Stereotactic Ablative Radiotherapy (SABR) in Patients with Medically Inoperable Peripheral Early Stage Lung Cancer: Outcomes for the First UK SABR Cohort

Clinical Oncology 28 (2016) 4e12 Contents lists available at ScienceDirect

Clinical Oncology journal homepage: www.clinicaloncologyonline.net

Original Article

Stereotactic Ablative Radiotherapy (SABR) in Patients with Medically Inoperable Peripheral Early Stage Lung Cancer: Outcomes for the First UK SABR Cohort L. Murray *, S. Ramasamy *, J. Lilley, M. Snee, K. Clarke, H.B. Musunuru, A. Needham, R. Turner, V. Sangha, M. Flatley, K. Franks St James’s Institute of Oncology, Leeds Cancer Centre, Leeds, UK Received 21 May 2015; received in revised form 1 September 2015; accepted 8 September 2015

Abstract Aims: To report outcomes for the first UK cohort treated for early stage peripheral lung cancer using stereotactic ablative radiotherapy (SABR). Materials and methods: Patients were included who received SABR between May 2009 and May 2012. Electronic medical records were reviewed for baseline characteristics, treatment details and outcomes. Patients were treated according to the UK SABR Consortium Guidelines. Univariate and multivariate Cox regression was used to determine factors that influenced overall survival and local control. Results: In total, 273 patients received SABR for 288 lesions in the time period examined. The median follow-up was 19.7 months. The median overall survival for all patients was 27.3 months, with 1, 2 and 3 year overall survival of 78.0, 54.9 and 38.6%, respectively. The 1, 2 and 3 year rates of local control were 98.2, 95.7 and 95.7%, respectively. All patients completed the planned course of treatment and rates of Common Toxicity Criteria grade 3þ toxicity were low. On multivariate analysis, patients with Medical Research Council (MRC) breathlessness scores of 3e5 had worse overall survival compared with patients with scores of 1e2 (hazard ratio: 2.10; 95% confidence interval: 1.25e3.59) and the presence of histological diagnosis conferred improved overall survival (hazard ratio: 0.54; 95% confidence interval: 0.31e0.93), probably reflecting that patients who are considered well enough to undergo biopsy are generally fitter overall. No factors were identified that significantly influenced local control. Conclusions: SABR is an effective and well-tolerated treatment option for patients with early stage peripheral lung cancer who are not suitable for surgery. No patient cohort was identified in whom SABR was considered inappropriate. This series adds to the existing positive data that support SABR for this patient group. Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Key words: Lung cancer; stereotactic ablative radiotherapy (SABR)

Introduction Stereotactic ablative radiotherapy (SABR) has become an increasingly recognised treatment option for patients with inoperable peripheral early stage lung cancer. High rates of local control have been shown in prospective and retrospective series [1e7]. Patients are often considered ‘medically inoperable’ as a result of severe or multiple comorbidities. As such, SABR is often used in patients who would otherwise receive no active treatment for lung cancer and yet, despite other significant health problems, it has Author for correspondence: K. Franks, St James’s Institute of Oncology, Leeds Cancer Centre, Beckett Street, Leeds LS9 7TF, UK. E-mail address: [email protected] (K. Franks). * Both authors contributed equally to this work.

been shown that SABR improves survival in this traditionally poor prognostic group [8,9]. St James’s Institute of Oncology (SJIO), Leeds, was the first UK centre to treat early stage peripheral lung cancer using SABR in May 2009. Since then, over 750 patients have received treatment. Here we present outcomes for the first 273 patients.

Materials and Methods Patient Selection Patients treated with SABR at SJIO from May 2009 to May 2012 were included. SJIO is a tertiary centre covering the West Yorkshire region. Patients were also referred from

http://dx.doi.org/10.1016/j.clon.2015.09.007 0936-6555/Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

other centres across the UK. All patients were discussed at the lung cancer multidisciplinary team at SJIO before treatment. In order to be considered for SABR, patients:  Had an multidisciplinary team agreed diagnosis of nonsmall cell lung cancer. This was histologically proven and/or based on the presence of a metabolically active lung lesion on positron emission tomography (PET) scan and/or based on a growing lesion on serial computed tomography imaging.  Had a T1eT2 (<5 cm) N0M0 tumour.  Had a PET scan that confirmed non-metastatic nodenegative disease or, in the case of PET uptake in hilar or medistinal nodes, negative endobronchial ultrasound or endoscopic ultrasound.  Were unsuitable for surgery due to comorbidity, technical inoperability or patient choice.  Had a peripheral tumour: defined as >2 cm in all directions from the central airways (trachea, carina and main bronchus up to the division of the second order bronchi). Suitability for SABR was not based on lung function. Patient Immobilisation and Data Acquisition For planning and treatment, the patient was positioned supine using a wing-board, Vac-lokÔ cushion (CIVCO Medical Solutions, USA) and a knee support. If the wingboard was not suitable (e.g. apical tumour), a thermoplastic immobilisation device was used. A four-dimensional helical computed tomography planning scan was carried out using 2 mm slices with the patient breathing freely and with intravenous contrast unless contraindicated. Stereotactic frames or gating were not used. A 12-phase respiratory cycle was produced and the reference phase was that where the tumour spent the most time. This was the 40% exhalation phase for most patients. This reference phase was used for organ at risk delineation and planning. Target Volume Definition The gross tumour volume (GTV) was contoured on lung windows. Mediastinal windows were also used for tumours adjacent to the chest wall. GTVs were contoured on the 40% exhalation (or equivalent reference phase), 100% exhalation and 0% exhalation (full inspiration) phases of the fourdimensional computed tomography dataset. No expansion was made to the clinical target volume (CTV), i.e. GTV ¼ CTV. An internal target volume (ITV) was contoured on the reference phase to encompass the GTV contoured here and on the 100% and 0% exhalation phases. The other respiratory phases were reviewed to ensure the GTV was encompassed within the ITV throughout the respiratory cycle. The ITV was auto-expanded by 5 mm in all directions to generate a planning target volume (PTV).

5

In the small proportion of cases where four-dimensional computed tomography scanning was unsuccessful (i.e. 5e10% of patients), patients received free-breathing helical computed tomography. Initially, as per the Radiation Therapy Oncology Group (RTOG) 0236 study [10], a 10 mm margin was added superiorly and inferiorly, and a 5 mm margin was added circumferentially to create the PTV. Fourdimensional cone beam computed tomography (CBCT) was then carried out on the treatment linear accelerator before planning to assess the degree of motion with respiration and make sure the tumour did not move outside the PTV. For later patients, four-dimensional CBCT was used to generate the ITV by contouring the GTV on the freebreathing dataset and creating multiple ITVs with universal expansion margins of 3, 5 and 7 mm. The most appropriate ITV was selected (i.e. the one that incorporates the motion envelope), if necessary adjusted, and a 5 mm universal margin added to create the PTV. Organs at Risk The organs at risk routinely contoured included the lungs, airways, spinal cord, oesophagus, heart and brachial plexus. The airways included the trachea and proximal bronchial tree: contouring of the lobar bronchi ceased at the point of segmental division into a second order bronchus. An artificial structure was created by expanding the airways structure by 2 cm. Any patients in whom the GTV encroached on this ‘no fly zone’ were excluded from SABR. Any patients in whom the PTV encroached on the ‘no fly zone’ could proceed, provided the airway constraint was met. Organ at risk dose constraints were based on the ROSEL study (Table 1) [11]. Minor deviations as per Table 1 could be accepted by the treating clinician upon review of the treatment plan. Treatment Planning SABR was planned using XIO (Elekta AB, Sweden) with a superposition algorithm. Plans were initially generated using a class solution involving seven to nine beams, and then individualised. Plans were optimised to conform the high Table 1 Organ at risk constraints (based on the ROSEL study [10]) Organ

Spinal cord

Volume Deviation given as cumulative absolute dose (Gy) (cm3)

Any point Oesophagus 1 Ipsilateral brachial 1 plexus Heart 1 Trachea and main 1 stem bronchus

Three fraction scheme

Five fraction scheme

None Minor

None Minor

18

>18e22 25

>25e28

24 24

>24e27 27 >24e26 27

>27e28.5 >27e29

24 30

>24e26 27 >30e32 32

>27e29 >32e35

6

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

dose area to the PTV, while minimising the surrounding volume treated to intermediate doses. Dose conformity constraints based upon the ROSEL study were used to evaluate plans [11]. Dose Prescription and Treatment Delivery A risk adapted fractionation scheme was used depending on tumour location (Figure 1):

 3  18 Gy e for lesions away from the chest wall and central structures;  5  11 Gy e for lesions in close proximity to the chest wall;  8  7.5 Gy e for lesions in close proximity to central structures to achieve organ at risk constraints. Doses were usually prescribed to the 80% isodose. As per the UK SABR Consortium Guidelines [12], at least 95% of the PTV received the prescription dose and at least 99% of the PTV received at least 90% of the prescription dose. SABR was delivered on an outpatient basis delivering two to three fractions per week with at least 40 h between fractions. All plans were reviewed by clinicians, dosimetrists, physicists, radiologists and radiographers in the weekly lung quality assurance meeting. Online image matching  correction using CBCT was carried out before each fraction. A match within 3 mm tolerance was required to proceed. Any shift was verified by further CBCT before treatment. CBCT was carried out at the end of each fraction and, if there was concern about intrafraction motion, further CBCTs could be carried out midway through treatment. For patients with significant tumour motion (>5 mm) or those tumours close to the diaphragm, four-dimensional CBCT was carried out rather than the standard three-dimensional CBCT. This aided matching and allowed tumour motion to be verified before treatment. In the case of synchronous primaries, each lesion was treated on alternate days. Treatment was delivered as per the UK SABR Consortium Guidelines [12]. Follow-up Patients were reviewed by a radiographer before each fraction. A clinical review was carried out in the event of acute toxicity. After SABR, patients were reviewed at 6 weeks, then 3 monthly for the first year, 6 monthly in year two and yearly thereafter. Patients referred from other tertiary centres were discharged back for local follow-up after SABR, and local follow-up letters were copied to the treatment centre’s electronic patient record. Chest X-rays were carried out at each follow-up visit. Computed tomography scans were scheduled at 6, 12 and 24 months after SABR. Additional scans were carried out at the clinician’s discretion or if there was clinical suspicion of disease recurrence. Data Collection and Statistics

Fig 1. Risk-adapted fractionation strategies for stereotactic ablative radiotherapy: (a) the tumour planning target volume (PTV) does not abut the chest wall or mediastinal structures: 54 Gy in three fractions, (b) the tumour PTV touches or extents into the ribs/pleural: 55 Gy in five fractions and (c) tumour where dose constraints for an organ at risk cannot be met: 60 Gy in eight fractions.

Electronic notes for all patients were retrospectively reviewed for demographics, toxicity (as per the Common Toxicity Criteria for Adverse Events version 4.0) and clinical outcomes. Survival outcomes were evaluated using KaplaneMeier analysis. All measures of survival and control were calculated from the date of SABR completion. For the calculation of relapse-free survival, lung cancer recurrence was considered an event, whereas deaths as a result of non-cancer causes (without previous cancer recurrence) were not included. For the calculation of overall survival, relapse-free survival, nodal

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

control and distant control, each patient was included in the analysis once (even if a patient had more than one lesion treated with SABR), and in the case of sequential primaries, overall survival was calculated from completion of the first SABR treatment. The calculation of local control was based on the total number of lesions treated. Univariate Cox regression was used to compare survival and local control between different patient groups based on age, performance status, Medical Research Council (MRC) breathlessness score, T stage, gender, histological confirmation of diagnosis, standardised uptake value (SUV) on pre-treatment FDG-PET, FEV1 and number of fractions. For the analysis, age was analysed as a continuous variable, performance status was grouped as 0/1 and 2/3, MRC breathlessness was grouped as grades 1e2 and 3e5, FEV1 was grouped as: <1.0 l and 1.0 l and SUV was dichotomised at the median value. Multivariate Cox regression, including the above factors, was also carried out to investigate impact on overall survival and local control. Statistics were analysed using SPSS version 21 software (IBM, USA).

Results Patient Characteristics In total, 273 patients received treatment between May 2009 and May 2012. The median follow-up was 19.68 months. Baseline characteristics are shown in Table 2. The median age was 74 years (range 47e90 years). In total, 288 lesions were treated in 273 patients with 14 patients (5.1%) having more than one lesion treated with SABR: eight patients (2.9%) had two synchronous lesions (treated on alternate days), five patients (1.8%) had sequential lesions and one patient (0.4%) had two lesions at first presentation, both treated with SABR, and a further primary at a later date, also treated with SABR. Twenty-five patients (9.2%) had received previous treatment for lung cancer: 11 patients had previous conventionally fractionated radical chest radiotherapy, 12 had previous lung cancer surgery (nine lobectomies, three pneumonectomies) and two had lobectomy as well as radical/adjuvant radiotherapy. Of 288 lesions treated with SABR, histological confirmation of malignancy was obtained in 100 (34.7%). All patients had positive FDG-PET scans and/or growing lesions on sequential computed tomography scans and were considered eligible for SABR by a lung multidisciplinary team as per UK SABR guidelines [12]. The vast majority of patients were unfit for surgery secondary to poor lung function and comorbidities. The median FEV1 was 1.20 l (range: 0.37e3.97 l) with 34% of patients with known lung function having pre-SABR FEV1 < 1.0 l. Survival The median overall survival for all patients was 27.3 months (95% confidence interval: 22.3e32.2 months;

7

Table 2 Baseline characteristics Factor

n

Total patient number 273 Total number of lesions 288 treated Number of treatment 279* episodes* Age (years) Median: 74 Gender Male 128 Female 145 (total 273) Performance status at time of SABR 0 9 1 102 2 130 3 37 Unknown 1 (total 279) MRC breathlessness score at time of SABR 1 55 2 83 3 63 4 66 5 9 Unknown 3 (total 279) T stage 1 140 2 130 Unknown 18 (total 288) Histology No 188 Yes 100 (total 288) SUV Median: 7.4 FEV1 at time of SABR <1 l 74 1 l 143 Unknown 62 (total 279) Number of fractions 3 (all received 54 Gy) 52 5 (all received 55 Gy) 186 8 (all but one received 50 (total 288) 60 Gy, one patient received 50 Gy)

% 100 100 100 Range: 47e90 46.9 53.1 3.2 36.6 46.6 13.3 0.4 19.7 29.7 22.6 23.7 3.2 1.1 48.6 45.1 6.3 65.3 34.7 Range: 1.4e27.4 26.5 54.8 22.2 18.1 64.5 17.3

SABR, stereotactic ablative radiotherapy; MRC, Medical Research Council; SUV, Standardised Uptake Value. * Of the 273 patients, eight had synchronous primaries treated at the same time (on alternate days), five had sequential primaries, all treated with SABR and one patient had two synchronous primaries (treated at the same time, on alternate days) and a subsequent sequential primary, also treated with SABR.

Figure 2a), with 1, 2 and 3 year overall survival of 78.0, 54.9 and 38.6%, respectively. The median relapse-free survival was not reached at the time of analysis, with 1, 2 and 3 year relapse-free survival of 87.2, 75.2 and 68.3%, respectively. At the time of analysis, 173 patients (63%) remained alive. Of the 100 patients who had died, the cause of death could be determined in 60: 32 patients died of lung cancer and 28 died of other causes. The 30 day mortality rate of our population was 0.4% (one patient), in whom the cause of death is unknown.

8

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

Fig 2. (a) Overall survival for all stereotactic ablative radiotherapy (SABR) patients, (b) local control for all SABR patients, (c) impact of Medical Research Council (MRC) breathlessness score on overall survival and (d) impact of presence or absence of histology on overall survival.

Patterns of Relapse At the time of analysis, there were seven episodes of local recurrence (isolated local relapse in three cases), all diagnosed with chest computed tomography  FDG-PET scan. The 1, 2 and 3 year rates of local control were 98.2, 95.7 and 95.7%, respectively (Figure 2b). There were 14 episodes of nodal relapse (isolated in six cases) with 1, 2 and 3 year rates of nodal control of 97.1, 92.1 and 84.4%, respectively. Of those patients with nodal relapse, seven had normal pre-treatment PET and computed tomography and so did not proceed to

nodal sampling using endobronchial ultrasound or mediastinoscopy. A further six patients had suspicious or indeterminate nodes on pre-treatment PET and so proceeded to nodal sampling, which was negative. The one remaining patient had undergone recent surgery for a more advanced contralateral lung cancer with positive nodes (N1 and N2) before proceeding to SABR without further nodal evaluation. In this case, the more advanced tumour was probably the source of the relapse, although this is unproven. Distant relapse occurred in 21 patients, with 1, 2 and 3 year rates of distant control of 91.9, 88.4 and 88.4%, respectively.

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

Predictors of Outcome: Overall Survival With respect to overall survival, univariate analysis revealed no significant difference in survival by performance status, age, T stage (i.e. T1 versus T2), pre-treatment SUV, FEV1 and gender (Table 3). MRC breathlessness grades 3e5 predicted worse overall survival compared with grades 1e2 (hazard ratio: 1.543, 95% confidence interval: 1.036e2.299, median survival 30.2 months for grades 1e2 versus 21.3 months for grades 3e5). The presence or absence of histology had a significant effect on overall survival, with patients with a confirmed diagnosis of lung cancer having better survival than those without (hazard ratio histology versus no histology: 0.649, 95% confidence interval: 0.424e0.992, median survival with histology: 28.3 months, median survival without histology: 21.5 months). Multivariate analysis for overall survival revealed that both MRC breathlessness grade and the presence or absence of histology remained significant (Table 3). Patients with grades 3e5 breathlessness had significantly worse outcome compared with patients with grades 1e2 breathlessness (hazard ratio: 2.120; 95% confidence interval: 1.251e3.593, P ¼ 0.005; Figure 2c). Patients with a confirmed histological diagnosis had improved overall survival compared with those without histological confirmation (hazard ratio 0.536, 95% confidence interval: 0.305e0.934, P ¼ 0.028; Figure 2d). It was noted that a significantly higher proportion of patients without histology had higher MRC breathlessness scores compared with those with histology (patients without histology and MRC breathlessness scores

9

3e5: 57% versus patients with histology and MRC scores 3e5: 36%, chi-squared test P ¼ 0.001), although, as above, both factors remained significant, independent predictors of overall survival in the multivariate model. Predictors of Outcome: Local Control Univariate analysis of the potential prognostic factors listed above with regard to local control did not identify any statistically significant factors (Table 3). In particular, the presence or absence of histology had no impact on local control. In addition, T stage had no impact on local control. Similarly on multivariate analysis, no significant factors remained in the final model when examining local control. There was also no difference in nodal and distant control in patients with and without a histological diagnosis (P ¼ 0.780 and P ¼ 0.724, respectively). Toxicities SABR treatment was well tolerated and all patients completed their planned treatment. Early (<6 weeks after treatment) grade 1e2 toxicities reported included cough, shortness of breath, pneumonitis, chest pain, fatigue, oesophagitis and skin reactions. During treatment, grade 3 toxicities included shortness of breath (one patient) and fatigue (four patients). No grade 4 toxicities were reported during treatment. At >6 weeks after treatment, grade 3 toxicities included cough (one patient), shortness of breath (10 patients), pneumonitis (one patient with Chronic

Table 3 Results of univariate and multivariate analysis Factor

Univariate analysis Age Gender MRC breathlessness grade: 3e5 versus 1þ2 T stage T2 versus T1 Performance status 2þ3 versus 0þ1 Histological diagnosis Yes versus no Number of fractions 5 versus 3 8 versus 3 SUV Above median versus below median Multivariate analysis MRC breathlessness score (score 3e5 versus 1þ2) Histology Yes versus no MRC, Medical Research Council. SUV, Standardised Uptake Value.

Overall survival

Local control

P value

P value

0.859 0.398 0.033

Hazard ratio where significant (95% confidence interval)

1.543 (1.036e2.299)

0.773 0.562 0.061

0.695

0.683

0.341

0.205

0.046

0.649 (0.424e0.992)

Hazard ratio where significant (95% confidence interval)

0.232

0.486 0.901

0.826 0.955

0.373

0.765

0.005

2.120 (1.251e3.593)

e

e

0.028

0.536 (0.305e0.934)

e

e

10

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

Obstructive Pulmonary Disease (COPD) and a pre-treatment MRC breathlessness score of 2 but no other notable clinical features) and fatigue (four patients). One episode of grade 4 fatigue was also reported at this time point. As above, one patient died within 30 days (on day 25) of SABR completion and the cause of death is unknown. No SABR-related symptoms were reported, but this event could potentially be grade 5 toxicity. Late (12 weeks after treatment) grade 3 toxicities were uncommon and included breathlessness (eight, five and two patients at 12 weeks, 6 months and 12 months, respectively), pneumonitis (one patient at 12 weeks), chest pain (three patients at 12 months) and fatigue (one patient at 12 months). Grade 4 shortness of breath was reported in one patient at 12 months. This is the only reported late grade 4 toxicity. One patient died at 12 weeks after SABR. She was a patient with pre-SABR MRC breathlessness grade 5 who developed increasing shortness of breath around 12 weeks after SABR. Chest computed tomography carried out less than 1 week before death showed no evidence of pneumonitis. The cause of death was attributed to severe COPD. Grade 3þ toxicities are summarised in Table 4. Shortness of breath as a consequence of treatment was difficult to objectively assess due to the poor pre-existing lung function in most patients, as evidenced by pre-SABR MRC breathlessness scores.

Discussion We present outcomes for the first series of lung cancer patients treated using SABR in the UK. As in other series [1e7,13], we observed excellent local control (>95% at 3 years) and low rates of nodal relapse. It has previously been shown that more complete PTV coverage (in terms of the volume of PTV receiving at least 100% of the prescription dose) results in improved local control [14], and although we

did not carry out a detailed dosimetric analysis here, our coverage requirements were that at least 95% of the PTV received the prescription dose, which probably contributed to the observed high rates of local control. Overall survival was 39% at 3 years, a figure towards the lower end of that reported in other series (3 year overall survival from 38 to 60% in [3,4,6,13,15]) and this probably reflects the severe comorbidities experienced by our real world, non-trial patient population. It is also noted that lung cancer patients in the UK have poorer survival outcomes compared with patients elsewhere and this may be a contributing factor, but is difficult to comment on further with specific regard to the dataset analysed here [16]. We identified that patients with poorer MRC breathlessness scores fared less well in terms of overall survival, although not in terms of local control. These patients should, however, still be considered for SABR: despite the presence of severe comorbidities, it has previously been shown that lung SABR improves survival in this patient group [8,9]. Furthermore, although no formal statistical comparisons can be carried out, registry data have shown that untreated patients with stage I non-small cell lung cancer have 3 year survival in the region of 25% for patients with T1 disease and 12% for patients with T2 disease, with median survivals of 15 and 12 months, respectively [17], figures that are worse than those shown for SABR patients in this cohort, including those with poor MRC breathlessness scores (3 year survival 38% overall, median survival 27 months overall, and 21 months in those with breathlessness scores of 3e5), again supporting the role of SABR in patients with comorbidities. No patient group was therefore identified for which we believe that SABR should not be considered. Treatment was well tolerated overall, and all patients completed their planned course of treatment. Our toxicity outcomes are also in keeping with those published for lung SABR elsewhere, with low rates of grade 3þ toxicity. The rates of grade 3þ cough (0.4% at 6 weeks) and pneumonitis (0.4% at 6 weeks and 0.4% at 12 weeks) were very low, and

Table 4 Grade 3þ toxicities: number of patients (and percentage)

Cough Breathlessness

Pneumonitis Chest pain Oesophagitis Skin toxicity Fatigue Death potentially related to treatment *

Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade

3 4 3 4 5 3 4 3 4 3 4 3 4 3 4 5

On treatment

6 weeks

12 weeks

6 months

12 months

0 0 1 (0.4) 0

1 (0.4) 0 10 (3.8) 0

0 0 5 (2.2) 0

0 0 2 (1.2) 1 (0.6)

0 0 0 0 0 0 0 0 4 (1.5) 0

1 (0.4) 0 0 0 0 0 0 0 4 (1.5) 1 (0.4) 1 (0.4) (on day 25)

0 0 8 (3.1) 0 1 (0.4)* 1 (0.4) 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0

0 0 3 (1.7) 0 0 0 0 0 1 (0.6) 0

Death was attributed to Chronic Obstructive Pulmonary Disease (COPD) rather than treatment.

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

compare favourably with published series. For example, Bral et al. [5] reported early and late grade 3 cough in 5% and 0% of patients, respectively, and early and late grade 3 pneumonitis in 5% and 22% of patients, respectively. Similarly, Taremi et al. [2] reported early grade 3 pneumonitis in 0% of patients and late grade 3 pneumonitis in 2% of patients [2] and Hayashi et al. [7] reported grade 3 pneumonitis in 2.5% of patients (time point not specified). The rate of late grade 3þ dyspnoea also seems to be similar to those published by Bral et al. [5] and Taremi et al. [2] [4% and 2 % (combined with cough), respectively] but, as noted above, shortness of breath was difficult to assess objectively given the poor lung function of most patients prior to SABR. Strategies have been suggested that aim to maximise the therapeutic ratio, and these should be adopted in an effort to ensure toxicity rates are as low as possible [18]. It has previously been postulated that the excellent local control rates experienced by patients receiving lung SABR could partly be the result of patients diagnosed based on radiology alone, and without histological confirmation of cancer. FDG-PET scanning, which is now routinely used to help differentiate benign form malignant lesions, was not widely used in some of the studies that suggested this [19]. In this current series, histological confirmation of cancer was obtained in 35% of patients, but all patients had PET scanning as part of the diagnostic work-up, adding confidence to the presumed diagnosis of malignancy. Indeed, more recent series, where a larger proportion of patients have received PET scanning, have found no difference in outcomes between patients with and without histological diagnosis [2]. In this current series, in terms of overall survival, we observed that patients with histological confirmation of disease had improved outcomes compared with those without, although local control was not influenced by the presence or absence of histology. It could be that lack of histology reflects patients with particularly severe comorbidities in whom biopsy is considered too high risk a procedure, and it is these comorbidities that affect overall survival rather than the lack of histology itself. We also observed that patients with poor MRC breathlessness scores had worse overall survival than those with the lowest scores, in part reflecting the effect of comorbidities on overall survival. This was not, however, reflected by FEV1 or performance status, which had no significant effect on overall survival, but which could also be considered reflective of comorbidities. On multivariate analysis the absence of histology and high MRC breathlessness scores remained significant in predicting poorer overall survival. Again, lack of histology may reflect patients considered unfit for biopsy, which is perhaps not adequately represented by measures such as performance status. An alternative measure, such as the Charleston Comorbidity Index, might help clarify this issue, but unfortunately these data were not available for this analysis. In addition, our sample size was relatively small and a larger sample size would be more appropriate for detailed statistical analysis. No factors were identified which had a significant effect on local control. This may be because the number of local failures in this series was very low.

11

The main limitation of this work is that it was retrospective and therefore subject to the inherent biases and problems that accompany this method of data collection. Our results are, however, similar to those reported in prospective series, which is reassuring. In addition, the number of patients involved was relatively small, and follow-up was relatively short. This is, however, the first reported series of UK lung SABR patients, as well as one of the larger series in the literature. Several series, including this one, show excellent outcomes from SABR in early peripheral lung cancer. Data suggest that outcomes are superior to those achieved with conventionally fractionated radiotherapy [9]. The main question that remains unanswered is the role of SABR in operable early stage lung cancer as an alternative to surgery. Studies that have attempted to compare outcomes between patients treated with SABR and those treated with surgery have shown mixed results, although there is a suggestion from matched analyses that SABR and surgery may be equivalent at least in terms of local control, if not also with respect to specific and overall survival [4,9,20,21]. Selection bias remains an issue in any non-randomised comparison, however, and randomised trials will be required to adequately answer this question. The SABRTOOTH trial (Stereotactic ablative radiotherapy with surgery in patients with peripheral stage I non-small cell lung cancer considered higher risk of complications from surgical resection) is a pilot study investigating SABR versus surgery, which will hopefully address this issue. It aims to assess the feasibility and acceptability of conducting a phase III randomised controlled trial comparing SABR with surgery in patients considered at high risk of surgical complications. Until this and similar studies provide evidence, it seems likely that in the UK, SABR will be reserved for medically inoperable patients and those who refuse surgery.

Conclusions In keeping with existing data, SABR has been shown to be an effective treatment option for medically inoperable patients with peripheral early stage lung cancer, resulting in high rates of local control. Treatment was well tolerated. Patients with high MRC breathlessness scores fared less well in terms of overall survival, although local control rates were comparable with other patients. The absence of histology also predicted worse overall survival, probably reflecting the poorer overall condition in those patients in whom biopsy was deemed too high risk. No patient group was identified in this cohort in whom SABR was felt to be inappropriate. These initial results for a UK population are highly encouraging and add to the existing positive data regarding the role of SABR in early stage peripheral lung cancer.

Acknowledgements Elekta have a research agreement with St James’s Institute of Oncology which provides funding for PhD work and provides support for travel to meetings.

12

L. Murray et al. / Clinical Oncology 28 (2016) 4e12

References [1] Lagerwaard FJ, Haasbeek CJ, Smit EF, Slotman BJ, Senan S. Outcomes of risk-adapted fractionated stereotactic radiotherapy for stage I non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2008;70(3):685e692. Epub 2008/01/01. [2] Taremi M, Hope A, Dahele M, et al. Stereotactic body radiotherapy for medically inoperable lung cancer: prospective, single-center study of 108 consecutive patients. Int J Radiat Oncol Biol Phys 2012 Feb 1;82(2):967e973. [3] Timmerman R, Paulus R, Galvin J, et al. Stereotactic body radiation therapy for inoperable early stage lung cancer. JAMA 2010;303(11):1070e1076. Epub 2010/03/18. [4] Grills IS, Mangona VS, Welsh R, et al. Outcomes after stereotactic lung radiotherapy or wedge resection for stage I nonsmall-cell lung cancer. J Clin Oncol 2010;28(6):928e935. Epub 2010/01/13. [5] Bral S, Gevaert T, Linthout N, et al. Prospective, risk-adapted strategy of stereotactic body radiotherapy for early-stage nonsmall-cell lung cancer: results of a Phase II trial. Int J Radiat Oncol Biol Phys 2011;80(5):1343e1349. Epub 2010/08/17. [6] Ricardi U, Frezza G, Filippi AR, et al. Stereotactic ablative radiotherapy for stage I histologically proven non-small cell lung cancer: an Italian multicenter observational study. Lung Cancer 2014;84(3):248e253. Epub 2014/04/01. [7] Hayashi S, Tanaka H, Kajiura Y, Ohno Y, Hoshi H. Stereotactic body radiotherapy for very elderly patients (age, greater than or equal to 85 years) with stage I non-small cell lung cancer. Radiat Oncol 2014;9:138. Epub 2014/06/18. [8] Louie AV, Rodrigues G, Hannouf M, et al. Withholding stereotactic radiotherapy in elderly patients with stage I nonsmall cell lung cancer and co-existing COPD is not justified: outcomes of a Markov model analysis. Radiother Oncol 2011;99(2):161e165. Epub 2011/05/31. [9] Shirvani SM, Jiang J, Chang JY, et al. Comparative effectiveness of 5 treatment strategies for early-stage non-small cell lung cancer in the elderly. Int J Radiat Oncol Biol Phys 2012;84(5):1060e1070. Epub 2012/09/15. [10] Radiation Therapy Oncology Group. RTOG 0236. A phase II trial of stereotactic body radiation therapy (SBRT) in the treatment of patients with medically inoperable stage I/II non-small cell lung cancer. [Internet] 2004 [updated 9 September 2009; cited 2 August 2015]. Available at: https:// www.rtog.org/ClinicalTrials/ProtocolTable/StudyDetails.aspx? study¼0236. [11] Hurkmans CW, Cuijpers JP, Lagerwaard FJ, et al. Recommendations for implementing stereotactic radiotherapy in

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

peripheral stage IA non-small cell lung cancer: report from the Quality Assurance Working Party of the randomised phase III ROSEL study. Radiat Oncol 2009;4:1. Epub 2009/01/14. UK SABR Consortium. Stereotactic ablative body radiotherapy (SABR): a resource. Version 4.1, April 2014. [Internet] 2014 [updated April 2014; cited 7 July 2014]. Available at: http:// www.actionradiotherapy.org/wp-content/uploads/2014/05/ UKSABRConsortiumGuidellinesv41.pdf. Andratschke N, Zimmermann F, Boehm E, et al. Stereotactic radiotherapy of histologically proven inoperable stage I nonsmall cell lung cancer: patterns of failure. Radiother Oncol 2011;101(2):245e249. Epub 2011/07/05. Thibault I, Poon I, Yeung L, et al. Predictive factors for local control in primary and metastatic lung tumours after four to five fraction stereotactic ablative body radiotherapy: a single institution’s comprehensive experience. Clin Oncol 2014; 26(11):713e719. Baumann P, Nyman J, Hoyer M, et al. Outcome in a prospective phase II trial of medically inoperable stage I nonsmall-cell lung cancer patients treated with stereotactic body radiotherapy. J Clin Oncol 2009;27(20):3290e3296. Epub 2009/05/06. Allemani C, Weir HK, Carreira H, et al. Global surveillance of cancer survival 1995e2009: analysis of individual data for 25,676,887 patients from 279 population-based registries in 67 countries (CONCORD-2). Lancet 2015;385(9972): 977e1010. Raz DJ, Zell JA, Ou SH, Gandara DR, Anton-Culver H, Jablons DM. Natural history of stage I non-small cell lung cancer: implications for early detection. Chest 2007;132(1):193e199. Epub 2007/05/17. Lo SS, Sahgal A, Chang EL, et al. Serious complications associated with stereotactic ablative radiotherapy and strategies to mitigate the risk. Clin Oncol 2013;25(6):378e387. Inoue T, Shimizu S, Onimaru R, et al. Clinical outcomes of stereotactic body radiotherapy for small lung lesions clinically diagnosed as primary lung cancer on radiologic examination. Int J Radiat Oncol Biol Phys 2009;75(3):683e687. Epub 2009/ 02/24. Crabtree TD, Denlinger CE, Meyers BF, et al. Stereotactic body radiation therapy versus surgical resection for stage I nonsmall cell lung cancer. J Thoracic Cardiovasc Surg 2010;140(2):377e386. Epub 2010/04/20. Shirvani SM, Jiang J, Chang JY, et al. Lobectomy, sublobar resection, and stereotactic ablative radiotherapy for earlystage non-small cell lung cancers in the elderly. JAMA 2014;149(12):1244e1253. Epub 2014/10/17.