During the past two decades, relevant progress has been achieved in multimodal treatment strategies for the management of localized NSCLC [25]. The recent introduction of consolidation durvalumab yielded 5‑year survival rates of 42.9% with RCT–immunotherapy in stage III [26, 27]. Boys et al. studied 126 patients with definitive RCT for stage III NSCLC and evaluated the eligibility for durvalumab consolidation in clinical practice [28]. A total of 56/126 patients (44.4%) were ineligible for consolidation [28]. The reasons included progressive disease or death (n = 14; 25%), RT dose < 54 Gy (n = 9; 16%), and radiation pneumonitis ≥ grade 2 (n = 8; 14%) [28]. Thus, optimization of RCT with precise target volume definition, sufficient RT dose, and avoidance of toxicities potentially has a beneficial influence and, thus, is of utmost relevance for the efficacy of multimodal treatment [29, 30]. Before the advent of immunotherapy in the treatment of stage III NSCLC, in a cohort study from 2010–2018 in Denmark, Meldgaard et al. reported improved outcomes in the more contemporary study period of 2016–2018 [2]. These improvements certainly reflect, to a major extent, advances in imaging modalities, RT planning, and RT techniques [5, 25]. In 2020, The German PET-Plan trial prospectively compared PET-CT-based planning vs. conventional planning in definitive RCT for NSCLC [6, 31]. Nestle et al. reported a potential for improved local control and good feasibility [6]. In 2017, in a secondary analysis of the RTOG 0617 study, Chun et al. found lower rates of toxicities with a dynamic RT technique (IMRT) [5]. Herein, we studied the specific clinical outcomes of patients who were treated with vs. without these refined strategies (PET-CT-based staging and dynamic RT techniques [IMRT/VMAT]) from 2008 to 2019.

In this study, PET-CT-staged patients presented with lower T stages (cT1–2 vs. cT3–4; p = 0.049) than conventionally staged patients. Comparably, when Vokes et al. analyzed the impact of PET-CT on survival in 598 NSCLC patients, stage IIIA (vs. IIIB) was diagnosed in 49% of the patients with PET-CT vs. 39% of patients without PET-CT [32]. The differences in the presented groups can be explained by our study design. We included patients who were referred to the radiation oncology department for curative RT/RCT. However, PET-CT vs. conventional CT-based staging can detect additional metastases in up to 10% of patients [33]. Thus, in patients undergoing PET-CT, upstaging might have led to a stage shift from localized cT3–4 tumors (Hao et al.; association of T stages with rates of metastasis [34]) to UICC stage IV. This might explain the overrepresentation of T1–2 patients relative to T3–4 patients in the current PET-CT study group [35].

When compared to conventionally staged patients, PET-CT-staged patients received higher RT doses (median 66 Gy vs. 60 Gy; maximum dose 70 Gy; p < 0.01), and had higher RT completion rates (87.1% vs. 77.7%; p = 0.04; ≥ 80% of the planned dose applied: 93.5% vs. 82.5%; p < 0.01). Additionally, with only a trend towards statistical significance, PET-CT-staged patients had higher rates of RCT vs. RT only (PET-CT: 84.7%; conventional: 76.7%; p = 0.1) and, in patients who received RCT, with concomitant cisplatin/vinorelbine vs. low-dose cisplatin (PET-CT: 41.7%; conventional: 25.3%; p = 0.08). Current studies aim at treatment intensification via RT dose escalation in adequately selected patients, tumor subsets, and treatment volumes and via optimization of systemic treatment [36]. Here, we have demonstrated relevant improvements in the PET-CT-staged study group. At the same time, rates of lung infections (≥ grade 2, p = 0.02; ≥ grade 3, p = 0.01) and leukopenia (≥ grade 1, p < 0.01) were higher in PET-CT-staged patients. These findings might be explained by intensified treatment in this group (previous studies; increased rates of hematologic toxicities/leukopenia with RCT vs. RT [37] and with cisplatin/vinorelbine vs. low-dose cisplatin [7, 38]; neutropenia as an important risk factor for infections [24]). In the present study, these relations were partly demonstrated for patients with leukopenia (see “Toxicities”; higher proportion of patients with RCT vs. RT only, with ≥ 80% of the planned RT dose applied, and with a higher applied RT dose). However, putatively due to the retrospective study design including patients within a long period of time, there were no significant differences in parameters for patients with lung infections. The coincidence of improved outcomes and higher toxicity rates in PET-CT-staged patients emphasizes the need for a detailed risk–benefit assessment during planning and application of treatment modalities [39].

Patients with PET-CT vs. conventional staging experienced better OS, PFS, LRPFS, and DPFS. Comparably, Vokes et al. found improved survival with PET-CT vs. conventional staging in RCT for localized NSCLC [32]. The influence on survival might, to a certain extent, be attributed to the detection of occult metastases and stage migration (in the presented study, possibly leading to overrepresentation of T1–2 patients in the PET-CT group, please see above) [32, 40]. The multivariable model with inclusion of parameters with a possible influence on outcomes was established to take these aspects into account. Remarkably, the survival advantages in PET-CT-staged patients were retained in multivariable analysis. Additionally, we found an association of previously reported baseline parameters (age [7, 41], Karnofsky index/performance status [42]) and of intensified treatment (applied RT dose > 60 Gy vs. ≤ 60 Gy; maximum dose of 70 Gy [43]) with survival. Taken together, with positive selection of PET-CT patients via detection of occult metastases, it can be concluded that PET-CT is an important basis for optimal RT/RCT indication setting [32, 40].

PET-CT-staged patients were treated more frequently in the more contemporary time period (12/2013–12/2019 vs. 01/2008–11/2013) when compared to conventionally staged patients. These findings reflect the increasing application of PET-CT in clinical routine (Bedir et al.: 2007–2010, initial implementation of PET-CT; 2011–2014, era in between; 2015–2018, PET-CT widely available and used) [44]. Next, patients in the PET-CT group were irradiated more frequently with modern RT techniques (IMRT/VMAT vs. 3D-CRT) when compared to conventionally staged patients. This reflects distinct parallels of RT technique and PET-CT introduction and further implementation in the clinical routine of NSCLC treatment [44].

In the presented study, we found a higher RT dose with IMRT/VMAT vs. 3D-CRT (median 66 Gy vs. 60 Gy; p = 0.03). Additionally, IMRT/VMAT-treated patients had higher rates (trend, not statistically significant) of intensified chemotherapy (concomitant cisplatin/vinorelbine: IMRT/VMAT, 50.6%, 3D-CRT, 34.5%, vs. low-dose cisplatin: IMRT/VMAT, 39.2%, 3D-CRT, 52.7%, see Table 4 for other types of chemotherapy, p=0.09). Comparable with the influence of PET-CT-based staging, this provides evidence that modern RT techniques enable treatment intensification. In contrast to findings with PET-CT, there were no effects of IMRT/VMAT on survival (OS, PFS, LRPFS, DPFS). Outcomes were comparable with IMRT/VMAT (3-year OS 26.9%) vs. 3D-CRT (3-year OS 19.3%). These results are in line with previous studies (no differences in OS, PFS, local failure, distant metastases) [5]. Chun et al. discussed that—via advantages in heart or lung sparing—dynamic RT techniques could be associated with superior survival in long-term outcomes [5]. However, a recent press release indicated that there were no differences at 5‑year follow-up [45].

Furthermore, with IMRT/VMAT vs. 3D-CRT, we found reduced lung exposure at higher dose levels (V40Gy) and increased exposure at lower dose levels (V5Gy and V10Gy; each, p < 0.01). These findings are as expected (general increase in low-dose bath with dynamic RT techniques [46]) and comparable with a planning study by Li et al. (higher dose levels, V30Gy, advantages of IMRT/VMAT; lower dose levels, V5Gy, highest exposure with VMAT) [47]. In line with previous studies [5], we found lower rates of pneumonitis with IMRT/VMAT (≥ grade 2, p < 0.01). Additionally, radiation dermatitis was reduced with IMRT/VMAT (≥ grade 1, p = 0.01). Reduced RT-induced skin reactions with dynamic RT techniques have previously been described for different tumor entities (IMRT vs. 3D-CRT in breast cancer [48], VMAT vs. 3D-CRT in anal cancer [49], and VMAT vs. 3D-CRT in rectal cancer [50]). Interestingly, in spite of more intensified systemic treatment in IMRT/VMAT-treated vs. 3D-CRT-treated patients (previous studies: higher rates of nausea with cisplatin/vinorelbine [7]; presented study: trend towards higher rates of concomitant cisplatin/vinorelbine vs. low-dose cisplatin), rates of nausea were lower with IMRT/VMAT (≥ grade 2, p = 0.045). Since IMRT/VMAT was more frequently applied in the later period (12/2013–12/2019 vs. 01/2008–11/2013), it might be hypothesized that general optimization of antiemetic treatment could have contributed to lower rates of nausea (first American Society of Clinical Oncology [ASCO] guideline on antiemetics published in 1999; updates in 2006, 2011, 2015, 2017, and 2020 [51]).

Finally, the limitations of the presented study are discussed. RT with IMRT/VMAT and staging with PET-CT can be considered as a current standard of care for NSCLC patients. Thus, conventional staging and 3D-CRT should not be regularly applied. At the same time, in clinical routine (e.g., when PET-CT is not available timely in patients with a high tumor burden who require urgent RT), PET-CT will not be realizable in certain patients. Additionally, we present a study with a detailed analysis of patient characteristics and clinical outcomes. Thus, the presented study adds relevant information on personalized treatment in localized NSCLC.