With this document, I aim to arm you with a working knowledge of key concepts relevant to RedBrick AI. I recommend consuming all the content, even if it doesn’t seem directly related to your role. This material is just a starting point for your own exploration in developing expertise in our market & product — please dive deep in areas you find useful to your & relevant to your work. A deep understanding of the market, problem, and customer profile is essential in all aspects of company building.
RedBrick AI is a SaaS platform for building medical imaging artificial intelligence. Our current focus is on offering tools for preparing high-quality annotated radiology datasets.
Leading healthcare organizations like Mass General Brigham Hospital, Qure AI, and Hyperfine use RedBrick AI to prepare their ground truth datasets. These datasets are then used to train deep-learning models for clinical environments. Our customers also use RedBrick AI to run validation studies to prove their AI systems' effectiveness to regulatory bodies like the FDA and CE.
Medical imaging is the process of imaging the interior of a body for clinical analysis. Everyone will be familiar with common types of medical imaging: CT scans, X-rays, MRIs, etc. Medical imagery is an incredibly important source of evidence in clinical analysis and medical intervention, so much so that medical imagery comprises about 90% of all data in healthcare 🤯.
This explosion of data, exponential progress in deep learning, and a shortage of radiologists (radiology practitioners) make using AI in radiology a promising use case. Modern AI systems are being used to improve radiologist workflows by assisting in diagnoses, triaging cases, and quantifying features in medical imagery.
Roentgen’s wife’s hand.
In 1895 physicist Wilhelm Roentgen discovered X-rays — aptly named to underline that their nature was unknown — the first key innovation in medical imagery.
X-rays are a penetrating radiation that images the body by passing through different body parts in different amounts. The variation in X-rays is measured, and a 2-dimensional image is produced from that.
Many years later, Godfrey Hounsfield used X-rays (taken from multiple different angles) to create a 3D representation of the object being imaged — this was the creation of the Computed Tomography (CT) scan. Hounsfield was awarded the Nobel prize for the invention in 1979.
The next key innovation in medical imaging/radiology was incorporating computers in practice. Until the 1970s, medical images were managed in hard copies, i.e., they were stored and viewed in physical form. With the introduction of the Picture Archiving and Communication System (PACS) and the Digital Imaging and Communications in Medicine (DICOM) standard, hospitals had a standardized & digital way of storing, sharing & viewing medical images.
The digitization, standardization, and proliferation of medical imagery paved the way for the current innovation — using Artificial Intelligence in radiology. Since the first FDA-cleared AI medical device in 1995, 500+ have been cleared for clinical use. RedBrick’s mission is to be the platform that all AI medical devices are built using.
The supply of AI tools for radiologists is extensive, and most radiology practices are eager to adopt them. The hype cycle has passed, and the industry is now quantifying the benefits of using these systems and working toward mass adoption. Of all the specializations, Radiology (and adjacent fields that use medical imagery) has seen the greatest adoption of AI.
The application of Artificial Intelligence in radiology can really be bucketed into 5 groups: