CUP-AI-Dx: A tool for inferring cancer tissue of origin and molecular subtype using RNA gene-expression data and artificial intelligence

Abstract

Background Cancer of unknown primary (CUP), representing approximately 3-5% of all malignancies, is defined as metastatic cancer where a primary site of origin cannot be found despite a standard diagnostic workup. Because knowledge of a patient's primary cancer remains fundamental to their treatment, CUP patients are significantly disadvantaged and most have a poor survival outcome. Developing robust and accessible diagnostic methods for resolving cancer tissue of origin, therefore, has significant value for CUP patients. Methods We developed an RNA-based classifier called CUP-AI-Dx that utilizes a 1D Inception convolutional neural network (1D-Inception) model to infer a tumor's primary tissue of origin. CUP-AI-Dx was trained using the transcriptional profiles of 18,217 primary tumours representing 32 cancer types from The Cancer Genome Atlas project (TCGA) and International Cancer Genome Consortium (ICGC). Gene expression data was ordered by gene chromosomal coordinates as input to the 1D-CNN model, and the model utilizes multiple convolutional kernels with different configurations simultaneously to improve generality. The model was optimized through extensive hyperparameter tuning, including different max-pooling layers and dropout settings. For 11 tumour types, we also developed a random forest model that can classify the tumour's molecular subtype according to prior TCGA studies. The optimised CUP-AI-Dx tissue of origin classifier was tested on 394 metastatic samples from 11 tumour types from TCGA and 92 formalin-fixed paraffin-embedded (FFPE) samples representing 18 cancer types from two clinical laboratories. The CUP-AI-Dx molecular subtype was also independently tested on independent ovarian and breast cancer microarray datasets Findings CUP-AI-Dx identifies the primary site with an overall top-1-accuracy of 98.54% in cross-validation and 96.70% on a test dataset. When applied to two independent clinical-grade RNA-seq datasets generated from two different institutes from the US and Australia, our model predicted the primary site with a top-1-accuracy of 86.96% and 72.46% respectively. Interpretation The CUP-AI-Dx predicts tumour primary site and molecular subtype with high accuracy and therefore can be used to assist the diagnostic work-up of cancers of unknown primary or uncertain origin using a common and accessible genomics platform. Funding NIH R35 GM133562, NCI P30 CA034196, Victorian Cancer Agency Australia.

Artificial Intelligence in Bulk and Single-Cell RNA-Sequencing Data to Foster Precision Oncology

“…An autoencoder-based DL approach combining supervised classification and unsupervised clustering revealed the presence of novel breast and bladder cancer subtypes associated with different prognosis [ 66 ]. Convolutional NNs have been employed to infer tumor’s primary tissue of origin of metastasis and to guide management of patients with cancer of unknown primary [ 67 ]. Again, integrating biological information (i.e., gene set enrichment analysis) in NNs resulted in an improved classification of individual colorectal and breast cancer subtypes relative to canonical ML approaches [ 18 ].…”

2020: A year to remember

“…This issue makes this type of cancer more difficult to treat. Yue Zhao and colleagues [6] from the Jackson laboratory (Farmington, CT, USA) developed an RNA-based classifier using artificial intelligence that identified the primary site in clinical datasets from Australia and the USA with a respective accuracy of 72% and 89%. The model and results are publicly available as a software package to allow reproduction of the results and application to new datasets.…”

Deep learning-based pan-cancer classification model reveals cancer-specific gene expression signatures

“…However, if two features contribute equally to the prediction, then they should have the same SHAP values. Compared to LIME [32] which used local surrogate models to interpret individual predictions, the SHAP method is more accurate and consistent with human intuition as well as has better computational efficiency [26]. We built 10 different DNN models using randomly split TCGA data and SHAP values for each of those models were computed using DeepExplainer, which is based on the combination of SHAP and deepLIFT [33], function from the python SHAP package.…”

“…Researchers have applied deep learning not only to medical images but also to various omics data types [20,21]. For example, some studies have employed deep learning to predict cancer types based upon mutation [22][23][24] and gene expression data [25][26][27]. Machine learning and deep learning have also been applied for cancer type classification and gene signatures identification, albeit with moderate levels of accuracy.…”