Back

Computed Tomography Radiomics Signatures: Sensitive biomarkers for clinical decision support in pancreatic cancer- a pilot study

Habibalahi, A.; Moses, D.; Campbell, J.; Mahbub, S.; Barbour, A.; Samra, J.; Haghighi, K.; Gebski, V.; Goldstein, D.; Goldys, E.

2021-12-05 oncology
10.1101/2021.12.03.21267217 medRxiv
Show abstract

AimTo evaluate if suitably designed computed tomography (CT) radiomic signatures are sensitive to tumour transformation, and able to predict disease free survival (DFS) and overall survival (OS) time in patients with pancreatic cancer. MethodEthical approval by UNSW review board was obtained for this retrospective analysis. This study consisted of 27 patients with pancreatic cancer. Unsupervised principal component analysis was employed to evaluate the sensitivity of radiomic signatures to cancer presence and treatment. Further, optimised radiomic signatures were discovered using swarm intelligence and assessed for their capability to predict DFS and OS based on Kaplan-Meier analysis and receiver-operator characteristics (ROC) curves. ResultsWe found that appropriate two radiomic signature are sensitive to cancer presence (area under the curve, AUC=0.95) and cancer treatment, respectively. Two other optimized radiomics signatures showed significant correlations with DFS and OS, respectively (p<0.05). ConclusionThe CT radiomics signatures are an independent biomarkers which are modified when cancer is present and can help to estimate DFS and OS in patients. These signatures have the potential to be used to support clinical decision-making in pancreatic cancer treatment.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.