人工智能辅助压缩感知加速技术对鼻咽癌MRI影像组学特征提取及分期诊断模型性能的影响

打开文本图片集

Abstract：ObjectiveToevaluatetheeffctofartificialinteligence-asistedcompressedsensing （ACS）accelerationonMRI radiomic featureextractionand performanceof diagnosticstagingmodels for nasopharyngeal carcinoma （NPC） incomparison withconventionalparallelimaging （PI）.MethodsAtotalof 64patients withnewlydiagnosedNPCunderwent3.0TMRIusing axialT1-weighted（T1W），-weighted（TW）andcontrast-enhancedT-weighted（CE-TW）sequences.BothPandA protocolswereperformedusing identicalimagingparameters.Thetotalscantimeforthe3sequencesinACS groupwas 227s， representinga 30% reduction from312sinthePIgroup.Eighteenfirst-orderand75texturefeatureswereextractedusing Pyradiomics.Intraclasscorelationcoeficients（ICCs）werecalculatedtoassesstheagreementbetweenthetwoacceleration methods.Afterfeatureselectionusingtheleastabsoluteshrinkageandselectionoperator（LASSO），randomforestregression modelswere constructed to distinguish early-stage （T1andT2）fromadvanced-stage （T3andT4）NPC.The diagnostic performance of themodels wasevaluated using thearea under thereceiver operatingcharacteristiccurve （AUC）and compared using the DeLong test. Results ACS-accelerated images demonstrated good radiomic reproducibility， with 86.0% （240/279） of features showing good agreement （ICC>0.75） ），withmeanICCsforT1W，T2WandCE-T1Wsequencesof0.91±0.09， 0.89±0.13 and 0.88±0.11. ， respectively. The staging prediction models achieved similar AUCs for ACS and PI （0.89 vs 0.90， P= 0.991）. Conclusion TheMRI radiomic features extracted using ACSand PItechniques are highlyconsistent，andthe ACSbasedmodelshowscompaabledagnosticpformancetothePasedmodel，butACSsignificantlyducesteantime and providesanefficientandreliableaccelerationstrategy forradiomicsinNPC.

Keywords： artificial inteligence; magneticresonance imaging;radiomics;agreement; nasopharyngealcarcinoma

鼻咽癌（NPC）是起源于鼻咽部黏膜上皮的恶性肿瘤，主要分布在东南亚及中国南部地区[12]。（剩余16283字）