1165PDeep learning in nasopharyngeal carcinoma: A retrospective cohort study of 3D convolutional neural networks on magnetic resonance imaging

M Y Qiang,X Lv,C F Li,K Y Liu,X Chen,X Guo
DOI: https://doi.org/10.1093/annonc/mdz252.057
IF: 51.769
2019-01-01
Annals of Oncology
Abstract:Background: The human microbiome has been suggested to be involved in the regulation of response to anticancer strategies, however, information about how commensal microbes in a cancer patient change during radiotherapy is limited and the relationship between the microbiome and response to radiotherapy is poorly studied.Methods: Sixty-two patients with nasopharyngeal carcinoma scheduled for radiotherapy-based treatment were prospectively enrolled.Nasopharyngeal swab samples were collected longitudinally during radiotherapy, and their microbial composition was assessed using 16S rRNA sequencing.All patients were followed up to 24 months after completion of radiotherapy.Results: The beta-diversity of nasopharyngeal microbial communities (np-microbiome) showed changes throughout radiotherapy-based treatment.The magnitude of changes in the np-microbiome was stably and statistically significantly different between early and late responders (P ¼ 0.02), with greater changes in early responders.This was confirmed in a temporal network analysis, where the networks varied widely in early responders, but were significantly more constrained in late responders (P ¼ 0.009).Operational taxonomic units (OTUs) mapped to Corynebacterrium, Staphylococcus and Anaerococcus showed increasing loss with treatment, while all other abundant OTUs were stable over treatment.Twenty-seven abundant OTUs differed statistically, significantly (P < 0.05) by patients' response throughout the treatment period.Conclusions: The nasopharyngeal microbiome in NPC patients changes during radiotherapy-based treatment.These changes are statistically, significantly associated with patients' response.
What problem does this paper attempt to address?