Opinions

Reply to Letter to the Editor: “Computer-aided diagnosis system for thyroid nodules on ultrasonography: diagnostic performance and reproducibility based on the experience level of operators”

by Eun Ju Ha, Eun Young Jeong & Hye Lin Kim (radhej@naver.com)

Computer-aided diagnosis system for thyroid nodules on ultrasonography: diagnostic performance and reproducibility based on the experience level of operators

We thank Dr Rojas-Rojas and colleagues for their comments on our recent article [1]. As they commented, we only used still images for the evaluation and cine loops were not used in this study. Since an experienced radiologist recorded still images of the each nodule in real-time US and provided images to the operators, we agree that the radiologist hability to identify nodules may affect the CAD system performance. Further studies are required to determine the diagnostic performance of the less-experienced operators in a real practice scenario.

Dr Rojas-Rojas and colleagues also suggested that one of the main limitations of the CAD system is its inability to recognize calcifications. As they commented, the CAD system (S-DetectTM ver. 1) integrated into a commercialized US platform (RS80A; Samsung Medison) cannot recognize calcifications. However, the recent CAD system (S-DetectTM ver. 2, RS85A; Samsung Medison) was developed to recognize calcifications automatically. Although the performance of this system has not been reported yet, we believe that the CAD system has scalability and the performance could be upgraded to incorporate reports on incorrect results in the future. We strongly agree that one of the main limitations of the current CAD system is its inability or inaccuracy, with regard to recognizing calcifications; therefore, this has to be undertaken by the radiologist at this stage.

References