This study analyzed 57 patients in LD-SCLC and 69 patients in ED-SCLC. Results suggest that in radiochemotherapy of SCLC the timing of irradiation was associated with differences in overall survival and progression-free survival. Despite the partial lack of significance, it was identified that in LD-SCLC very late start of irradiation is superior to early start with respect to OS and PFS. In contrast, no superiority of early or late start of irradiation could be established in terms of OS and PFS. Toxicity in LD-SCLC was dependent on sex and mean lung dose. Analysis of the ED-SCLC group showed that starting irradiation after the second cycle of chemotherapy is associated with prolonged OS and PFS. This was the case for patients with Karnofsky performance status \(\ge\) 80 as well.

For LD-SCLC, existing data concerning the impact of start of radiation on OS remain inconclusive. The results of Wang et al. [22] confirmed superiority in OS and response of patients with complete or partial remission after two or three cycles of initial chemotherapy. Further studies have shown superiority of an early start of thoracic radiation over a late or very late start. Patients receiving irradiation during the first or second cycle of chemotherapy presented prolonged median OS [7, 23], PFS [23], 2-, 3 and 5-year OS [7, 24] and better local control [24].

However, when one of these studies was replicated, an increased OS in patients receiving radiation concurrently with their sixth cycle of chemotherapy compared to those receiving radiation concurrently with their first cycle (15.1 months vs 13.7 months) was reported [25]. However, due to the wide CI range of 0.72 to 1.28 results were deemed as insignificant. In contrast, the working group of Perry et al. found significantly increased rates of complete remission and 2-year OS as well as 2-year-failure-free-OS in the group with late start of radiation (starting during the fourth cycle of chemotherapy vs starting during the first cycle) [26]. However, treatments used in this study are hardly comparable to therapies applied in the clinical setting today. Other publications on this subject were unable to detect differences in OS and incidence of recurrence and presented no recommendation on the timing of irradiation [27, 28]. Our study identified increased OS and PFS in patients beginning irradiation after the start of the fourth cycle of chemotherapy. Although only multivariate analysis of PFS was significant, all results indicated the trend of superiority of very late start compared with early and late start of radiation in terms of OS and PFS. This could be explained by the fact that SCLC is very chemotherapy-sensitive [29]. After initial volume reduction by chemotherapy, radiation can target the reduced tumor tissue more intensively. It is safe to administer several courses of chemotherapy before applying radiotherapy. In addition, it should be considered whether the patients receiving late TRT were in better health conditions at the time of diagnosis. This could have led to the radiation being delayed even further or only during a recurrence. Therefore, these patients had a better chance of long-term OS and PFS from the start. Although, the sensitivity analysis performed in this study showed similar values in OS and PFS after diagnosis compared to OS and PFS after first irradiation, which suggests this bias can be objected. But it must be noted that due to the wide confidence interval in these results, a negative effect on OS and PFS cannot be ruled out. Previous publications have already identified further prognostic factors associated with prolonged OS, like female sex [30,31,32,33], KPS ≥ 70 [34] respectively Eastern Cooperative Oncology Group (ECOG) performance status 0–1 [33, 35] or N-Stage 0–2 [34] as well as therapy-related factors like radiation dose > 52 Gy [34] or PCI [36]. In our analysis, a trend of protective influence of female sex, KPS 80–100 in OS were detected but insignificant.

Since the influence of mean lung dose on toxicity dose was only significant after adjusting for PTV, it can be concluded that the TRT dose to the lung irrespective of the PTV is disadvantageous, while the contribution of the PTV itself to the lung dose is of little effect in terms of toxicity. Timing of irradiation had no influence on toxicity, a consistent finding with results of other publications [24, 27]. Takada et al. [7] suggested an increased hematological toxicity and esophagitis in patients with an early start of irradiation compared to a late start treatment. It should be mentioned that the frequency of chemotherapy varied in the study. Patients with early start TRT received cytostatics in an interval of three weeks and patients with late start TRT in an interval of four weeks. This may have an impact especially on hematological toxicity. Another study supported the results of Takada et al. by detecting increased rates of neutropenia in patients undergoing a simultaneous start of chemotherapy and radiation compared to patients starting radiation after three cycles of chemotherapy [26]. As demonstrated by Singh et al. [31] toxicity could also depend on sex. In their analysis women were more likely to suffer from hematological and gastrointestinal side effects like vomiting and stomatitis as well as infection. Our study also identified sex as a factor influencing toxicity, but in contrast to Singh et al. different toxicities were observed. Unlike in the study by Singh et al., female sex is identified as protective factor in this analysis. Analysis of outcome variable occurrence of pneumonitis detected an insignificant trend of female sex and lower mean lung dose as protective factors as well. However, although insignificant, late and very late start of irradiation appeared to be associated with an increased incidence of pneumonitis. This could be explained by the fact of a higher cumulative dose of chemotherapy at the start of radiation.

Skarlos et al. [28] compared start of irradiation with the first or with the fourth cycle of chemotherapy with the similar result of an increased occurrence of pneumonitis in patients with the late start of irradiation (not significant).

For ED-SCLC a role of thoracic irradiation is still controversial. As demonstrated by Slotman et al. [37] thoracic radiotherapy in addition to PCI is recommended in every patient with response after initial chemotherapy. Although they could not prove increased 1-year OS, 2-year OS was significantly prolonged. After TRT tumor progression was less likely and six months after radiotherapy PFS was better in the irradiated group than in the control group. As reported by further publications, it is recommended that TRT should be added to chemotherapy to reduce local recurrence [38] and to prolong survival [24, 39]. Shang et al. [40] detected that in ED-SCLC patients with distant metastasis, TRT improves OS, especially in those with only one metastatic site.

However, evidence of timing of TRT in ED-SCLC is still insufficient. The present study shows increased OS in patients receiving TRT during the third cycle of initial chemotherapy or later. This can also be explained by chemotherapy-sensitive SCLC [29]. Effective chemotherapy initially often results in rapid responses and noticeable improvement in symptoms [41]. The tumor tissue should primarily react to cytostatics. Since ED-SCLC is not entirely located in the lungs, but also in nodal and distant organ metastases, chemotherapy plays a bigger role than TRT at the beginning of therapy. This allows the TRT to have a more intensive effect on the remaining tissue. While mean OS was 8 to 13 months [17,18,19,20] in the pre-immunotherapy era, 39.4 months in late start irradiation represents a significant increase, but the wide 95% CI for OS in late start from 0.0 to 92.8 must be mentioned. The sensitivity analysis shows that good OS in patients with late start is not affected by the fact that the patient may be in a better health condition at the date of diagnosis receiving late TRT and patients in worse health conditions start radiation earlier, because OS after first irradiation give similar values. Values in PFS differ by T-stage, N-stage and timepoint of start of radiation. A possible explanation is that patients with a small tumor or low metastatic tendency are irradiated later on, or only after the tumor begins progression, and so they tend to continue to experience tumor progression immediately after their first irradiation, however, overall they have better chances for long PFS after diagnosis.

Although the influence on OS could be proven, evidence for prolonged PFS was only significant in univariate analysis but with a trend identified in multivariate analysis. In contrast to LD-SCLC, there was only one more study investigating the impact of timing of TRT in ED-SCLC. Luo et al. [10] defined early TRT as irradiation during the third cycle of chemotherapy or earlier and compared it to late TRT. Despite lacking significance, improved OS and PFS of patients receiving late TRT and in contrast, better locoregional recurrence-free survival of patients receiving early TRT was detected.

Another beneficial prognostic factor in ED-SCLC is Karnofsky performance status ≥ 80, influencing both OS and PFS, as demonstrated in this study. Further publications confirmed this factor or the equivalent ECOG 0–1. [32, 33, 35, 42] Although an advantage in OS of female sex could not be determined in this study, it has already been reported in other publications [31, 33]. Previously recognized harmful prognostic factors are tumor-related, including large tumor size, multiple metastatic sites at diagnosis [43] and patient-related factors such as smoking index ≥ 400 (number of cigarettes smoked per day * years of tobacco smoking) and age ≥ 70 [17, 44].

The retrospective nature of this analysis is its major limitation, leading to a lack of unity in chemotherapy regimens, dose and target volumes. In ED, multidrug chemotherapy includes different types of cytostatics and current immunotherapy as well as second line chemotherapy. Patients received chemotherapy over the course of five years, however, during this time the therapy recommendations have changed. All data were collected from clinical documents by referring hospitals. Some information on psychological toxicity, such as fatigue, is dependent on subjective assessment of patients. In addition, 126 patients were divided into two analysis groups (LD and ED), which therefore represented limited numbers and would need to be expanded to confirm the results. Differences in radiation dose, methods and chemotherapy and immunotherapy regimens can contribute to the bias of this study.