Objective To develop a neural network-based deep learning model for predicting postoperative recurrence in thyroid tumor patients and validate the model with external datasets for providing clinicians with a reliable decision support tool.Methods An artificial neural network structure was adopted in the study,with thyroid tumor data from the SEER database serving as the training set.External validation was conducted with open-source data from the University of California,Irvine(UCIrvine),and the data from 100 patients at a general tertiary hospital in Hunan province.The model's accuracy and reliability in predicting recurrence were evaluated through multiple performance metrics.Results Experimental results showed that the model outperformed Logistic model in recurrence prediction,with accuracy,recall rate,precision and F1 score reaching 0.915 3,0.981 8,0.921 1 and 0.947 4 in internal validation.Moreover,the model achieved accuracies,recall rates,precisions,F1 scores and ROC_AUC values of 0.832 9,0.945 5,0.841 4,0.890 4 and 0.78 on the UCIrvine validation set,while 0.870 0,0.880 0,0.862 7,0.871 3 and 0.80 on the local validation set.Conclusion This neural network-based predictive model exhibits excellent performance in thyroid tumor recurrence prediction,providing clinicians with a valuable decision support tool that can help optimize postoperative treatment plans and improve patient prognosis management.

Objective To develop a neural network-based deep learning model for predicting postoperative recurrence in thyroid tumor patients and validate the model with external datasets for providing clinicians with a reliable decision support tool.Methods An artificial neural network structure was adopted in the study,with thyroid tumor data from the SEER database serving as the training set.External validation was conducted with open-source data from the University of California,Irvine(UCIrvine),and the data from 100 patients at a general tertiary hospital in Hunan province.The model's accuracy and reliability in predicting recurrence were evaluated through multiple performance metrics.Results Experimental results showed that the model outperformed Logistic model in recurrence prediction,with accuracy,recall rate,precision and F1 score reaching 0.915 3,0.981 8,0.921 1 and 0.947 4 in internal validation.Moreover,the model achieved accuracies,recall rates,precisions,F1 scores and ROC_AUC values of 0.832 9,0.945 5,0.841 4,0.890 4 and 0.78 on the UCIrvine validation set,while 0.870 0,0.880 0,0.862 7,0.871 3 and 0.80 on the local validation set.Conclusion This neural network-based predictive model exhibits excellent performance in thyroid tumor recurrence prediction,providing clinicians with a valuable decision support tool that can help optimize postoperative treatment plans and improve patient prognosis management.

Bivalves are species-rich mollusks with prominent protective roles in coastal ecosystems.Across these ancient lineages,colony-founding larvae anchor themselves either by byssus produc-tion or by cemented attachment.The latter mode of sessile life is strongly molded by left-right shell asymmetry during larval development of Ostreoida oysters such as Crassostrea hongkongensis.Here,we sequenced the genome of C.hongkongensis in high resolution and compared it to reference bivalve genomes to unveil genomic determinants driving cemented attachment and shell asymmetry.Importantly,loss of the homeobox gene Antennapedia(Antp)and broad expansion of lineage-specific extracellular gene families are implicated in a shift from byssal to cemented attachment in bivalves.Comparative transcriptomic analysis shows a conspicuous divergence between left-right asymmetrical C.hongkongensis and symmetrical Pinctada fucata in their expression profiles.Especially,a couple of orthologous transcription factor genes and lineage-specific shell-related gene families including that encoding tyrosinases are elevated,and may cooperatively govern asymmet-rical shell formation in Ostreoida oysters.