SpecialReport
Today’sradiologistsmeettomorrow’sAI:thepromises,pitfalls,andunbridledpotentialDianwenNg1^,HaoDu1^,MelissaMin-SzuYao2,3^,RussellOliverKosik3,WingP.Chan2,3,4#^,MenglingFeng1,4#^
1SawSweeHockSchoolofPublicHealth,NationalUniversityHealthSystem,NationalUniversityofSingapore,Singapore;2DepartmentofRadiology,WanFangHospital,TaipeiMedicalUniversity,Taipei;3DepartmentofRadiology,SchoolofMedicine,CollegeofMedicine,TaipeiMedicalUniversity,Taipei;4MedicalInnovationDevelopmentCenter,WanFangHospital,TaipeiMedicalUniversity,Taipei
#Theseauthorscontributedequallytothiswork.
^ORCID:DianwenNg,---;HaoDu,---;MelissaMin-SzuYao,---;WingP.Chan,---;MenglingFeng,---.
Correspondenceto:WingP.Chan,MD.DepartmentofRadiology,WanFangHospital,TaipeiMedicalUniversity,HsingLongRoad,Section3,Taipei.Email:wingchan
tmu.edu.tw;MenglingFeng,PhD.SawSweeHockSchoolofPublicHealth,NationalUniversityHealthSystem,NationalUniversityofSingapore,TahirFoundationBuilding12ScienceDrive2,#10-01,Singapore.Email:ephfmnus.edu.sg.Abstract:Advancesininformationtechnologyhaveimprovedradiologists’abilitiestoperformanincreasingvarietyoftargeteddiagnosticexams.However,duetoagrowingdemandforimagingfromanagingpopulation,thenumberofexamscouldsoonexceedthenumberofradiologistsavailabletoreadthem.However,artificialintelligencehasrecentlyresoundingsuccessinseveralcasestudiesinvolvingtheinterpretationofradiologicexams.Assuch,theintegrationofAIwithstandarddiagnosticimagingpracticestorevolutionizemedicalcarehasbeenproposed,withtheultimategoalbeingthereplacementofhumanradiologistswithAI‘radiologists’.However,the