Correspondence
Details of the directly relevant patterns of intra-thoracic failure (ie, within, marginally within, or outside the radiotherapy field) would have been infor mative and will have implications for the design of future radiotherapy trials of management strategies for this disease.
We are concerned with the increasing trend to extrapolate claims of important patient benefit when this is not supported by data. This over-interpretation is apparent even in abstracts from respected journals. We urge authors, reviewers, and editors to provide a balanced portrayal of data.
I declare no competing interests.
We declare no competing interests.
Michael J McKay
*Simon P Hart, Alyn H Morice
[email protected]
[email protected]
University of Sydney, Sydney, NSW, Australia; and North Coast Cancer Institute, Lismore, NSW, Australia
Hull York Medical School, Respiratory Medicine, Castle Hill Hospital, Cottingham, East Riding of Yorkshire HU16 5JQ, UK
1
1
2
For the NTR1527 trial see http:// www.trialregister.nl/trialreg/ admin/rctview.asp?TC=1527
1292
Slotman BJ, van Tinteren H, Praag JO, et al. Use of thoracic radiotherapy for extensive stage small-cell lung cancer: a phase 3 randomised controlled trial. Lancet 2015; 385: 36–42. van Meerbeeck JP, Ball D. Small-cell lung cancer: local therapy for a systemic disease? Lancet 2015; 385: 9–10.
We disagree with the conclusions made by Ben Slotman and colleagues1 in their study of thoracic radiotherapy for patients with small-cell lung cancer. Overall survival at 1 year was not significantly improved in patients receiving thoracic radiotherapy compared with controls, and we question whether the authors arbitrarily chose points from the Kaplan-Meier curve to support presumed benefits of thoracic radiotherapy. The authors state that they had planned to look at overall survival at 2 years, but this is not mentioned in their protocol recorded in the Netherlands trial register (NTR1527). After 2 years, only 19 of the original 495 patients remained in the trial. Although the effect of thoracic therapy was statistically significant, we would argue that it is clinically invalid to highlight an arbitrarily chosen group of patients who represent less than 5% of the original cohort. We agree that it would be interesting to establish whether there was an increase in long-term survival, but this was not an objective of the study. The authors’ concluding suggestion to consider the universal application of post-chemotherapy thoracic radiotherapy is inappropriate.
Slotman BJ, van Tinteren H, Praag JO, et al. Use of thoracic radiotherapy for extensive stage small-cell lung cancer: a phase 3 randomised controlled trial. Lancet 2015; 385: 36–42.
Authors’ reply Which patients with extensive stage small-cell lung cancer (ES SCLC) are most likely to benefit from thoracic radiotherapy after chemotherapy? We thank the correspondents for their interest in our trial of thoracic radiotherapy in patients with ES SCLC.1 The results showed that thoracic radiotherapy did not significantly improve overall survival (p=0·066). However, thoracic radiotherapy improved 2-year survival (p=0·004), and as pointed out in the accompanying Comment,2 this finding is consistent with the delayed survival benefit reported in a previous practice-changing meta-analysis of thoracic radiotherapy for limited-stage SCLC.3 Yet, the failure of our study to meet the primary endpoint of survival at 1 year has raised a number of questions and comments. Encouse Golden and colleagues commented on the possible benefit of early chemoradiotherapy with higher doses and the possible effect of thoracic radiotherapy on the immune system. Although we cannot exclude that a subgroup of patients with ES SCLC could benefit from early chemoradiotherapy or higher radiotherapy doses, we are unable to identify these patients at present, and in view of the poor prognosis of
ES SCLC, such treatment would not be appropriate for most patients. We agree that radiation-induced cell death could lead to an immunological anti-tumour response in addition to the local effect of the thoracic radiotherapy.4 This area of research merits further investigation. Michael McKay pointed out that treatment for progressive disease could have affected treatment outcomes, as patients might have remained fitter for longer after thoracic radiotherapy, thereby increasing their chance of receiving second-line or third-line chemotherapy. After receiving prophylactic cranial irradiation, patients with ES SCLC are more likely to receive chemotherapy for disease progression than controls (68% vs 45%).5 However we did not record treatment for disease progression in our study1 of thoracic radiotherapy. Our study had 80% power to detect a 10% difference in the overall survival rate at 1 year, and not 71% power, as suggested by Aaron Mansfield and Sumithra Mandrekar. Power calculations were performed using nQuery Advisor v7.0 (3 year accrual and 1 year follow-up; 2 × 230 patients needed). With 5% dropout, 80% power requires 483 patients. In our study,1 495 patients were included in the intention-to-treat (ITT) analysis. Mansfield and Mandrekar further comment on 2-year survival as endpoint and the lack of maturity of the data. We also calculated that to show a survival difference of 10% after 1 year, 416 recorded deaths were needed. At the time of our analysis, 425 patients had died. This number of events is irrespective of the duration of accrual or follow-up time. We find their comment, “median follow-up was only 24 months”, inappropriate, since the median survival was 8 months, and only 13% of patients in the thoracic radiotherapy group and 3% of patients in the control group were still alive at 24 months. The number needed to treat is given with 95% confidence interval and is based on the survival estimates at 2 years. www.thelancet.com Vol 385 April 4, 2015
Simon Hart and Alyn Morice postulated that the use of arbitrarily chosen points from the Kaplan-Meier curve could lead to exaggerated conclusions. Although 2-year survival was not chosen arbitrarily, we have taken this important point into consideration and have performed an additional analysis to improve guidance to clinicians for routine use of thoracic radiotherapy. One stratification factor in the study was the presence or absence of residual intrathoracic disease after chemotherapy, which was assessed in 97% of all randomised patients using a CT scan of the thorax.1 We did an additional analysis of overall survival and progression-free survival in patients with and without residual intrathoracic disease. Of 495 patients included in the ITT analysis, 434 had residual intrathoracic disease at baseline (of whom 215 were allocated to the thoracic radiotherapy group and 219 to the control group), and 61 patients had no residual intrathoracic disease (of whom 32 were allocated to the thoracic radiotherapy group and 29 to the control group). Since residual intrathoracic disease was a stratification factor, we found no differences in patient characteristics between the two groups. In patients with residual intrathoracic disease, the overall survival was significantly longer in the thoracic radiotherapy group (hazard ratio 0·81, 95% CI 0·66–1·00, p=0·044). Survival rates at 1 year were 33% (95% CI 27–40) in the thoracic radiotherapy group and 26% (95% CI 21–33) in the control group. At 2 years, the survival rates were 12% (95% CI 8–19) in the thoracic radiotherapy group and 3% (95% CI 1–8) in the control group. Progression-free survival was significantly longer in the thoracic radiotherapy group (0·70, 0·57–0·85, p=0·0002). Intrathoracic progression, either with or without progression elsewhere, was reported in 43·7% of the thoracic radiotherapy group and in 81·3% in the control group (p
Details of the directly relevant patterns of intra-thoracic failure (ie, within, marginally within, or outside the radiotherapy field) would have been infor mative and will have implications for the design of future radiotherapy trials of management strategies for this disease.
We are concerned with the increasing trend to extrapolate claims of important patient benefit when this is not supported by data. This over-interpretation is apparent even in abstracts from respected journals. We urge authors, reviewers, and editors to provide a balanced portrayal of data.
I declare no competing interests.
We declare no competing interests.
Michael J McKay
*Simon P Hart, Alyn H Morice
[email protected]
[email protected]
University of Sydney, Sydney, NSW, Australia; and North Coast Cancer Institute, Lismore, NSW, Australia
Hull York Medical School, Respiratory Medicine, Castle Hill Hospital, Cottingham, East Riding of Yorkshire HU16 5JQ, UK
1
1
2
For the NTR1527 trial see http:// www.trialregister.nl/trialreg/ admin/rctview.asp?TC=1527
1292
Slotman BJ, van Tinteren H, Praag JO, et al. Use of thoracic radiotherapy for extensive stage small-cell lung cancer: a phase 3 randomised controlled trial. Lancet 2015; 385: 36–42. van Meerbeeck JP, Ball D. Small-cell lung cancer: local therapy for a systemic disease? Lancet 2015; 385: 9–10.
We disagree with the conclusions made by Ben Slotman and colleagues1 in their study of thoracic radiotherapy for patients with small-cell lung cancer. Overall survival at 1 year was not significantly improved in patients receiving thoracic radiotherapy compared with controls, and we question whether the authors arbitrarily chose points from the Kaplan-Meier curve to support presumed benefits of thoracic radiotherapy. The authors state that they had planned to look at overall survival at 2 years, but this is not mentioned in their protocol recorded in the Netherlands trial register (NTR1527). After 2 years, only 19 of the original 495 patients remained in the trial. Although the effect of thoracic therapy was statistically significant, we would argue that it is clinically invalid to highlight an arbitrarily chosen group of patients who represent less than 5% of the original cohort. We agree that it would be interesting to establish whether there was an increase in long-term survival, but this was not an objective of the study. The authors’ concluding suggestion to consider the universal application of post-chemotherapy thoracic radiotherapy is inappropriate.
Slotman BJ, van Tinteren H, Praag JO, et al. Use of thoracic radiotherapy for extensive stage small-cell lung cancer: a phase 3 randomised controlled trial. Lancet 2015; 385: 36–42.
Authors’ reply Which patients with extensive stage small-cell lung cancer (ES SCLC) are most likely to benefit from thoracic radiotherapy after chemotherapy? We thank the correspondents for their interest in our trial of thoracic radiotherapy in patients with ES SCLC.1 The results showed that thoracic radiotherapy did not significantly improve overall survival (p=0·066). However, thoracic radiotherapy improved 2-year survival (p=0·004), and as pointed out in the accompanying Comment,2 this finding is consistent with the delayed survival benefit reported in a previous practice-changing meta-analysis of thoracic radiotherapy for limited-stage SCLC.3 Yet, the failure of our study to meet the primary endpoint of survival at 1 year has raised a number of questions and comments. Encouse Golden and colleagues commented on the possible benefit of early chemoradiotherapy with higher doses and the possible effect of thoracic radiotherapy on the immune system. Although we cannot exclude that a subgroup of patients with ES SCLC could benefit from early chemoradiotherapy or higher radiotherapy doses, we are unable to identify these patients at present, and in view of the poor prognosis of
ES SCLC, such treatment would not be appropriate for most patients. We agree that radiation-induced cell death could lead to an immunological anti-tumour response in addition to the local effect of the thoracic radiotherapy.4 This area of research merits further investigation. Michael McKay pointed out that treatment for progressive disease could have affected treatment outcomes, as patients might have remained fitter for longer after thoracic radiotherapy, thereby increasing their chance of receiving second-line or third-line chemotherapy. After receiving prophylactic cranial irradiation, patients with ES SCLC are more likely to receive chemotherapy for disease progression than controls (68% vs 45%).5 However we did not record treatment for disease progression in our study1 of thoracic radiotherapy. Our study had 80% power to detect a 10% difference in the overall survival rate at 1 year, and not 71% power, as suggested by Aaron Mansfield and Sumithra Mandrekar. Power calculations were performed using nQuery Advisor v7.0 (3 year accrual and 1 year follow-up; 2 × 230 patients needed). With 5% dropout, 80% power requires 483 patients. In our study,1 495 patients were included in the intention-to-treat (ITT) analysis. Mansfield and Mandrekar further comment on 2-year survival as endpoint and the lack of maturity of the data. We also calculated that to show a survival difference of 10% after 1 year, 416 recorded deaths were needed. At the time of our analysis, 425 patients had died. This number of events is irrespective of the duration of accrual or follow-up time. We find their comment, “median follow-up was only 24 months”, inappropriate, since the median survival was 8 months, and only 13% of patients in the thoracic radiotherapy group and 3% of patients in the control group were still alive at 24 months. The number needed to treat is given with 95% confidence interval and is based on the survival estimates at 2 years. www.thelancet.com Vol 385 April 4, 2015
Simon Hart and Alyn Morice postulated that the use of arbitrarily chosen points from the Kaplan-Meier curve could lead to exaggerated conclusions. Although 2-year survival was not chosen arbitrarily, we have taken this important point into consideration and have performed an additional analysis to improve guidance to clinicians for routine use of thoracic radiotherapy. One stratification factor in the study was the presence or absence of residual intrathoracic disease after chemotherapy, which was assessed in 97% of all randomised patients using a CT scan of the thorax.1 We did an additional analysis of overall survival and progression-free survival in patients with and without residual intrathoracic disease. Of 495 patients included in the ITT analysis, 434 had residual intrathoracic disease at baseline (of whom 215 were allocated to the thoracic radiotherapy group and 219 to the control group), and 61 patients had no residual intrathoracic disease (of whom 32 were allocated to the thoracic radiotherapy group and 29 to the control group). Since residual intrathoracic disease was a stratification factor, we found no differences in patient characteristics between the two groups. In patients with residual intrathoracic disease, the overall survival was significantly longer in the thoracic radiotherapy group (hazard ratio 0·81, 95% CI 0·66–1·00, p=0·044). Survival rates at 1 year were 33% (95% CI 27–40) in the thoracic radiotherapy group and 26% (95% CI 21–33) in the control group. At 2 years, the survival rates were 12% (95% CI 8–19) in the thoracic radiotherapy group and 3% (95% CI 1–8) in the control group. Progression-free survival was significantly longer in the thoracic radiotherapy group (0·70, 0·57–0·85, p=0·0002). Intrathoracic progression, either with or without progression elsewhere, was reported in 43·7% of the thoracic radiotherapy group and in 81·3% in the control group (p
