Abstract
The abundance and/or location of tumor infiltrating lymphocytes (TILs), especially CD8+ T cells, in solid tumors can serve as a prognostic indicator in various types of cancer. However, it is often difficult to select an appropriate threshold value in order to stratify patients into well-defined risk groups. It is also important to select appropriate tumor regions to quantify the abundance of TILs. On the other hand, machine-learning approaches can stratify patients in an unbiased and automatic fashion. Based on immunofluorescence (IF) images of CD8+ T lymphocytes and cancer cells, we develop a machine-learning approach which can predict the risk of relapse for patients with Triple Negative Breast Cancer (TNBC). Tumor-section images from 9 patients with poor outcome and 15 patients with good outcome were used as a training set. Tumor-section images of 29 patients in an independent cohort were used to test the predictive power of our algorithm. In the test cohort, 6 (out of 29) patients who belong to the poor-outcome group were all correctly identified by our algorithm; for the 23 (out of 29) patients who belong to the good-outcome group, 17 were correctly predicted with some evidence that improvement is possible if other measures, such as the grade of tumors, are factored in. Our approach does not involve arbitrarily defined metrics and can be applied to other types of cancer in which the abundance/location of CD8+ T lymphocytes/other types of cells is an indicator of prognosis.
Highlights
In most cancer types, it has been demonstrated that patients with higher numbers of tumor infiltrating lymphocytes (TILs) in their solid tumors usually have better prognosis in term of the overall survival as well as the disease-free survival (Gooden et al, 2011)
If we assume that any patient whose Rtn < Rc is predicted to have the poor outcome, we will achieve some degree of accuracy of the prediction (Ac) by the trained MXNet
Since the 20% of small patches from the CH cohort are randomly selected and the convolutional neural network (CNN) can have some randomness, the Rtn for each patient can vary for different realizations
Summary
In most cancer types, it has been demonstrated that patients with higher numbers of tumor infiltrating lymphocytes (TILs) in their solid tumors usually have better prognosis in term of the overall survival as well as the disease-free survival (Gooden et al, 2011). In colorectal cancer and melanoma (Pagès et al, 2009; Galon et al, 2016), the ratio of T-cell density in the core of a tumor (CT) to that at the invasive margin (IM), i.e., the Immunoscore, has demonstrated its power to indicate prognosis. Due to the heterogeneity of the abundance of TILs within tumors, selection of the threshold-value for defining patient categories can be ambiguous. There have recently been a few successful applications of machine-learning approaches in cancer research: Agarap (2017) compared six machine-learning (ML) algorithms on the Wisconsin Diagnostic Dataset for a binary prediction problem of benign vs. malignant tumor; Heidari et al (2018) developed a machine-learning approach to predict short-term cancer risk by comparing asymmetry of the left vs. right breasts; Saltz et al (2018) trained a convolutional neural network (CNN) to recognize TILs in the H&E histological images from the TCGA database and generated TIL maps of TCGA samples; here the authors showed that TIL densities and spatial structure can be associated with features such as tumor types, immune subtypes, and tumor molecular subtypes
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
