Clinical Oncology 27 (2015) 205e212 Contents lists available at ScienceDirect
Clinical Oncology journal homepage: www.clinicaloncologyonline.net
Original Article
Factors Affecting the Use of Single-Fraction Radiotherapy for the Palliation of Bone Metastases in Australia A.N. Petrushevski *, G.S. Gabriel y, T.P. Hanna *y, S. Allen y, R.W. Allison z, M.B. Barton *y * Department
of Radiation Oncology, Liverpool Hospital, Liverpool, New South Wales, Australia Collaboration for Cancer Outcomes Research and Evaluation (CCORE), Ingham Institute for Applied Medical Research, University of New South Wales, Liverpool, New South Wales, Australia z Cancer Care Services, Royal Brisbane and Women’s Hospital, Brisbane, Australia y
Received 3 June 2014; received in revised form 13 October 2014; accepted 30 November 2014
Abstract Aims: Palliative radiotherapy for bone metastases remains an important treatment in patients with metastatic malignancy. Previous studies have indicated a reluctance to adopt single-fraction treatment despite considerable evidence. This study aims to describe the factors determining the use of palliative radiotherapy in patients with bone metastases and assess whether fractionation patterns have changed over time with emerging evidence. Materials and methods: A retrospective review of radiotherapy databases at Liverpool/Macarthur Cancer Therapy Centre and the Royal Brisbane and Women’s Hospital was conducted for the period 1997e2009. Patients receiving palliative radiotherapy for bony metastases were identified and treatment sites were grouped into ‘spine’, ‘limb’, ‘multiple’ or ‘other’. Treatment courses were divided into single- or multiple-fraction treatments. The effects of socioeconomic and geographical factors on radiotherapy utilisation and fractionation were assessed. Results: In total, 5683 patients were identified in the cohort; they received a total of 8211 bone treatments. The overall proportion of single-fraction radiotherapy was 29%, with significant variation over the study period (P < 0.001). Age under 70 years and spine or multiple treatment sites were all associated with lower usage of single-fraction radiotherapy on multivariate analysis. Prostate and lung primary sites were associated with higher usage of single-fraction treatment. The proportion of single-fraction treatment remained low (35%), even for patients who survived less than 22 days from their last treatment. Socioeconomic and geographical factors had little effect on the number of fractions used. Conclusions: The rate of single-fraction radiotherapy for bone metastases has remained low in two large Australian institutions, despite considerable evidence that single-fraction treatment provides equivalent pain relief to fractionated therapy. This trend towards fractionated treatment was largely maintained, even in patients with limited life expectancy. Further measures to increase the rate of single-fraction therapy are needed. Crown Copyright Ó 2014 Published by Elsevier Ltd on behalf of The Royal College of Radiologists. All rights reserved.
Key words: Bone; fractionation; metastasis; palliative; radiotherapy
Introduction Radiotherapy is an important palliative treatment modality for patients with bony metastases. Previous work has shown that radiotherapy to bone sites accounted for over 16% of all radiotherapy treatment courses in several Australian centres [1]. Many randomised controlled trials [2e9] and two subsequent systematic reviews [10,11] have shown no Author for correspondence: A.N. Petrushevski, Department of Radiation Oncology, Liverpool Hospital, Elizabeth St, Liverpool, NSW 2170, Australia. Tel: þ61-2-9828-5189; Fax: þ61-2-9828-5189. E-mail address: [email protected] (A.N. Petrushevski).
significant difference between single-fraction radiotherapy (SFRT) (typically 8e10 Gy) and multiple-fraction courses (typically 20 Gy in five fractions or 30 Gy in 10 fractions) in terms of overall pain response and complete response. Furthermore, no significant difference between the treatment groups was seen in terms of toxicity, time to pain response or durability of analgesic effect. The most recent meta-analysis from Chow et al. [12] included 5000 patients from 16 randomised trials and confirmed the findings of the earlier data. SFRT is strongly supported over multiplefraction radiotherapy (MFRT) for uncomplicated bone metastases in most practice guidelines [13e15]. SFRT has several potential benefits for patients, their care givers and the health system. It offers significant logistical
http://dx.doi.org/10.1016/j.clon.2014.11.027 0936-6555/Crown Copyright Ó 2014 Published by Elsevier Ltd on behalf of The Royal College of Radiologists. All rights reserved.
206
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212
advantages, particularly in patients with limited life expectancy or geographical factors precluding fractionated treatment [16,17]. Previous studies have shown wide variety in the utilisation of palliative radiotherapy for bone metastases, with evidence that both patient and health care system-related factors affect patterns of care [16,18]. Practice surveys of clinicians have also revealed marked variance in practice and choice of fractionation regimen for various clinical scenarios [19e22]. The aims of our study were to assess the effect of various clinical and demographic factors on radiotherapy utilisation and fractionation for the palliation of bony metastases in Australia, and to assess whether fractionation patterns have changed over time with emerging evidence and practice guidelines.
Materials and Methods Sources of Data A retrospective review was conducted using routinely collected administrative data from two large Australian public cancer centres, the Liverpool and Macarthur Cancer Therapy Centres (LM, New South Wales, Australia) and the Royal Brisbane and Women’s Hospital (RBH, Queensland, Australia). The databases used for the analysis contained details of all patients undergoing radiotherapy treatment in these institutions between 1997 and 2009, including patient demographics, diagnosis date and primary cancer site. Treatment date, site and dose information were collected for each individual treatment. Variables extracted for analysis were standardised across both institutions to allow direct comparison. Study Population Patients aged 18 years and older with a histologically proven malignancy who had received palliative radiotherapy to a bone site at one of the two treatment institutions between 1997 and 2009 were included in the study. ICD-10 and morphology codes were used to identify the primary cancer diagnosis and morphology. Primary bone tumours were excluded from the analysis to exclude potentially curative treatments. Treatment sites were grouped into five categories: primary, bone, brain, soft tissue and multiple. Treatments assigned to the ‘multiple’ treatment code that consisted of at least one identified bone site were selected for the cohort, along with all ‘bone’ category treatments. Where radiotherapy dose was available, any treatment dose over 40 Gy was excluded, to avoid potentially curative treatments from being included in the analysis. Patients with cutaneous primaries and ambiguous treatment site information (for example, ‘thigh’ or ‘shoulder’) were also excluded to avoid any skin treatments being included. Due to difficulty ascertaining which malignancy a treatment course was assigned to, patients with multiple primary cancers were not included in the study.
Study Variables Patient variables extracted for analysis included age, gender, primary cancer diagnosis, diagnosis date and death date. Treatment year, treatment site and number of bone treatments were also determined. Treatment site was categorised into spine, limb, other or multiple (a combination of a bone and non-bone site treated simultaneously). Treatment start and end dates were used to classify treatment episodes as either single- or multiple-fraction treatments (i.e. if a treatment episode started and finished on the same day it was considered a single-fraction treatment). Subsequent bone treatments were also identified in patients selected for the study. If treatment start and end dates were more than 7 days apart, they were considered separate treatment episodes. Treatment end dates and patient death date were used to determine the time from the last bone treatment to death. The effect of these variables on fractionation regimen was assessed. Geographical data were collected on the patient cohort, using the postcode of the patient’s residential address. This was used to determine the Socioeconomic Index for Areas (SEIFA) score, which is determined for each postcode using Australian census data. This score was then used to divide the cohort into five quintiles based on the Index of Relative Socioeconomic Disadvantage (IRSD) [23]. The patient’s postcode data were used to determine geographical remoteness using the Accessibility/Remoteness Index of Australia (ARIA). The ARIA scores were grouped into five groups, ranging from highly accessible to very remote. Statistical Analysis Statistical Package for Social Science (SPSS) version 21 was used to carry out a data analysis [24]. Categorical variables were compared using Pearson chi-squared tests, and P values were determined. Logistic regression analysis was used to determine variables independently associated with fractionation choice. The patient’s street address or postcode and hospital address were geocoded to allow calculation of the travelling distance from the patient’s postcode to the nearest treatment facility using ArcGIS software [25]. Road distance was subsequently used to analyse whether travelling distance was associated with fractionation regimen.
Results Study Population Between 1997 and 2009, 5683 patients received at least one course of palliative bone radiotherapy, and among them 8211 individual bone treatment episodes were identified. Table 1 presents the baseline patient characteristics of the cohort. Fifty-nine per cent of the patients were men and the median age of the study population was 67 years, with a range of 18e97 years. Patients diagnosed with lung (30%),
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212 Table 1 Baseline patient characteristics of the study cohort Total number of Total number of individual patients treatment courses (%) (%) Gender Male Female Age 79 Primary site Breast Prostate Lung Colorectal Head and neck Upper gastrointestinal Gynaecological Melanoma Other genitourinary Other Treating institution RBH LM Year of treatment 1997e2000 2001e2004 2005e2009 Total
3367 (59%) 2316 (41%)
4763 (58%) 3448 (42%)
1806 1551 1624 702
(32%) (27%) (29%) (12%)
2819 2223 2245 924
(35%) (27%) (27%) (11%)
1110 1078 1685 283 159 132 100 186 386 564
(19%) (19%) (30%) (5%) (3%) (2%) (2%) (3%) (7%) (10%)
1834 1790 2198 358 216 161 135 239 530 750
(22%) (22%) (27%) (4%) (3%) (2%) (2%) (3%) (6%) (9%)
4416 (78%) 1267 (22%)
6384 (78%) 1827 (22%)
2330 (41%) 2046 (36%) 1307 (23%) 5683
3249 (40%) 2970 (36%) 1992 (24%) 8211
RBH, Royal Brisbane and Women’s Hospital; LM, Liverpool and Macarthur Cancer Therapy Centres.
prostate (19%) or breast (19%) cancers accounted for more than two thirds of the cohort. The overall mean number of bone treatments per patient was 1.45, with a higher mean number of treatments for prostate (1.66) and breast (1.65)
Fig 1. Frequency distribution of the number of bone treatment episodes given per patient between 1997 and 2009.
207
cancers. Patients under 60 years of age had a mean number of 1.56 bone treatments each, compared with 1.31 treatments for patients aged 80 years and over (P < 0.001). Figure 1 shows the number of bone radiotherapy treatment episodes per patient. Forty-five per cent of patients underwent only one bone treatment, 30% received a subsequent second bone treatment, with diminishing numbers for subsequent bone treatments, although a small number of patients received over five separate bone treatments during the study period. Demographic Factors and Number of Treatments Of the 5683 patients, a valid postcode was available for 5402 patients (who received 7818 treatments); this was used for the analysis of socioeconomic status and remoteness. Table 2 shows the effect of demographic factors on the mean number of treatment episodes per patient. Socioeconomic status (IRSD) had a small effect on the mean number of bone treatments, with patients living in higher socioeconomic areas being more likely to receive subsequent bone treatments (mean number of treatments 1.49 for IRSD 4/5 patients versus 1.40 for IRSD 1, P < 0.001). Similarly, patients living in highly accessible areas had a higher number of bone treatments compared with those living in very remote regions (1.49 versus 1.33 respectively, P < 0.001). Analysis of Fractionation The overall rate of SFRT was 29% for the cohort over the study period (Figure 2), with marked variance over time (P < 0.001). The rate of SFRT was lowest in the first year of the study (8%), but slowly increased each year before reaching 38% in 2003e2004. The rate of SFRT subsequently plateaued, averaging 31% for the period after 2004. 20 Gy in five fractions (47.2%) and 30 Gy in 10 fractions (11.4%) were the most common MFRT regimens (mean dose 21.5 Gy for LM versus 22.5 Gy for RBH). The patient and disease-related factors associated with the use of SFRT are summarised in Table 3. Men were more likely (P < 0.001) to receive single-fraction treatments (31%) than women (25%). Age was significantly associated with choice of fractionation (P < 0.001), with patients 80 years and older at the time of treatment (36%) being more likely to receive single fractions compared with those under 60 years of age (24%). Prostate (37%) and lung (31%) cancer patients had higher proportions of single-fraction treatments, whereas those with breast cancers (22%), melanoma (20%) or other genitourinary tumours (21%) were less likely to receive single fractions. Fractionation varied between the two institutions, with LM (36.7%) reporting a higher rate of SFRT than RBH (26.5%). The site of treatment was associated with marked variation in the use of SFRT (P < 0.001), with ‘limb’ (40%) and ‘other’ (41%) categories having higher rates of single-fraction treatment compared with ‘spine’ (25%) and ‘multiple’ sites (22%). For bone treatments given to patients who died during the study period (n ¼ 5254), the group who died
208
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212
Table 2 Effect of socioeconomic status and geographical remoteness on the mean number of bone treatments Patients
Socioeconomic status IRSD-1 (Lowest) 1180 IRSD-2 1015 IRSD-3 1063 IRSD-4 1074 IRSD-5 (highest) 1070 Geographical remoteness (ARIA) Highly accessible 3700 Moderately accessible 1340 Accessible 253 Remote 76 Very remote 33 Total 5402
Bone treatments
Mean number of bone treatments per patient
1651 1448 1530 1599 1590
1.40 1.43 1.44 1.49 1.49
5507 1824 338 105 44 7818
1.49 1.36 1.34 1.38 1.33 1.45
On multivariate logistic regression analysis (Table 4), age under 70 years (odds ratio 0.81) was independently associated with lower utilisation of SFRT. Patients who received their treatment at RBH were significantly less likely to receive SFRT (odds ratio 0.53). Spine and multiple treatment sites were also associated with lower SFRT usage (odds ratio 0.44 and 0.34, respectively). Patients who survived more than 21 days after their last bone treatment had significantly lower odds of receiving single-fraction treatment (odds ratio 0.70). Prostate (odds ratio 2.03) and lung (odds ratio 1.34) were the only primary sites significantly associated with a higher proportion of SFRT, with no primary sites significantly associated with lower SFRT use. Gender was not significantly associated with fractionation choice on multivariate analysis.
Discussion
IRSD, Index of Relative Socioeconomic Disadvantage; ARIA, Accessibility/Remoteness Index of Australia.
Proportion of bone treatments
within 21 days of their last bone treatment had a higher proportion of SFRT (35% versus 29%, P ¼ 0.006). However, MFRT was still the most common treatment in patients who died within 21 days of last treatment (65% of treatments). Socioeconomic status was not associated with fractionation regimen for those patients who received treatment (P ¼ 0.4). Although there was a small statistically significant trend towards patients living in remote and very remote areas receiving multiple fractions (80% in remote/very remote versus 70% in highly accessible areas, P ¼ 0.002), there was no significant difference between the groups when travelling distance to the treatment facility was examined (P ¼ 0.3).
Radiotherapy is an important palliative modality in the treatment of patients with bony metastases. This study is one of the largest studies assessing patterns of fractionation since the emergence of evidence-based guidelines for the palliation of bone metastases. It also covers a large time period, during which the outcome of several randomised trials were published, all underlining the equal effectiveness of single and multiple fractions on bone pain. Several population-based studies have previously documented significant variation in patterns of practice in the treatment of bone metastases worldwide. Elderly patients and those living in remote areas are particularly underserviced in terms of access to palliative radiotherapy [16,18,26]. There is also evidence that health financing and departmental culture may influence fractionation [20]. This variation in practice is supported by numerous practice surveys that have been published, indicating that
Year of treatment Single Fraction
Multiple Fraction
Fig 2. Proportion of single-fraction radiotherapy for bone metastases by year of treatment.
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212 Table 3 Patient and tumour factors associated with fractionation regimen Number of Number of P value single-fraction multipletreatments fraction treatments Gender Male 1491 (31%) 3272 Female 872 (25%) 2576 Age 79 702 (36%) 924 Primary site Breast 412 (22%) 1422 Prostate 660 (37%) 1130 Lung 688 (31%) 1510 Colorectal 93 (26%) 265 Head and neck 56 (26%) 160 Upper gastrointestinal 49 (30%) 112 Gynaecological 30 (22%) 105 Melanoma 47 (20%) 192 Other genitourinary 113 (21%) 417 Other 215 (29%) 536 Treating institution RBH 1692 (27%) 4692 LM 671 (37%) 1156 Bone treatment site Spine 797 (25%) 2405 Limb 392 (40%) 587 Other 614 (41%) 872 Multiple 560 (22%) 1984 Time from last bone radiotherapy to death 0e7 days 90 (35%) 168 8e21 days 176 (34%) 336 >21 days 1295 (29%) 3189 Socioeconomic status (SEIFA) IRSD-1 475 (29%) 1176 IRSD-2 416 (29%) 1032 IRSD-3 406 (27%) 1124 IRSD-4 467 (29%) 1132 IRSD-5 465 (29%) 1125 Geographic remoteness (ARIA) Highly accessible 1634 (30%) 3873 Accessible 483 (26%) 1341 Moderately accessible 82 (24%) 256 Remote and very remote 30 (20%) 119 Travelling distance to treatment facility 100 km 945 (28%) 2486
(69%)
Clinical Oncology journal homepage: www.clinicaloncologyonline.net
Original Article
Factors Affecting the Use of Single-Fraction Radiotherapy for the Palliation of Bone Metastases in Australia A.N. Petrushevski *, G.S. Gabriel y, T.P. Hanna *y, S. Allen y, R.W. Allison z, M.B. Barton *y * Department
of Radiation Oncology, Liverpool Hospital, Liverpool, New South Wales, Australia Collaboration for Cancer Outcomes Research and Evaluation (CCORE), Ingham Institute for Applied Medical Research, University of New South Wales, Liverpool, New South Wales, Australia z Cancer Care Services, Royal Brisbane and Women’s Hospital, Brisbane, Australia y
Received 3 June 2014; received in revised form 13 October 2014; accepted 30 November 2014
Abstract Aims: Palliative radiotherapy for bone metastases remains an important treatment in patients with metastatic malignancy. Previous studies have indicated a reluctance to adopt single-fraction treatment despite considerable evidence. This study aims to describe the factors determining the use of palliative radiotherapy in patients with bone metastases and assess whether fractionation patterns have changed over time with emerging evidence. Materials and methods: A retrospective review of radiotherapy databases at Liverpool/Macarthur Cancer Therapy Centre and the Royal Brisbane and Women’s Hospital was conducted for the period 1997e2009. Patients receiving palliative radiotherapy for bony metastases were identified and treatment sites were grouped into ‘spine’, ‘limb’, ‘multiple’ or ‘other’. Treatment courses were divided into single- or multiple-fraction treatments. The effects of socioeconomic and geographical factors on radiotherapy utilisation and fractionation were assessed. Results: In total, 5683 patients were identified in the cohort; they received a total of 8211 bone treatments. The overall proportion of single-fraction radiotherapy was 29%, with significant variation over the study period (P < 0.001). Age under 70 years and spine or multiple treatment sites were all associated with lower usage of single-fraction radiotherapy on multivariate analysis. Prostate and lung primary sites were associated with higher usage of single-fraction treatment. The proportion of single-fraction treatment remained low (35%), even for patients who survived less than 22 days from their last treatment. Socioeconomic and geographical factors had little effect on the number of fractions used. Conclusions: The rate of single-fraction radiotherapy for bone metastases has remained low in two large Australian institutions, despite considerable evidence that single-fraction treatment provides equivalent pain relief to fractionated therapy. This trend towards fractionated treatment was largely maintained, even in patients with limited life expectancy. Further measures to increase the rate of single-fraction therapy are needed. Crown Copyright Ó 2014 Published by Elsevier Ltd on behalf of The Royal College of Radiologists. All rights reserved.
Key words: Bone; fractionation; metastasis; palliative; radiotherapy
Introduction Radiotherapy is an important palliative treatment modality for patients with bony metastases. Previous work has shown that radiotherapy to bone sites accounted for over 16% of all radiotherapy treatment courses in several Australian centres [1]. Many randomised controlled trials [2e9] and two subsequent systematic reviews [10,11] have shown no Author for correspondence: A.N. Petrushevski, Department of Radiation Oncology, Liverpool Hospital, Elizabeth St, Liverpool, NSW 2170, Australia. Tel: þ61-2-9828-5189; Fax: þ61-2-9828-5189. E-mail address: [email protected] (A.N. Petrushevski).
significant difference between single-fraction radiotherapy (SFRT) (typically 8e10 Gy) and multiple-fraction courses (typically 20 Gy in five fractions or 30 Gy in 10 fractions) in terms of overall pain response and complete response. Furthermore, no significant difference between the treatment groups was seen in terms of toxicity, time to pain response or durability of analgesic effect. The most recent meta-analysis from Chow et al. [12] included 5000 patients from 16 randomised trials and confirmed the findings of the earlier data. SFRT is strongly supported over multiplefraction radiotherapy (MFRT) for uncomplicated bone metastases in most practice guidelines [13e15]. SFRT has several potential benefits for patients, their care givers and the health system. It offers significant logistical
http://dx.doi.org/10.1016/j.clon.2014.11.027 0936-6555/Crown Copyright Ó 2014 Published by Elsevier Ltd on behalf of The Royal College of Radiologists. All rights reserved.
206
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212
advantages, particularly in patients with limited life expectancy or geographical factors precluding fractionated treatment [16,17]. Previous studies have shown wide variety in the utilisation of palliative radiotherapy for bone metastases, with evidence that both patient and health care system-related factors affect patterns of care [16,18]. Practice surveys of clinicians have also revealed marked variance in practice and choice of fractionation regimen for various clinical scenarios [19e22]. The aims of our study were to assess the effect of various clinical and demographic factors on radiotherapy utilisation and fractionation for the palliation of bony metastases in Australia, and to assess whether fractionation patterns have changed over time with emerging evidence and practice guidelines.
Materials and Methods Sources of Data A retrospective review was conducted using routinely collected administrative data from two large Australian public cancer centres, the Liverpool and Macarthur Cancer Therapy Centres (LM, New South Wales, Australia) and the Royal Brisbane and Women’s Hospital (RBH, Queensland, Australia). The databases used for the analysis contained details of all patients undergoing radiotherapy treatment in these institutions between 1997 and 2009, including patient demographics, diagnosis date and primary cancer site. Treatment date, site and dose information were collected for each individual treatment. Variables extracted for analysis were standardised across both institutions to allow direct comparison. Study Population Patients aged 18 years and older with a histologically proven malignancy who had received palliative radiotherapy to a bone site at one of the two treatment institutions between 1997 and 2009 were included in the study. ICD-10 and morphology codes were used to identify the primary cancer diagnosis and morphology. Primary bone tumours were excluded from the analysis to exclude potentially curative treatments. Treatment sites were grouped into five categories: primary, bone, brain, soft tissue and multiple. Treatments assigned to the ‘multiple’ treatment code that consisted of at least one identified bone site were selected for the cohort, along with all ‘bone’ category treatments. Where radiotherapy dose was available, any treatment dose over 40 Gy was excluded, to avoid potentially curative treatments from being included in the analysis. Patients with cutaneous primaries and ambiguous treatment site information (for example, ‘thigh’ or ‘shoulder’) were also excluded to avoid any skin treatments being included. Due to difficulty ascertaining which malignancy a treatment course was assigned to, patients with multiple primary cancers were not included in the study.
Study Variables Patient variables extracted for analysis included age, gender, primary cancer diagnosis, diagnosis date and death date. Treatment year, treatment site and number of bone treatments were also determined. Treatment site was categorised into spine, limb, other or multiple (a combination of a bone and non-bone site treated simultaneously). Treatment start and end dates were used to classify treatment episodes as either single- or multiple-fraction treatments (i.e. if a treatment episode started and finished on the same day it was considered a single-fraction treatment). Subsequent bone treatments were also identified in patients selected for the study. If treatment start and end dates were more than 7 days apart, they were considered separate treatment episodes. Treatment end dates and patient death date were used to determine the time from the last bone treatment to death. The effect of these variables on fractionation regimen was assessed. Geographical data were collected on the patient cohort, using the postcode of the patient’s residential address. This was used to determine the Socioeconomic Index for Areas (SEIFA) score, which is determined for each postcode using Australian census data. This score was then used to divide the cohort into five quintiles based on the Index of Relative Socioeconomic Disadvantage (IRSD) [23]. The patient’s postcode data were used to determine geographical remoteness using the Accessibility/Remoteness Index of Australia (ARIA). The ARIA scores were grouped into five groups, ranging from highly accessible to very remote. Statistical Analysis Statistical Package for Social Science (SPSS) version 21 was used to carry out a data analysis [24]. Categorical variables were compared using Pearson chi-squared tests, and P values were determined. Logistic regression analysis was used to determine variables independently associated with fractionation choice. The patient’s street address or postcode and hospital address were geocoded to allow calculation of the travelling distance from the patient’s postcode to the nearest treatment facility using ArcGIS software [25]. Road distance was subsequently used to analyse whether travelling distance was associated with fractionation regimen.
Results Study Population Between 1997 and 2009, 5683 patients received at least one course of palliative bone radiotherapy, and among them 8211 individual bone treatment episodes were identified. Table 1 presents the baseline patient characteristics of the cohort. Fifty-nine per cent of the patients were men and the median age of the study population was 67 years, with a range of 18e97 years. Patients diagnosed with lung (30%),
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212 Table 1 Baseline patient characteristics of the study cohort Total number of Total number of individual patients treatment courses (%) (%) Gender Male Female Age 79 Primary site Breast Prostate Lung Colorectal Head and neck Upper gastrointestinal Gynaecological Melanoma Other genitourinary Other Treating institution RBH LM Year of treatment 1997e2000 2001e2004 2005e2009 Total
3367 (59%) 2316 (41%)
4763 (58%) 3448 (42%)
1806 1551 1624 702
(32%) (27%) (29%) (12%)
2819 2223 2245 924
(35%) (27%) (27%) (11%)
1110 1078 1685 283 159 132 100 186 386 564
(19%) (19%) (30%) (5%) (3%) (2%) (2%) (3%) (7%) (10%)
1834 1790 2198 358 216 161 135 239 530 750
(22%) (22%) (27%) (4%) (3%) (2%) (2%) (3%) (6%) (9%)
4416 (78%) 1267 (22%)
6384 (78%) 1827 (22%)
2330 (41%) 2046 (36%) 1307 (23%) 5683
3249 (40%) 2970 (36%) 1992 (24%) 8211
RBH, Royal Brisbane and Women’s Hospital; LM, Liverpool and Macarthur Cancer Therapy Centres.
prostate (19%) or breast (19%) cancers accounted for more than two thirds of the cohort. The overall mean number of bone treatments per patient was 1.45, with a higher mean number of treatments for prostate (1.66) and breast (1.65)
Fig 1. Frequency distribution of the number of bone treatment episodes given per patient between 1997 and 2009.
207
cancers. Patients under 60 years of age had a mean number of 1.56 bone treatments each, compared with 1.31 treatments for patients aged 80 years and over (P < 0.001). Figure 1 shows the number of bone radiotherapy treatment episodes per patient. Forty-five per cent of patients underwent only one bone treatment, 30% received a subsequent second bone treatment, with diminishing numbers for subsequent bone treatments, although a small number of patients received over five separate bone treatments during the study period. Demographic Factors and Number of Treatments Of the 5683 patients, a valid postcode was available for 5402 patients (who received 7818 treatments); this was used for the analysis of socioeconomic status and remoteness. Table 2 shows the effect of demographic factors on the mean number of treatment episodes per patient. Socioeconomic status (IRSD) had a small effect on the mean number of bone treatments, with patients living in higher socioeconomic areas being more likely to receive subsequent bone treatments (mean number of treatments 1.49 for IRSD 4/5 patients versus 1.40 for IRSD 1, P < 0.001). Similarly, patients living in highly accessible areas had a higher number of bone treatments compared with those living in very remote regions (1.49 versus 1.33 respectively, P < 0.001). Analysis of Fractionation The overall rate of SFRT was 29% for the cohort over the study period (Figure 2), with marked variance over time (P < 0.001). The rate of SFRT was lowest in the first year of the study (8%), but slowly increased each year before reaching 38% in 2003e2004. The rate of SFRT subsequently plateaued, averaging 31% for the period after 2004. 20 Gy in five fractions (47.2%) and 30 Gy in 10 fractions (11.4%) were the most common MFRT regimens (mean dose 21.5 Gy for LM versus 22.5 Gy for RBH). The patient and disease-related factors associated with the use of SFRT are summarised in Table 3. Men were more likely (P < 0.001) to receive single-fraction treatments (31%) than women (25%). Age was significantly associated with choice of fractionation (P < 0.001), with patients 80 years and older at the time of treatment (36%) being more likely to receive single fractions compared with those under 60 years of age (24%). Prostate (37%) and lung (31%) cancer patients had higher proportions of single-fraction treatments, whereas those with breast cancers (22%), melanoma (20%) or other genitourinary tumours (21%) were less likely to receive single fractions. Fractionation varied between the two institutions, with LM (36.7%) reporting a higher rate of SFRT than RBH (26.5%). The site of treatment was associated with marked variation in the use of SFRT (P < 0.001), with ‘limb’ (40%) and ‘other’ (41%) categories having higher rates of single-fraction treatment compared with ‘spine’ (25%) and ‘multiple’ sites (22%). For bone treatments given to patients who died during the study period (n ¼ 5254), the group who died
208
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212
Table 2 Effect of socioeconomic status and geographical remoteness on the mean number of bone treatments Patients
Socioeconomic status IRSD-1 (Lowest) 1180 IRSD-2 1015 IRSD-3 1063 IRSD-4 1074 IRSD-5 (highest) 1070 Geographical remoteness (ARIA) Highly accessible 3700 Moderately accessible 1340 Accessible 253 Remote 76 Very remote 33 Total 5402
Bone treatments
Mean number of bone treatments per patient
1651 1448 1530 1599 1590
1.40 1.43 1.44 1.49 1.49
5507 1824 338 105 44 7818
1.49 1.36 1.34 1.38 1.33 1.45
On multivariate logistic regression analysis (Table 4), age under 70 years (odds ratio 0.81) was independently associated with lower utilisation of SFRT. Patients who received their treatment at RBH were significantly less likely to receive SFRT (odds ratio 0.53). Spine and multiple treatment sites were also associated with lower SFRT usage (odds ratio 0.44 and 0.34, respectively). Patients who survived more than 21 days after their last bone treatment had significantly lower odds of receiving single-fraction treatment (odds ratio 0.70). Prostate (odds ratio 2.03) and lung (odds ratio 1.34) were the only primary sites significantly associated with a higher proportion of SFRT, with no primary sites significantly associated with lower SFRT use. Gender was not significantly associated with fractionation choice on multivariate analysis.
Discussion
IRSD, Index of Relative Socioeconomic Disadvantage; ARIA, Accessibility/Remoteness Index of Australia.
Proportion of bone treatments
within 21 days of their last bone treatment had a higher proportion of SFRT (35% versus 29%, P ¼ 0.006). However, MFRT was still the most common treatment in patients who died within 21 days of last treatment (65% of treatments). Socioeconomic status was not associated with fractionation regimen for those patients who received treatment (P ¼ 0.4). Although there was a small statistically significant trend towards patients living in remote and very remote areas receiving multiple fractions (80% in remote/very remote versus 70% in highly accessible areas, P ¼ 0.002), there was no significant difference between the groups when travelling distance to the treatment facility was examined (P ¼ 0.3).
Radiotherapy is an important palliative modality in the treatment of patients with bony metastases. This study is one of the largest studies assessing patterns of fractionation since the emergence of evidence-based guidelines for the palliation of bone metastases. It also covers a large time period, during which the outcome of several randomised trials were published, all underlining the equal effectiveness of single and multiple fractions on bone pain. Several population-based studies have previously documented significant variation in patterns of practice in the treatment of bone metastases worldwide. Elderly patients and those living in remote areas are particularly underserviced in terms of access to palliative radiotherapy [16,18,26]. There is also evidence that health financing and departmental culture may influence fractionation [20]. This variation in practice is supported by numerous practice surveys that have been published, indicating that
Year of treatment Single Fraction
Multiple Fraction
Fig 2. Proportion of single-fraction radiotherapy for bone metastases by year of treatment.
A.N. Petrushevski et al. / Clinical Oncology 27 (2015) 205e212 Table 3 Patient and tumour factors associated with fractionation regimen Number of Number of P value single-fraction multipletreatments fraction treatments Gender Male 1491 (31%) 3272 Female 872 (25%) 2576 Age 79 702 (36%) 924 Primary site Breast 412 (22%) 1422 Prostate 660 (37%) 1130 Lung 688 (31%) 1510 Colorectal 93 (26%) 265 Head and neck 56 (26%) 160 Upper gastrointestinal 49 (30%) 112 Gynaecological 30 (22%) 105 Melanoma 47 (20%) 192 Other genitourinary 113 (21%) 417 Other 215 (29%) 536 Treating institution RBH 1692 (27%) 4692 LM 671 (37%) 1156 Bone treatment site Spine 797 (25%) 2405 Limb 392 (40%) 587 Other 614 (41%) 872 Multiple 560 (22%) 1984 Time from last bone radiotherapy to death 0e7 days 90 (35%) 168 8e21 days 176 (34%) 336 >21 days 1295 (29%) 3189 Socioeconomic status (SEIFA) IRSD-1 475 (29%) 1176 IRSD-2 416 (29%) 1032 IRSD-3 406 (27%) 1124 IRSD-4 467 (29%) 1132 IRSD-5 465 (29%) 1125 Geographic remoteness (ARIA) Highly accessible 1634 (30%) 3873 Accessible 483 (26%) 1341 Moderately accessible 82 (24%) 256 Remote and very remote 30 (20%) 119 Travelling distance to treatment facility 100 km 945 (28%) 2486
(69%)
