[Clinical Validation of Chest X-ray Educational Content for Radiography Students Using Gaze Information].

PURPOSE: Radiography training for students in colleges of radiology should be based on real clinical situations. The purpose of this study was to verify the clinical validity of our originally developed scenarios for chest X-ray training and the instructional contents using gaze information of experienced radiology technologists (RTs). METHODS: We divided 8 RTs with different experiences into an evaluator group (3 RTs) and a subject group (5 RTs). The evaluator group created a validation model consisting of 31 items, a chest X-ray scenario, instructional contents, and gaze attention objects during the scenario. The subject group simulated chest X-ray wearing an eye tracker. The evaluator group evaluated fit rates of the validation model to subjects' procedures based on gaze information to verify the clinical validity of the validation model. RESULTS: The subject group procedures did not deviate from the scenario. We obtained a fit rate of 91.6±6.70%. CONCLUSION: Our validation model showed more than 90% fitting with the chest X-ray techniques of five RTs with different backgrounds. This result suggested that the scenario and instructional contents in this study had clinical validity.

Evaluation of factors associated with the effectiveness of radiation protection glasses.

The effects of lead equivalent and lens area of radiation-protective eyewear on lens exposure control were examined. The simulated patient underwent 10-min X-ray fluoroscopy, and the lens dose of the simulated surgeon wearing radiation protection glasses was measured using lens dosemeters attached to the corner of the eye and eyeball. In total, 10 types of radiation protection glasses were selected for measurement. Correlation analysis of the equivalent dose in the lens of the eye with lead equivalence and lens area was performed. The equivalent dose in the lens of the eye of the corner of the eye was negatively correlated with the area of the lens. The equivalent dose in the lens of the eye and the eyeball showed a strong negative correlation with lead equivalence. Lens dosemeters worn at the corner of the eye may overestimate the equivalent dose in the lens of the eye. Moreover, the reduction in exposure of the lens was significantly influenced by the lead equivalent.

Extracting clinical terms from radiology reports with deep learning.

Extracting clinical terms from free-text format radiology reports is a first important step toward their secondary use. However, there is no general consensus on the kind of terms to be extracted. In this paper, we propose an information model comprising three types of clinical entities: observations, clinical findings, and modifiers. Furthermore, to determine its applicability for in-house radiology reports, we extracted clinical terms with state-of-the-art deep learning models and compared the results. We trained and evaluated models using 540 in-house chest computed tomography (CT) reports annotated by multiple medical experts. Two deep learning models were compared, and the effect of pre-training was explored. To investigate the generalizability of the model, we evaluated the use of other institutional chest CT reports. The micro F1-score of our best performance model using in-house and external datasets were 95.36% and 94.62%, respectively. Our results indicated that entities defined in our information model were suitable for extracting clinical terms from radiology reports, and the model was sufficiently generalizable to be used with dataset from other institutions.

End-to-End Approach for Structuring Radiology Reports.

Radiology reports include various types of clinical information that are used for patient care. Reports are also expected to have secondary uses (e.g., clinical research and the development of decision support systems). For secondary use, it is necessary to extract information from the report and organize it in a structured format. Our goal is to build an application to transform radiology reports written in a free-text form into a structured format. To this end, we propose an end-to-end method that consists of three elements. First, we built a neural network model to extract clinical information from the reports. We experimented on a dataset of chest X-ray reports. Second, we transformed the extracted information into a structured format. Finally, we built a tool that enabled the transformation of terms in reports to standard forms. Through our end-to-end method, we could obtain a structured radiology dataset that was easy to access for secondary use.