Automated Detection and Classification of Desmoplastic Reaction at the Colorectal Tumour Front Using Deep Learning

SIMPLE SUMMARY: Desmoplastic reaction (DR) has previously been shown to be a promising prognostic factor in colorectal cancer (CRC). However, its manual reporting can be subjective and consequently consistency of reporting might be affected. The aim of our study was to develop a deep learning algorithm that would facilitate the objective and standardised DR assessment. By applying this algorithm on a CRC cohort of 528 patients, we demonstrate how deep learning methodologies can be used for the accurate and reproducible reporting of DR. Furthermore, this study showed that the prognostic significance of DR was superior when assessed through the use of the deep learning classifier than when assessed manually. In this study, we demonstrate how the application of machine learning approaches can help by not only identifying complex patterns present within histopathological images in a standardised and reproducible manner, but also report a more accurate patient stratification. ABSTRACT: The categorisation of desmoplastic reaction (DR) present at the colorectal cancer (CRC) invasive front into mature, intermediate or immature type has been previously shown to have high prognostic significance. However, the lack of an objective and reproducible assessment methodology for the assessment of DR has been a major hurdle to its clinical translation. In this study, a deep learning algorithm was trained to automatically classify immature DR on haematoxylin and eosin digitised slides of stage II and III CRC cases (n = 41). When assessing the classifier’s performance on a test set of patient samples (n = 40), a Dice score of 0.87 for the segmentation of myxoid stroma was reported. The classifier was then applied to the full cohort of 528 stage II and III CRC cases, which was then divided into a training (n = 396) and a test set (n = 132). Automatically classed DR was shown to have superior prognostic significance over the manually classed DR in both the training and test cohorts. The findings demonstrated that deep learning algorithms could be applied to assist pathologists in the detection and classification of DR in CRC in an objective, standardised and reproducible manner.