Words matter. Images matter. The Scientific Inquirer needs your support. Help us pay our contributors for their hard work. Visit our Patreon page and discover ways that you can make a difference. http://bit.ly/2jjiagi
An observational study including 3.9 million cancer cases in seven high-income countries between 1995-2014, published in The Lancet Oncology journal, finds that survival of seven cancers is generally improving, although the overall level and pace of improvement varies between countries and for each cancer type.
The authors say that the cancer stage at diagnosis, timely access to effective treatment, and co-occurring health conditions are likely the main determinants of patient outcomes, and future studies will be needed to assess the impact of these factors and better understand these international disparities.
Lead author Dr Melina Arnold of the International Agency for Research on Cancer (IARC) says: “As part of the SURVMARK-2 project and the second phase of the International Cancer Benchmarking Partnership (ICBP), we investigated incidence, one- and five-year survival, and mortality for seven cancers in seven countries over a 20-year period. While cancer survival and prognosis continue to improve across these high-income countries, the disparities we see are likely due to stage of the disease at diagnosis, the time it takes to get effective treatment and the effect of other concomitant health conditions.”
She continues: “The improvements in cancer survival observed are likely a direct consequence of healthcare reforms and technological advances that enable earlier diagnosis, more effective and tailored treatment and better patient management. Improvements in surgical techniques and new guidelines including preoperative radiotherapy as well as better diagnosis and scanning, enabling better staging of cancers and selection for targeted therapies, have all improved patient outcomes.” 
Over the last two decades, there have been consistent improvements in cancer survival in high-income countries, yet there continue to be differences between countries.
The new study, by the International Cancer Benchmarking Partnership (ICBP – an international, multidisciplinary partnership of clinicians, academics and policymakers), investigated these differences. This phase of the project used data from 3.9 million cases of oesophageal, stomach, colon, rectum, pancreatic, lung or ovarian cancer from population-based registries in seven countries with high-quality cancer registries, and universal health care (Australia, Canada, Denmark, Ireland, New Zealand, Norway, and the UK).
They calculated age-standardised survival at one and five years after diagnosis by site, age group, and period of diagnosis. They also mapped changes in incidence and mortality to changes in survival to assess progress in cancer control.
Results by cancer type (see Table 2 in paper for age-standardised five-year net survival by cancer site, country, and period of diagnosis):
* Between 2010-2014, five-year survival of oesophageal cancer was highest in Australia (23.5%) and lowest in Denmark (14.7%). The country that saw the greatest improvements in survival from 1995 to 2014 was Ireland (11 percentage point increase, from 10.9% in 1995-99 to 21.9% in 2010-14), while Canada saw the slowest improvement (2.8 percentage points, from 13.5% to 16.8%). Overall, survival improvements were more pronounced in patients aged less than 75 years, compared to those aged 75 or older.
* For stomach cancer, the UK had the lowest five-year survival (20.8%) while Australia had the highest (32.8%). New Zealand saw the lowest improvements in survival (3.2 percentage points improvement, from 21.3% in 1995-99 to 24.5% in 2010-14) while Ireland saw the greatest (11.1 percentage points, from 17.3% to 28.4%), notably in patients under 75 years old (where five-year survival increased from 19.7 to 32.8%).
* Colon cancer had relatively high five-year survival – with 70.8% of patients in Australia living for five years after diagnosis (highest), compared with 58.9% in the UK (lowest). Improvements in survival were more variable – ranging from 2.8 percentage points (New Zealand – increasing from 59.3% in 1995-99 to 62.1% in 2010-14) to 16.6 percentage points (Denmark – from 49.1% to 65.7%).
* The highest five-year survival was seen for rectal cancer, which also saw the largest improvements in between 1995-2014. Australia had the highest survival (70.8%), while the UK had the lowest (62.1%), and Denmark saw the greatest improvements in survival (21 percentage points improvement, from 48.1% to 69.1%) while New Zealand saw the lowest (9.1 percentage points improvement, from 56.3% to 65.4%).
* Pancreatic cancer had the lowest five-year survival of all – ranging from 7.9% in the UK (lowest) to 14.6% in Australia (highest). Progress in this cancer was largest in Australia, with survival increasing by 8.2 percentage points (from 6.4% in 1995-99 to 14.6% in 2010-14), but static in New Zealand (-0.6% percentage points – from 8.8% to 8.2% survival).
* For lung cancer, Canada had the highest five-year survival (21.7%) while the UK had the lowest (14.7%), and improvements in survival between 1995-2014 varied from 4-10.7 percentage points (4 percentage points improvement in New Zealand from 11.5% in 1995-99 to 15.5% in 2010-14, and 10.7 percentage points improvement in Denmark from 8.2% to 18.9%).
* Ovarian cancer survival was highest in Norway (46.2%) and lowest in Ireland (36%). New Zealand saw the lowest improvements over the 20-year period (4.4 percentage points, from 31.9% in 1995-99 to 36.3% in 2010-14), while Denmark saw the greatest improvements (10.1 percentage points – from 32% to 42.1%).
The authors note several limitations, including that differences in how cancer registries collect data may affect the results, although considerable effort was made to ensure the data were comparable.
The percentage of cases eligible for analysis varied by country and depended heavily on the amount of cases identified by death certificate only, which were excluded from the analysis. Data from Canada and Australia were not national, though covered most of their populations – 76% and 70% respectively – thus the datasets are reasonably representative.