Classification of Precancerous Colorectal Lesions via ConvNeXt on Histopathological Images

In this translational study, the classification of precancerous colorectal lesions is performed by the ConvNeXt method on MHIST histopathological imaging dataset. The ConvNeXt method is a modernized ResNet-50 architecture having some training tricks inspired by Swin Transformers and ResNeXt. The performance of the ConvNeXt models are benchmarked on different scenarios such as ‘full data’, ‘gradually increasing difficulty based data’ and ‘k-shot data’. It is shown that the ConvNeXt model outperformes almost all the other studies in the literature which are applied on MHIST by using ResNet models, vision transformers, weight distillation, self-supervised learning and curriculum learning strategy based on different scenarios and metrics. The ConvNeXt model trained with ‘full data’ yields the best result with the score of 0.8890 for accuracy, 0.9391 for AUC, 0.9121 for F1 and 0.7633 for cohen’s cappa. The power of ConvNeXt is found as promising for classifying precancerous histopathological images and may be a good base line for miscellaneous tasks of computational pathology field with respect to the classical convolutional neural networks and vision transformers..