Anonymizing Radiographs Using an Object Detection Deep Learning Algorithm

Abstract

Radiographic markers contain protected health information that must be removed before public release. This work presents a deep learning algorithm that localizes radiographic markers and selectively removes them to enable de-identified data sharing. The authors annotated 2000 hip and pelvic radiographs to train an object detection computer vision model. Data were split into training, validation, and test sets at the patient level. Extracted markers were then characterized using an image processing algorithm, and potentially useful markers (eg, "L" and "R") without identifying information were retained. The model achieved an area under the precision-recall curve of 0.96 on the internal test set. The de-identification accuracy was 100% (400 of 400), with a de-identification false-positive rate of 1% (eight of 632) and a retention accuracy of 93% (359 of 386) for laterality markers. The algorithm was further validated on an external dataset of chest radiographs, achieving a de-identification accuracy of 96% (221 of 231). After fine-tuning the model on 20 images from the external dataset to investigate the potential for improvement, a 99.6% (230 of 231, P = .04) de-identification accuracy and decreased false-positive rate of 5% (26 of 512) were achieved. These results demonstrate the effectiveness of a two-pass approach in image de-identification. Keywords: Conventional Radiography, Skeletal-Axial, Thorax, Experimental Investigations, Supervised Learning, Transfer Learning, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2023 See also the commentary by Chang and Li in this issue.

Keywords: Conventional Radiography; Convolutional Neural Network (CNN); Experimental Investigations; Skeletal-Axial; Supervised Learning; Thorax; Transfer Learning.

Figure 1:

Overview of the proposed pipeline. * The patient identifiers were in the same place, but they were taken out and replaced with fake numbers, using the same font settings.

Figure 2:

(A) Marker detection on pelvic radiographs. Note that the marker areas were obfuscated to ensure patient privacy. (B, C) Results of the localizer model's performance on out-of-domain chest radiographs before (B) and after (C) fine-tuning. The red boxes are zoomed-in versions of the marker areas that were detected. As shown, the fine-tuned model draws tighter bounding boxes with fewer nonmarker areas and also has more specificity, not mistaking nonmarker densities, like electrocardiography leads, compared with the initial model.

Figure 3:

Cases of model failure on the (A) internal and external datasets (B) before and (C) after fine-tuning.