Anonymization and validation of three-dimensional volumetric renderings of computed tomography data using commercially available T1-weighted magnetic resonance imaging-based algorithms

Purpose

Previous studies have demonstrated that three-dimensional (3D) volumetric renderings of magnetic resonance imaging (MRI) brain data can be used to identify patients using facial recognition. We have shown that facial features can be identified on simulation-computed tomography (CT) images for radiation oncology and mapped to face images from a database. We aim to determine whether CT images can be anonymized using anonymization software that was designed for T1-weighted MRI data.

Approach

Our study examines (1) the ability of off-the-shelf anonymization algorithms to anonymize CT data and (2) the ability of facial recognition algorithms to identify whether faces could be detected from a database of facial images. Our study generated 3D renderings from 57 head CT scans from The Cancer Imaging Archive database. Data were anonymized using AFNI (deface, reface, and 3Dskullstrip) and FSL's BET. Anonymized data were compared to the original renderings and passed through facial recognition algorithms (VGG-Face, FaceNet, DLib, and SFace) using a facial database (labeled faces in the wild) to determine what matches could be found.

Results

Our study found that all modules were able to process CT data and that AFNI's 3Dskullstrip and FSL's BET data consistently showed lower reidentification rates compared to the original.

Conclusions

The results from this study highlight the potential usage of anonymization algorithms as a clinical standard for deidentifying brain CT data. Our study demonstrates the importance of continued vigilance for patient privacy in publicly shared datasets and the importance of continued evaluation of anonymization methods for CT data.

Three-Dimensional Micro-Computed Tomography of the Adult Mouse Ovary

In the mouse ovary, folliculogenesis proceeds through eight main growth stages, from small primordial type 1 (T1) to fully grown antral T8 follicles. Most of our understanding of this process was obtained with approaches that disrupted the ovary three-dimensional (3D) integrity. Micro-Computed Tomography (microCT) allows the maintenance of the organ structure and a true in-silico 3D reconstruction, with cubic voxels and isotropic resolution, giving a precise spatial mapping of its functional units. Here, we developed a robust method that, by combining an optimized contrast procedure with microCT imaging of the tiny adult mouse ovary, allowed 3D mapping and counting of follicles, from pre-antral secondary T4 (53.2 ± 12.7 μm in diameter) to antral T8 (321.0 ± 21.3 μm) and corpora lutea, together with the major vasculature branches. Primordial and primary follicles (T1–T3) could not be observed. Our procedure highlighted, with unprecedent details, the main functional compartments of the growing follicle: granulosa, antrum, cumulus cells, zona pellucida, and oocyte with its nucleus. The results describe a homogeneous distribution of all follicle types between the ovary dorsal and ventral regions. Also, they show that each of the eight sectors, virtually segmented along the dorsal-ventral axis, houses an equal number of each follicle type. Altogether, these data suggest that follicle recruitment is homogeneously distributed all-over the ovarian surface. This topographic reconstruction builds sound bases for modeling follicles position and, prospectively, could contribute to our understanding of folliculogenesis dynamics, not only under normal conditions, but, importantly, during aging, in the presence of pathologies or after hormones or drugs administration.

Learning Acetabular Fracture Classification using a Three-Dimensional Interactive Software: A Randomized Controlled Trial

Acetabular fractures are a real challenge for junior doctors as well as experienced orthopedic surgeons. Correct fracture classification is crucial for appreciating the fracture type, surgical planning, and predicting prognosis. Although three-dimensional (3D) tutorial is believed to improve the understanding of the complex anatomy structure, there have been few applications and randomized controlled trials to confirm it in orthopedics. This study aims to develop a 3D interactive software system for teaching acetabular fracture classification and evaluate its efficacy. Participants were randomly but evenly allocated into either the experimental group (who learned the acetabular fracture classification using a 3D software) or the control group (who used a traditional two-dimensional [2D] tutorial). Both groups were then tasked to classify 10 acetabular fractures and complete a five-point Likert scale on their satisfaction of each learning modality. To calculate significance (P < 0.05), independent t-test was used for normally distributed data whereas Mann-Whitney U test for non-normally distributed data. The experimental group significantly outperformed the control group (t (28) = 2.526, P = 0.017) with identifying correct acetabular fracture classification. Moreover, Likert scale score in the experimental group was also significantly higher than in the control group (Z = 2.477, P = 0.013). This 3D classification software has objectively and subjectively showed an advantage over the traditional 2D tutorial, resulting in an improved classification accuracy and higher Likert scale score. The 3D software has the potential to improve both clinical knowledge as well as identifying correct patient management in orthopedics.