SEARCH

SEARCH BY CITATION

Keywords:

  • acute myeloid leukemia;
  • age;
  • Ontario Cancer Registry;
  • survival outcomes;
  • treatment patterns

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

BACKGROUND:

Acute myeloid leukemia (AML) is associated with a poor prognosis, particularly in older patients. To the authors' knowledge, few population-based studies of AML treatment patterns and outcomes exist to date.

METHODS:

The authors used the Ontario Cancer Registry to identify all patients diagnosed with AML between 1965 and 2003. Referral to specialized cancer centers (SCCs) and receipt of chemotherapy were examined as quality of care indicators. Survival outcomes were examined using logistic regression at 30 days, 1 year, and 3 years.

RESULTS:

A total of 9365 patients (mean age, 58.1 years; range, 0 to 103 years) developed AML between 1965 and 2003. Overall, 75.1%, 32.9%, and 17.3% of patients survived to 30 days, 1 year, and 3 years, respectively. Although survival improved over time among patients aged 19 to 59 years, similar improvements were not seen among older patients. The proportion of patients receiving chemotherapy declined with age (59.0% vs 29.3% among patients ages 19-59 vs ≥60 years). Fewer patients aged ≥60 years were referred to a SCC compared with younger patients (20.8% vs 29.9%). Younger age, less comorbidity, later year of diagnosis, receipt of chemotherapy, and being referred to a SCC were associated with better 30-day and long-term survival in multivariate models.

CONCLUSIONS:

Although the prognosis has improved over time among younger adults, it remains poor among those aged ≥60 years. Fewer older patients were referred to SCCs or treated with chemotherapy compared with younger patients, whereas both factors were associated with improved survival. Opportunities may exist to improve the quality of care and outcomes among older adults with AML. Cancer 2009. © 2009 American Cancer Society.

Acute myeloid leukemia (AML) is characterized by abnormal proliferation of immature myeloid cells with secondary hematopoietic insufficiency.1 AML incidence increases with age, with >60% of patients being aged ≥60 years at diagnosis.2 Current treatment of AML consists of intensive chemotherapy with an anthracycline and cytosine arabinoside.3

Several population-based studies have reported improvements in outcomes for younger people with AML within the last 30 years.4, 5 This is in contrast to older patients, generally referring to those aged ≥60 years, in whom the prognosis is poor and has remained virtually unchanged over the past several decades.4 Even in selected series, 3-year and 5-year survival rates among older adults are reported to be only 10%6 and 5% to 8%, respectively.7 The poor prognosis in older adults is because of a combination of factors, including a larger proportion of cases with adverse-risk cytogenetics and other deleterious genetic and epigenetic changes,8-12 lower rates of achieving complete remission with intensive chemotherapy,12 higher risk of disease recurrence after achieving remission,6, 13 greater comorbidity,14 inability to tolerate conventional allogeneic bone marrow transplantation regimens,15 and other adverse prognostic factors.12, 16 In part because of these aforementioned factors, older adults with AML receive chemotherapy less often than younger patients, and clinicians appear to be more pessimistic with respect to the management and outcomes for these patients than for their younger counterparts.17 In turn, lower levels of aggressive treatment of older adults with AML may compound prognostic differences associated with disease biology.

The median overall survival of older patients with AML was reported to be as low as 2 months in 1 population-based study.18 In contrast, several large randomized trials and case series from major institutions have reported 1-year survival rates of ≥50%.19-21 A likely reason for the discrepancy is referral bias of more medically fit and stable patients to specialized institutions.22 An alternate possibility is the higher rate of aggressive treatment and high-quality care of older patients with AML in specialized centers. However, to our knowledge, there are few high-quality, population-based studies of AML treatment patterns to address this issue. Menzin et al studied 2657 patients aged ≥65 years with AML using linked Surveillance, Epidemiology, and End Results (SEER)-Medicare data between 1991 and 1996.18 The investigators found that only 30% of patients received any chemotherapy within 2 years of diagnosis.18 In the only other community-based study of AML, by Xie et al, using SEER data from 1974 to 1993,4 the 5-year survival rate was 30%, 13%, and 3% for patients ages 20 to 44 years, 45 to 64 years, and ≥65 years, respectively, but receipt of chemotherapy was not examined. More current data are needed to better understand the burden of AML, particularly among older adults. In addition, AML is a relatively uncommon malignancy, and may benefit from specialized, regionalized care. Prior studies of AML have not reported on whether referral patterns to specialized cancer centers vary with age or other factors, or whether outcomes vary by treatment center. Such variations, if they exist, may represent an important area for study to identify ways to enhance the care of older adults with AML.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Study Overview

This was a retrospective analysis of provincial cancer registry data for all patients with newly diagnosed AML between 1965 and 2003. The study received ethics approval from the Research Ethics Board at the University Health Network. Because of the anonymized nature of the data, requirements for individual patient consent were waived.

Study Database

The patient cohort was assembled from the Ontario Cancer Registry (OCR). The OCR is the most comprehensive cancer database in Ontario. Data for the OCR are derived from 4 main sources. First, hospital discharge abstract data are received for all inpatient or same day surgery stays in every acute care hospital in Ontario for patients with cancer. Second, information is received on the diagnosis and management of patients seen at all 9 specialized cancer centers in Ontario. Third, the OCR collects patient information from death certificates and pathological reports from all Ontario hospitals. Nearly 90% of OCR cancer diagnoses are histologically verified. However, results of cytogenetic tests are not contained in the OCR. The OCR features well-defined algorithms and trained data coding personnel to ensure accurate capture rates.23 Finally, vital status information is obtained from the Registrar General's office.

Treatment information on chemotherapy is available from 2 sources, including procedure codes from hospital stays captured in discharge abstracts and from treatment at regional cancer centers. We followed the procedure of Menzin et al18 to define receipt of inpatient chemotherapy using specific diagnostic and procedure codes. This was combined with chemotherapy treatment indicators from specialized cancer centers. Although these indicators generally capture only intravenous chemotherapy, neither source allows reliable differentiation between curative and palliative forms of intravenous chemotherapy.

Patient Population

We included all patients with a first diagnosis of AML (defined using International Classification of Disease Ninth Edition codes 205.0, 206.0, or 207.0) between January 1, 1964 and December 31, 2003. Sociodemographic information, all hospital separations, treatment information, and vital status information were obtained for each patient. Comorbidity was measured using the Charlson-Deyo Index.24, 25

Statistical Analyses

We used simple descriptive statistics to analyze the number of patients diagnosed in each 5-year period from 1965 to 2003. Analyses for outcomes, including survival, referral to a specialized cancer center, and receipt of chemotherapy, were assessed using multivariate logistic regression models. These models evaluated several possible covariates, including age, sex, socioeconomic status (neighborhood income quartile), comorbidity, health planning, referral to a specialized cancer center, and receipt of chemotherapy as appropriate. There are 8 health planning regions across the province. Each 1 includes at least 1 specialized cancer center, but regions vary in population density and geographic size. Region A had both the highest overall population and the highest population density. Regions B, C, G, and H had moderate population sizes and were a mix of primarily urban and suburban regions. Regions D and E had relatively small populations and large geographic sizes, consisting of mostly rural areas. Region F was a mix of urban and suburban, but was the second smallest in terms of population size. Analyses of referral patterns were restricted to patients aged ≥19 years, because children with acute leukemia are treated in specialized pediatric tertiary care hospitals.

We examined 30-day, 1-year, and 3-year survival. Thirty-day survival reflects both disease severity and treatment-related toxicity. One-year and 3-year survival are common benchmarks in AML trials. Potential predictor variables were assessed in univariate logistic regression models and included in multivariate models regardless of P value, unless there were too few events to allow for model stability. Goodness of fit for all logistic models was assessed with the Hosmer-Lemeshow statistic,26 and model discrimination was evaluated with the c-statistic. We repeated these analyses for patients diagnosed in 1995 onward, as the results are more reflective of the current treatment era. Similar to our analyses of referral patterns, all regression analyses of survival were restricted to patients aged ≥19 years. However, we report raw estimates of 30-day, 1-year, and 3-year survival for patients aged 0 to 18 years.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Baseline Characteristics

A total of 9613 patients were diagnosed with AML during the study period. The mean age at diagnosis was 58.2 years (median, 64 years; range, 0 to 103 years), with 53.5% male. Of the entire cohort, 6.9% were ages 0 to 18 years, 35.5% were ages 19 to 59 years, and 57.5% were aged ≥60 years at diagnosis.

There was a steady increase in the number of newly diagnosed cases per 5-year interval, from 491 cases between 1965 and 1970 to 1814 cases between 1995 and 2000, predominantly among the patients aged ≥60 years (Fig. 1). Mean age at diagnosis increased steadily over time, from 49.1 years in 1965 to 62.5 years in 2003 (P < .001 for linear trend).

thumbnail image

Figure 1. Incident cases of acute myeloid leukemia by age group and 5-year period are shown.

Download figure to PowerPoint

The level of comorbidity increased with age group, with a mean Charlson score of 0.27, 0.43, and 0.72 for patients aged 0 to 18 years, 19 to 59 years, and ≥60 years, respectively (P < .001). Similarly, the proportion of patients with prior cancer increased from 9.5% for patients aged 0 to 18 years to 11.2% and 15.8% for patients aged 19 to 59 years and ≥60 years, respectively (P < .001).

Referral Patterns and Receipt of Chemotherapy

Overall, 39.2% of patients aged ≥19 years were referred to a specialized cancer center. This proportion has changed with time and differs by age group. Over time, the proportion of adult patients referred to a specialized cancer center has increased from 27.9% in 1965 through 1970 to 52.2% for patients diagnosed in 2001 through 2003. This increase was primarily seen in patients ages 19 to 59 years (data not shown). There was a strong negative relation between increasing age and likelihood of referral to a specialized cancer center, whether age was analyzed as a dichotomous variable (age 19-59 years vs age ≥60 years; P < .001) or by decade (P < .001). This is illustrated in Figure 2. There was a weak relation noted between comorbidity and referral to a specialized cancer center, such that patients with a Charlson score of 1, 2, 3, or higher had an odds ratio of referral of 0.60 (95% confidence interval [95% CI], 0.51-0.72), 0.92 (95% CI, 0.81-1.05), and 0.82 (95% CI, 0.68-0.99) compared with patients with a Charlson score of 0. There was no relation noted between either sex (P = .74) or income quartile (P = .30) and referral to a specialized cancer center.

thumbnail image

Figure 2. Proportion of patients referred to a regional cancer center by age group is shown. Y indicates yes; N,no.

Download figure to PowerPoint

Overall, 41.7% of patients were treated with chemotherapy. Fewer patients aged ≥60 years received chemotherapy compared with patients ages 19 to 59 years (29.3% vs 59.0%; P < .001). When age was considered per decade, the proportion of patients receiving chemotherapy decreased, particularly among those aged ≥60 years compared with younger patients (Table 1). There was no relation noted between sex and receipt of chemotherapy (P = .24). There were significant differences in the proportion of patients receiving chemotherapy by geographic region, ranging from 33.5% to 47.4% (Table 1). Patients referred to a specialized cancer center were significantly more likely to receive chemotherapy (62.0% vs 28.7%; P < .001), even after adjustment for age. Specifically, among patients ages 19 to 59 years, 73.0% of those referred to a specialized cancer center received chemotherapy, as opposed to 43.0% who were not referred. Among patients aged ≥60 years, the proportions were 49.5% versus 20.5%, respectively. Over time, there was an increase in the proportion of patients treated with chemotherapy, among both younger and older patients, reaching a plateau around 1995 for younger patients and a decade earlier for patients aged ≥60 years. This is illustrated in Figure 3. When patients were restricted to those diagnosed in 1995 onward, the above differences persisted, although absolute proportions treated with chemotherapy increased in all categories (Table 1).

thumbnail image

Figure 3. Proportion of patients who received chemotherapy by age group over time is shown.

Download figure to PowerPoint

Table 1. Receipt of Chemotherapy Among Strata of Specific Variables
Variable% Treated (1965-2003)% Treated (1995-2003)
Age, y  
 19-2955.5%87.7%
 30-3961.6%83.3%
 40-4960.4%86.4%
 50-5958.0%80.5%
 60-6945.0%61.1%
 70-7927.5%34.2%
 80-899.3%10.2%
 ≥900.7%0.0%
Sex  
 Men41.2%55.4%
 Women42.4%55.4%
Charlson comorbidity score
 042.5%62.0%
 140.6%41.1%
 238.3%49.8%
 341.2%41.4%
Seen at specialized cancer center
 Yes62.0%77.6%
 No28.7%38.3%
Prior cancer  
 Yes39.7%51.8%
 No42.0%56.0%
Region of Ontario  
 Region A43.2%57.1%
 Region B45.0%56.6%
 Region C34.2%40.9%
 Region D40.3%52.8%
 Region E33.5%48.6%
 Region F47.4%67.7%
 Region G40.9%54.3%
 Region H40.2%57.7%
Year of diagnosis  
 1965-19700.4% 
 1971-19751.0% 
 1976-198024.6% 
 1981-198546.7% 
 1986-199043.9% 
 1991-199555.5% 
 1996-200056.7%56.7%
 2001-200352.2%52.2%

Survival

For the cohort as a whole, 30-day, 1-year, and 3-year survival rates were 75.1%, 32.9%, and 17.3%, respectively. Across age groups, survival decreased with increasing age. Among patients aged 0 to 18 years, 19 to 59 years, and ≥60 years, the 30-day survival was 90.1%, 84.9%, and 67.2%, respectively. One-year survival was 58.0%, 48.8%, and 20.1%, respectively, whereas 3-year survival was 35.8%, 28.0%, and 8.4%, respectively. Over time, there was a steady increase in 30-day and 1-year survival, but this improvement was noted exclusively in younger patients with no improvement in survival observed among patients aged ≥60 years (Fig. 4).

thumbnail image

Figure 4. (A) Thirty-day survival by 5-year time period is shown. (B) One-year survival by 5-year time period is shown.

Download figure to PowerPoint

Analyses of predictors of survival were restricted to patients aged ≥19 years. On univariate analyses, predictors of 30-day mortality included age (P < .001), Charlson score (P < .001), referral to a specialized cancer center (P < .001), receiving chemotherapy (P < .001), and geographic region (P < .001). On multivariate logistic regression analysis, significant predictors of 30-day mortality included age, Charlson score, prior cancer, referral to a specialized cancer center, receipt of chemotherapy, and geographic region (Table 2).

Table 2. Multivariate Logistic Regression Models of Survival Among Patients Aged 19 Years and Older With Acute Myeloid Leukemia
Predictor30-Day Survival1-Year Survival3-Year Survival
  • *

    A Charlson comorbidity score of 0 is the referent.

  • Region A is the referent.

Age (per 10 y)0.83 (0.80-0.86)0.70 (0.68-0.72)0.65 (0.63-0.68)
Sex0.98 (0.89-1.09)0.91 (0.82-1.00)0.89 (0.79-1.01)
Diagnosis year1.00 (0.99-1.01)1.01 (1.01-1.02)1.05 (1.04-1.06)
Charlson comorbidity score*   
 10.49 (0.41-0.59)0.49 (0.39-0.61)0.50 (0.38-0.67)
 20.40 (0.31-0.52)0.58 (0.43-0.79)0.63 (0.43-0.92)
 30.35 (0.26-0.46)0.38 (0.26-0.54)0.32 (0.20-0.51)
Prior cancer2.32 (1.78-3.01)1.28 (0.93-1.76)1.23 (0.81-1.85)
Referral to a specialized cancer center (yes)2.35 (2.07-2.67)1.66 (1.49-1.85)1.29 (1.13-1.48)
Receipt of chemotherapy (yes)3.37 (2.94-3.86)2.31 (2.06-2.59)1.47 (1.27-1.70)
Geographic region   
 B (central west)0.86 (0.74-1.01)0.99 (0.85-1.16)0.99 (0.82-1.20)
 C (east)0.80 (0.67-0.95)0.90 (0.74-1.08)0.86 (0.68-1.08)
 D (northeast)1.25 (0.98-1.58)1.01 (0.81-1.25)1.27 (1.00-1.63)
 E (northwest)0.87 (0.62-1.24)1.34 (0.97-1.85)1.99 (1.40-2.85)
 F (south)0.60 (0.45-0.79)0.89 (0.68-1.15)0.82 (0.58-1.16)
 G (southeast)0.84 (0.69-1.04)0.89 (0.72-1.09)0.79 (0.60-1.04)
 H (southwest)0.74 (0.63-0.86)0.90 (0.77-1.05)0.91 (0.75-1.09)
Model c-statistic0.750.770.77

Univariate predictors of 1-year survival included age (P < .001), sex (P = .011), year of diagnosis (P < .001), Charlson score (P < .001), prior cancer (P < .001), referral to a specialized cancer center (P < .001), and receiving chemotherapy (P < .001). On multivariate logistic regression analysis, significant predictors of 1-year mortality included age, year of diagnosis, Charlson score, referral to a specialized cancer center, and receipt of chemotherapy (Table 2).

With respect to 3-year survival, univariate predictors included age (P < .001), sex (P = .018), year of diagnosis (P < .001), Charlson score (P < .001), prior cancer (P < .001), referral to a specialized cancer center (P < .001), receipt of chemotherapy (P < .001), and geographic region (P < .001). Significant predictors of 3-year survival in multivariate logistic regression were age (P < .001), year of diagnosis (P < .001), Charlson score (P < .001), referral to a specialized cancer center (P < .001), receipt of chemotherapy (P < .001), and geographic region (P < .001) (Table 2). Model fit for all 3 multivariate models was good, with c-statistics ranging from 0.75 to 0.77 (Table 2).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

We examined patterns of care of AML using population-based data spanning nearly 40 years. Our findings can be summarized as follows. First, an increasing proportion of patients with AML have been treated with chemotherapy over time, particularly among people ages 19 to 59 years. Second, several factors are associated with referral to a specialized cancer center and receipt of chemotherapy, including age, comorbidity, and geographic region, but not sex. Third, outcomes have been improving over time, mostly among younger patients. Consistent predictors of better survival at 30 days, 1 year, and 3 years include younger age, a later year of diagnosis, less comorbidity, referral to a specialized cancer center, and receipt of chemotherapy.

Before discussing these findings in detail, it is important to recognize several important limitations to our analyses. First, the OCR does not contain information regarding cytogenetics or other disease markers, nor does it contain details regarding chemotherapy agent or dose. As such, we can only make high-level comments about referral patterns, treatment patterns, and outcomes. Second, administrative data do not include information concerning patient preferences. Many older patients, when faced with the poor prognosis and significant toxicity of intensive chemotherapy, may choose best supportive care. Conversely, we cannot determine what proportion of patients who did not receive chemotherapy was offered this option by the treating physician. This is clearly an important issue. Third, receipt of chemotherapy is likely to be undercoded in our database, such that our observed rates may be lower than in practice. In addition, we could not separately capture receipt of transplantation, either bone marrow or peripheral stem cell. However, undercoding of chemotherapy is not likely to change with age, meaning that our observed age-based patterns of care are likely to be robust. In addition, our results are remarkably similar to studies using SEER-Medicare data.18, 27 Moreover, our results for receipt of chemotherapy are also similar to Luke et al, who analyzed treatment and survival patterns associated with myeloid leukemia in South Australia using hospital registry data.5 These results provide some external validity for our findings. That being said, given the limitations of the data, the current study should be viewed as a first step to understanding patterns of care among patients with AML. Regardless of concerns regarding undercoding of chemotherapy, our survival data are robust and provide relatively current, population-based survival data for this disease that complement data from trials and cohort studies.

The current study also features several strengths. First, it is a large, population-based analysis. Second, the data are fairly current, and span nearly 40 years, providing an almost-unparalleled opportunity to understand AML over time. Third, there is universal healthcare insurance in Ontario, which minimizes any economic impact on diagnosis, referral, or treatment patterns.

With respect to our main findings, we noted an increase in the proportion of patients treated with chemotherapy over time, particularly among younger patients. In an updated analysis of data originally published by Menzin et al,18 Lang et al found similar results among 3439 older patients with AML using SEER-Medicare data, with 29% of patients diagnosed in 1991 treated with chemotherapy compared with 38% diagnosed in 1999.27 Luke et al reported an increase from 57.5% of patients of any age with AML treated with chemotherapy between 1977 and 1986 to 85.8% for patients diagnosed between 1995 and 2002.5 However, these investigators used a broad definition of chemotherapy, including agents such as interferon and imatinib. By using a narrower definition of chemotherapy, we found relatively little change in the use of chemotherapy among patients aged ≥60 years between 1985 and 2003. To our knowledge, there are no other population-based analyses of treatment of AML.

The current study is the first, to our knowledge, to analyze referral patterns to specialized cancer centers among patients with AML. Similar to Lang et al27 and Luke et al,5 we observed a decreasing proportion of cases receiving treatment with increasing age. This was seen regardless of year of diagnosis, and differences in the proportion of patients treated with chemotherapy persisted regardless of whether patients were referred to a specialized cancer center. However, referral to a specialized cancer center was associated with higher proportions of patients treated with chemotherapy regardless of age.

Perhaps the most provocative of our findings relate to treatment outcomes and the role of a specialized cancer center. Consistent with prior literature, we demonstrated that older people with AML had a poorer short-term and long-term prognosis compared with younger people.4, 5, 27 Comorbidity was consistently associated with a poorer prognosis independent of age, receipt of treatment, and other factors. The importance of comorbidity on survival outcomes has been well recognized among other malignancies,28, 29 and recently has been described among patients with AML.30 In addition, receipt of chemotherapy was associated with improved short-term and long-term survival. Two factors may explain this observation. First, intensive chemotherapy is clearly associated with improved survival in younger people with AML,31 and is likely associated with improved outcomes among people aged ≥60 years,19 although the latter point remains controversial, particularly among patients aged ≥70 years or those with unfavorable cytogenetics.32, 33 Receipt of chemotherapy also included patients who underwent allogeneic transplantation, which is associated with improved survival in several subgroups of patients.31 Second, selection bias may explain a component of the improved survival among patients treated with chemotherapy. This is because patients who are more fit and have more favorable disease characteristics are more likely to be treated. How much of a role each of these 2 factors is playing in explaining the improved survival cannot be determined from our data.

We also found referral to a specialized cancer center was associated with improved short-term and long-term outcomes, after adjusting for age, sex, comorbidity, year of diagnosis, and receipt of chemotherapy. One possible interpretation of this finding is that AML is a relatively uncommon disease that requires specialized assessment and treatment, similar to other complex malignant conditions for which regionalization of surgical care has been recommended.34-36 A contributing factor is access to allogeneic transplantation, which is associated with improved survival and is only available in specialized cancer centers. Conversely, more healthy patients with better cytogenetics or other disease characteristics that make them more likely to benefit from chemotherapy may be more likely to be referred to specialized cancer centers. This selection bias may explain the better outcomes seen among patients at these centers. The effect of referral to a specialized cancer center, with odds ratios for improved survival ranging from 1.29 to 2.35, are of potential clinical importance. At the same time, effects of referral appear to be attenuated over time, with less pronounced effects on 3-year survival compared with 30-day survival (Table 2). In addition, we were only able to partially adjust for differences in patient characteristics and treatment by considering comorbidity, a prior history of cancer, geographic region, and receipt of chemotherapy in our models. As such, our data must be viewed as hypothesis-generating, and further analyses involving more detailed clinical and treatment-related information are required to understand these factors.

We found that geographic region was generally not a significant predictor of mortality, particularly beyond 30-day mortality. Geographic region was important in several univariate models, but effects were attenuated in multivariate models, suggesting that much of the impact of geographic region is because of a combination of other variables, such as age, referral to a specialized cancer center, and receipt of chemotherapy. Living in more rural areas does not appear to be a major factor in determining outcome after a diagnosis of AML, provided patients have access to appropriate expertise.

In summary, the number of diagnosed cases of AML has been rising in the past 40 years, primarily among patients aged ≥60 years. Chemotherapy is being increasingly used in the treatment of AML, but utilization remains lower among those aged ≥60 years. Survival has been improving over time, particularly among younger patients, and important predictors of better survival include younger age, a more recent year of diagnosis, less comorbidity, referral to a specialized cancer center, and receipt of chemotherapy.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

We thank Cancer Care Ontario and the Ontario Cancer Registry for preparing the study file.

Conflict of Interest Disclosures

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Dr. Shabbir M. H. Alibhai is a Research Scientist of the Canadian Cancer Society Research Institute.

References

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References