The digital model (.stl file or other acceptable formats) should accurately represent source DICOM data. Segmentation techniques are categorized into one of the three methodology types as described below. Poorly rendered models that deviate too far from the source image will be rejected.

De-identified image datasets in the Digital Imaging and Communications in Medicine (DICOM) format are required to be submitted with the model, both for the peer-review process as well as to provide a reference for segmentation analysis. A 3D imaging dataset which is free of artifact with resolution between 1 and 1.5 mm isovoxels is ideal. Datasets with resolution or artifacts that do not allow for accurate reproduction of anatomy will not be accepted due to the increased risk of false data transmission to the digital model.

Sharing case study details and raw medical image data is critical to the Library. Extra care must be taken to protect patient privacy when making clinical data available to the public. Concern for HIPAA compliance may create a barrier for clinicians who wish to submit their 3D case reports. We have extensively evaluated HIPAA requirements to ensure that contributions to the library do not violate patient privacy. If no Personal Health Information (PHI) is included in the data, then the data does not fall under the HIPAA Privacy Rule [15]. All content contributed to the Heart Library will not contain PHI. The submitting clinician is responsible for reviewing and removing any PHI prior to uploading data for peer review. Two methods to de-identify PHI are the “Safe Harbor” method and/or the use of a statistician to limit risk. The former requires removal of all 18 identifiers enumerated at section 164.514(b)(2) of the regulation (Table 17.2) and for the purposes of the Heart Library is the only acceptable method for submission.

Utmost care must be taken by the submitting physician or organization in de-identifying DICOM data. The NIH 3D Print Exchange also has tools in place to automatically remove PHI that is named in the HIPAA Privacy Rule (see Table 17.2). Having redundancy in checking for privacy data will significantly minimize liability.

The peer-review aspect of the Library is essential to providing verification that a digital 3D model is a true representation of the source 2D medical image series. In the field of congenital heart disease, expertise in both congenital heart imaging and 3D model creation is not well-established. It is therefore difficult to define quality in an emerging field such as this. In collaboration with recognized leaders in 3D printing and modeling of cardiac anatomy, we established standards and methods for model creation and evaluation.