- Email: [email protected]
A Population-Based Study of the Fractionation of Postlumpectomy Breast Radiation Therapy
International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Breast Cancer
A Population-Based Study of the Fractionation of Postlumpectomy Breast Radiation Therapy Allison Ashworth, MD, MSc, FRCPC,*,y Weidong Kong, MSc,* Timothy Whelan, BM, BCh, MSc, FRCPC,z and William J. Mackillop, MB, ChB, FRCPC, FRCR* *Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, Kingston, Ontario, Canada; y Cancer Center of Southeastern Ontario, Kingston, Ontario, Canada; and zJuravinski Cancer Center, Hamilton, Ontario, Canada Received Jul 31, 2012, and in revised form Nov 23, 2012. Accepted for publication Dec 18, 2012
Summary This retrospective cohort study was conducted by linking electronic treatment records to a populationbased cancer registry and included all patients who underwent lumpectomy for invasive breast cancer in Ontario between 1984 and 2008. It shows that the use of shorter schedules of postlumpectomy radiation therapy in Ontario increased after completion of the Ontario Clinical Oncology Group (OCOG) trial but wide intercenter variations in fractionation persisted after the OCOG trial was published in 2002.
Purpose: The optimal fractionation schedule of post lumpectomy radiation therapy remains controversial. The objective of this study was to describe the fractionation of postlumpectomy radiation therapy (RT) in Ontario, before and after the seminal Ontario Clinical Oncology Group (OCOG) trial, which showed the equivalence of 16- and 25-fraction schedules. Methods and Materials: This was a retrospective cohort study conducted by linking electronic treatment records to a population-based cancer registry. The study population included all patients who underwent lumpectomy for invasive breast cancer in Ontario, Canada, between 1984 and 2008. Results: Over the study period, 41,747 breast cancer patients received post lumpectomy radiation therapy to the breast only. Both 16- and 25-fraction schedules were commonly used throughout the study period. In the early 1980s, shorter fractionation schedules were used in >80% of cases. Between 1985 and 1995, the proportion of patients treated with shorter fractionation decreased to 48%. After completion of the OCOG trial, shorter fractionation schemes were once again widely adopted across Ontario, and are currently used in about 71% of cases; however, large intercenter variations in fractionation persisted. Conclusions: The use of shorter schedules of post lumpectomy RT in Ontario increased after completion of the OCOG trial, but the trial had a less normative effect on practice than expected. Ó 2013 Elsevier Inc.
Reprint requests to: Dr William J. Mackillop, Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, 10 Stuart St, Level 2, Kingston, ON K7L 3N6 Canada. Tel: (613) 5336895; E-mail: [email protected] Int J Radiation Oncol Biol Phys, Vol. 86, No. 1, pp. 51e57, 2013 0360-3016/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2012.12.015
Conflict of interest: none. Supplementary material www.redjournal.org.
for
this
article
can
be
found
at
52
Ashworth et al.
Introduction Breast conservation therapy, involving lumpectomy and postoperative radiation therapy (RT), is currently the treatment of choice for many women with early breast cancer. Postlumpectomy RT reduces the risk of local recurrence and the need for salvage mastectomy, and also reduces long-term breast cancer mortality (1-3). The effectiveness of post lumpectomy RT was initially demonstrated in several independent trials that used different fractionation schedules: NSABP-06 used 50 Gy/25 fractions (4), the Uppsala trial used 54 Gy/17 (5), and the Ontario trial used 40 Gy/16 (6). These fractionation schemes all achieved similar rates of local control and cosmetic outcomes (4-6). Optimal fractionation has therefore remained controversial, and fractionation has been shown to vary widely in routine practice (7-10). To address this controversy, the Ontario Clinical Oncology Group (OCOG) conducted a randomized trial comparing 50 Gy/25 with 42.5 Gy/16, after lumpectomy in women with T1, N0 or T2, N0 breast cancer (1). This showed that shorter fractionation achieved outcomes that were equivalent to those of longer fractionation in terms of local control and cosmesis at 10 years (11). The UK START trial has confirmed these findings (12). Not surprisingly, therefore, many clinical practice guidelines now recognize that both shorter and longer fractionation schemes may be medically appropriate (13, 14). Although the state of knowledge now appears to justify the routine use of the more convenient and less expensive, shorter fractionation schedules, it is not known to what extent this practice has been adopted. The objectives of the present study were to describe the fractionation of post lumpectomy RT in Ontario between 1984 and 2008, and to determine whether the publication of the results of the OCOG trial were associated with an increase in the use of the shorter fractionation schedules.
Methods and Materials Study design This was a retrospective cohort study. The study population included all women diagnosed with breast cancer in Ontario between January 1, 1984, and August 31, 2008, who underwent partial mastectomy followed by adjuvant postoperative RT to the breast only. Patients who received nodal irradiation were excluded. Electronic records of surgery and RT were linked to the Ontario Cancer Registry to describe patterns of fractionation in Ontario before and after publication of the landmark OCOG trial.
Data sources The Ontario Cancer Registry (OCR) is a passive, populationbased cancer registry operated by the provincial cancer agency, Cancer Care Ontario. The operation of the OCR has been described in detail elsewhere (15). The registry collects information about the demographic and clinical characteristics of all patients diagnosed with cancer in Ontario. The registry is reported to be 98% complete and >98% accurate, based on a capturee recapture analysis (16). Ontario’s population of 13.5 million is demographically quite representative of Canada’s overall
International Journal of Radiation Oncology Biology Physics population of 34.9 million (17). The age distribution of the Ontario population is very similar to that of the nation as a whole: Ontario has a higher ratio of urban to rural residents (Ontario 85:15; Canada 80:20), and Ontarians have a slightly higher than average median household income (Ontario $69,800; Canada $68,400). Ontario’s population includes a higher proportion of persons born outside of Canada (Ontario 28.3%; Canada 19.8%), and a lower proportion of aboriginal people (Ontario 2%; Canada 3.8%) (17). The incidence of breast cancer in Ontario women is similar to the Canadian average (Ontario 100/100,000; Canada 96/100,000), and breast cancer mortality in Ontario is also similar to that of Canada as a whole (Ontario 20/100,000; Canada 19/ 100,000) (18). The OCR provided records of all female patients diagnosed with breast cancer between 1984 and 2008, including the following: primary site (ICD 10), histology (ICD-O), date of diagnosis, date of birth, and place of residence (postal code). The OCR did not compile information about cancer stage over this period. The Canadian Institute of Health Information (CIHI) compiles electronic separation records from all Ontario hospitals. CIHI provided records of all surgical procedures, coded according to the Canadian Classification of Procedures, and these were linked to the OCR to identify lumpectomy cases. All RT in Ontario is provided by a network of publicly funded cancer centers. Each of these centers provided its complete electronic RT records including the following variables: total dose, number of fractions, treatment intent, dates of treatment, body region irradiated, and number of fields. RT records were complete to the end of 2009, allowing us to report on the use of postoperative RT within 1 year of diagnosis of cases diagnosed up to 2008. A “course” is defined as all of the fractions delivered sequentially to a specific anatomical area without an interruption of more than 30 days. Boost fractions were identified based on the body region treated, usually labeled as “boost,” or by identification of a second RT course prescribed to the same breast, consisting of either 4 or 5 fractions, initiated within 7 days of completion of the first course. Patients treated with palliative intent were excluded from the analysis.
Data analysis SAS, version 9.1, was used for all statistical analyses. Distance from the patient’s residence to the RT center was estimated by the linear distance between the geometric center of her postal code and the treatment center. Rates of use of shorter fractionation were compared using the Chi-square test. Inter center rate variations were quantified by their coefficient of variation. Multivariable Poisson regression was used to identify factors independently associated with the use of shorter fractionation.
Results Study population Figure 1 summarizes how the patients who underwent postoperative RT to the breast only after lumpectomy were selected from the total population of patients diagnosed with breast cancer in Ontario between 1984 and 2008. Table 1 (Column 1) shows the characteristics of the study population of 41,747 patients.
Volume 86 Number 1 2013
Fractionation of post lumpectomy breast RT
53
Fig. 1. Identification of the Study Population. Flow chart illustrating how patients treated with post lumpectomy radiation therapy to the breast only were identified.
Fractionation variations Figure 2 shows a frequency distribution of the total number of fractions per course, including boost fractions, for the entire population over the whole study period. A small proportion of patients (1.3%) who received <15 fractions delivered were deemed not to have completed treatment. These are not shown here and were excluded from the subsequent analyses. Schedules of 16 or 25 fractions were most frequently used. The smaller proportion of courses administered in 20, 21, or 30 fractions represent courses of 16 or 25 fractions to the whole breast, plus 4 to 5 boost fractions. As shown in Figure 3, there was substantial variation in fractionation among cancer centers. Longer fractionation schedules
were used more frequently at centers B and I, whereas shorter schedules were used more frequently by centers C, D, F, and H. Overall, 16.0% of patients received a boost, and this proportion also varied from center to center.
Factors affecting the use of shorter fractionation To explore factors affecting the choice of shorter or longer fractionation, cases were dichotomized into those patients who received less than or more than 21 fractions. Table 1 shows the results of a Poisson regression analysis in which we examined the effects of patient-related and health care system-related factors on the use of shorter fractionation. Older patients were much more likely to receive shorter fractionation, and this was statistically
54
International Journal of Radiation Oncology Biology Physics
Ashworth et al.
Table 1 Study population and factors associated with use of hypofractionated radiation therapy
Characteristic Age, y <50 50-59 60-69 70-79 80 Period 1984-1986 1987-1989 1990-1992 1993-1995 1996-1998 1999-2001 2002-2004 2005e2008 Center A B C D E F G H I Distance <10 km 10-50 km >50 km All cases combined
N
Percentage with 21 fractions
RR (95% CI)
10,172 11,307 11,458 7429 1381
52.7% 56.9% 61.4% 66.6% 78.8%
Reference 1.07 (1.05-1.10) 1.15 (1.13-1.17) 1.25 (1.22-1.27) 1.45 (1.40-1.50)
987 1979 4208 6095 7001 6394 6879 8204
79.5% 63.5% 60.9% 51.7% 43.1% 61.7% 65.5% 68.7%
0.90 (0.86-0.93) 0.83 (0.81-0.86) 0.81 (0.79-0.83) 0.69 (0.67-0.71) 0.60 (0.58-0.62) 0.95 (0.92-0.97) 0.96 (0.94-0.98) Reference
4582 9412 7261 2525 3968 9240 1310 1556 1893
45.4% 35.4% 68.7% 90.8% 46.6% 82.5% 56.6% 87.4% 31.5%
0.52 (0.50-0.54) 0.40 (0.39-0.41) 0.78 (0.77-0.80) 1.03 (1.00-1.05) 0.53 (0.51-0.55) Reference 0.62 (0.59-0.65) 0.99 (0.96-1.01) 0.35 (0.33-0.37)
14,658 16,385 10,692 41,747
61.5% 57.5% 60.1% 59.6%
Reference 1.01 (0.99-1.03) 1.05 (1.03-1.07) -
Abbreviations: CI Z confidence interval; RR Z risk ratio.
highly significant. Patients who lived farther from a cancer center were significantly more likely to be treated with shorter fractionation, but the magnitude of this association was small. As shown in Table 1, there was wide variation in the use of shorter fractionation among the cancer centers (range, 31.5%-90.8%), and these large differences remained highly significant after controlling for patient-related factors in the model. Multivariate analysis also showed that there was a highly significant decrease in the use of hypofractionation in the first decade of the study, followed by a significant increase in the second decade.
Fig. 2. Fractionation of post lumpectomy radiation therapy. Histogram showing the frequency distribution of the total number of fractions per course of post lumpectomy radiation therapy to the breast in Ontario between 1984 and 2008. Courses that included a boost are shown as pale gray bars, and other courses as dark gray bars. fractionation decreased to <50%. This decline occurred after the publication of the NSABP-06 trial in 1985 (4), the first large randomized trial to describe the efficacy of post lumpectomy breast irradiation using 50 Gy in 25 fractions. The use of shorter schedules continued to decline during the 1990s, reaching a nadir of 42% in 1996/1997. After the presentation of the OCOG trial results at the American Society of Clinical Oncology (ASCO) meeting in 2000, there was a rapid rise in the use of shorter schedules, from 45% in 1999 to 68% in 2000. The proportion of patients treated with shorter schedules stabilized at this level for the next 4 years, and did not increase further after the print publication of the OCOG trial in 2002. In 2004, the use of shorter schedules in Ontario declined again to 57% before returning to 68% in 2007. This increase coincided with the presentation of new data supporting the use of shorter fractionation schedules from the UK START trial at ASCO in June, 2007 (19) and the presentation of the long-term outcomes from the OCOG trial in 2007 at the San Antonio Breast Cancer Symposium (SABCS). (20). A supplementary figure (Fig. S1, available online) demonstrates that the temporal trend in fractionation varied among the cancer centers. At centers A, B, G, and I, a marked shift toward the use of shorter fractionation schedules occurred after first presentation of the OCOG trial in 2000. Center C adopted shorter schedules earlier in 1998. At centers D, F, and H, shorter fractionation schedules were used consistently before presentation of the results of the OCOG trial, and no further change was observed thereafter. At center E, the use of shorter schedules declined steadily after 1992.
Trends in fractionation in relation to changes in the state of knowledge
Contemporary variations in fractionation
To determine whether the publication of the results of the OCOG trial was associated with an increase in the use of shorter fractionation schedules, we examined the temporal changes in fractionation in Ontario over the study period in relation to the dates of the key clinical trials (Fig. 4). In the early 1980s, shorter fractionation schedules were used in >80% of cases, whereas during the 1990s the proportion of cases treated with shorter
Figure 5 shows that substantial intercenter variations in fractionation practices persisted in the 2006 to 2008 period, well after publication of the results of the OCOG trial. Centers A, B, C, D, F, G, and H used shorter schedules in the majority of cases, whereas centers E and I continued to use the longer schedules in more than 50% of cases. The magnitude of the intercenter variation in the mean number of fractions per case decreased slightly
Volume 86 Number 1 2013
Fractionation of post lumpectomy breast RT
55
Fig. 3. Intercenter variations in fractionation, 1984-2008. Histograms illustrating variations in the fractionation of post lumpectomy radiation therapy among the 9 cancer centers in Ontario over the entire study period. Courses that included a boost are shown as pale gray bars, and other courses as dark gray bars. after the results of the OCOG trial were reported, but the difference was not statistically significant. The coefficient of intercenter variation in the mean number of fractions was 0.15 (95% confidence interval [CI] Z 0.10-0.29) for the 1984-1999 period and 0.11 (95% CI Z 0.07-0.21) for the 2000 to 2008 period (PZ.35).
Discussion The main findings of this study are that there was an overall increase in the use of shorter fractionation in Ontario after the
Fig. 4. Trends in the use of shorter fractionation. Graph illustrating temporal changes in the proportion of patients treated with shorter fractionation schedules of post lumpectomy radiation therapy in Ontario between 1984 and 2008. Dates of key trials and publications are shown for reference.
OCOG trial, but large intercenter variations in fractionation persisted after the dissemination of its results. This study has a number of limitations. First, it was confined to a single Canadian province, and our findings may not be generalizable to other communities; the OCOG trial may well have had less influence on practice beyond where the study was conceived. Second, the available data were slightly out of date, and our results, therefore, do not necessarily reflect current practice. Third, and most importantly, this study provides no explanation for the wide intercenter variation in practice that persisted after the OCOG trial results were published; other approaches will be required to understand what factors influence the treatment choices of individual physicians. The temporal changes in fractionation observed in the decade after the OCOG trial proved to be quite complex. It is of note that the initial increase in the use of shorter fractionation that we observed coincided with first presentation of the results of the OCOG study, rather than its subsequent publication. Moreover, the later publication of a Canadian guideline, supporting the use of shorter fractionation, had no additional impact on practice. This failure of the guideline to modify practice is consistent with the findings of a recent study of temporal trends in the use of postmastectomy RT in the United States (21). After the initial increase in the use of shorter fraction after publication of the OCOG results, the proportion of patients treated with shorter fractionation schedules unexpectedly decreased. We speculate that this may have been due to recurring concerns about the potential for greater long-term toxicity with the use of larger fractions. The subsequent rebound in the use of shorter fractionation at the end of the study period may have been a response to the presentation of the reassuring long-term results of the OCOG trial in 2007 (20). It may also reflect the presentation of the early results of the UK START trial (19), which confirmed the OCOG findings and demonstrated the equivalence of shorter fractionation in a broader range of patients.
56
Ashworth et al.
International Journal of Radiation Oncology Biology Physics
Fig. 5. Contemporary intercenter variations in fractionation. Histograms illustrating the fractionation of post lumpectomy radiation therapy at 9 individual cancer centers in Ontario between 2006 and 2008. Courses that included a boost are shown as pale gray bars, and other courses as dark gray bars. One reason that shorter fractionation schemes have not been more uniformly adopted may be that the Canadian practice guideline did not offer any guidance as to which treatment choice is more appropriate in cases in which shorter and longer fractionation schemes are equally acceptable medically (13). Similarly, the recent excellent American Society for Radiation Oncology (ASTRO) guideline concludes unequivocally that shorter and longer fractionation schemes are “equally effective for in-breast tumor control and comparable in long-term side effects” for patients who meet 4 criteria (ie, age 50 years, pT1-T2, pN0, no chemotherapy, and minimum and maximum dose plus or minus 7% of the prescribed dose); but the authors do not go so far as to recommend that shorter fractionation be used in preference to longer fractionation in patients who meet these criteria (14). In the final paragraph of their report, the authors state that even in such patients, the longer fractionation schemes may still be preferred “on the basis of the abundance of long-term data” (14). This guideline may therefore be reassuring for those who already use shorter fractionation schemes, but it provides little incentive for others to adopt this approach. We would argue that when 2 therapeutic options have been demonstrated to be medically equivalent, the choice of the preferred option should be based on convenience and cost. Daily RT treatments over several weeks can be very inconvenient for many patients, particularly in Canada, where patients often live a long distance from a cancer treatment center. Shorter fractionation schemes would minimize the time that patients are away from their families, homes, and jobs and would also minimize the cost associated with travel, lodging, and time away from work (22). The use of shorter fractionation schedules also has the potential to make RT more widely available and increase rates of use (23). Although randomized trials remain the only certain way of establishing the efficacy of cancer treatments, population-based observational studies are also necessary to evaluate the impact of knowledge on practice, and to identify opportunities for improving
practice (24). We have previously shown that the publication of the results of clinical trials that have demonstrated improvements in survival was associated with widespread and rapid adoption of the more effective treatment in our cancer system in Ontario (24). In contrast, we not did find here a similarly complete adoption of the more efficient approach after the demonstration that the 2 different fractionation schemes provided medically equivalent outcomes.
Conclusion In conclusion, our results suggest that, even in Ontario, which played a key role in the evaluation of alternative fractionation schemes, the results of the OCOG trial did not result in uniform adoption of shorter fractionation. The fact that this trial had only a modest normative effect on practice indicates that the state of scientific knowledge is not the only thing that determines patterns of fractionation in the real world. Further studies will be required to ascertain what additional factors influence the decisions made by individual physicians about the use of shorter fractionation schemes in this context.
References 1. Whelan T, MacKenzie R, Julian J, et al. Randomized trial of breast irradiation schedules after lumpectomy for women with lymph nodenegative breast cancer. J Natl Cancer Inst 2002;94:1143-1150. 2. Fisher B, Anderson S, Bryant J, et al. Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N Engl J Med 2002;347:1233-1241. 3. Clarke M, Collins R, Darby S, et al. Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005;366:2087-2106.
Volume 86 Number 1 2013 4. Fisher B, Bauer M, Margolese R, et al. Five-year results of a randomized clinical trial comparing total mastectomy and segmental mastectomy with or without radiation in the treatment of breast cancer. N Engl J Med 1985;312:665-673. 5. Liljegren G, Holmberg L, Bergh J, et al. 10-Year results after sector resection with or without postoperative radiotherapy for stage I breast cancer: a randomized trial. J Clin Oncol 1999;17:2326-2333. 6. Clark RM, McCulloch PB, Levine MN, et al. Randomized clinical trial to assess the effectiveness of breast irradiation following lumpectomy and axillary disection for node-negative breast cancer. J Natl Cancer Inst 1992;84:683-689. 7. Whelan T, Marcellus D, Clark R, et al. Adjuvant radiotherapy for early breast cancer: patterns of practice in Ontario. Clin Oncol 1993;149: 1273-1277. 8. Priestman TJ, Bullimore JA, Godden TP, et al. The Royal College of Radiologists’ Fractionation Survey. Clin Oncol 1989;1:39-46. 9. Ceilley E, Jagsi R, Goldberg S, et al. Radiotherapy for invasive breast cancer in North America and Europe: results of a survey. Int J Radiat Oncol Biol Phys 2005;61:365-373. 10. Jagsi R, Makris A, Goldberg S, et al. Intra-European differences in the radiotherapeutic management of breast cancer: a survey study. Clin Oncol 2006;18:369-375. 11. Whelan TJ, Pignol JP, Levine MN, et al. Long-term results of hypofractionated radiation therapy for breast cancer. N Engl J Med 2010; 362:513-520. 12. The START Trialists’ Group, Bentzen SM, Agrawal RK, Aird EGA. The UK Standardisation of Breast Radiotherapy (START) Trial B of radiotherapy hypofractionation for treatment of early breast cancer: a randomised trial. Lancet 2008;371:1098-1107. 13. Whelan T, Olivotto I, Levine M. Clinical practice guidelines for the care and treatment of breast cancer: breast radiotherapy after breastconserving surgery (summary of the 2003 update). Can Med Assoc J 2003;168:437-439.
Fractionation of post lumpectomy breast RT
57
14. Smith BD, Bentzen SM, Correa CR, et al. Fractionation for whole breast irradiation: an American Society for Radiation Oncology (ASTRO) evidence-based guideline. Int J Radiat Oncol Biol Phys 2011;81:59-68. 15. Clarke EA, Marrett LD, Kreiger N. Cancer registration in Ontario: a computer approach. Toronto, Ontario, Canada: IARC Scientific Publications; 1991;95:246-257. 16. Robles SC, Marrett LD, Clarke EA, et al. An application of capturerecapture methods to the estimation of completeness of cancer registration. J Clin Epidemiol 1988;41:495-501. 17. Statistics Canada. Available at: www.statscan.gc.ca. Last date of access: July 15, 2012. 18. Canadian Cancer Society. Available at: www.cancer.ca. Date last accessed July 15, 2012. 19. Dewar JA, Haviland JS, Agrawal RK, et al. Hypofractionation for early breast cancer: first results of the UK Standardization of Breast Radiotherapy (START) trials abstract. J Clin Oncol 2007;25: LBA518. 20. Whelan TJ, Pignol JP, Julian J, et al. Long-term results of a randomized trial of accelerated hypofractionated whole breast irradiation following breast conserving surgery in women with node negative breast cancer. Breast Cancer Res Treat 2007;106:S6. 21. Shirvani SM, Pan I-W, Buchholz TA, et al. Impact of evidence-based clinical guidelines on the adoption of postmastectomy radiation in older women. Cancer 2011;117:4595-4605 (Abstract). 22. Shelley W, Brundage M, Hayter C, et al. A shorter fractionation schedule for postlumpectomy breast cancer patients. Int J Radiat Oncol Biol Phys 2000;47:1219-1228. 23. Dixon PF, Mackillop WJ. Could changes in clinical practice reduce waiting lists for radiotherapy? J Health Serv Res Policy 2001;6:70-77. 24. Booth CM, Mackillop WJ. Translating new medical therapies into societal benefit&the role of population-based outcome studies. JAMA 2008;300:2177-2179.
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Breast Cancer
A Population-Based Study of the Fractionation of Postlumpectomy Breast Radiation Therapy Allison Ashworth, MD, MSc, FRCPC,*,y Weidong Kong, MSc,* Timothy Whelan, BM, BCh, MSc, FRCPC,z and William J. Mackillop, MB, ChB, FRCPC, FRCR* *Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, Kingston, Ontario, Canada; y Cancer Center of Southeastern Ontario, Kingston, Ontario, Canada; and zJuravinski Cancer Center, Hamilton, Ontario, Canada Received Jul 31, 2012, and in revised form Nov 23, 2012. Accepted for publication Dec 18, 2012
Summary This retrospective cohort study was conducted by linking electronic treatment records to a populationbased cancer registry and included all patients who underwent lumpectomy for invasive breast cancer in Ontario between 1984 and 2008. It shows that the use of shorter schedules of postlumpectomy radiation therapy in Ontario increased after completion of the Ontario Clinical Oncology Group (OCOG) trial but wide intercenter variations in fractionation persisted after the OCOG trial was published in 2002.
Purpose: The optimal fractionation schedule of post lumpectomy radiation therapy remains controversial. The objective of this study was to describe the fractionation of postlumpectomy radiation therapy (RT) in Ontario, before and after the seminal Ontario Clinical Oncology Group (OCOG) trial, which showed the equivalence of 16- and 25-fraction schedules. Methods and Materials: This was a retrospective cohort study conducted by linking electronic treatment records to a population-based cancer registry. The study population included all patients who underwent lumpectomy for invasive breast cancer in Ontario, Canada, between 1984 and 2008. Results: Over the study period, 41,747 breast cancer patients received post lumpectomy radiation therapy to the breast only. Both 16- and 25-fraction schedules were commonly used throughout the study period. In the early 1980s, shorter fractionation schedules were used in >80% of cases. Between 1985 and 1995, the proportion of patients treated with shorter fractionation decreased to 48%. After completion of the OCOG trial, shorter fractionation schemes were once again widely adopted across Ontario, and are currently used in about 71% of cases; however, large intercenter variations in fractionation persisted. Conclusions: The use of shorter schedules of post lumpectomy RT in Ontario increased after completion of the OCOG trial, but the trial had a less normative effect on practice than expected. Ó 2013 Elsevier Inc.
Reprint requests to: Dr William J. Mackillop, Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, 10 Stuart St, Level 2, Kingston, ON K7L 3N6 Canada. Tel: (613) 5336895; E-mail: [email protected] Int J Radiation Oncol Biol Phys, Vol. 86, No. 1, pp. 51e57, 2013 0360-3016/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2012.12.015
Conflict of interest: none. Supplementary material www.redjournal.org.
for
this
article
can
be
found
at
52
Ashworth et al.
Introduction Breast conservation therapy, involving lumpectomy and postoperative radiation therapy (RT), is currently the treatment of choice for many women with early breast cancer. Postlumpectomy RT reduces the risk of local recurrence and the need for salvage mastectomy, and also reduces long-term breast cancer mortality (1-3). The effectiveness of post lumpectomy RT was initially demonstrated in several independent trials that used different fractionation schedules: NSABP-06 used 50 Gy/25 fractions (4), the Uppsala trial used 54 Gy/17 (5), and the Ontario trial used 40 Gy/16 (6). These fractionation schemes all achieved similar rates of local control and cosmetic outcomes (4-6). Optimal fractionation has therefore remained controversial, and fractionation has been shown to vary widely in routine practice (7-10). To address this controversy, the Ontario Clinical Oncology Group (OCOG) conducted a randomized trial comparing 50 Gy/25 with 42.5 Gy/16, after lumpectomy in women with T1, N0 or T2, N0 breast cancer (1). This showed that shorter fractionation achieved outcomes that were equivalent to those of longer fractionation in terms of local control and cosmesis at 10 years (11). The UK START trial has confirmed these findings (12). Not surprisingly, therefore, many clinical practice guidelines now recognize that both shorter and longer fractionation schemes may be medically appropriate (13, 14). Although the state of knowledge now appears to justify the routine use of the more convenient and less expensive, shorter fractionation schedules, it is not known to what extent this practice has been adopted. The objectives of the present study were to describe the fractionation of post lumpectomy RT in Ontario between 1984 and 2008, and to determine whether the publication of the results of the OCOG trial were associated with an increase in the use of the shorter fractionation schedules.
Methods and Materials Study design This was a retrospective cohort study. The study population included all women diagnosed with breast cancer in Ontario between January 1, 1984, and August 31, 2008, who underwent partial mastectomy followed by adjuvant postoperative RT to the breast only. Patients who received nodal irradiation were excluded. Electronic records of surgery and RT were linked to the Ontario Cancer Registry to describe patterns of fractionation in Ontario before and after publication of the landmark OCOG trial.
Data sources The Ontario Cancer Registry (OCR) is a passive, populationbased cancer registry operated by the provincial cancer agency, Cancer Care Ontario. The operation of the OCR has been described in detail elsewhere (15). The registry collects information about the demographic and clinical characteristics of all patients diagnosed with cancer in Ontario. The registry is reported to be 98% complete and >98% accurate, based on a capturee recapture analysis (16). Ontario’s population of 13.5 million is demographically quite representative of Canada’s overall
International Journal of Radiation Oncology Biology Physics population of 34.9 million (17). The age distribution of the Ontario population is very similar to that of the nation as a whole: Ontario has a higher ratio of urban to rural residents (Ontario 85:15; Canada 80:20), and Ontarians have a slightly higher than average median household income (Ontario $69,800; Canada $68,400). Ontario’s population includes a higher proportion of persons born outside of Canada (Ontario 28.3%; Canada 19.8%), and a lower proportion of aboriginal people (Ontario 2%; Canada 3.8%) (17). The incidence of breast cancer in Ontario women is similar to the Canadian average (Ontario 100/100,000; Canada 96/100,000), and breast cancer mortality in Ontario is also similar to that of Canada as a whole (Ontario 20/100,000; Canada 19/ 100,000) (18). The OCR provided records of all female patients diagnosed with breast cancer between 1984 and 2008, including the following: primary site (ICD 10), histology (ICD-O), date of diagnosis, date of birth, and place of residence (postal code). The OCR did not compile information about cancer stage over this period. The Canadian Institute of Health Information (CIHI) compiles electronic separation records from all Ontario hospitals. CIHI provided records of all surgical procedures, coded according to the Canadian Classification of Procedures, and these were linked to the OCR to identify lumpectomy cases. All RT in Ontario is provided by a network of publicly funded cancer centers. Each of these centers provided its complete electronic RT records including the following variables: total dose, number of fractions, treatment intent, dates of treatment, body region irradiated, and number of fields. RT records were complete to the end of 2009, allowing us to report on the use of postoperative RT within 1 year of diagnosis of cases diagnosed up to 2008. A “course” is defined as all of the fractions delivered sequentially to a specific anatomical area without an interruption of more than 30 days. Boost fractions were identified based on the body region treated, usually labeled as “boost,” or by identification of a second RT course prescribed to the same breast, consisting of either 4 or 5 fractions, initiated within 7 days of completion of the first course. Patients treated with palliative intent were excluded from the analysis.
Data analysis SAS, version 9.1, was used for all statistical analyses. Distance from the patient’s residence to the RT center was estimated by the linear distance between the geometric center of her postal code and the treatment center. Rates of use of shorter fractionation were compared using the Chi-square test. Inter center rate variations were quantified by their coefficient of variation. Multivariable Poisson regression was used to identify factors independently associated with the use of shorter fractionation.
Results Study population Figure 1 summarizes how the patients who underwent postoperative RT to the breast only after lumpectomy were selected from the total population of patients diagnosed with breast cancer in Ontario between 1984 and 2008. Table 1 (Column 1) shows the characteristics of the study population of 41,747 patients.
Volume 86 Number 1 2013
Fractionation of post lumpectomy breast RT
53
Fig. 1. Identification of the Study Population. Flow chart illustrating how patients treated with post lumpectomy radiation therapy to the breast only were identified.
Fractionation variations Figure 2 shows a frequency distribution of the total number of fractions per course, including boost fractions, for the entire population over the whole study period. A small proportion of patients (1.3%) who received <15 fractions delivered were deemed not to have completed treatment. These are not shown here and were excluded from the subsequent analyses. Schedules of 16 or 25 fractions were most frequently used. The smaller proportion of courses administered in 20, 21, or 30 fractions represent courses of 16 or 25 fractions to the whole breast, plus 4 to 5 boost fractions. As shown in Figure 3, there was substantial variation in fractionation among cancer centers. Longer fractionation schedules
were used more frequently at centers B and I, whereas shorter schedules were used more frequently by centers C, D, F, and H. Overall, 16.0% of patients received a boost, and this proportion also varied from center to center.
Factors affecting the use of shorter fractionation To explore factors affecting the choice of shorter or longer fractionation, cases were dichotomized into those patients who received less than or more than 21 fractions. Table 1 shows the results of a Poisson regression analysis in which we examined the effects of patient-related and health care system-related factors on the use of shorter fractionation. Older patients were much more likely to receive shorter fractionation, and this was statistically
54
International Journal of Radiation Oncology Biology Physics
Ashworth et al.
Table 1 Study population and factors associated with use of hypofractionated radiation therapy
Characteristic Age, y <50 50-59 60-69 70-79 80 Period 1984-1986 1987-1989 1990-1992 1993-1995 1996-1998 1999-2001 2002-2004 2005e2008 Center A B C D E F G H I Distance <10 km 10-50 km >50 km All cases combined
N
Percentage with 21 fractions
RR (95% CI)
10,172 11,307 11,458 7429 1381
52.7% 56.9% 61.4% 66.6% 78.8%
Reference 1.07 (1.05-1.10) 1.15 (1.13-1.17) 1.25 (1.22-1.27) 1.45 (1.40-1.50)
987 1979 4208 6095 7001 6394 6879 8204
79.5% 63.5% 60.9% 51.7% 43.1% 61.7% 65.5% 68.7%
0.90 (0.86-0.93) 0.83 (0.81-0.86) 0.81 (0.79-0.83) 0.69 (0.67-0.71) 0.60 (0.58-0.62) 0.95 (0.92-0.97) 0.96 (0.94-0.98) Reference
4582 9412 7261 2525 3968 9240 1310 1556 1893
45.4% 35.4% 68.7% 90.8% 46.6% 82.5% 56.6% 87.4% 31.5%
0.52 (0.50-0.54) 0.40 (0.39-0.41) 0.78 (0.77-0.80) 1.03 (1.00-1.05) 0.53 (0.51-0.55) Reference 0.62 (0.59-0.65) 0.99 (0.96-1.01) 0.35 (0.33-0.37)
14,658 16,385 10,692 41,747
61.5% 57.5% 60.1% 59.6%
Reference 1.01 (0.99-1.03) 1.05 (1.03-1.07) -
Abbreviations: CI Z confidence interval; RR Z risk ratio.
highly significant. Patients who lived farther from a cancer center were significantly more likely to be treated with shorter fractionation, but the magnitude of this association was small. As shown in Table 1, there was wide variation in the use of shorter fractionation among the cancer centers (range, 31.5%-90.8%), and these large differences remained highly significant after controlling for patient-related factors in the model. Multivariate analysis also showed that there was a highly significant decrease in the use of hypofractionation in the first decade of the study, followed by a significant increase in the second decade.
Fig. 2. Fractionation of post lumpectomy radiation therapy. Histogram showing the frequency distribution of the total number of fractions per course of post lumpectomy radiation therapy to the breast in Ontario between 1984 and 2008. Courses that included a boost are shown as pale gray bars, and other courses as dark gray bars. fractionation decreased to <50%. This decline occurred after the publication of the NSABP-06 trial in 1985 (4), the first large randomized trial to describe the efficacy of post lumpectomy breast irradiation using 50 Gy in 25 fractions. The use of shorter schedules continued to decline during the 1990s, reaching a nadir of 42% in 1996/1997. After the presentation of the OCOG trial results at the American Society of Clinical Oncology (ASCO) meeting in 2000, there was a rapid rise in the use of shorter schedules, from 45% in 1999 to 68% in 2000. The proportion of patients treated with shorter schedules stabilized at this level for the next 4 years, and did not increase further after the print publication of the OCOG trial in 2002. In 2004, the use of shorter schedules in Ontario declined again to 57% before returning to 68% in 2007. This increase coincided with the presentation of new data supporting the use of shorter fractionation schedules from the UK START trial at ASCO in June, 2007 (19) and the presentation of the long-term outcomes from the OCOG trial in 2007 at the San Antonio Breast Cancer Symposium (SABCS). (20). A supplementary figure (Fig. S1, available online) demonstrates that the temporal trend in fractionation varied among the cancer centers. At centers A, B, G, and I, a marked shift toward the use of shorter fractionation schedules occurred after first presentation of the OCOG trial in 2000. Center C adopted shorter schedules earlier in 1998. At centers D, F, and H, shorter fractionation schedules were used consistently before presentation of the results of the OCOG trial, and no further change was observed thereafter. At center E, the use of shorter schedules declined steadily after 1992.
Trends in fractionation in relation to changes in the state of knowledge
Contemporary variations in fractionation
To determine whether the publication of the results of the OCOG trial was associated with an increase in the use of shorter fractionation schedules, we examined the temporal changes in fractionation in Ontario over the study period in relation to the dates of the key clinical trials (Fig. 4). In the early 1980s, shorter fractionation schedules were used in >80% of cases, whereas during the 1990s the proportion of cases treated with shorter
Figure 5 shows that substantial intercenter variations in fractionation practices persisted in the 2006 to 2008 period, well after publication of the results of the OCOG trial. Centers A, B, C, D, F, G, and H used shorter schedules in the majority of cases, whereas centers E and I continued to use the longer schedules in more than 50% of cases. The magnitude of the intercenter variation in the mean number of fractions per case decreased slightly
Volume 86 Number 1 2013
Fractionation of post lumpectomy breast RT
55
Fig. 3. Intercenter variations in fractionation, 1984-2008. Histograms illustrating variations in the fractionation of post lumpectomy radiation therapy among the 9 cancer centers in Ontario over the entire study period. Courses that included a boost are shown as pale gray bars, and other courses as dark gray bars. after the results of the OCOG trial were reported, but the difference was not statistically significant. The coefficient of intercenter variation in the mean number of fractions was 0.15 (95% confidence interval [CI] Z 0.10-0.29) for the 1984-1999 period and 0.11 (95% CI Z 0.07-0.21) for the 2000 to 2008 period (PZ.35).
Discussion The main findings of this study are that there was an overall increase in the use of shorter fractionation in Ontario after the
Fig. 4. Trends in the use of shorter fractionation. Graph illustrating temporal changes in the proportion of patients treated with shorter fractionation schedules of post lumpectomy radiation therapy in Ontario between 1984 and 2008. Dates of key trials and publications are shown for reference.
OCOG trial, but large intercenter variations in fractionation persisted after the dissemination of its results. This study has a number of limitations. First, it was confined to a single Canadian province, and our findings may not be generalizable to other communities; the OCOG trial may well have had less influence on practice beyond where the study was conceived. Second, the available data were slightly out of date, and our results, therefore, do not necessarily reflect current practice. Third, and most importantly, this study provides no explanation for the wide intercenter variation in practice that persisted after the OCOG trial results were published; other approaches will be required to understand what factors influence the treatment choices of individual physicians. The temporal changes in fractionation observed in the decade after the OCOG trial proved to be quite complex. It is of note that the initial increase in the use of shorter fractionation that we observed coincided with first presentation of the results of the OCOG study, rather than its subsequent publication. Moreover, the later publication of a Canadian guideline, supporting the use of shorter fractionation, had no additional impact on practice. This failure of the guideline to modify practice is consistent with the findings of a recent study of temporal trends in the use of postmastectomy RT in the United States (21). After the initial increase in the use of shorter fraction after publication of the OCOG results, the proportion of patients treated with shorter fractionation schedules unexpectedly decreased. We speculate that this may have been due to recurring concerns about the potential for greater long-term toxicity with the use of larger fractions. The subsequent rebound in the use of shorter fractionation at the end of the study period may have been a response to the presentation of the reassuring long-term results of the OCOG trial in 2007 (20). It may also reflect the presentation of the early results of the UK START trial (19), which confirmed the OCOG findings and demonstrated the equivalence of shorter fractionation in a broader range of patients.
56
Ashworth et al.
International Journal of Radiation Oncology Biology Physics
Fig. 5. Contemporary intercenter variations in fractionation. Histograms illustrating the fractionation of post lumpectomy radiation therapy at 9 individual cancer centers in Ontario between 2006 and 2008. Courses that included a boost are shown as pale gray bars, and other courses as dark gray bars. One reason that shorter fractionation schemes have not been more uniformly adopted may be that the Canadian practice guideline did not offer any guidance as to which treatment choice is more appropriate in cases in which shorter and longer fractionation schemes are equally acceptable medically (13). Similarly, the recent excellent American Society for Radiation Oncology (ASTRO) guideline concludes unequivocally that shorter and longer fractionation schemes are “equally effective for in-breast tumor control and comparable in long-term side effects” for patients who meet 4 criteria (ie, age 50 years, pT1-T2, pN0, no chemotherapy, and minimum and maximum dose plus or minus 7% of the prescribed dose); but the authors do not go so far as to recommend that shorter fractionation be used in preference to longer fractionation in patients who meet these criteria (14). In the final paragraph of their report, the authors state that even in such patients, the longer fractionation schemes may still be preferred “on the basis of the abundance of long-term data” (14). This guideline may therefore be reassuring for those who already use shorter fractionation schemes, but it provides little incentive for others to adopt this approach. We would argue that when 2 therapeutic options have been demonstrated to be medically equivalent, the choice of the preferred option should be based on convenience and cost. Daily RT treatments over several weeks can be very inconvenient for many patients, particularly in Canada, where patients often live a long distance from a cancer treatment center. Shorter fractionation schemes would minimize the time that patients are away from their families, homes, and jobs and would also minimize the cost associated with travel, lodging, and time away from work (22). The use of shorter fractionation schedules also has the potential to make RT more widely available and increase rates of use (23). Although randomized trials remain the only certain way of establishing the efficacy of cancer treatments, population-based observational studies are also necessary to evaluate the impact of knowledge on practice, and to identify opportunities for improving
practice (24). We have previously shown that the publication of the results of clinical trials that have demonstrated improvements in survival was associated with widespread and rapid adoption of the more effective treatment in our cancer system in Ontario (24). In contrast, we not did find here a similarly complete adoption of the more efficient approach after the demonstration that the 2 different fractionation schemes provided medically equivalent outcomes.
Conclusion In conclusion, our results suggest that, even in Ontario, which played a key role in the evaluation of alternative fractionation schemes, the results of the OCOG trial did not result in uniform adoption of shorter fractionation. The fact that this trial had only a modest normative effect on practice indicates that the state of scientific knowledge is not the only thing that determines patterns of fractionation in the real world. Further studies will be required to ascertain what additional factors influence the decisions made by individual physicians about the use of shorter fractionation schemes in this context.
References 1. Whelan T, MacKenzie R, Julian J, et al. Randomized trial of breast irradiation schedules after lumpectomy for women with lymph nodenegative breast cancer. J Natl Cancer Inst 2002;94:1143-1150. 2. Fisher B, Anderson S, Bryant J, et al. Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N Engl J Med 2002;347:1233-1241. 3. Clarke M, Collins R, Darby S, et al. Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005;366:2087-2106.
Volume 86 Number 1 2013 4. Fisher B, Bauer M, Margolese R, et al. Five-year results of a randomized clinical trial comparing total mastectomy and segmental mastectomy with or without radiation in the treatment of breast cancer. N Engl J Med 1985;312:665-673. 5. Liljegren G, Holmberg L, Bergh J, et al. 10-Year results after sector resection with or without postoperative radiotherapy for stage I breast cancer: a randomized trial. J Clin Oncol 1999;17:2326-2333. 6. Clark RM, McCulloch PB, Levine MN, et al. Randomized clinical trial to assess the effectiveness of breast irradiation following lumpectomy and axillary disection for node-negative breast cancer. J Natl Cancer Inst 1992;84:683-689. 7. Whelan T, Marcellus D, Clark R, et al. Adjuvant radiotherapy for early breast cancer: patterns of practice in Ontario. Clin Oncol 1993;149: 1273-1277. 8. Priestman TJ, Bullimore JA, Godden TP, et al. The Royal College of Radiologists’ Fractionation Survey. Clin Oncol 1989;1:39-46. 9. Ceilley E, Jagsi R, Goldberg S, et al. Radiotherapy for invasive breast cancer in North America and Europe: results of a survey. Int J Radiat Oncol Biol Phys 2005;61:365-373. 10. Jagsi R, Makris A, Goldberg S, et al. Intra-European differences in the radiotherapeutic management of breast cancer: a survey study. Clin Oncol 2006;18:369-375. 11. Whelan TJ, Pignol JP, Levine MN, et al. Long-term results of hypofractionated radiation therapy for breast cancer. N Engl J Med 2010; 362:513-520. 12. The START Trialists’ Group, Bentzen SM, Agrawal RK, Aird EGA. The UK Standardisation of Breast Radiotherapy (START) Trial B of radiotherapy hypofractionation for treatment of early breast cancer: a randomised trial. Lancet 2008;371:1098-1107. 13. Whelan T, Olivotto I, Levine M. Clinical practice guidelines for the care and treatment of breast cancer: breast radiotherapy after breastconserving surgery (summary of the 2003 update). Can Med Assoc J 2003;168:437-439.
Fractionation of post lumpectomy breast RT
57
14. Smith BD, Bentzen SM, Correa CR, et al. Fractionation for whole breast irradiation: an American Society for Radiation Oncology (ASTRO) evidence-based guideline. Int J Radiat Oncol Biol Phys 2011;81:59-68. 15. Clarke EA, Marrett LD, Kreiger N. Cancer registration in Ontario: a computer approach. Toronto, Ontario, Canada: IARC Scientific Publications; 1991;95:246-257. 16. Robles SC, Marrett LD, Clarke EA, et al. An application of capturerecapture methods to the estimation of completeness of cancer registration. J Clin Epidemiol 1988;41:495-501. 17. Statistics Canada. Available at: www.statscan.gc.ca. Last date of access: July 15, 2012. 18. Canadian Cancer Society. Available at: www.cancer.ca. Date last accessed July 15, 2012. 19. Dewar JA, Haviland JS, Agrawal RK, et al. Hypofractionation for early breast cancer: first results of the UK Standardization of Breast Radiotherapy (START) trials abstract. J Clin Oncol 2007;25: LBA518. 20. Whelan TJ, Pignol JP, Julian J, et al. Long-term results of a randomized trial of accelerated hypofractionated whole breast irradiation following breast conserving surgery in women with node negative breast cancer. Breast Cancer Res Treat 2007;106:S6. 21. Shirvani SM, Pan I-W, Buchholz TA, et al. Impact of evidence-based clinical guidelines on the adoption of postmastectomy radiation in older women. Cancer 2011;117:4595-4605 (Abstract). 22. Shelley W, Brundage M, Hayter C, et al. A shorter fractionation schedule for postlumpectomy breast cancer patients. Int J Radiat Oncol Biol Phys 2000;47:1219-1228. 23. Dixon PF, Mackillop WJ. Could changes in clinical practice reduce waiting lists for radiotherapy? J Health Serv Res Policy 2001;6:70-77. 24. Booth CM, Mackillop WJ. Translating new medical therapies into societal benefit&the role of population-based outcome studies. JAMA 2008;300:2177-2179.