Abstract
Cellular profiling with multiplexed immunofluorescence (MxIF) images can contribute to a more accurate patient stratification for immunotherapy. Accurate cell segmentation of the MxIF images is an essential step. We propose a deep learning pipeline to train a Mask R-CNN model (deep network) for cell segmentation using nuclear (DAPI) and membrane (Na+K+ATPase) stained images. We used two-stage domain adaptation by first using a weakly labeled dataset followed by fine-tuning with a manually annotated dataset. We validated our method against manual annotations on three different datasets. Our method yields comparable results to the multi-observer agreement on an ovarian cancer dataset and improves on state-of-the-art performance on a publicly available dataset of mouse pancreatic tissues. Our proposed method, using a weakly labeled dataset for pre-training, showed superior performance in all of our experiments. When using smaller training sample sizes for fine-tuning, the proposed method provided comparable performance to that obtained using much larger training sample sizes. Our results demonstrate that using two-stage domain adaptation with a weakly labeled dataset can effectively boost system performance, especially when using a small training sample size. We deployed the model as a plug-in to CellProfiler, a widely used software platform for cellular image analysis.
Highlights
Cellular profiling with multiplexed immunofluorescence (MxIF) images can contribute to a more accurate patient stratification for immunotherapy
There are some disadvantages of this approach: (1) the use of these methods usually involves manually fine-tuning parameters, which is often laborintensive and time-consuming[12], (2) methods built from those conventional algorithms often do not generalize well for batch processing or processing images from different datasets, (3) they may not be able to achieve satisfactory performance for a pplication[6]
We present a deep learning-based pipeline for instance cell segmentation for MxIF images of DAPI and Na+K+ATPase stained images by using two-stage domain adaptations to train a Mask R-convolutional neural network (CNN) model
Summary
Cellular profiling with multiplexed immunofluorescence (MxIF) images can contribute to a more accurate patient stratification for immunotherapy. Meijering[6] published a comprehensive review of the literature on methodologies for nuclear/cell segmentation, covering many conventional algorithms (i.e. non-deep learning), including thresholding, filtering, morphological operations, region accumulation, and model fitting. These methods used alone are seldom able to achieve satisfactory results. Deep learning, fully convolutional networks (FCN)[13] based methods, such as U-Net14, DoGnet[15], and DeepCell[16] have come into wide use for cell/nuclear segmentation. They demonstrate better accuracy in performance for cell nuclear s egmentation[14,16,17] with much better generalization capacity and are suitable for batch processing. Since the core concept of deep learning is to train a model using labeled samples, a large quantity of labeled data that covers an appropriately broad range of sample variation is necessary to achieve satisfactory results
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
