Consolidation radiotherapy for patients with extended disease small cell lung cancer in a single tertiary institution: impact of dose and perspectives in the era of immunotherapy

Of Majority (47.5%) of patients received all 6 planned cycles of chemotherapy, 66 patients (35.3%) received less than 4 cycles of ChT. Etoposide with platinum was the most frequent combination (83.8%), the rest received anthracycline based ChT. In the group with cRT were less patients who received 4 ChT cycles or less.

Radiotherapy with linear accelerators (photon beam 6-10MV), based on 3D CT-based conformal radiation therapy planning, started after ChT. There was no difference in frequency of patients who started before (29 patients) and after 4 weeks (25 patients) of ChT completion. Prophylactic cranial irradiation was delivered with two opposed lateral fields with the dose of 25 Gy in 10 fractions using 2D planning and 6MV photon beam energy.

The primary endpoints in this analysis were mOS, 1-year and 2-year OS of ED-SCLC patients treated with ChT only versus patients treated with ChT and cRT and those receiving higher vs. lower dose of cRT. Median OS was calculated from the time of diagnosis to the time of death due to any cause or last follow up visit. Kaplan-Meier (KM) method and log-rank test were used for comparison of survival curves between different groups. Cox proportional hazards algorithm was used for univariate and multivariate analysis. Association between subgroups and clinico-pathological characteristics of patients were tested using chi-square method. All p values reported were based on 2 side hypothesis. The statistical analysis was computed using SPSS v.20 statistical package.

This survey was approved by Institutional Ethics Committee and Institutional Review Board in December 2017.

Baseline characteristics of 187 patients, divided to those with ChT only and those who also received cRT are presented in Table 1. The two groups were balanced regarding gender, age, T and N stage and metastatic locations. However, lower number of patients received 4 or less cycles of ChT and had 2 or more metastases present at diagnosis in ChT plus cRT group.

Table 2 present baseline characteristics of 59 patients who received > 30 Gy cRT, comparing those with higher dose (45 Gy) cRT and lower dose (30–36 Gy). In summary, median age was 63 years, more than half were men. Majority of patients were younger than 65 years. Unfortunately, reliable PS could not be retrieved from medical records for half of the patients and more than 10% of patients had PS 2-3 before cRT. Non-significantly more patients had larger tumors (T3-4) and more extended lymph node disease (N3) in the group treated with lower dose RT. For more than 10% of patients with central tumors, the size of tumor (T) or nodal status could not be determined. Fifty-eight percent of patients had one metastatic site. The most frequent site of metastases were liver. Less than third of patients had PCI.

Median OS of patients who had either BSC or RT only was poor, 1.86 and 2.42 months, respectively. Patients who had any form of additional chest irradiation (173 patients) had significantly better mOS than 113 patients with ChT only (9.9m vs. 7.6m, p = 0.002).

Consolidation RT was delivered to 74 patients. Those patients had significantly longer mOS compared to 113 patients with ChT only as presented in Figure 2, 11.1 months (CI 10.1–12.0) vs. 7.6 months (CI 6.9–8.5), p < 0.001. They also had significantly longer 1-year OS (44% vs. 23%, p = 0.0025), but non significantly longer 2-year OS (10% vs. 5%, p = 0.19).

Figure 2

Univariate survival analysis (UVA) for patients with or without cRT included the following variables: cRT, gender, age, number of ChT cycles, T and N stage, metastatic location, number of metastatic locations and PCI. Presence of cRT, female gender, number of ChT cycles (4 or less and more than 4) and PCI were significant in univariate analysis and were tested in multivariate analysis (MVA) (Table 3). Except for gender, they were all independent predictors of better survival.

In the group of 59 patients irradiated with cRT ≥ 30 Gy patients irradiated with 45 Gy had better mOS compared to patients irradiated with doses 30–36 Gy, 17.2 months vs. 10.3 months, p = 0.03. (Figure 3) Patients with higher dose of consolidation RT had significantly longer 1-year OS (68%) than those with lower dose (30%), p = 0.01, but non-significantly longer 2-year OS (18% vs. 5%, p = 0.11).

Figure 3

In the group of patients with cRT, we made another analysis. We included gender, age categories, PS before RT, RT dose, T and N stage, metastatic locations, number of ChT cycles, number of metastatic lesions, PCI and timing of RT in UVA. Statistically significant predictors of longer mOS were PCI irradiation and higher RT dose. Both were analyzed in MVA (Table 4) and remained independent predictors of improved survival. (PCI HR = 0.51, 95% CI 0.27–0.96; higher RT dose HR = 0.47, 95% CI 0.25–0.87).

Our analysis showed that cRT significantly improved mOS compared to patients who had ChT only, 11.1 months vs. 7.6 months. Those patients also had significantly longer 1-year OS (44% vs. 23%) and non-significantly longer 2-year OS (10% vs. 5%). Apart from cRT, independent predictors of survival were also PCI and higher number of ChT cycles delivered. Unfortunately, the response to ChT could not be included to our analysis, as due to retrospective nature of this study the response to ChT was not uniformly evaluated.

How results of our study compares to others is presented in Table 5. In a retrospective study of 119 patients by Zhu et al., survival results were much better than in our analysis, with mOS of 17 months for patients in ChT plus cRT group and 9.3 months for those with ChT only, and 2-year OS of 35% and 17%, respectively. They delivered higher cRT dose (range 40–60 Gy) and had comparable mOS (17.2 months) as our group of patients irradiated with 45 Gy.10 Study by Yee et al. included only 33 patients, all with PCI and cRT (40 Gy), but their reported mOS of 8.3 months is lower than ours.11 Another small retrospective study of 19 patients with cRT 40 Gy in 15 fraction reported mOS 14 months with 1-year and 2-year OS 58% and 14%.12 Difference in the results of these studies show that survival benefit could not be attributed to RT only, but also to the increased chances of those patients who remained in a better shape and fitter at the time of disease progression to receive subsequent lines of chemotherapy. Data from SEER analysis on almost 7000 patients also provide evidence that radiotherapy for thoracic lesion and any metastatic sites could significantly improve the OS, except for brain metastasis.13

Three prospective randomized trials researched impact of RT on survival in ED SCLC.8, 9,18 Trial by Jeremic et al. differs in many ways from more recently reported studies. They used accelerated hyperfractionation (54 Gy in 36 fractions) with concomitant ChT after 3 cycles of induction ChT and additional 2 cycles after RT in one group or after 5 cycles of ChT in another, both groups also eligible for PCI. They studied combined modality treatment rather than cRT. The reported mOS was excellent for those who received RT early (17 months) as compared to those who received late RT (6–8 months).18 Another concern regarding hyperfractionated RT is that is delivered twice daily (BID) and is technically challenging for patients with bilateral mediastinal lesions, which represented the majority in our population. Further, patients selected for combined modality treatment, which incorporates BID RT must have excellent performance status and baseline pulmonary function. In our study more than 10% of patients had PS 2-3 before cRT and unfortunately in more than half of patients PS could not be reliably retrieved from medical records.

Phase III EORTC study (CREST) included patents with PS 0–2 without brain and pleural metastases. Responders after 4–6 cycles of ChT and residual disease in the thorax were treated with irradiation of 30 Gy in 10 fractions.8 Contrary to our results, no benefit was shown for added RT after ChT regarding mOS, which reported to be 8 months in both groups and for 1-year OS (33% for ChT with cRT vs. 28% for ChT group only). However, they reported significant difference in 2-year OS 13% vs. 3% (p = 0.004). It should, however, be noted that mOS was calculated from the randomization while mOS from diagnosis (as calculated in our analysis) was 12 months.

More aggressive thoracic irradiation was given in RTOG 0937 trial with 45 Gy in 15 fractions.9 Reported median OS (15.8 months) was better than anticipated and much better than in CREST and our study. Unlike all other studies, they reported better mOS for ChT only group (15.8 months) than for ChT plus RT group (13.8 months), though the difference was not statistically significant. 1-year OS was similar, surprisingly higher for ChT only than for ChT plus cRT group (60.1% vs. 50.8%).

Two meta-analyses were published. The first, published by Palma et al. in 2015 included 2 studies with 604 patients, while the second published in 2019 by Rathod et al. added also 86 patients from prematurely closed RTOG 0937 data.19, 20 First meta-analysis found increased OS (p = 0.01), while the second failed to show improvement in overall survival by adding cRT to ChT, (p = 0.36).

We found that patients who had been irradiated with higher dose (45 Gy in 18 fractions) had better mOS compared to those who received lower doses 30–36 Gy (in 10–12 fractions), 17.2 months vs. 10.3 months. Patients with higher dose of cRT had better 1-year OS (68%) than those with lower dose (30%) and also better 2-year OS (18% vs. 5%).

Not many studies looked into dose difference for cRT. In retrospective study including 306 patients of whom 170 received cRT, those with higher RT dose (BED > 50 Gy) had longer 2y-OS, 32.3% vs. 17% (p < 0.001), respectively.21 In recently published retrospective analysis of National Cancer Database that included 3280 patients they also reported that patients treated with the dose at least 45 Gy had better survival; 1-year OS was 58.1% and 2-year OS was 25.2% compared to 43.8% and 15.1% for lower dose.22 Our results for 1-year OS compare favorable, but 2-year OS data are lower, suggesting our subsequent treatments were not as effective.

In CREST study, cRT dose used was 30 Gy in 10 fractions. The relative high intrathoracic failure rate of 42% indicated that this dose might be insufficient to eliminate all the residual disease. In additional analysis from CREST study, for patients with complete intrathoracic response no benefit of TRT was observed. They concluded that TRT should be offered to patients with a good or partial response after chemotherapy, but not to those without residual disease in the thorax. It appears that the greater the volume of the residual disease in the thorax is, the higher dose is needed to eliminate the tumor.

However, dose restrictions to the organs at risk and consequent toxicity limit the actual received dose.

Number of metastases was not predictive factor for survival in our analysis. Contrary to that, in recent retrospective publications it was shown that tumor burden of metastatic disease should be taken into account when treating ED SCLC patients, since those with ≥2 metastases had significantly worse outcome than those with only one metastasis.23, 24

No difference of timing was found in our survival analysis if RT started before or after 4 weeks after ChT completion. In RTOG 0937 trial and one retrospective Chinese study also no difference was found in survival for patients who received RT early or late.9, 25 On the contrary, meta-analysis for limited SCLC, showed that earlier or shorter RT brings 7.7% advantage in 5-year survival.26

In our study PCI was independent predictor of better survival, although only 21.9% of patients received one. Our previous publication, focused on impact of PCI on survival in patients with LD-SCLC, also showed that only low number of patients (6%) actually received PCI in routine clinical setting, nevertheless OS was improved with PCI.27 As our analysis is retrospective, this reflects real clinical situation. However, the reason why such a low number of patients actually received PCI is unclear. PCI as independent predictor of survival was reported also in retrospective study by Xu et al.21 PCI in ED-SCLC was studied in EORTC conducted prospective study that showed reduced incidence of symptomatic brain metastases and improved 1-year OS (27% vs. 13.3%, HR 0.68, p = 0.003). That study, however, was highly criticized due to the insufficient imaging prior to PCI.9 Japanese prospective study evaluated 224 patients with ED-SCLC who performed MRI prior to randomization to PCI or observation with MRI.28 The study was terminated prematurely due to lower rate of brain metastases in PCI arm (40%) vs. MRI observation only (64%), but they found no significant difference in 1-year OS. None of our patients had MRI prior to PCI and only one third had CT evaluation, indicating that imaging in routine clinical practice should improve. In CREST study PCI dose was not uniform (20–30 Gy in 5–15 fractions) with unusual hypofractionated dose (20 Gy in 5 fraction) used in majority of patients (62%).8 It was delivered concurrently with thoracic irradiation in 88% of patients, while other studies used sequential approach and uniform dose of 25 Gy in 10 fractions.9, 18 Difference in pre-PCI imaging and dose delivered as well as timing of PCI show diversified approach on this not fully researched area.29

Immunotherapy (IT) has been successfully incorporated into the treatment of metastatic non-small cell lung cancer (NSCLC) either as combination of ChT and IT or as mono-IT and lately also in stage III as consolidation treatment after concomitant chemoradiotherapy.30, 31, 32, 33, 34, 35, 36, 37, 38, 39

Recently, two randomized studies confirmed efficacy of IT also for the treatment in ED-SCLC. IMpower 133 study was the first to show improved OS in patients treated with atezolizumab combined with ChT (12.3 months) as compared to ChT plus placebo (10.3 months). 1-year OS rate was 51.7% in the atezolizumab group and 38.2% in the placebo group.16 Consolidation RT was not permitted, while patients could have PCI. The same criteria about cRT and PCI were also applied in CASPIAN study with durvalumab.17 Again, IT combination showed increased results, mOS in ChT-IT arm was 13 months and 10.3 months in ChT only arm and 1-year OS was 54% vs. 40%, respectively. Though PCI was allowed in the non IT group, only 8% of patients received it. If the inclusion of immuno-therapy would prove to reduce the incidence of brain metastases in ES-SCLC considerably in future trials as suggested from present studies, then PCI and consequently neurotoxic sequels could be omitted in the future. The decision about skipping cRT might be more challenging. Survival data from current studies has not shown superior survival in first line treatment with ChT-IT in ED-SCLC compared to studies with ChT and cRT. Could cRT be combined with IT during the consolidation phase? First reported data indicate that the combination is tolerable, however trials are still ongoing and safety as well as survival results are expected in the future.40 As previously reported, the use of thoracic RT may enhance the effect of IT by influencing the immune system and its interactions with cancer cells and tumors, recruiting anti-tumor immune cells, increasing the exposure of tumor antigens, and improving cross-presentation of these antigens to the adaptive immune system.41, 42, 43

Beside retrospective nature of our analysis we should acknowledge several other limitations of our research. The irradiation dose was not specified by the protocol or any other department regulation and the decision was under the discretion of treating physician. Larger tumors (T3-4, N3) were more frequently irradiated with lower dose, but this does not necessarily mean that larger tumors would not be feasible to the treatment with larger doses. Unfortunately, we were not able to retrieve reliable information about PS before RT in half of patients, reflecting real clinical practice. This would be valuable information as treatment decision in clinical practice is greatly influenced by PS and consequently might influence survival data. Due to the fact that not all patients were treated with ChT in our institution, PS before ChT could not be included in UVA and MVA. Also, the response to initial Cht as one of the main prognostic factors of cRT efficacy according to the published data, is missing, since not all the patients were treated at our institution. However, all the patients were discussed at the MTB before the treatment which at least partially reduces this shortcoming.