Stereotactic Radiotherapy for Early Stage Non-Small Cell Lung Cancer

Current Standards and Ongoing Research

Eugenia Vlaskou Badra; Michael Baumgartl; Silvia Fabiano; Aurélien Jongen; Matthias Guckenberger


Transl Lung Cancer Res. 2021;10(4):1930-1949. 

In This Article

Current Standard of SBRT for Early-stage NSCLC

Epidemiology, History, and Development of SBRT as Standard of Care

Lung cancer is among the most frequent malignancies and the leading cause of cancer-related deaths worldwide,[1] with non-small cell lung cancer (NSCLC) accounting for approximately 5% of all cancer-related mortality. Around 16% of patients with NSCLC are diagnosed at early stages, which are characterized by a small primary tumor and lack of lymph node metastases (stages T1–2, N0).[2] This proportion of early stage NSCLC is expected to increase in health care systems with implementation of CT-based lung cancer screening.[3,4]

Early-stage (ES) NSCLC has traditionally been managed by lobectomy and systematic hilar and mediastinal lymph node dissection. An overall survival of 60–92% at 5 years[5] indicates this tumor stage as a curable disease. A significant number of patients is however medically inoperable due to their comorbidities and this proportion of inoperable or high-risk patients is growing due to an aging population.[6] For this group of patients, traditional treatment options have been best supportive care, limited/extra-anatomical resection, and radiotherapy. Conventionally fractionated external beam radiotherapy (EBRT) with total irradiation doses of 60–66 Gy had been an established curative treatment option for such medically inoperable patients. However, several studies reported a dose-response relationship for radiation doses beyond this range, with improved local tumor control and survival for higher radiation doses.[7]

Stereotactic body radiation therapy (SBRT) is defined as a form of EBRT that accurately delivers a high dose of radiation to an extracranial target in a single or few fraction(s).[8] Developed in the early 1990s,[9] SBRT was further adapted and advanced by multiple groups and is nowadays a well-established and guideline-recommended component of modern radiotherapy. In some publications, SBRT is referred to as stereotactic ablative radiation therapy (SABR).

Multiple, methodologically and technically diverse studies on SBRT in early-stage NSCLC have consistently shown favorable outcomes in terms of high local control rates (74–100%), preserved quality of life, and low treatment-related toxicity.[10–21] Recent randomized clinical trials (RCTs) comparing EBRT and SBRT have shown comparable results in terms of progression-free (PFS)[17,20] and improved overall survival (OS)[20] in favor of SBRT. Today, SBRT is established as the gold standard for medically inoperable patients with ES NSCLC,[22–27] with increasing use, due to aging populations in many societies. Figure 1 illustrates an exmple of ES NSCLC treated at our institution.

Figure 1.

Patient case. Figure shows a patient treated at our institution with 3×13.5 Gy@65% ILD. (A) Pre-treatment; (B) treatment plan; (C) 1 year post-treatment.

Guideline Perspective of SBRT for ES NSCLC

Clinical Practice Guidelines. International guidelines[22–24,27] recommend treating node-negative ES NSCLC with surgical lobectomy, if pulmonary and cardiac comorbidities allow it. In patients considered medically inoperable based on an interdisciplinary discussion, as well as in those unwilling to undergo surgery, SBRT is the treatment of choice. The European Society for Radiotherapy and Oncology (ESTRO) Advisory Committee in Radiation Oncology Practice (ACROP) consensus guidelines also suggest a minimum performance status of ECOG 3 and a minimal estimated life expectancy of one year for SBRT patient selection.[24]

For assessment of patient operability, guidelines agree on a multidisciplinary patient assessment. Pre-treatment evaluation before SBRT or surgery includes (but is not limited to) pulmonary function testing, bronchoscopy, mediastinal lymph node evaluation, PET/CT staging[22–25] while some also recommend cranial MRI in stages IB (optional) to IIA.[23] While SBRT is the treatment of choice in inoperable patients according to the above-mentioned guidelines, there is no commonly accepted definition of patient inoperability. The perioperative risk can be estimated using validated systems especially considering cardio-pulmonary function—however, none have yet been prospectively validated in NSCLC patients.

Pre-SBRT biopsy confirmation is strongly recommended but not a prerequisite for patients unwilling to undergo invasive biopsy or patients with an excessively high periprocedural risk.[23–25] The challenge of clinically diagnosed ES NSCLC will be discussed later in this article.

The main failure pattern after treatment of ES NSCLC is distant, with about 20–30%[28,29] of the patients developing metastatic disease during follow-up. Some guidelines, therefore, recommend the evaluation of adjuvant chemotherapy after SBRT in patients with high-risk features, such as poor tumor differentiation, vascular invasion, pleural involvement, and unknown lymph node status[25] while others do not.[23]

Follow-up after SBRT should consist of clinical visits and CT imaging every 3–6 months for at least two years. Distinguishing between post-therapeutic fibrosis and persistent or recurrent NSCLC is as pivotal as complex. For CT-based follow-up imaging, high-risk features, such as bulging margin, craniocaudal extension, and linear margin disappearance have been identified to more accurately differentiate between vital tumor and progressive fibrosis.[30,31] FDG-PET/CT scans are not routinely recommended but should be used in patients where differentiation between post-SBRT fibrosis and tumor recurrence is otherwise difficult.[23,24,32]

Guidelines do not routinely endorse SBRT as primary treatment in patients deemed to be at a "standard operable-risk", despite concerns regarding surgical mortality and morbidity. Literature has demonstrated 3 year-OS rates of 76–86%[15,32–36] after SBRT in selected cohorts of operable patients unwilling to undergo surgery. A meta-analysis of 4850 patients within 40 SBRT studies and 7,071 patients within 23 surgical studies in ES NSCLC reported no significant difference in OS or disease-free survival (DFS) when adjusting for age and comorbidities.[37] Another meta-analysis of 23 studies reported improved outcome in terms of overall- and cancer-free-survival after surgery compared to SBRT in both the matched and unmatched group.[38] All retrospective and cross-study comparisons suffer from insufficient matching of surgical and SBRT patient cohorts because relevant prognostic factors are frequently unavailable for the matching process. Additionally, such studies have been shown to be prone to interpretation bias.[39] Randomized prospective trials are therefore needed to properly address this important clinical question.

Medical Physics Practice Guidelines. In order to describe the technical requirements of treatment units for safe and effective SBRT of ES NSCLC, six national and international guidelines, recommendations, and an expert review group consensus were reviewed.[22,24,40–42]

The ESTRO ACROP consensus has been released with the key aspects on SBRT treatment delivery for ES NSCLC, discussing in detail the minimum machine performance.[22] The ASTRO guideline provides a detailed overview about the clinical part only[24] whereas the listed American Association of Physicists in Medicine (AAPM) reports cover the technical requirements for SRS/SBRT in general.[22,24,40–43] The Deutsche Gesellschaft für Medizinische Physik (DGMP) expert review gives a fair overview of technical specifications necessary for SBRT/SRS treatments in general.[42]

There is a strong agreement in implementing an end-to-end test during the commissioning phase of the linear accelerator, not only to check the accuracy and reliability of the system before a first SBRT treatment, but also to conduct regular machine quality assurance (QA) checks to guarantee a stable performance of the treatment unit afterwards. End-to-end tests are powerful tools in QA protocols to ensure the reliability of the entire treatment chain through sufficient imaging protocols for the planning CT, image reconstruction, data transfer, treatment planning system performance, motion management, and irradiation of dummy treatment plans on QA phantoms and comparing calculated with measured data. Most importantly, the equipment specific QA has to be extended and pass stricter criteria than standard IMRT QA protocols.[43]

Teaching of the medical staff involved in SBRT treatment, continuous training, credentialing, setting up standard operating procedures and clinical protocols are all essential and indispensable to be conducted and implemented before and while providing SBRT treatments in general.[22,24,40–44]

Despite the fact that dedicated SBRT treatment devices such as the CyberKnife® or Vero® are compelling and well established technologies in radiation therapy, their added value in comparison to standard linear accelerators (linacs) is uncertain.[22] For most of the radiation oncology centers, standard linacs represent the most accessible, affordable, and efficient treatment units. Most modern machines are equipped with necessary SBRT quality requirements: high-resolution multi-leaf collimators (MLC) <10 mm, volumetric image-guided radiation therapy (IGRT) technology, and 4D-CT. They can therefore be used for standard and more sophisticated treatment techniques such as SBRT for ES NSCLC.[22]

Current Challenges in SBRT for ES NSCLC

Patients With Centrally/Ultra-centrally Located NSCLC. SBRT in ES NSCLC is a well-tolerated and efficient treatment with high rates of local control when applied to peripherally located lesions. However, as high ablative doses are needed in order to achieve optimal tumor control, SBRT in tumors located close to critical structures (such as major bronchi, esophagus, large vessels, and brachial plexus) is potentially associated with a higher risk of organs at risk (OAR) damage. Although commonly used in literature as well as in clinical practice, there is no uniformly accepted definition of the terms "central" or "ultracentral".

Timmerman et al. initially reported "excessive toxicity" in patients with central tumors treated with 3 fractions of 20–23 Gy. Tumor location was the strongest predictive factor for toxicity, with up to 11-fold increased risk of grade 3 or higher toxicities.[45] Those results led to the development of the first "no-fly" zone definition, adapted by the RTOG 0236 trial and still in use by the ASTRO guideline,[24] which defined central tumor location as "2 cm in all directions around the proximal bronchial tree (PBT)".[45]

The RTOG 0813 trial tumors[46] was designed to evaluate SBRT outcomes in centrally located NSCLC and added to the RTOG 0236 definition as follows "the zone […] of RTOG 0236, with the addition of tumors which are immediately adjacent to mediastinal or pericardial pleura (PTV touching the pleura)" .[46] The International Association for the Study of Lung Cancer (IASLC) has a broader definition for central tumor location "within 2 cm in all directions of any mediastinal critical structure, including the bronchial tree, esophagus, heart, brachial plexus, major vessels, spinal cord, phrenic nerve, and recurrent laryngeal nerve" .[32] While patients with centrally located ES NSCLC are at a higher risk of toxicity from SBRT, surgery in this population is also associated with worse outcomes.[47]

The term "ultracentral" has been established more recently and is also lacking a uniform definition. It often refers to tumors directly abutting or invading the PBT or esophagus.[24] All definitions have in common that anatomical location and not radiotherapy doses to critical organs at risk are the basis for risk stratification.

A 2013 systematic review[48] analyzed findings of 20 trials, including 315 ES NSCLC tumors out of a total of 563 centrally located lung tumors. They reported SBRT-related mortality of 2.7% and grade 3 or higher toxicities at 9%. The OS did not differ between central and peripheral tumors, but the heterogeneity of treatment delivery did not allow the determination of an optimal dose/fractionation regime. Since then, several single-center-studies have been published, mostly,[12,32,49–51] but not exclusively,[52,53] reporting central tumor location as a predictor for increased toxicity.

The RTOG 0813 trial—a seamless phase I/II trial—evaluated fractionation schedules of 5 fractions every two to three days up to a total dose ranging from 50–60 Gy, escalating in 0.5 Gy per fraction steps. With a median follow-up of 37.9 months, they reported a maximal tolerated dose of 5×12.0 Gy/fx, with an accompanying probability of 7.2% dose-limiting toxicity. Local control at 2 years in the 11.5 Gy/fx and 12.0 Gy/fx cohorts was 89.4% and 87.9%, respectively, while OS was reported at 67.9% and 72.7% and therefore comparable to outcomes of peripheral tumors.[46] However, there were relevant numbers of adverse events with a total of 13 out of 70 patients (19%) experiencing toxicity graded 3 and higher, while grade 5 toxicity was reported for 6 patients.

The treatment of central NSCLC in inoperable patients or those unwilling to undergo surgery has been evaluated prospectively within the multicentric EORTC LungTech trial.[54,55] Endpoints include treatment efficiency and toxicity. They defined central tumor location as " located within 2 cm or touching the zone of the proximal bronchial tree or immediately adjacent to the mediastinal or pericardial pleura, with a PTV expected to touch or include the pleura".[54] The study closed early due to slow recruitment: Two potentially treatment related deaths were observed after inclusion of 39 patients.

The Nordic HILUS trial, a phase-II-multicenter trial on SBRT to central tumors, included primary NSCLC as well as metastatic disease. They defined central location as tumors located within "≤1 cm from the proximal bronchial tree". Forty-two out of 74 patients had tumors located close to the main bronchus (arm A) while 31 patients had tumors located close to a lobar bronchus (arm B). Toxicity graded 3 or higher was reported in 28% of patients, while 9% (a total of seven patients, six patients in arm A and one patient in arm B) experienced grade 5 toxicity (lethal hemoptysis and pneumonitis).[12,21]

Recent data on ultracentral tumors show comparable rates of local control, but with sometimes substantial toxicity rates.[56–58] While the possibility of potentially fatal toxicity in high-risk cohorts remains present, literature reports reasonable outcomes, particularly with protracted fractionation schedules. International guidelines therefore recommend SBRT in patients with central ES NSCLC using risk-adapted fractionation regimes,[22,24] however an optimal fractionation schedule has not been recommended. An ongoing multicentric phase I dose escalation study, the SUNSET trial, is currently evaluating maximal tolerated dose in this setting in order to identify a safe and efficient fractionation regime, starting out at 60 Gy in 8 daily fractions[59] (Table 1, Figure 2).

Figure 2.

Centrally located NSCLC. Color Coding: pink = esophagus, orange = trachea, green = main bronchus, yellow = internal target volume (ITV), red = planning target volume (PTV).

Patients Without Histopathological Confirmation of Cancer. Obtaining histologic confirmation of solid pulmonary nodules or masses by biopsy is highly recommended in all practice guidelines.[22–25,68] However, a relevant proportion of patients has been undergoing SBRT without biopsy confirmation.[55] This is in patients considered as being at a too high-risk for performing trans-thoracic or trans-bronchial biopsy confirmation. The probability of malignancy is then estimated using clinical scores considering clinical and imaging factors such as smoking status, lesion size and growth rate, CT morphological criteria such as spiculae, and FDG- PET activity.[69]

Retrospective data on clinically-diagnosed ES NSCLC lesions treated with SBRT, as opposed to histologically-proven ones, showed no significant difference regarding OS and local control while similar rates of DFS and distant failure between pathologically confirmed and presumed NSCLC[70–74] were observed.

A recent prospective observational study of 62 patients undergoing SBRT without histologic confirmation of malignancy (median follow-up of 55 months) reported a 3-year OS of 83.3% for all patients and 94.7% for those under the age of 74. Local, locoregional, and distant failure was reported at rates of 6.4%, 4.8%, and 11.7%, respectively. Eight patients experienced toxicity graded 3 and 4, and there were no grade 5 toxicities.[75]

A systematic review and meta-analysis of 11,047 patients treated with SBRT in 47 cohorts showed a more favorable outcome in terms of DFS and cause-specific survival in clinically staged patients compared to biopsy-proven ones. Regarding OS, they showed better outcomes for the clinically staged patients at 3 years, however 5-year OS did not differ significantly.[76]

It is important to note that the treatment of lung lesions without prior histological confirmation, whether with SBRT or surgery, represents a risk of overtreatment of benign lesions. In patients ineligible for, or refusing biopsy, SBRT is a guideline-recommended option for suspected malignancies.[22–25,68]

Patients With Coexisting Interstitial Lung Disease. Interstitial lung diseases (ILDs) are a heterogeneous group of diffuse parenchymal lung disorders with various patterns of inflammation and extents of fibrosis.[77,78] Idiopathic pulmonary fibrosis (IPF) is the most common form of ILD and describes a chronic and progressive condition of fibrosing of lung tissue, with a median survival of 2–3 years.[77,78] IPF has a poor outcome by itself, but it is also associated with increased rates of lung cancer[79] and treatment-related toxicity[80] or ineligibility for treatment.

Although normo-fractionated radiotherapy of either curative or palliative intent in patients with (sub)clinical ILD has traditionally been associated with a relevant risk of severe radiation-induced pneumonitis,[81–83] retrospective studies on the SBRT have shown mixed results.

While some studies evaluated SBRT in ES NSCLC alone,[84,85] others evaluated ES NSCLC and metastatic lung disease alike.[83,86] All except for one[87] reported significantly higher rates of radiation pneumonitis in their ILD cohorts, with reported incidence of underlying ILD of 6–16% of patients.[84–86,88] The incidence of radiation pneumonitis graded ≥2 and ≥3 in ILD patients was reported at significantly higher rates of 19–55% and 10–32%, respectively in all[84–86] but one[87] study. Grade 5 radiation pneumonitis was reported at rates of 7.6–20%.[85,86]Therefore, the risk of severe toxicity and mortality after SBRT for ES NSCLC needs to be carefully balanced with the risk of the underlying cancer and the pulmonary diseases.

Patients With Local Recurrence After Initial SBRT. Management of local recurrence after SBRT is limited by thorough patient selection. Surgical resection is barely an option in medically inoperable patients, but presents as a salvage treatment option in those previously unwilling to undergo surgery.[89–91] Evidence on repeat SBRT after initial SBRT is limited in terms of patient volume and its retrospective nature. The largest cohort to date has been reported by Ogawa et al.,[92] consisting of 31 patients (n=23 with NSCLC; n=8 with lung metastasis) with either radiologically (n=17) and histologically (n=14) proven local recurrence after initial SBRT. The initial SBRT treatments were mainly performed with 48–52 Gy in 4 fractions, while repeat SBRT doses were mainly either the same or 60 Gy in 8 fractions. The reported OS, PFS and local control at three years for NSCLC patients was 27%, 40%, and 40%, respectively, for central location and 31%, 25%, and 52%, respectively, for peripheral location, with no toxicity graded 3 or higher reported.[84] Smaller studies have reported comparable results.[93–99] The available data suggests that salvage SBRT with BED >100 Gy10 appears to be well tolerated and safely applicable in carefully selected patients with peripheral tumor location; repeat SBRT should be evaluated only very carefully in centrally located tumors.[95] However, as grade 5 toxicity[95,97,100] has been reported in this setting and due to the limited availability of data, more studies evaluating fractionation and dose constraints and normal tissue tolerance are warranted.