Artificial intelligence quantified tumour-stroma ratio is an independent predictor for overall survival in resectable colorectal cancer

Abstract

Background An artificial intelligence method could accelerate the clinical implementation of tumour-stroma ratio (TSR), which has prognostic relevance in colorectal cancer (CRC). We, therefore, developed a deep learning model for the fully automated TSR quantification on routine haematoxylin and eosin (HE) stained whole-slide images (WSI) and further investigated its prognostic validity for patient stratification. Methods We trained a convolutional neural network (CNN) model using transfer learning, with its nine-class tissue classification performance evaluated in two independent test sets. Patch-level segmentation on WSI HE slides was performed using the model, with TSR subsequently derived. A discovery (N=499) and validation cohort (N=315) were used to evaluate the prognostic value of TSR for overall survival (OS). Findings The CNN-quantified TSR was a prognostic factor, independently of other clinicopathologic characteristics, with stroma-high associated with reduced OS in the discovery (HR 1.72, 95% CI 1.24-2.37, P=0.001) and validation cohort (2.08, 1.26-3.42, 0.004). Integrating TSR into a Cox model with other risk factors showed improved prognostic capability. Interpretation We developed a deep learning model to quantify TSR based on histologic WSI of CRC and demonstrated its prognostic validity for patient stratification for OS in two independent CRC patient cohorts. This fully automatic approach allows for the objective and standardised application while reducing pathologists' workload. Thus, it can potentially be of significant aid in clinical prognosis prediction and decision-making. Funding National Key Research and Development Program of China, National Science Fund for Distinguished Young Scholar, and National Science Foundation for Young Scientists of China.

Deep Learning Prediction of Metastasis in Locally Advanced Colon Cancer Using Binary Histologic Tumor Images

“…In recent years, convolutional neural networks (CNN) and other deep learning algorithms have been established as state-of-the-art methods for a wide field of image classification tasks. CNNs for the quantification of tumor stroma proportion have already been developed [ 18 ], and it has been shown that tumor budding can be validly determined using machine learning methods [ 19 ]. For a comprehensive review of deep learning in colon cancer, we refer to Pacal et al [ 20 ].…”

The role of artificial intelligence to quantify the tumour-stroma ratio for survival in colorectal cancer

“…In this article of EBioMedicine , Ke Zhao and colleagues [1] show that it is possible to quantify the tumour-stroma percentage by artificial intelligence, using a convolutional neural network on whole-slide images (WSI). Moreover, the prognostic effect of the tumour-stroma ratio (TSR) for overall survival was confirmed, showing the robustness of the TSR method.…”

“…Research is exploring possibilities of developing new algorithms to support the pathologists in daily practice and to reduce their workload. Most of the algorithms are still only used in research setting, but look quite promising, as is the algorithm developed by Zhao et al [1] .…”

“…The basis of the TSR algorithm for WSI analysis is set and has shown to be a prognostic factor for overall survival. Scoring TSR with the automated method used by Zhao et al [1] should preferentially be validated in a larger, prospective study. For the TSR in colon cancer a prospective European validation study, the UNITED study, is currently underway [7] .…”

“…This study consists of an E-learning module with quality control, and will compare visual with automated TSR scoring, with the scope to be implemented in daily routine pathology practise. The article of Zhao et al [1] is anticipating to the need for automation of the routine pathology process, deep-learning algorithms for TSR on WSI is a good first step.…”