The joint International Association of Cancer Registries and North American Association of Central Cancer Registries conference 2019, Vancouver

written by Cong Chen

The IACR is an organisation based in France which coordinates cancer registration practices and rules around the world, publishes outputs such as “Cancer Incidence in Five Continents” and supports the development in technological and social terms of upcoming registries around the world. Because it’s a global organisation the conference rotates around the five continents with cancer registries on an annual basis. This year it took place in Vancouver, and when it takes place in North America it is held jointly with the NAACCR, which has representatives from each central state registry of the US and Canadian provincial cancer registries. Next year it will take place in New Caledonia in Oceania.

The talks fell into many tracks, from higher-tech approaches taken by more established registries to enhance mature datasets to the more robust approaches needed to estimate statistics when data is less complete or available. The division into streams meant that each of the team could only see a small fraction of the talks, but selected slides should be made available on the conference website.

A visual highlight was a talk about presenting cancer data in the Australian Cancer Atlas and how this can help people to understand geographical variation. Advanced uses of linked treatment datasets, such as presentations about using cancer registration data to measure adherence to treatment guidelines in the Netherlands, using state-level registry data to study drug outcomes and using treatment outcomes to build clinical decision-making systems were connected to registries’ inclination to expand their purpose by capturing treatment and outcomes data.

I attended alongside colleagues from Public Health England’s National Cancer Registration and Analysis Service, presenting Health Data Insight’s Simulacrum and Brian Shand’s work on pseudonymisation and how this has helped bring in linked prescriptions and molecular data. (Our presentations can be seen at the end of this blog post).

The main topic of discussion at this conference was the use of Natural Language Processing in interpreting and automating ingestion of pathology and molecular testing text reports. Automated tools have also been used to identify potential cancer clusters by a number of registries or to identify correlation between variables for potential cancer risks, though the presentations indicated that there was still a challenge in moving from observed patterns in data to testable hypotheses.

Cancer recurrence was the hot topic closing the conference, with discussion on how to ascertain recurrent tumours without overburdening registry staff, appropriate tools and methodologies (Artificial Intelligence and otherwise) for ascertaining recurrence from routine treatment and outcomes data, and the need for appropriate follow-up in terms of treatment data to carry out those methodologies.

My colleague from Public Health England, Dr Luke Hounsome, also attended this conference. You can read his observations here.

Slideshares from this conference:

The Simulacrum, a Synthetic Cancer Dataset

This presentation describes the applications of synthetic data to cancer registries’s efforts to support understanding of and research based on cancer while re…

Pseudonymised Linkage of Cancer Registry Data in England

This presentation describes OpenPseudonymiser and extensions which allow flexible linkage of data without a trusted third party and without revealing membershi…