suggests a selection effect based on disease virulence. On the other hand, in the study by Gressen et al., time interval (≥10 months vs. <10 months) did not appear to influence outcome. Dose response was explicitly evaluated by Jackson and Ball  and, for the doses they employed (median doses of 20 Gy and 30 Gy), was not found to affect outcomes. Initial stage at diagnosis may influence survival after reirradiation, with stage I and II patients having better survival then those with IIIA or IIIB and with no difference between survival for patients with IIIA and IIIB. Last, symptom relief did not lead to improved survival according to the study by Gressen et al.

Overview of study limitations 

Valid comparisons between the results of the studies reviewed were limited by a number of factors, including but not limited to obvious patient numbers, selection and heterogeneity in the characteristics of patients treated, and nonuniform EBRT dosing and delivery. With respect to patient factors, few studies documented the characteristics of original population of initially treated patients from which the reirradiated patients were drawn, with inconsistent descriptions of age, sex, performance, and initial staging. At recurrence, a minority of studies reported that patients were restaged, and again, there was inconsistent documentation of relevant patient variables such as performance. It was noteworthy that some studies differentiated symptomatic from asymptomatic patients at the time of retreatment, even in the setting of palliative dose schedules.

The hallmark of these studies was the lack of uniformity in how RT was delivered prior to recurrence and then at the time of retreatment. Thus, initial EBRT doses ranged from 25 to 66 Gy, while those administered at the time of recurrence ranged from 6 to 57 Gy, with cumulative doses ranging, therefore, from 43 to 122 Gy. Typically, the reports varied by the amount of RT documentation provided: there was no consistent reporting of original RT delivered, cumulative RT doses, original fractionations, or original treatment fields. The form of EBRT at the time of reirradiation was better described, especially with respect to energies, design of targets, and field setup, but obvious lacunae include relating cumulative doses delivered to initial and reirradiated volumes. In that regard, most authors were clear that at reirradiation, treatment portals generally included only “visible” disease with a margin of 1 to 2 cm and no elective nodal treatment, and this was in distinction to the larger initial treatment fields. Last, there were no reports using radiobiological parameters such as biologically equivalent dose to characterize or describe the RT delivered prior to or at the time of reirradiation. The inconsistent use and interaction of CHT with repeat EBRT with respect to efficacy or toxicity was not addressed in any study.

The survival data presented in these reports was limited and no doubt reflected the palliative intent at the time of reirradiation. That said, it was interesting to note that two studies did not report details of symptom relief for their reirradiated patients. Despite the fact that all studies provided data for time intervals between initial RT and reirradiation, only three studies  provided survival data from initial RT. There was little documentation of local progression-free survival after reirradiation or distant metastasis-free survival. Few studied provided comprehensive pattern-of-failure documentation. Finally, only four studies attempted to evaluate survival according to potential prognostic factors such as interval between first and second RT, PS, age, size, location, or response

Discussion 

Overall, chest reirradiation clearly achieves symptom control in a substantial proportion of cases, and interestingly, longer survival times might be produced when higher doses are delivered at the time of reirradiation. Importantly, this retreatment also appears to be accompanied by low rates of high-grade toxicity (3%–5%), which appeared all the more remarkable in the setting that reported cumulative doses in the range of 85 to 100 Gy and sometimes delivered with concurrent chemotherapy as well. Certain treatment trends also were notable as the studies became more contemporaneous. Modern planning techniques were being used to design treatment fields for treating recurrences. Administering concurrent CHT during reirradiation appeared feasible based on the low toxicity rates in studies reporting its use. There was a trend to better reporting of not only early but also late toxicity occurring during and after reirradiation and for using validated toxicity scoring systems.

These studies also support the truism that intrinsic tumor behavior is likely deterministic. In other words, the time interval between first and second RT seemed to be an important factor for understanding outcomes after treating recurrent disease. The study by Tada suggested that time interval was an important factor influencing treatment outcome, as did the study by Cetingoz et a, who performed a multivariate analysis showing that time interval between initial RT and reirradiation was the only independent prognosticator influencing overall survival. These findings may be a reflection of a variety of mechanisms interacting at the tumor level: indolently behaving tumors, differential radioresponsiveness with better initial local control, and independent dose-response effects. On the other hand, one might also speculate that the efficacy of reirradiation is independently associated with the characteristics of the initial RT. Interestingly, in a large systematic review of palliative thoracic RT a nonsignificant trend was noted for more retreatment to the thorax after low-dose palliative RT compared to that after high-dose RT

This review therefore leads to the recommendation that as novel diagnostic and therapeutic tools become increasingly available, it is becoming even more critical to address clinical questions on the efficacy and safety of reirradiation within the context of well-designed prospective trials designed for appropriately stratified homogeneous cohorts of patients. It is in that context then that recent advances in RT delivery may be highly relevant to the setting of reirradiation. Reports of the use of modalities such as SBRT and IMRT are now being published . As noted earlier, Beavis et al.had reported a planning study of IMRT in the retreatment of a patient with NSCLC. IMRT appears particularly suited to the setting of retreatment as dose can be highly conformed to limit exposure of normal tissues. Although the dose distributions to organs at risk achieved by IMRT were similar to those with conventional techniques, Beavis  found that the target coverage given by the conventional treatment option was also clearly inferior to that offered by the IMRT plan.

More recently, two reports of the use of SBRT for recurrent NSCLC were published almost simultaneously in 2008. Chang et al.used 4D CT-based planning to deliver 40 to 50 Gy to 14 patients with isolated recurrent NSCLC previously treated with definitive RT with or without CHT or surgical resection before SBRT. With a median follow-up of 17 months (range, 6–40 months), the crude local control rate at the treated site was 100% for patients treated with 50 Gy. Four (29%) patients developed grade 2 pneumonitis, but no patient experienced grade ≥3 toxicity. Coon et al. reported on the use of a similar fractionated stereotactic RT regimen but using a CyberKnife delivery system. For 12 patients with recurrent tumors, a dose of 60 Gy was given in 3 fractions. Importantly, for the majority of patients, pretreatment positron-emission tomography (PET)-CT scans were performed to aid in delineation of tumor volume, and at the time of follow-up, all patients received scheduled CT or PET-CT imaging to assess responses. Overall response rate was 75%, while 17% of patients experienced stable disease. At a median follow-up of 11 months, the local control rate at the site treated was 92%, and overall survival was 67%. Only 1 (2%) patient experienced grade 2 pneumonitis.

Conclusions 

For selected patients, chest reirradiation appears to be feasible, safe, and effective at relieving symptoms associated with recurrence. Depending on patient and disease presentation at relapse, long-term control also may be possible by increasing RT dose. Modern RT techniques likely enhance the ability to deliver any treatment for patients who previously received irradiation to the chest. Well-designed prospective clinical trials are the most appropriate setting in which to determine the optimal way of integrating current technologies into the care of patients with locally recurrent lung cancer.