Opinions

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 María M. Rojas-Rojas & Gustavo A. Triana (gustavotrianamd@gmail.com )

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

To the Editor,

We read the paper by Jeong et al entitled “Computer-aided diagnosis system for thyroid nodules on ultrasonography: diagnostic performance and reproducibility based on the experience level of operators”. According to this study, a computer-aided diagnosis (CAD) system using ultrasonography (US) may have a potential role in the thyroid cancer diagnosis.

In this study, an experience radiologist performed all US examinations and recorded still images of each nodule. Later, the experience radiologist and three other operators with different levels of US experience obtained CAD data. The diagnostic performance of the CAD system based on the operators’ experience was calculated.

They found that the sensitivity and accuracy of the CAD system were not significantly different from those of the experienceradiologist [1]. Yet this results raise some queries: first, the experience radiologist located the nodules and selected the margins, so the CAD system accuracy was based on the radiologist hability to identify thyroid nodules and define its margins.

CAD data from three other operators was obtainedbased on the images selected by the experience radiologist, althought their hability to identify nodules was not evaluated and we believe that this could affect the CAD system performance in a real practice scenario.

Also, presence of microcalcifications increases the risk of malignancy in thyroid nodules and is one of the most specific US findings [2, 3, 4]. Microcalcifications are defined as as tiny, hyperechoic foci (< 1 mm) within the solid portion of the nodule [5]. Defining its presence in a nodule is a challenging feature of thyroid ultrasound and could change patient management.

One of the main limitations of the CAD system is its inability to recognize calcifications, so this feature must be established by the radiologist, affecting the accuracy of the CAD system. Regarding less experience operators in this study, we wonder how microcalcifications were determined and how this affected the diagnostic performance given that the detection differences were not evaluated.

Finally, we consider, further studies are aimed to determine the diagnostic performance of the CAD system in a real practice scenario.

References