Visualise and Diagnose Patients in 2D & 3D

using 3DICOM for Diagnostic Review

Disclaimer: 3Dicom MD is FDA510(k) cleared for Diagnostic Review in the USA. In all other jurisdictions, it is only intended for research, scientific & educational purposes only.

Create training datasets for medical AI models with medical image labelling & annotations

The use of artificial intelligence in medicine and particularly medical imaging is growing a massive Compound Annual Growth Rate (CAGR) of 30.4%.

Before AI models can make it to market, vast datasets of labelled medical images are required to train convolutional neural networks to identify and automatically segment organs-at-risk, various pathologies and anatomical structures.

3Dicom R&D allows for creation of highly accurate segmentation of anatomy with both manual and rule-based segmentation and island removal which can then be exported as multi-class and binary masks for use in these training datasets as well as for the creation of physical 3D printed anatomical models.

Our development team are currently working on APIs and SDKs to allow researchers to integrate and test the outputs of their machine learning and AI models inside of the R&D software for a full end-to-end medical image research tool.

Leverage existing scan databases and teach medical students with virtual models of real anatomy and pathology

Cadaveric studies provide medical students with an unparalleled practical exposure to human anatomy however, the cost of cadavers is high and limits their learning to one patient’s body and pathologies.

By combining 3Dicom R&D’s advanced 3D rendering, realistic colour rendering and leveraging the thousands of open source CT & MRI scans with diverse pathologies from databases such as The Cancer Imaging Archive, you can provide your students with wider exposure to radiological images with virtual anatomy and pathology.

Further enhance virtual pathological scans with 2D & 3D labels provided by our annotation tool and segmenting pathology.

Convert 2D radiological images into 3D printable anatomical models within minutes

A full suite of segmentation tools allow for researchers, educators and even students to segment particular anatomical structures with different colours and labels.

Using semi-automated techniques such as threshold flood-fill, level tracing, and island removal, scans can be rapidly segmented with small edits made manually.

Whilst in the software, segments can be viewed in 3D and overlaid on the initial scan to provide a contextual understanding of that anatomy, pathology or even implant.

Segmented anatomy and pathology can be exported to STL, OBJ or PLY file types for use in modeling software and also for the manufacture of physical 3D anatomical models using 3D printing or traditional manufacturing.

In-Silico Modeling & Integration of Medical AI Models

We’re constantly working on new integrations with 3rd party programs and toolkits to allow for in-silico modeling with full 3D volumetric and materials analysis.

Users also have the ability to use the in-built Medical Computer Aided Design (MCAD) functionality to import, manipulate and position medical devices from screws to patient-specific implants ‘into’ the virtual patient’s anatomy.

Test your hypotheses and the fit and design of your next medical device with virtual patients created from real patient’s radiological images.

Conceptual Image for 2D Measurement Tool

Cutting Edge Colour Surface Rendering with Raytracing & 3D Transfer Functions

Leveraging advanced post processing techniques in graphics and game development, the 3D volume rendering of greyscale MRI and CT scans now features RayTracing and 3D transfer functions to ‘map’ an accurate colour to every voxel in an image.

Whilst not suitable for diagnostic imaging, adding colour to standard MRI & CT scans in 3D provides a greater level of understanding for patients and is a greater addition to your next publication, presentation and marketing materials.

As with all features in 3Dicom, the colour rendering of scans works on both Windows and Mac OS and has been optimised to reduce the required graphical processing power.