The integrative analysis of histopathological images and genomic data has received increasing attention for survival prediction of human cancers. However, the existing studies always hold the assumption that full modalities are available. As a matter of fact, the cost for collecting genomic data is high, which sometimes makes genomic data unavailable in testing samples. A common way of tackling such incompleteness is to generate the genomic representations from the pathology images. Nevertheless, such strategy still faces the following two challenges: (1) The gigapixel whole slide images (WSIs) are huge and thus hard for representation. (2) It is difficult to generate the genomic embeddings with diverse function categories in a unified generative framework. To address the above challenges, we propose a Conditional Latent Differentiation Variational AutoEncoder (LD-CVAE) for robust multimodal survival prediction, even with missing genomic data. Specifically, a Variational Information Bottleneck Transformer (VIBTrans) module is proposed to learn compressed pathological representations from the gigapixel WSIs. To generate different functional genomic features, we develop a novel Latent Differentiation Variational AutoEncoder (LD-VAE) to learn the genomic and function-specific posteriors for the genomic embeddings with diverse functions. Finally, we use the product-of-experts technique to integrate the genomic posterior and image posterior for the joint latent distribution estimation in LD-CVAE. We test the effectiveness of our method on five different cancer datasets, and the experimental results demonstrate its superiority in both complete and missing modality scenarios. The code is released†.

Multi-task learning (MTL) has become a research hotspot for the analysis of whole-slide histopathological images (WSIs) since it can capture the shared representations of different tasks for the improvement of individual tasks. However, the shared representations learned by MTL are always dominated by the tasks appearing in the training set that is difficult to directly apply it on the unseen (new) tasks, especially when the unseen tasks are significantly different from the known tasks. To address the above issues, we develop a Task-Agnostic Feature-Learner (TAFL) for efficient adaptation to unseen clinical tasks, which can leverage useful image information from the existing tasks for new clinical trials with minimal task-specific modifications. Specifically, we firstly develop a neural architecture search (NAS) module that can design the network architectures of TAFL automatically. Then, a novel task-level meta-learning algorithm is developed to extract efficient and universal information from the known tasks for improving the prediction performance on the unseen tasks. We evaluate our method on three publicly available datasets derived from The Cancer Genome Atlas (TCGA) for various clinical prediction tasks (i.e., staging, cancer subtyping and survival prediction), and the experimental results indicate that our TAFL can effectively adapt to unseen tasks with better prediction performance.

With the remarkable success of digital histopathology and the deep learning technology, many whole-slide pathological images (WSIs) based deep learning models are designed to help pathologists diagnose human cancers. Recently, rather than predicting categorical variables as in cancer diagnosis, several deep learning studies are also proposed to estimate the continuous variables such as the patients' survival or their transcriptional profile. However, most of the existing studies focus on conducting these predicting tasks separately, which overlooks the useful intrinsic correlation among them that can boost the prediction performance of each individual task. In addition, it is sill challenge to design the WSI-based deep learning models, since a WSI is with huge size but annotated with coarse label. In this study, we propose a general multi-instance multi-task learning framework (HistMIMT) for multi-purpose prediction from WSIs. Specifically, we firstly propose a novel multi-instance learning module (TMICS) considering both common and specific task information across different tasks to generate bag representation for each individual task. Then, a soft-mask based fusion module with channel attention (SFCA) is developed to leverage useful information from the related tasks to help improve the prediction performance on target task. We evaluate our method on three cancer cohorts derived from the Cancer Genome Atlas (TCGA). For each cohort, our multi-purpose prediction tasks range from cancer diagnosis, survival prediction and estimating the transcriptional profile of gene TP53. The experimental results demonstrated that HistMIMT can yield better outcome on all clinical prediction tasks than its competitors.

Survival analysis is to estimate the survival time for an individual or a group of patients, which is a valid solution for cancer treatments. Recent studies suggested that the integrative analysis of histopathological images and genomic data can better predict the survival of cancer patients than simply using single bio-marker, for different bio-markers may provide complementary information. However, for the given multi-modal data that may contain irrelevant or redundant features, it is still challenge to design a distance metric that can simultaneously discover significant features and measure the difference of survival time among different patients. To solve this issue, we propose a Feature-Aware Multi-modal Metric Learning method (FAM3L), which not only learns the metric for distance constraints on patients' survival time, but also identifies important images and genomic features for survival analysis. Specifically, for each modality of data, we firstly design one feature-aware metric that can be decoupled into a traditional distance metric and a diagonal weight for important feature identification. Then, in order to explore the complex correlation across multiple modality data, we apply Hilbert-Schmidt Independence Criterion (HSIC) to jointly learn multiple metrics. Finally, based on the learned distance metrics, we apply the Cox proportional hazards model for prognosis prediction. We evaluate the performance of our proposed FAM3L method on three cancer cohorts derived from The Cancer Genome Atlas (TCGA), the experimental results demonstrate that our method can not only achieve superior performance for cancer prognosis, but also identify meaningful image and genomic features correlating strongly with cancer survival.

The tumor-infiltrating lymphocytes (TILs) and its correlation with tumors have shown significant values in the development of cancers. Many observations indicated that the combination of the whole-slide pathological images (WSIs) and genomic data can better characterize the immunological mechanisms of TILs. However, the existing image-genomic studies evaluated the TILs by the combination of pathological image and single-type of omics data (e.g., mRNA), which is difficulty in assessing the underlying molecular processes of TILs holistically. Additionally, it is still very challenging to characterize the intersections between TILs and tumor regions in WSIs and the high dimensional genomic data also brings difficulty for the integrative analysis with WSIs. Based on the above considerations, we proposed an end-to-end deep learning framework i.e., IMO-TILs that can integrate pathological image with multi-omics data (i.e., mRNA and miRNA) to analyze TILs and explore the survival-associated interactions between TILs and tumors. Specifically, we firstly apply the graph attention network to describe the spatial interactions between TILs and tumor regions in WSIs. As to genomic data, the Concrete AutoEncoder (i.e., CAE) is adopted to select survival-associated Eigengenes from the high-dimensional multi-omics data. Finally, the deep generalized canonical correlation analysis (DGCCA) accompanied with the attention layer is implemented to fuse the image and multi-omics data for prognosis prediction of human cancers. The experimental results on three cancer cohorts derived from the Cancer Genome Atlas (TCGA) indicated that our method can both achieve higher prognosis results and identify consistent imaging and multi-omics bio-markers correlated strongly with the prognosis of human cancers.