Back

ScanTecc classifies primary cancers via cell-free extrachromosomal circular DNA in peripheral blood

Fang, J.; Xu, Y.; Wang, J.; Shan, B.; Ding, K.; Li, S.; Yu, Q.; Zhang, W.; shao, Y.; Yang, J.; Liu, K.; Xu, G.; Yao, X.; Sun, R.; Zhang, M.; Zhang, Z.; Pan, Y.; Guo, C.; Qu, K.

2022-06-27 oncology
10.1101/2022.06.24.22276859 medRxiv
Show abstract

Extrachromosomal circular DNA (eccDNA) is gaining substantial attention in basic and clinical research for its contribution to tumor heterogeneity, biogenesis, and evolution. However, its presence and molecular features in peripheral blood from cancer patients are poorly understood. Here, we demonstrated the existence of eccDNA molecules in cancer patients plasma using atomic-force microscopy and high throughput sequencing technology. We then developed a method named ScanTecc (Screening cancer Types with cell-free eccDNA) that integrates cell-free circular DNA sequencing and machine learning approaches for cancer patient screening and classification. We applied ScanTecc on a total of 413 patients with multiple types of human primary cancers and 52 healthy individuals and found a significantly greater number and more extended size of eccDNAs in patients peripheral blood. ScanTecc accurately distinguished cancer patients (including early stages I and II) from healthy individuals with an overall prediction rate of 0.85 and identified distinct cancer types with an overall prediction rate of 0.84. Our study provides evidence for a non-invasive approach potentially applicable for early detection of cancer patients in clinics.

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.