Deep Learning Improves Pancreatic Cancer Diagnosis Using RNA-Based Variants

Abstract

For optimal pancreatic cancer treatment, early and accurate diagnosis is vital. Blood-derived biomarkers and genetic predispositions can contribute to early diagnosis, but they often have limited accuracy or applicability. Here, we seek to exploit the synergy between them by combining the biomarker CA19-9 with RNA-based variants. We use deep sequencing and deep learning to improve differentiating pancreatic cancer and chronic pancreatitis. We obtained samples of nucleated cells found in peripheral blood from 268 patients suffering from resectable, non-resectable pancreatic cancer, and chronic pancreatitis. We sequenced RNA with high coverage and obtained millions of variants. The high-quality variants served as input together with CA19-9 values to deep learning models. Our model achieved an area under the curve (AUC) of 96% in differentiating resectable cancer from pancreatitis using a test cohort. Moreover, we identified variants to estimate survival in resectable cancer. We show that the blood transcriptome harbours variants, which can substantially improve noninvasive clinical diagnosis.

NetRank Recovers Known Cancer Hallmark Genes as Universal Biomarker Signature for Cancer Outcome Prediction

“…This is probably due to the complexity of the genetic component of pancreatic cancer, which makes it not easily explainable. It is widely accepted that other low penetrance genes play a role in pancreatic cancer ( Milne et al, 2009 ; Klein and Westenberger, 2012 ; Al-Fatlawi et al, 2021 ). When we looked deeper into biomarker studies, including both pancreatic and ovarian cancer, we realized that both pancreatic and ovarian cancer are similar in terms of tissue structure, and both are located in the endocrine system.…”

“…Identifying highly accurate biomarkers is a complex problem. CA19-9, for example, is well established in pancreatic cancer but has only an accuracy of 70–80% ( Al-Fatlawi et al, 2021 ), which means that it is not suitable for diagnosis on its own, but only to monitor relapse after surgery. One way to improve accuracy and robustness of the diagnoses is to employ biomarker signatures instead of only using single biomarkers.…”

Deep Learning Applications in Pancreatic Cancer

“…Similarly, Al-Fatlawi et al demonstrated that in conjunction with Ca 19-9, RNA-based variants can be utilized to differentiate between resectable PC, non-resectable PC, and chronic pancreatitis through DL with an AUC of 0.9. Additionally, their neural network was able to identify two mutations, B4GALT5 and GSDMD, which are closely related to PC progression and survival [26]. With the addition of genomic variants, not only can we diagnose PC at an earlier stage, but we can potentially offer patients a more personalized treatment regimen and prognostication as novel targeted therapeutic options are developed.…”

Predicting Lung Cancer Survival after Curative Surgery Using Deep Learning of Diffusion MRI

“…The convolutional neural network (CNN) is a branch of deep learning which was successfully applied in the image analysis such as the detection, classi cation, segmentation, and prediction tasks using medical image data such as pathology, X-ray, CT and MRI. The medical images are also quite a promising eld of research by using CNN in the detection and classi cation of pathology and the prediction of clinically relevant outcomes [15][16][17][18][19][20][21][22][23][24][25][26] . In our knowledge, there have been no reports of deep learning using CNN in the analysis of DWI and prognosis of lung cancer.…”