The LUNA25 Challenge Banner
Join

Clinical Problem 🔍

Lung cancer is the leading cause of cancer-related deaths worldwide, with 2.5 million people diagnosed and 1.8 million deaths occurring annually worldwide (Bray et al., 2024). Two landmark randomized controlled trials, the National Lung Cancer Screening Trial (NLST) and the Dutch-Belgian lung cancer screening trial (NELSON), have provided strong evidence that lung cancer mortality can be reduced by repeated screening with low-dose chest CT of high-risk individuals (Aberle. et al., 2011, de Koning et al., 2020). Lung cancer screening could therefore play an important role in reducing lung cancer mortality. However, worldwide implementation of lung cancer screening will further increase the already high workload on radiologists, so there is a growing demand for AI technology that helps to lower the burden on radiologists.

Need for Benchmarks 📈

Numerous AI algorithms for nodule analysis have been described in the scientific literature and more than 15 commercial products have obtained CE marking for the European market. These algorithms can help radiologists with nodule detection, nodule volumetric measurement, nodule tracking, and nodule risk estimation. However, it is at present challenging to adequately validate and benchmark the increasing amount of AI algorithms being developed. For this, Grand challenges, which are international public competitions, offer the means to compare and validate AI algorithms and therefore address the lack of adequate validation among AI solutions (Leeuwen et al., 2021). LUNA16 (Setio et al., 2017) and the Kaggle 2017 Data Science Bowl (Jacobs et al., 2021) are examples of previous competitions to evaluate and validate AI algorithms in the field of lung nodule analysis.

The LUNA25 Challenge and Reader Study 👩‍💻🧑‍⚕️

LUNA25 is a new grand challenge with over 4,000 carefully-annotated low-dose chest CT exams to develop and validate modern AI algorithms and estimate radiologist' performance at lung nodule malignancy risk estimation. The design of LUNA25 has been established in collaboration with an international, multi-disciplinary scientific advisory board (15 experts in radiology, pulmonology, epidemiology, AI, and a lung cancer patient representative) to ensure meaningful validation of lung cancer AI models towards clinical translation.

The 2025 edition of LUNA will focus on validating AI and radiologists' performance at lung nodule malignancy risk estimation at low-dose chest CT.

LUNA25 primarily consists of two sub-studies:

  • AI Study: An annotated dataset consisting of retrospective, multi-center, low-dose chest CT scans is made publicly available for all participating teams and the research community. The dataset aims to represent scans seen during lung cancer screening trials. Teams can use this dataset to develop AI models and submit their algorithms (in Docker containers) for evaluation. At the end of this development phase, all algorithms are ranked based on their performance on a hidden testing cohort of unseen cases (exact number to be decided, pending final decisions from a few data providers).

  • Reader study: The LUNA25 reader study will include unseen cases from multiple institutions that will be read by a panel of international readers with varying levels of expertise (exact number also to be decided, pending final decisions from a few data providers).

Prizes 🏆

The top 5 participating teams will be invited to join the LUNA25 consortium, and in turn, they will be listed as consortium members on an upcoming journal paper summarizing the findings of our challenge.

Furthermore, they will receive the following cash prizes, as per their ranking:
🥇1st place: €1000
🥈2nd place: €500
🥉3rd place: €250
🏵️4th place: €150
🏵️5th place: €100