Machine-learning for Optimal detection of iNflammatory cells in the KidnEY¶

The Banff classification is the gold standard for histopathologic assessment of transplant kidney biopsies. It consists of 17 Banff Lesion Scores (BLS), ten focusing on the presence and extent of inflammatory cells in different kidney compartments. Most BLS are graded semi-quantitatively as mild, moderate, and severe based on the amount of inflammatory cells within the corresponding compartment. As the diagnosis and subsequent treatment decision depends on the result of the different BLS, the assessment of these individual BLS must be objective and consistent. However, Banff scoring has mediocre reproducibility and is time-consuming in daily practice. Therefore, the development of automated biopsy assessment holds great potential to reduce pathologists' workload and increase scoring consistency.

Challenge Overview¶

The challenge will have two different leaderboards:

• Leaderboard 1: Detection of mononuclear, inflammatory cells (mononuclear leukocytes (MNLs))
• Leaderboard 2: Detect and distinguish inflammatory cells: monocytes and lymphocyte

We will provide a baseline and a tutorial on loading the data and training a basic model. The submission format will be coordinate text files (one for the MNLs, one for the monocytes, and one for the lymphocytes).

The top three winners of the test phase will receive a monetary award. The top-ranking teams will be invited to co-author the publication resulting from the Monkey Challenge (a maximum of three members per team).

The model development and validation phases run concurrently. The winners will be decided on a final test set.