Cytoreductive Stereotactic Body Radiotherapy with Tyrosine Kinase Inhibitors for Intrathoracic Oligoprogressive EGFR-mutated Lung Cancer

Background: Epidermal growth factor receptor (EGFR) mutated non-small cell lung cancer (NSCLC) shows an impressive initial response to EGFR tyrosine kinase inhibitors (EGFR-TKI). However, resistance invariably develops, commonly involving the site of initial gross disease. Cytoreductive stereotactic body radiotherapy (SBRT) for thoracic oligoprogressive disease (OPD) may effectively delay progression through EGFR-TKI therapy. Methods: From a prospectively maintained IRB-approved institutional registry, we identied 23 patients consecutively treated between 2011-2019 with thoracic SBRT and received EGFR-TKI within 6-months of SBRT. Radiographic progression-free (PFS) and overall survival (OS) were estimated using Kaplan-Meier analysis. Toxicity and patient-reported Edmonton Symptom Assessment Scale (ESAS) scores were reviewed. Results: Median follow-up after SBRT was 20-months (range, 4-100), and the median age was 68-years (range, 33-89). Most patients were females (n=21;91.3%). RT dose was 50-60 Gy in 5-10 fractions. EGFR-TKI administered were erlotinib, osimertinib and getinib in 15, 5, and 3 patients, respectively. Median PFS and OS following SBRT were 8-months and 31-months, respectively. 1-year PFS and OS were 34.8% and 78.3%. The median duration of EGFR-TKI therapy was 26-months (1-91). Most patients progressed in new distant sites, most commonly bones (n=5;21.7%) and distant lung (n=4;17.4%), with only 2/23 patients having initial progression within the SBRT eld. Grade-2 pneumonitis (n=2) and rib fracture (n=1) were noted radiation-related toxicities. Dominant ESAS symptoms were fatigue (21.7%), pain (8.7%), and loss of appetite (8.7%). OPD on EGFR-TKI, resulted in changes in subsequent patterns of failure, lengthened PFS, and prolongs the duration of initial TKI therapy.


Introduction
Multiple oncogenic driver mutations, including epidermal growth factor receptor (EGFR) have been identi ed to play a vital role in the oncogenesis of NSCLC [1]. Their discovery led to the development of targeted therapies such as tyrosine kinase inhibitors (TKIs), that changed the landscape of systemic therapy for advanced NSCLC. However, following initial impressive response to EGFR-TKIs, acquired resistance is inevitable after 6-13 months of therapy initiation and remains the major challenge in achieving long-term control [2][3][4].
Traditionally, even a single site of metastasis was considered incurable, and treatment was palliative. However, oligometastatic disease (≤5 metastatic sites) treatment paradigm of NSCLC has recently shifted. Local cytoreductive therapy (LCT) has shown to improve progression free survival (PFS) and overall survival (OS) for oligometastatic NSCLC [5][6][7]. Although progression is often inevitable, progressive metastatic disease limited to few cancer deposits i.e. oligo-progressive disease (OPD) may harbor unique biologic subclones that confers resistance to initial line of therapy. Pattern-of-failure analyses in all-comer lung cancer patients on systemic therapy demonstrated that progression at intrathoracic disease site is the predominant site of initial failure [8][9][10][11]. A recent study showed that the majority of EGFR-mutant patients treated with osimertinib developed progressive disease initially involving the original site of gross disease and were amenable to stereotactic body radiotherapy (SBRT) [9]. Ablative treatment to these lesions may act synergistically with EGFR-TKIs to eradicate the dedifferentiated resistant subclones, confer overall treatment sensitivity, and allow patients to continue EGFR-TKIs for a prolonged duration.
SBRT offers a radiobiological advantage of delivering ablative radiation dose to the tumor to achieve a high rate of durable control, that could potentially result in additional clinical bene t of targeted therapy [12][13][14]. A recent study assessed upfront SBRT combined with targeted therapy in treatment naïve oligometastatic patients, however, the effectiveness of SBRT in thoracic OPD patient that progressed through EGFR-TKI therapy is still investigational [15]. We present our long-term experience, which to the best of our knowledge is one of the largest data evaluating role of SBRT in EGFR-mutated NSCLC with limited intrathoracic oligoprogression through EGFR-TKI therapy. We hypothesized that addition of SBRT in this patient group is safe, lead to improved PFS compared to historical data for patients receiving EGFR-TKI alone, and may change subsequent patterns of failure.

Patient Population:
We collected data from a prospectively-maintained IRB approved registry which include 1764 consecutive patients treated with lung SBRT. Eligible patients were >18 years, with Karnofsky performance score >60, biopsy proven EGFR-mutant NSCLC, had intra-thoracic oligoprogression through EGFR-TKI therapy (1st, 2nd, or 3rd generation) and received thoracic SBRT, de ned as 5-10 fractions. Between 06/2010 -06/2019, 23 EGFR-mutant NSCLC patients were identi ed, who received EGFR-TKIs within 6-months of SBRT. Oligoprogression was identi ed by positron emission tomography (PET) and seen on correlative computed tomography (CT) within 8 weeks before the initiation of SBRT. Patients with non-thoracic disease progression and progression at multiple distant sites were excluded from this study. Patients with brain metastasis who were controlled after excision or stereotactic radiosurgery were included.

Radiation Technique
Patients were simulated in supine position within a customized immobilization device (Body-x, Smithers Medical Products Inc., North Canton, OH), to ensure daily set-up reproducibility. All patients underwent a treatment planning 4D-CT simulation to consider respiratory motion. Radiotherapy target volumes included the progressing lesion on imaging. The gross tumor volume (GTV) was outlined on lung window with no expansion for microscopic disease. An internal target volume (iGTV) was created to account for respiratory motion. The planning target volume was de ned as a 3-5 mm isotropic expansion on the iGTV.
The most common SBRT fractionation was 50-60 Gy in 5 daily fractions at 10 -12 Gy per fraction. A conservative "SBRT-like" hypo-fractionated regimen (5-7.5 Gy per fraction) was used for lesions near critical structures. Target coverage followed principles of SBRT planning such that 99% of PTV received a minimum 90% of the prescription dose. Radiotherapy was planned and delivered with IMRT/VMAT in all cases with daily cone-beam CT image guidance to facilitate accurate targeting. The dosimetry data is outlined in supplemental table 1.

Assessments
Patients were assessed during SBRT treatment and every 3 months thereafter with CT chest/abdomen, physical examination, and laboratory tests per institutional guidelines. Further evaluation was based on suspicion of progression or toxicity. Patients were followed by thoracic medical oncologist and continued EGFR-TKI until progression or per discretion of medical oncologist. All disease response evaluations were based on Response Evaluation Criteria in Solid Tumor (RECIST) criteria (version 1.1). Brain imaging were performed for patients with brain metastases at baseline or signs for brain metastases. Patient lled ESAS questionnaire at each visit, and toxicity data was collected prospectively. Toxicities were evaluated using the National Cancer Institute's Common Terminology Criteria for Adverse Events (version 5.0).

Statistical Analysis:
Descriptive statistics were used to describe pre-treatment, treatment, toxicity and dosimetric characteristics. Primary endpoints were radiographic PFS and duration of EGFR-TKI therapy. Time-toprogression were calculated from completion of SBRT to the date when any disease progression was rst detected. Survival was calculated from SBRT to the date of death; patients alive at the date of last contact were censored at that time. Actuarial local control rates (LCR), PFS and OS were calculated by the Kaplan-Meier method. Local failure (in-eld and marginal) was de ned per RTOG SBRT trial de nition of progressive consolidation on CT within 1-cm of SBRT eld, not consistent with benign radiation-induced changes [16]. The log-rank test and univariable analysis of Cox proportional hazards models were used to evaluate the impact of clinical factors on disease outcome. Multivariable analysis was not performed because of the relatively small number of patients. A p-value of <0.05 was considered signi cant. All levels of signi cance were 2-sided. Data were analyzed using the statistical software SPSS for Windows (version 24.0, IBM Corp., Armonk, N.Y., USA).
Toxicity and ESAS Assessment: Toxicity and ESAS data are illustrated in Table 2. All patients completed planned treatment, and no acute SBRT-related grade 2-5 adverse effects were noted. Grade 2 pneumonitis was observed in 2 (8.7%) patients at 3 and 6 months (both received erlotinib). Patient who developed pneumonitis at 3-months also received chemoradiation for stage IIIB disease. One patient developed treatment related rib fracture.
EGFR-mutant NSCLC patients that progress through EGFR-TKIs experience initial oligoprogression predominantly (50-80%) involving lungs [9,11,21]. These oligoprogressive sites harbor a population of TKI-resistant and dedifferentiated subclones with the potential to re-seed systemically leading to widespread disease progression [22,23]. Aggressive SBRT to ablate these resistant subclones can delay progression and need to switch (or discontinue) therapy. While emerging evidence from oligometastatic disease shows upfront LCT delay progression and improve survival, there is little data about SBRT in patients that progressed through modern EGFR-TKI therapy [5][6][7]. A retrospective study of 23 EGFRmutant patients on 1st generation EGFR-TKIs demonstrated an initial median PFS of 10.3 months, however, in 10 patients that received subsequent LCT for oligoprogression, showed additional PFS bene t of 6.2 months [14]. Another retrospective study of 18 patients with EGFR-mutated NSCLC that acquired resistance to 1st generation TKIs received elective LCT with continued EGFR-TKI showed a prolonged PFS of 10 months and OS of 41 months with good tolerance [12]. Our study supports that lung SBRT in EGFRmutated NSCLC patients progressing through EGFR-TKI can extend median PFS by additional 8 months, and median OS by 31 months [4,24]. Despite patients in this study already had an initial progression prior to SBRT, and PFS was calculated from SBRT, results are still comparable to some studies that evaluated adjuvant EGFR-TKI alone in treatment naïve NSCLC [4,[24][25][26]. To put in perspective, osimertinib provides an 8-months PFS bene t compared to older generation TKIs in the FLAURA study [26]. However, this bene t is seen in a treatment naïve patient population. The bene t of SBRT in our study is in a population that has already showed resistance to targeted therapy through progression making an 8-months additional PFS bene t signi cant for this treatment group.
Our results also showed that SBRT enabled patients to effectively continue on EGFR-TKI therapy for a prolonged duration with median therapy duration of 26-months (1-85 months). This gure is longer than many historical studies using EGFR-TKI alone for recurrent and even some studies in adjuvant setting [12,20]. Recently reported ADAURA study, a phase III, randomized placebo-controlled trial that showed resected EGFR-mutated NSCLC (stage IB-IIIA), demonstrated encouraging results with median duration of adjuvant osimertinib of 22 months [27]. This reiterate how local therapy may bene t in combination with targeted treatment.
Our study expands on the work of a recent observational study in which EGFR-mutated NSCLC patient treated with osimertinib, showed 80% patients developed progressive disease within the initially involved thoracic sites, most of which were a potential SBRT candidate [9]. Although this study brought forth an important concept to use SBRT to delay progression, no patients were actually treated with SBRT. Our study is among largest experiences in this patient population evaluating modern EGFR-TKIs plus SBRT for intrathoracic OPD. Our results demonstrated that lung SBRT dramatically alter the typical pattern of subsequent failures. Most common sites of initial failure following SBRT were distant, and only 2 patients (8.7%) had initial recurrence within the treated site [ Figure 1]. This contrasts with pattern of failure studies in literature for patients on EGFR-TKI, that showed approximately 50-80% had lungs as the initial site of progressive disease [5,9,21].
With maturing evidence of combination of EGFR-TKIs plus SBRT, particularly with newer-generation TKI e.g. osimertinib, it is pragmatic to explore toxicity data. A phase II study with erlotinib and SBRT in metastatic NSCLC showed 29 grade 3-5 toxicities observed in 24 patients, four of which could be attributed to SBRT (pneumonitis and vertebral fracture) [13]. Some retrospective studies have shown higher rates of severe radiation pneumonitis with combination of radiotherapy to rst-generation EGFR-TKIs [28]. With careful patient selection (disease size and location) and precise planning and delivery of SBRT, dose to normal lung can be effectively reduced mitigating the risk of pneumonitis. Our patients tolerated treatment well with no treatment related death or dose limiting toxicity. Grade 2 pneumonitis was observed in two patients, while one developed treatment related grade 2 (rib fracture) chest wall toxicity. Patient reported ESAS outcome analysis showed good tolerance to therapy, with fatigue being most common symptom.
Given disease heterogeneity and relative rarity of presentation (EGFR-mutant, oligoprogressive on TKIs) designing a randomized trial will always be challenging and time consuming; and lack of a uniform comparator group will limit valid interpretation. Therefore, the evidence needs to be established with assemblage of data from institutional retrospective studies and small phase II trials. A single-arm phase II trial (NCT01941654) from China is evaluating preemptive LCT to residual metabolically active oligometastases in 34 patients with EGFR-mutant NSCLC who had an initial response to rst-line TKI therapy. The fact that this study has only accrued 18/34 patients since 2013 underscores this di culty with accruing large patient numbers in such study. A single arm, multicenter phase II study (NCT02314364) SBRT for stage IV oncogene-driven NSCLC receiving TKIs is ongoing. The target accrual is 30 patients between 2013-2021 (7-years), with primary outcome to study patterns of distant failures following SBRT. Preliminary report of an ongoing prospective, multicenter, randomized study (SINDAS Trial, NCT02893332) evaluating EGFR-TKIs ± upfront SBRT in therapy naïve EGFR-mutant metastatic NSCLC patients showed that SBRT signi cantly improved PFS and OS compared to TKI alone [15]. The HALT trial (NCT03256981), is a randomized, multi-center, phase II/III study ongoing in UK, is the only ongoing trial directly evaluating the role of SBRT + TKI in patients with oligoprogressive oncogeneaddicted NSCLC (n=110) [29].
There are several limitations of this study that should be acknowledged. First, this was a single institution study with relatively small number of patients. Most data evaluating role of SBRT in oligoprogressive EGFR-mutant NSCLC on TKIs is in the form of small single-institution experiences. The small target accrual of ongoing prospective studies over extended time periods emphasizes the complexity of such studies and relative rarity and heterogeneity of this clinical scenario, especially in context of newergeneration EGFR-TKIs. To the best of our knowledge, this study is still one of the largest series evaluating effectiveness of thoracic SBRT with modern EGFR-TKI in oligoprogressive EGFR-mutant NSCLC patients. Second, given the retrospective design there were no pre-de ned inclusion criteria. However, at our institution we prospectively collect all data for lung cancer patients including toxicity and patient reported outcome. Last, superior outcome reported in this study could potentially be due to predominant women (n=21, 91.3%) in this study, and it has been shown that females fare better with treatment than males [30].
This study demonstrates the e cacy and safety of a novel therapy indication of thoracic SBRT in oligoprogressive EGFR-mutant NSCLC with EGFR-TKI prolonging PFS and OS and allowing continuation of targeted therapy longer than historic data for EGFR-TKI alone. More information will emerge after completion of ongoing prospective studies, until then retrospective evidence should be explored to establish more evidence. Funding: This research did not receive any speci c grant from funding agencies in the public, commercial, or not-for-pro t sectors.
Ethical Approval and Consent to participate: All patients signed informed consent for the treatment. Data were collected retrospectively from Mo tt Cancer Center IRB approved registry.