HISTOKT: CROSS KNOWLEDGE TRANSFER IN COMPUTATIONAL PATHOLOGY

The lack of well-annotated datasets in computational pathology (CPath) obstructs the application of deep learning techniques for classifying medical images. Many CPath workflows involve transferring learned knowledge between various image domains through transfer learning. Currently, most transfer learning research follows a model-centric approach, tuning network parameters to improve transfer results over few datasets. In this paper, we take a data-centric approach to the transfer learning problem and examine the existence of generalizable knowledge between histopathological datasets. First, we create a standardization workflow for aggregating existing histopathological data. We then measure inter-domain knowledge by training ResNet18 models across multiple histopathological datasets, and cross-transferring between them to determine the quantity and quality of innate shared knowledge. Additionally, we use weight distillation to share knowledge between models without additional training. We find that hard to learn, multi-class datasets benefit most from pretraining, and a two stage learning framework incorporating a large source domain such as ImageNet allows for better utilization of smaller datasets. Furthermore, we find that weight distillation enables models trained on purely histopathological features to outperform models using external natural image data.

Automated summary

In this paper, the lack of well-annotated datasets in computational pathology is addressed as a barrier to the application of deep learning techniques in classifying medical images. The study takes a data-centric approach to the transfer learning problem in order to examine the potential existence of generalizable knowledge between histopathological datasets. The results show that pretraining is most beneficial for hard to learn, multi-class datasets and a two-stage learning framework with a large source domain allows for better utilization of smaller datasets. The use of weight distillation also enables models trained on purely histopathological features to outperform models using external natural image data.