The Research Ethics Board of our institution approved this retrospective single institution study.

This retrospective study was based on a cohort of patients used for another research project that investigated the prediction of survival outcomes of SBRT treated lung cancer patients based on radiomics analysis [14]. This study included lung cancer patients treated with SBRT in a single institution (Fig. 1). All the patients included in the study had baseline staging PET/CT imaging before the SBRT treatment. Clinical data about tumor histology, primary tumor size, prescribed biological effective dose, age, gender, initial stage, were obtained from the institutional database and are summarized in Table 1.

The flow chart describes the number of patients that underwent SBRT in our institution between 2008 and 2012 and the inclusion and exclusion criteria of the study

One hundred and thirty-one tumors were biopsy proven based on transbronchial or CT-guided biopsy. Twelve tumors underwent inconclusive biopsy, and the remaining 40 tumors did not undergo biopsy and were judged to be non-small cell lung cancer by consensus in multidisciplinary tumor board meetings based on serial CT and PET-CT findings. The median follow-up period after SBRT was 29 months (range: 8–72).

The SBRT technique used at our institution has been previously described [14, 15]. Patients were immobilized using one of 2 techniques: the Elekta BlueBAG vacuum cushion (Elekta AB, Stockholm, Sweden) with an abdominal compression plate, or the full Elekta BodyFIX system. A four dimensional (4D)-CT was acquired with phase-binning reconstruction software. The gross tumor volume (GTV) was delineated by the radiation oncologist on the 0% (peak inspiratory), 50% (peak expiratory), and maximum intensity projection (MIP) image sets, and their combined volume was used to generate the internal target volume (ITV). There was no expansion for microscopic disease. A 5-mm isotropic margin was added to form the planning target volume (PTV). The radiotherapy plan was calculated on the CT average image set and optimized using 7–10 beam angles. Intensity-modulated radiation therapy (IMRT) was used since 2009. The institutional policy was to deliver 48–52 Gy/4 fractions (fx) for peripheral NSCLC tumors (48 Gy if ≤3 cm, 52 Gy if > 3 cm) and 50 Gy/5fx for all central tumors (defined as tumors immediately adjacent to the esophagus, trachea, main stem bronchi, great vessels, and/or heart), regardless of size or histology. Plans were optimized to aim for ≥99% of the ITV to receive the prescription dose (ITV V100 ≥ 99%), and ≥ 99% of the PTV to receive 95% of the prescription dose (PTV V95 ≥ 99%). Radiotherapy plans were corrected for tissue inhomogeneity using the collapsed cone convolution algorithm.

Treatment was delivered using the Elekta Synergy units (Elekta AB, Stockholm, Sweden) equipped with the Elekta Synergy Beam Modulator (high resolution 4 mm multi-leaf collimator), a kilovoltage cone-beam CT (CBCT) image-guidance system and the Hexapod robotic couch permitting 6 degrees of freedom patient positioning.

Patients were scheduled for chest CT follow-up in intervals of 3–4 months after SBRT for the first 3 years and every 6 months thereafter.

Chest CT studies were conducted using GE LightSpeed Plus or GE Lightspeed VCT 64 MDCT. The parameters used were the following: 120 kVp with tube current adjusted automatically, the beam pitch was 0.984:1, reconstruction thickness 0.625 mm, reconstruction interval 2.5 mm, scan field of view (FOV) 50 cm and display FOV adjusted to patient size, matrix 512 × 512 (pixel spacing: 0.933).

Chest CTs were analyzed by a cardiothoracic radiologist with 15 years’ experience in thoracic imaging, a second-year cardiothoracic imaging fellow and a 4th-year medical student in consensus. We sought to classify the lung tumors on pre-treatment types and post-treatment CT patterns according to the Fleischner Society glossary of terms for thoracic imaging [16] and relevant literature on post radiation pulmonary CT patterns [6, 8, 17].

1) solid nodule: rounded or irregular opacity with homogeneous soft tissue attenuation, well or poorly defined, measuring up to 3 cm in diameter (mass, if > 3 cm), 2) cavitary nodule: a nodule or mass with a gas-filled space seen as lucency or low attenuation area, measuring up to 3 cm in diameter (mass, if > 3 cm), 3) pure GGO nodule: rounded or irregular lesion with hazy increased attenuation in the lung that does not obliterate the bronchial and vascular margins, measuring up to 3 cm in diameter (mass, if > 3 cm), 4) mixed solid/GGO nodule: rounded or irregular lesion which consists of both ground glass and solid soft tissue attenuation components, measuring up to 3 cm in diameter (mass if > 3 cm).

1) diffuse consolidation: homogeneous increase in pulmonary parenchymal attenuation that obscures the margins of vessels and airway walls occasionally with air-bronchogram, with no intervening areas of spared parenchyma, 2) peribronchial/patchy consolidation: patchy areas of consolidation located around the bronchovascular bundles with intervening areas of spared parenchyma, 3) consolidation and ground glass opacities (GGO): combination of areas with increased parenchymal attenuation that obscures the margins of vessels and airway walls and that does not obliterate the bronchial and vascular margins, 4) diffuse GGO: increased parenchymal attenuation that does not obliterate the bronchial and vascular margins with no intervening areas of spared parenchyma, 5) patchy GGO: increased parenchymal attenuation that does not obliterate the bronchial and vascular margins with intervening areas of spared parenchyma, 6) modified conventional pattern: consolidation associated with volume loss, traction bronchiectasis and architectural distortion, similar to, but less extensive than conventional radiation fibrosis, 7) nodule/mass-like pattern: focal consolidation limited around the treated tumor with a well-defined or irregular rounded morphology smaller than 3 cm (nodule) or larger than 3 cm (mass), 8) scar-like pattern: linear or band-like opacity in the region of the treated tumor associated with loss of volume, 9) “orbit-like pattern”: central solid nodule or mass surrounded by relatively normal lung parenchyma (skipped area) and more distally there is a peripheral circumferential band of consolidation, mimicking the appearance of the gravitational trajectory of an object around a central planet [18].

Descriptive statistics frequency and percentages for categorical variables, median and range for continuous variables were used to describe the sample. The frequencies and percentages of pre-treatment lesions and post-treatment CT patterns were calculated. The time points of start and end of increase or decrease in size of the primary targeted lesions post SBRT were recorded. Similarly, the time points of start and end of increase or decrease of the overall volume of radiation injury - when this was inseparable from the targeted lesion - were recorded. The time points of start of each pattern post SBRT and duration of each pattern were also recorded. The period until stabilization without evidence of local or distant recurrence was calculated for each lesion. Stabilization of a CT pattern was defined as absence of change in morphology and size of radiation injury pattern over at least two consecutive follow-up CTs (8 months) until the end of the follow up period of the study. The time point that radiation-related rib fractures, local recurrence, distant or regional metastatic disease occurred post SBRT was also recorded. The statistical software SAS 9.4 was used for data manipulations and analyses.