Introduction

Non-small cell (NSCLC) and small cell lung cancer (SCLC) are tumor entities characterized by both a high incidence and a high mortality rate [1]. Risk factors for both include tobacco smoking and exposure to chemicals. SCLC accounts for approximately 15% of all lung cancers and is an aggressive tumor entity characterized by a rapid doubling time and early dissemination [2]. Since there is no effective screening method so far and symptoms are vague, about two-thirds of newly diagnosed patients present with locally advanced disease or even distant metastases, the latter in particular being the case in SCLC [3, 4]. Treatment of early-stage NSCLC consists of surgical tumor resection and lymphadenectomy in medically fit patients or (stereotactic) radiotherapy in frail patients or those refusing surgery [5]. Treatment of irresectable locally advanced NSCLC and limited disease SCLC consists of a combination of radiotherapy and platinum-based chemotherapy, preferably given concurrently, but also sequentially in patients with impaired medical condition [6,7,8].

PET imaging for staging in lung cancer

Typically, primary lung tumors are diagnosed on a planar X-ray acquired for diagnosis of pulmonary complaints, i.e., suspicion of pneumonia, work-up prior to cardiac procedures or rheumatologic disease. Thereafter, CT is the method of choice to depict the extent of the primary tumor and intrathoracic lymph node metastases in clinical routine [9]. However, staging on the basis of additional functional imaging by means of FDG-PET/CT has been shown to outperform that of anatomical imaging, and thus has found its way into numerous guidelines as well as clinical practice. This is true for detecting the primary tumor, particularly in the presence of atelectasis, for selecting the affected lymph nodes to be treated and for diagnosing distant metastasis.

The value of FDG-PET/CT for staging of locally advanced NSCLC has recently been summarized by Grootjans et al. [10]. A thorough review of the literature on FDG-PET/CT in staging of SCLC is thus far lacking and therefore given here.

Carter et al. [9] found an upstaging from limited to extensive disease in 19% of SCLC patients and a downstaging from extensive to limited SCLC in 8% of patients using FDG-PET in combination with conventional imaging methods. In addition, in a proof-of-principle study by Saima et al. [11], the rate of agreement between CT and FDG-PET/CT was calculated using Cohen’s kappa. They found a strong (κ = 0.82) correlation for determination of primary tumor, a fair (κ = 0.24) value for lymph nodes and only poor (κ = 0.12) agreement rates for metastases for both imaging techniques. In their population of 23 patients, 47% were upstaged with visceral and bone metastases not detected by means of only CT. Furthermore, PET imaging is more sensitive and specific for detection of distant metastases, except for brain metastases, compared to CT or magnetic resonance imaging (MRI) [9, 12, 13]. Overall, these findings support the important role of FDG-PET/CT for improvement of initial SCLC staging with corresponding prognostic implications.

In SCLC, additional functional imaging techniques, e.g., a 99mTc-labeled vasopressin conjugate as a potential radiopharmaceutical for imaging of oncogene receptors overexpressed, have only been reported in single case studies or in preclinical investigations, but were thus far not translated into clinical routine [14].

PET imaging for radiation treatment planning

Since primary radiochemotherapy is the mainstay of treatment for patients with locally advanced NSCLC or limited disease SCLC and appropriate performance status, functional imaging could provide important information, especially on target volumes definition for irradiation [6, 8, 15, 16] (Fig. 1).

Fig. 1
figure 1

a Passive scattered proton beam therapy plan (in transverse, sagittal and frontal direction) for irradiation of a cT4N2M0 non-small cell carcinoma patient. b depicts the respective functional and anatomical imaging information gathered by 18Fluorodeoxyglucose-PET-CT at initial diagnosis and taken into account for radiation treatment planning

Full size image

The added value of FDG-PET for delineating the primary tumor over CT alone has been elegantly shown by Steenbakkers et al. [17]. The standard deviation of interobserver variation of the gross tumor volume (GTV) contoured by 11 experts in thoracic radiotherapy significantly decreased when offering the participants an FDG-PET/CT as opposed to a CT scan only (both without intravenous contrast agent). Moreover, a good correlation between the macroscopic tumor extension determined in the resection specimen and the size of the tumor on pre-treatment has been reported [18, 19]. Using FDG-PET/CT, atelectasis surrounding the tumor as well as tumor infiltration of the mediastinum can be better depicted largely affecting the precision of defining the target volume, and decreasing the irradiated volume as well as the dose to tumor-surrounding radiation-sensitive organs at risk.

In the era of two- and three-dimensional radiotherapy, being based on planar X-ray imaging or computed tomography, respectively, increasing the radiation dose from 50 to 66 Gy in NSCLC was merely hampered by the large, elective mediastinal volume irradiated. That is why in the early 2000s, there were increasing efforts to identify the affected lymph nodes and subsequently irradiate these selectively and to a higher dose. De Ruysscher et al. [20] and Belderbos et al. [21] pioneered this approach in locally advanced NSCLC patients both using a FDG-PET/CT-based selective nodal irradiation (SNI) using a 3D radiation technique. In the first study, 44 locally advanced NSCLC patients were irradiated to doses of 61.2–64.8 Gy (1.8 Gy fractions b.i.d.) and after a median follow-up period an isolated nodal recurrence rate of 2.3% was reported [20]. The same incidence of regional recurrences was reported by Belderbos et al. [21] in a cohort of 88 NSCLC patients treated with escalating radiation doses. Hypothesizing that the incidental dose to the non-selected regional lymph nodes may at least in part be responsible for these favorable results, we recently retrospectively analyzed the regional recurrence rate of locally advanced NSCLC patients having undergone SNI in the era of intensity-modulated radiation therapy (IMRT) [22]. Based on treatment plans and outcome data of 183 patients, the isolated nodal recurrence rate was 1.6% and the combined locoregional recurrence rate 2.2%. Thus, the concept of SNI is still valid in the era of high-conformal radiotherapy. The guideline proposed by Senan et al. [23] is still applicable nowadays, stating that FDG-PET positive lymph nodes should be included in the GTV as should PET-negative lymph nodes with either a large necrotic core on CT or with pathological conformation on cytology.

In contrast to NSCLC, where SNI has been accepted as gold standard, there is no final consensus on nodal irradiation volumes in SCLC, even though most institutions nowadays adhere to the selective concept. In a small prospective study on omission of elective nodal irradiation (ENI) in patients with SCLC, noteworthy based on CT scans only, an unexpectedly high rate of isolated nodal failures (11%) occurred [24]. In contrast, another prospective phase II study on CT-based omission of ENI in 38 SCLC patients revealed no isolated nodal recurrence [25]. These results are supported by further retrospective data on CT-based SNI in SCLC [26]. Conversely, radiation treatment planning based on FDG-PET/CT scans reported by van Loon et al. [27] and Shirvani et al. [28] revealed considerably lower rates of isolated nodal failures in the order of 2–3%. These data seem to support the use of SNI also in limited-stage SCLC when incorporating FDG-PET/CT thus resulting in reduction of radiation dose to organs at risk and potentially increase dose to the target [29, 30].

Regarding this, the results of the recently published CONVERT trial are also of major interest [31]. This multicenter randomized phase III superiority trial compared the currently used twice-daily irradiation (45 Gy in 1.5 Gy fractions b.i.d. [8]) to a once-daily radiotherapy with a higher total dose in line with locally advanced NSCLC (66 Gy in 2 Gy fractions), both with concurrent chemotherapy. Interestingly, the survival outcomes did not differ significantly between both regimes and toxicity was both lower than expected and similar in both groups. In this study, FDG-PET/CT was non-mandatory for staging and radiation treatment planning including the definition of the target volume for mandatory SNI. Since the results are somewhat unexpected, the authors discussed that potentially the prolonged overall treatment time may be the cause and that dose escalation to a biologically defined sub-volume may be of benefit, again underlining possible advantages of specific functional imaging techniques in SCLC.

Prognostic value of PET imaging before, during and after treatment

In addition to treatment planning and staging accuracy, functional imaging, especially different FDG-PET parameters, may also have prognostic value in patients with lung cancer. In locally advanced NSCLC patients, repeat FDG-PET/CT imaging has been conducted for two reasons—first to predict treatment outcome by comparing pre- to per-treatment functional imaging, and second to correlate the high uptake volume prior to treatment with that of the recurrent disease.

A recent review has summarized the role of FDG-PET/CT for early response evaluation in NSCLC [32]. From this it can be appreciated, that results differ widely, some of which are given here. Van Elmpt et al. [33] performed FDG-PET/CT scans before and in the second week of radiochemotherapy in 34 consecutive locally advanced NSCLC patients. They found that CT-based tumor volume did not correlate with overall survival (OS), whereas a decrease in mean standardized uptake value (SUVmean) of 20 ± 21% was found in those patients surviving 2 years as opposed to an increase by 2 ± 22% in non-survivors. In a prospective imaging study, 28 locally advanced NSCLC patients underwent FDG-PET/CT before treatment, at the end of the second week of treatment, and 2 weeks and 3 months after completion of radiochemotherapy [34]. These FDG-PET/CT scans were evaluated regarding maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG). The authors found pre-treatment TLG to be a prognostic factor for worse progression-free survival (PFS). Moreover, a decrease in TLG of more than 38% between pre- and per-treatment imaging was associated with a significantly longer PFS. Grootjans et al. [35] evaluated the use of TLG derived from different automatic segmentations algorithms for early response monitoring (i.e., prior to and in second week of treatment) in 27 of these NSCLC patients undergoing radiochemotherapy. The authors reported pre-treatment TLG of the tumor to be predictive for PFS and OS, and the addition of the TLG of the metastatic lymph nodes to improve assessment. Moreover, the difference between pre- and per-treatment TLG of the summed primary tumor and lymph nodes was again statistically significantly associated with PFS and OS. Bearing these results in mind, patient selection for treatment intensification may be a next step. However, it would be appealing to define the target for dose escalation to spare the tumor-surrounding organs at risk.

To find a putative target volume for dose escalation, Aerts et al. [36] assessed 22 locally advanced stage NSCLC patients with persistent FDG-PET/CT after radio(chemo)therapy taken from a total cohort of 55 patients. The authors found that the high FDG uptake volume (standardized uptake value 50%; SUV50%) largely corresponds with that prior to treatment. For all patients, the hotspot of the residual volume (SUV90%) was completely inside the GTV, and largely overlapped with the pre-radio(chemo)therapy SUV50% volume. This finding was validated in an independent patient cohort [37] and led to the design of a prospective multicenter phase II clinical trial, called the PET-boost trial. In this trial, the radiation dose in locally advanced NSCLC patients was increased to either the entire GTV or the SUV50% volume keeping within the tolerance doses for radiation-sensitive organs at risk [38]. The study was finished in October 2017 after completion of patient accrual and results are eagerly being awaited.

In SCLC, several studies on the prognostic value of FDG-PET/CT imaging have been published during the last years and reporting conflicting results. Starting with the SUVmax, as the most frequently used PET parameter, an association between high pre-treatment SUVmax and worse OS or PFS was found in some investigations [39,40,41] whereas other authors could not confirm these correlations [42, 43]. One possible explanation could be the lacking correlation between SUVmax of the primary tumor and disease stage [9]. In contrast, volume-based PET parameters seem to have a more homogeneous association with patient outcome. In particular, the MTV has been shown to provide significant prognostic information on OS and PFS in patients with SCLC [41, 43, 44].

Nevertheless, most of these studies have been conducted retrospectively in heterogeneous patient cohorts limiting the studies’ validity. Some authors even propagate the use of these parameters only in sub-groups, e.g., patients with limited-stage SCLC. Lee et al. [45] evaluated tumor metabolic activities using FDG-PET/CT and their relationship with markers of biological behavior (e.g., lactate dehydrogenase, glucose transporter 1) in SCLC patients. In multivariate analyses, pre-treatment PET parameters in combination with some biological markers remain significant. These results suggest better prognostic values by combining functional imaging with additional biomarkers.

There is only limited data on the significance of FDG-PET/CT for validity of early response or post-therapeutic restaging in SCLC patients. One study on response to first-line chemotherapy in extensive disease SCLC patients suggested that greater MTV and TLG may correlate with poor response [46]. The authors concluded that both PET parameters may be used for therapy decisions in patients who are not suitable for first-line chemotherapy [46]. An earlier investigation on the same topic revealed early metabolic response in CT and FDG after start of chemotherapy as significant prognostic factor for survival in patients with SCLC [47]. One final aim out of these data could be patient selection according to predicted outcome for individualization of treatment.

The use of FDG-PET for post-therapeutic restaging in SCLC patients is currently not a clinical routine. In some studies, comparing this functional imaging modality to CT, 20–57% of the patients were found to have more tumor mass and 14–38% less disease burden than expected using CT alone [9]. Despite this heterogeneity, FDG-PET is suitable for evaluation of residual and/or recurrent disease and may thus be offered to symptomatic patients in good general condition amenable to first-line treatment if residual or recurrent disease is confirmed.

Current developments

During the past few years, radiomics, i.e., automated analyses of large amounts of imaging features, has been shown to provide new potential for personalized medicine [48,49,50]. With the help of such unique information about imaging-based tumor characteristics, in particular non-small cell, lung carcinoma, individualized, multidisciplinary strategies for improved patient outcome seem possible [51]. However, up till now no prospective clinical studies on this approach can be found on www.clinicaltrials.gov thus far, a prerequisite for this experimental approach being introduced in clinical practice.

Except for FDG-PET, other PET tracers depicting tumor characteristics, such as hypoxia and tumor cell proliferation have been investigated in NSCLC. HX4, a hypoxia-related PET tracer, was found to represent overlapping, but also distinct tumor subvolumes in NSCLC patients undergoing both HX4- and FDG-PET imaging [52]. So, depending on the findings of the above-mentioned PET-boost study, HX4 may be a different target for dose escalation. Moreover, the value of repeat 18fluorothymidine (FLT-)PET for outcome prediction in NSCLC patients undergoing radio(chemo)therapy was assessed in 37 locally advanced patients [53]. Paradoxically, stable FLT-PET readings in the second week of treatment were associated with longer overall survival and progression-free survival, as opposed to the hypothesis that patients with decreased FLT-PET uptake would do better, as found in other solid tumors [54, 55]. Thus, the use of this tracer for scientific purposes has been abandoned.

With the introduction of combined MR photon-based linear accelerators in the field of radiation oncology, the value of MRI for identification of affected lymph nodes in the hilum and mediastinum has been assessed in a recent meta-analysis [56]. In the per-patient and per-nodal analysis, this study confirmed that in particular functional MR imaging may augment selective nodal irradiation. Thus far, however, this approach has not been tested in the context of a clinical trial.

Conclusion

In summary, the introduction of FDG-PET/CT in staging, radiation treatment planning and response evaluation has immensely altered the management and target volume definition in lung cancer patients. Currently, studies incorporating metabolic tumor information in defining the boost target volume are being conducted and results are eagerly awaited.