FLORENCE were represented with a poster at the annual symposium of the Association of the Nordic Cancer Registries (ANCR) in Bodø, Norway, 28th -30th of August 2024.

The object of the Association is to provide a continuing, organized framework for the activities of the Nordic cancer registries to facilitate exchange of both scientific and technical information, to contribute to the uniformity of definitions used by the registries, and to facilitate the organisation of inter-Nordic studies and studies proposed by international organisations and various groups of individuals.

Senior Advisor at the Cancer Registry of Norway, Kristin Oterholt Knudsen summarized the conference with three points:

  • Increasing focus on cancer prevention, where cancer registries can play a central role
  • Increasing focus on sharing data across the Nordic countries and how we can utilize the useful data
  • Bodø is a very nice city

You can find Kristins great poster bellow or read the abstract at the bottom of this article.

Federated Learning Using OMOP Modelling of Health Data for Elevating Colorectal Cancer Care in the Nordic Countries (FLORENCE)

Kristin Oterholt Knudsen3 , Siri Larønningen3 , Espen Enerly1 , Tina Sture1 , Vegar Johansen Dagenborg4 , Severin Elvatun2 , Narasimha Raghavan2 , Svein Erling Tysvær2 , Gintaras Pikelis2 , and Jan F. Nygård2

1 Department of Research, Cancer Registry of Norway, Norwegian Institute of Public Health, Oslo, Norway.

2 Department of Registry Informatics, Cancer Registry of Norway, Norwegian Institute of Public Health, Oslo, Norway.

3 Department of Registration, Cancer Registry of Norway, Norwegian Institute of Public Health, Oslo, Norway.

4 Department of Surgical Oncology, Radium Hospital, Oslo University Hospital, Oslo, Norway

 

Introduction

Colorectal cancer is the second most common cancer in Norway, considering both genders. Surgical removal of the tumor is considered curative for most patients. One in four colorectal cancer patients (CRC) experience post-operative complications, with an equal proportion faced with cancer recurrence. Both with implications for quality-of-life, survival and health economy. Risk stratifying and improving CRC patients with poor health prior to surgery has been shown to side 53 reduce complications post-operatively. Better utilization of registry and hospital data can provide clinicians iin pre-operative risk prior to surgery enabling prehabilitation and ensure that the patient is in an optimal shape, which reduces the risk of post-operative complications. Florence is a collaboration between the Cancer Registry of Norway, Center for Surgical Science, Oslo University Hospital, Lund University, and Technical University of Denmark.

 

Methods

The OMOP Common Data Model, a generic method for standardising the structure and content of observational health data, will be used to harmonize registry data across data partners. Federated learning will be established using Vantage6 technology. The data is collected from the Cancer Registry of Norway and Norwegian Registry for gastrointestinal Surgery (Norgast).

 

Results

Data from ~30 000 colorectal cancer patients that has undergone surgery has been harmonized to OMOP. In parallel Vantage6 has been tested with synthetic data to set up the infrastructure needed for federated learning.

 

Conclusions

The project is underway to reach its aim to lay the foundation for a decision-support tool that can improve the treatment of colorectal cancer patients.