Reporting of Incidence and Outcome of Bone Metastases in Clinical Trials Enrolling Patients with Epidermal Growth Factor Receptor Mutated Lung Adenocarcinoma—A Systematic Review

Simple Summary Around 30–60% of the patients with lung cancer develop bone metastases, which are associated with decreased survival, bone pain, and skeletal-related events such as need for radiation. Patients with an epidermal growth factor mutation (EGFR), a subgroup of the patients with lung cancer, seem to develop more bone metastases than other patients with lung cancer. Due to prolonged survival of these patients, they live longer with bone metastases and/or skeletal-related events, therefore optimal management is warranted. The aim of our systematic review is to gain more insight in reporting of bone metastases, skeletal-related events, and bone-specific outcome of treatment in clinical trials enrolling patients with EGFR-mutated lung cancer. We found that data on bone metastases and bone-related outcomes are largely lacking in clinical trials. There should be more focus on reporting and preventing of skeletal-related events in these patients. Abstract Bone metastases, occurring in 30–60% of patients with non-small cell lung cancer (NSCLC), are associated with decreased survival, cancer-induced bone pain, and skeletal-related events (SREs). Those with an activating epidermal growth factor mutation (EGFR+) seem to be more prone to develop bone metastases. To gain more insight into bone metastases-related outcomes in EGFR+ NSCLC, we performed a systematic review on Pubmed (2006–2021). Main inclusion criteria: prospective, phase II/III trials evaluating EGFR-tyrosine kinase inhibitors, ≥10 EGFR+ patients included, data on bone metastases and/or bone-related outcomes available. Out of 663 articles, 21 (3176 EGFR+ patients) met the eligibility criteria; 4 phase III (one double blind), 17 phase II trials (three randomized) were included. In seven trials dedicated bone imaging was performed at baseline. Mean incidence of bone metastases at diagnosis was 42%; 3–33% had progression in the bone upon progression. Except for one trial, it was not specified whether the use of bone target agents was permitted, and in none of the trials, occurrence of SREs was reported. Despite the high incidence of bone metastases in EGFR+ adenocarcinoma, there is a lack of screening for, and reporting on bone metastases in clinical trials, as well as permitted bone-targeted agents and SREs.


Introduction
Activating epidermal growth factor mutations (EGFR) are found in approximately 10% of the Caucasian and 50% of the Asian patients with lung adenocarcinoma [1][2][3]. EGFR mutations are prognostic and are also predictive for efficacy of EGFR-tyrosine kinase inhibitors (TKI). For patients with advanced lung adenocarcinoma and an EGFR mutation, treated with epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKIs), the five-year survival rate is 40-50% [4,5]. This is more favorable than the historical (i.e., before the introduction of immune checkpoint inhibitors) 5.8% five-year survival rate of patients without oncogenic-driven non-small-cell lung carcinoma (NSCLC) [6]. Immune checkpoint inhibitors (ICI) as monotherapy result in disappointing outcomes in patients with EGFR-mutated lung adenocarcinoma, with low responses rates and low survival and should only be considered after exhaustion of other systemic therapies [7]. ICI combined with EGFR-TKI has been evaluated in clinical studies but the combinations were either too toxic or did not provide an advantage over EGFR-TKI alone [7].
The currently available clinical trials evaluating bone-targeted agents (BTAs) and clinical guidelines (European Society for Medical Oncology (ESMO), National Comprehensive Cancer Network (NCCN)) providing recommendations for the management of bone metastases and skeletal-related events (SREs), do not focus on specific molecular subgroups of lung adenocarcinoma [14][15][16][17][18]. The guidelines state that it is advised to treat patients with bone metastases and a favorable survival (specified as at least three months) with BTAs. However, a more personalized advise is important, as in clinical trials especially the patients with a prolonged overall survival (OS) (i.e., historically mainly patients with metastatic breast and prostate cancer) benefit most from BTAs (i.e., significant reduction of SREs) such as bisphosphonates or denosumab [19][20][21]. As the clinical behavior of EGFR-mutated lung adenocarcinoma resembles metastatic breast and prostate cancer, with a real possibility for prolonged survival, data for this subgroup is also needed. However, to the best of our knowledge, clinical trials evaluating BTAs specifically in EGFRmutated lung adenocarcinoma do not exist.
The risk of a negative influence on quality of life (QoL) and OS, caused by SREs, could be significant in patients with EGFR-mutated lung adenocarcinoma and bone metastases [22,23]. In one retrospective case-control study (n = 189, no use of BTAs) survival postbone metastases diagnosis was superior for patients with an EGFR mutation compared to patients with a KRAS mutation or those without an EGFR/KRAS mutation, while time to first SRE was not significantly different [12]. As a result, these patients live longer with SREs. Therefore, optimal management, treatment, and outcome of bone metastases in this specific patient population is necessary and should be further evaluated.
As large prospective series on bone metastases-related outcomes are lacking for patients with EGFR-mutated lung adenocarcinoma, we performed a systematic review to gain more insight in the reporting of bone metastases and/or SREs, and bone-specific outcomes in patients with EGFR-mutated lung adenocarcinoma included in phase II/III EGFR-TKI trials. Improved knowledge about bone-related events in EGFR-mutated tumors can lead to better advice about the use of BTA in this subgroup of patients.

Search Strategy and Selection Criteria
A systematic search was performed using the Pubmed database. The search period was limited to January 2006 until January 2021 (search data January 8th 2021). The start date of January 2006 was chosen, as in 2006 EGFR-TKIs were approved by the United States of America's Food and Drug Administration (FDA), and became standard treatment for patients with lung adenocarcinoma and an EGFR mutation. Published studies were identified using a search strategy based on the patient intervention comparator outcome (PICO) method (shown in Table S1) [24]. Because the outcome variable of interest (i.e., bone metastases and SRE incidence) was an undefined endpoint in EGFR-TKI trials, we decided to exclude this outcome variable in the search strategy to prevent missing data. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2009 checklist for systematic reviews is shown in Table S2. Furthermore, the control intervention was not included in the search strategy to include single arm EGFR-TKI trials. Trials had to include a minimum of ten patients with non-squamous NSCLC and an EGFR mutation, as trials in the beginning of the TKI era also included patients with wildtype EGFR. Only prospective, phase II and III trials were included. All inclusion criteria are summarized in Table 1.
To minimize missing articles, in the same time period, we searched for relevant articles in the meeting libraries of the American Society of Clinical Oncology (ASCO), European Society for Medical Oncology (ESMO) and International Association for the Study of Lung Cancer (IASLC). Bone metastases and SREs at baseline or during the course of the disease, and/or their outcome. Regardless of whether they were primary, secondary, or no pre-specified endpoint of the trial Abbreviations: NSCLC, non-small-cell lung cancer; EGFR, epidermal growth factor receptor; EGFR-TKI, epidermal growth factor receptor tyrosine kinase inhibitor; SREs, skeletal-related events.

Study Selection
The titles of the retrieved studies, and the abstracts of the eligible studies based on title screening, were evaluated independently by two reviewers (A.B. and S.D.). The same reviewers independently examined the full text of the remaining articles regarding the inclusion criteria. Studies were included if they met the pre-specified inclusion criteria as shown in Table 1. To complete the search, the references of all eligible articles were manually searched for additional relevant articles. In case of disagreement during study inclusion, consensus was sought.

Data Selection
Two reviewers (A.B. and S.D.) independently extracted relevant characteristics of each eligible study. When available and if applicable, the following data were extracted: year of publication; phase II or III trial, number of study arms, randomization method, blinding method, duration of study and follow-up, histological diagnosis, method of staging, number of patients and number of patients with an EGFR+ mutation, intervention (i.e., type, dose, duration, route and frequency of administration of TKIs), bone metastases (i.e., incidence, outcome, treatment), SREs (i.e., incidence, outcome, treatment), secondary and primary objectives of the trial, and OS. The Jadad scale was used to assess the methodological quality of the included trials [25]. We did not perform a formal test of heterogeneity because of the heterogeneous type of trials included in the systemic review, with three quarter of the included trials being single arm (i.e., per definition high risk of bias).

Study Selection
The literature search identified 663 unique articles in total. About 317 articles were excluded because of non-relevant titles. About 160 of the 346 remaining articles were excluded because they did not fulfill the inclusion criteria based on the abstract. The full text of the remaining 186 articles was screened; 166 articles were excluded due to: no information about bone metastases or SREs (n = 139), unknown EGFR mutational status (n = 10), unknown if the patients with NSCLC and bone metastases were patients with an EGFR mutation or if they were wildtype patients (n = 9), insufficient patients with an
The number of patients included in the studies ranged from 10 [38] to 556 (Flaura [29]), leading to in total 3176 patients with advanced NSCLC and an EGFR mutation included in this review. One trial also enrolled patients with advanced non-squamous NSCLC and EGFR wildtype, the results were specified per molecular subgroup [46]. The inclusion criteria were generally similar across the included trials (i.e., pathological proven locally advanced NSCLC not suitable for treatment with radical radio-chemotherapy or metastatic NSCLC). All studies, except one in which also patients with non-squamous NSCLC and EGFR wildtype were included [46], enrolled exclusively patients with an activating EGFR mutation. The other exception in inclusion criteria was systemic treatment history of the patients. The exclusion criteria concerning comorbidities were comparable among all studies. The Aura 3 was the only trial in which explicitly a statement about bone-targeted agents (BTAs) was added, it was permitted for patients to use medication (e.g., denosumab) for painful bone metastases [26]. The other trials provided no information about BTAs [26][27][28][29][30][31][32][33][34][36][37][38][39][40][41][42][43][44][45][46].

Imaging and Incidence of Bone Metastases during Follow-Up
In two out of 21 studies dedicated bone imaging during follow-up was performed: with FDG-PET-CT scan [45,46] or with a bone scintigraphy [33]. Bone scintigraphy was performed in the study of Reguart when clinically indicated [42].
In ten studies (total 2378 patients) the incidence of bone metastases as site of progressive disease (PD) was reported ( [27,30,33,38,39,46], Flaura [29], Aura 2 [32], Aura 3 [26], Aspiration study [41]). Three to 26% of the patients had development or progression of bone metastases as site of PD (215 of 2378 patients). In none of the included studies data were provided whether bone progression was the only site of progression or not.

Skeletal Related Events
In none of the included studies information about SREs was provided. In Table 3 a summary of the reported imaging, incidence of bone metastases and SREs is shown.

Discussion
Bone metastases with their risk of SRE development and resulting impact on QoL, can become a clinically relevant problem in patients with lung adenocarcinoma and an EGFR mutation because of their prolonged post-bone metastases diagnosis survival [4,12,23,48,49]. To gain more insight into bone metastases and their outcomes, we performed a systematic review focusing on screening, treatment, and reporting of bone metastases and/or SREs and bone-specific outcomes in EGFR-TKI trials.
In none of the trials, primary or secondary outcomes related to bone metastases and/or its complications were mentioned. A 42% median baseline incidence of bone metastases in patients with NSCLC and an EGFR mutation was reported, which is slightly lower compared to the 54% baseline incidence reported in the Dutch nationwide database study and other retrospective studies [9,12,50]. Of note, in only seven of the included trials specific bone imaging was performed at baseline, possibly resulting in an underestimation of the real incidence of bone metastases. Up to 26% of patients had progression in the bone upon PD. This probably is also underestimated as in only two of the trials standardized follow-up bone imaging was performed. Patients in the Aura 2 trial were permitted to use BTAs in case of painful bone metastases, but further information on actual BTA use and outcome was not provided. In all other trials, all data regarding BTA was lacking.
EGFR-TKIs have a high efficacy in patients with lung adenocarcinoma and an EGFR mutation and bone metastases [51,52]. Their efficacy in bone is mediated by blockade of receptor activator of NF-κB ligand (RANKL)-mediated osteoclast activation and by inhibiting epidermal growth factor (EGF) signaling in bone stromal cells [53]. Therefore, it could be that in this specific patient population, bone metastases do not frequently lead to SREs. Indeed, a retrospective study in patients with lung adenocarcinoma and bone metastases (n = 410) reported a preventive effect of EGFR-TKIs on the development of SREs: 23.5% of the patients with lung adenocarcinoma who were treated with EGFR-TKIs experienced SREs compared with 61.7% of patients without EGFR-TKI treatment (information about specific treatment in this group is not provided) (p < 0.001) [54]. However, even with EGFR-TKI use almost a quarter of the patients experienced SREs in this study, and in other studies, the reported frequency of SREs is even higher (37.3% to 58%) in patients with EGFR-mutated lung adenocarcinoma mainly treated with EGFR-TKIs [12,49,[54][55][56][57]. A recently published retrospective study, which evaluated the type and frequency of SREs in patients with EGFR-mutated lung adenocarcinoma and bone metastases (n = 274, of which 148 treated with EGFR-TKI), showed that one-third of these patients developed their first SRE before start of EGFR-TKI treatment, the other two-third of the patients developed SREs in the first year of EGFR-TKI treatment [49]. The above summarized SRE percentages were observed in patients with EGFR-mutated lung adenocarcinoma, treated with first or second generation EGFR-TKIs. To the best of our knowledge, no data are available for the different generation EGFR-TKIs (i.e., first/second versus third) regarding efficacy specifically on bone metastases. Mouse models were set up to investigate the efficacy of osimertinib with or without bevacizumab on bone metastases of NSCLC. Treatment with osimertinib (with and without bevacizumab), showed tumor regression and bone remodeling [58]. Based on these results, it is not clear whether osimertinib is superior to earlier generation TKI in humans in the treatment of bone metastases.
In the abovementioned retrospective studies, use of BTAs varied from 0 to 65% [12,49,[54][55][56][57]. Interestingly, in vitro and in vivo studies showed that bisphosphonates can act synergistically with EGFR-TKIs [54,55,59]. The in vitro study of Chang on the HCC827 NSCLC cell line expressing mutated EGFR, suggested that the combination of gefitinib and zoledronic acid caused more tumor suppression [59]. A small retrospective study of Cui et al. (n = 38) studied the efficacy of bisphosphonates in patients with EGFR-mutated lung adenocarcinoma and bone metastases, treated with EGFR-TKIs. They showed a significant additive effect of bisphosphonates on OS post-bone metastases diagnosis: postbone metastases OS in EGFR-TKI + bisphosphonate group: 28.3 months versus 22.0 months in the EGFR-TKI only group, p = 0.0587 [55]. Another small retrospective study studied the effects of bisphosphonates in patients with EGFR-mutated lung adenocarcinoma and bone metastases (n = 62) and found comparable results (PFS and OS prolonged in the bisphosphonate + EGFR-TKI group compared with the EGFR-TKI group) [54]. As these are retrospective, small series, these data are only hypothesis generating. To the best of our knowledge, no data on denosumab combined with EGFR-TKI are available and it would be interesting to prospectively evaluate this.
Due to the increasing number of treatment options (e.g., EGFR-TKI in combination with chemotherapy, or combination of EGFR-TKI with angiogenesis inhibition), survival is further improving for patients with an EGFR mutation [47,51,52,60,61]. In the Flaura trial, the median OS with first line osimertinib was 38.6 months, in the NEJ009 trial (combination gefitinib with carboplatinum/pemetrexed versus gefitinib alone), median OS was 50.9 months [51,60]. Five-year survival rates for these trials have not been reported yet, but with a median OS of 50 months, 5-year survival rates resemble that of advanced breast or prostate cancer in which 28.1-30.2% of the patients are alive five years after the diagnosis [48,62,63]. This is important, as it is suggested in retrospective series that patients with EGFR-mutated NSCLC have a long post-bone metastatic survival of 15.5 to 28.0 months [12,49], implying that these patients live long with SREs.
Despite the similarities in the incidence and nature of EGFR-mutated lung adenocarcinoma and breast cancer bone metastases, current guidelines (e.g., ESMO, National Comprehensive Cancer Network [NCCN], and National Institute for Health and Care Excellence [NICE], ASCO) provide different recommendations for screening of bone metastases for different primary tumors [18,[64][65][66][67]. The most important difference between the guidelines is the recommendation to screen all breast cancer patients, whereas for NSCLC only the Lung Cancer South East French Guidelines recommend to screen for bone metastases in NSCLC [18,[64][65][66][67][68]. The French guideline also recommends to evaluate each bone metastasis for pain, neurological risk, and fracture risk to aid in defining the optimal bone metastasis management in harmony with the oncological treatment [68]. BTAs demonstrated benefit in reducing SREs and providing better pain control, in advanced breast patients diagnosed with bone metastases [19]. In the ESMO guideline on advanced breast cancer it is recommended to use BTAs in these patients (level of evidence I, grade of recommendation A [65]. Guidelines for lung cancer are less clear in their recommendations: the NCCN NSCLC guideline advises to consider BTAs in patients with NSCLC and bone metastases [16]. The ESMO guideline on bone health further specifies and recommends using BTAs in patients with a life expectancy of >3 months (i.e., almost all patients with an EGFR mutation) [18,67]. No specific recommendations for patients with EGFRmutated lung adenocarcinoma were found in these guidelines. Probably because of the historically poor OS of NSCLC compared to advanced breast cancer, only 15-33% of patients with NSCLC and bone metastases are treated with BTAs in daily practice [69,70]. To the best of our knowledge no trials are ongoing that evaluate BTAs in patients with an EGFR mutation, although trials in patients without an oncogenic driver are ongoing [(NCT03669523 trial: denosumab in combination with nivolumab, NCT01951586 trial: denosumab in combination with chemotherapy (recently finished, results are not published)].
Drawbacks for this systematic review are: (1) The heterogeneity of the included trials with differences in populations (e.g., ethnicity) and/or follow-up which could have led to the observed differences in reported incidences of SREs; (2) the lack of primary or secondary outcomes related to bone metastases and/or related complications in studies could have led to underreporting of these outcomes.

Conclusions
Despite long (post-bone metastatic) OS of patients with EGFR-mutated NSCLC, and the high incidence of bone metastases in this patient population, occurrence of SRE and outcome of bone metastases is barely reported in clinical trials. Based on in vitro data found and retrospective series there might be synergistic activity of EGFR-TKI and BTA. However, prospective research is needed to validate these observations. Furthermore, the results of this systematic review stress on the importance of screening for bone metastases and reporting of clinical outcomes of treatment on bone metastases future trials for patients with EGFR-mutated NSCLC.