BACKGROUND: Monitoring the progress of cancer survival in a population over time is an important part of cancer surveillance. DATA AND METHODS: Data are from the Canadian Cancer Registry with mortality follow-up through record linkage to the Canadian Vital Statistics Death Database and tax files. Net survival (NS) was derived using the Pohar Perme method. Predicted estimates of NS for the period from 2012 to 2014 were calculated using the period method. Age-standardized and age-specific changes in five-year NS between the periods from 1992 to 1994 and 2012 to 2014 were determined for 30 individual cancers. RESULTS: Predicted five-year NS for 2012 to 2014 ranged from 98% for thyroid cancer to 7% for mesothelioma. Between 1992 to 1994 and 2012 to 2014, improvements in five-year age standardized NS were greatest for chronic myeloid leukemia (23.9 percentage points), though a large majority of the increase occurred in the first decade. Increases exceeding 15.0 percentage points were also observed for non-Hodgkin lymphoma (19.5), cancer of the small intestine (17.4) and multiple myeloma (16.9). In contrast, little to no improvement was observed for cancers of the anus, larynx, soft tissue or uterus, or for mesothelioma. Increases in five-year NS were greatest for chronic myeloid leukemia in each age group with the exception of those aged 75 to 84 years (thyroid). DISCUSSION: This study reveals important areas of progress in cancer outcomes in Canada since the early 1990s. It also sheds light on cancers for which there has seemingly been no improvement in five-year net survival over a 20-year period.