How Many Annotators Do We Need? -- A Study on the Influence of Inter-Observer Variability on the Reliability of Automatic Mitotic Figure Assessment
Authors:
Frauke Wilm,
Christof A. Bertram,
Christian Marzahl,
Alexander Bartel,
Taryn A. Donovan,
Charles-Antoine Assenmacher,
Kathrin Becker,
Mark Bennett,
Sarah Corner,
Brieuc Cossic,
Daniela Denk,
Martina Dettwiler,
Beatriz Garcia Gonzalez,
Corinne Gurtner,
Annika Lehmbecker,
Sophie Merz,
Stephanie Plog,
Anja Schmidt,
Rebecca C. Smedley,
Marco Tecilla,
Tuddow Thaiwong,
Katharina Breininger,
Matti Kiupel,
Andreas Maier,
Robert Klopfleisch
, et al. (1 additional authors not shown)
Abstract:
Density of mitotic figures in histologic sections is a prognostically relevant characteristic for many tumours. Due to high inter-pathologist variability, deep learning-based algorithms are a promising solution to improve tumour prognostication. Pathologists are the gold standard for database development, however, labelling errors may hamper development of accurate algorithms. In the present work…
▽ More
Density of mitotic figures in histologic sections is a prognostically relevant characteristic for many tumours. Due to high inter-pathologist variability, deep learning-based algorithms are a promising solution to improve tumour prognostication. Pathologists are the gold standard for database development, however, labelling errors may hamper development of accurate algorithms. In the present work we evaluated the benefit of multi-expert consensus (n = 3, 5, 7, 9, 11) on algorithmic performance. While training with individual databases resulted in highly variable F$_1$ scores, performance was notably increased and more consistent when using the consensus of three annotators. Adding more annotators only resulted in minor improvements. We conclude that databases by few pathologists and high label accuracy may be the best compromise between high algorithmic performance and time investment.
△ Less
Submitted 8 January, 2021; v1 submitted 4 December, 2020;
originally announced December 2020.
