LifeVoxel.AI White Paper: Interactive Streaming & AI Cloud Platform
The field of medical imaging today faces a number of challenges. Medical imaging market has the stringent requirement for privacy, reliability and accuracy requirement that demands utmost performance, precision and quality. These problems can be rectified via the deployment of powerful technology alternatives. Data is stored at an unprecedented amount but knowledge within it is unattainable:
- Downloading large data to end user is required and time consuming for viewing or analysis
- Computational resources for generating views or analysis of large data is expensive
- Bandwidth and latency varies to reliably deliver dynamic content interactively
The image below outlines the various issues that are present in the healthcare imaging space. Market leaders aren’t offering a viable solution and continuing to market their existing two decades old outdated RIS PACS Software technology.
When it comes to the consumer side, the issue of cost is a major factor. Often many consumers settle for lower grade options due to a misconstrued notion that the best technology is unaffordable for them. The problem of compatibility across devices and platforms also needs to be addressed.
Medical Imaging Problem definition
In United States of America, they spend $100 Billion on 500 Million medical scans every year for these purposes. Insurance reimbursements provide financial support towards the cost of these scans.
Radiology, Cardiology, Oncology and Pathology are medical specialties that employ the use of imaging to both diagnose and treat disease visualized within the human body. Practices use an array of imaging technologies such as X-ray radiography, ultrasound, computed tomography (CT), nuclear medicine, positron emission tomography (PET) and magnetic resonance imaging (MRI) to obtain images of the internal organ.
When a patient visits his/her primary care physician or specialist, the doctor orders an imaging procedure when further information about the patient’s morphology is needed. A diagnostic imaging scan is performed at an imaging center or hospital, and presented to a radiologist, who then interprets the images for the purpose of diagnosing and treating diseases. The radiologist creates a report of the finding, and the radiologist sends the result to the patient’s doctor who ordered the scan. The doctor subsequently incorporates the findings in the prognosis of the patient’s condition.
Radiology images enables the detection of diseases occurring within one’s body, and early detection of a serious illness can ensure longevity. Visualization is the key.
The misdiagnosis rate is at an all-time high:
- Researchers believe that 15% of diagnoses are wrong.
- More than 1 in 3 cranial CTs are misread by Emergency Department doctors.
- As many as 12.7% of stroke admissions each year are read missions after a previous misdiagnosis.
- Research has shown that for some types of cancers, misdiagnosis can be as high as 28% to 44%.
Hence, as the U.S. works to reform the way we pay for medical care—eliminating incentives to do more and paying providers based on quality—focusing on accurately diagnosing the patient should be paramount.
The possibilities on the Cloud
In the past, medical scans were on films. As you can imagine, this became a burden to healthcare providers. So, the concept of Picture Archiving and Communication Systems (PACS) was introduced in the 1980’s and became widespread in the 1990s. Ancillary software applications emerged, which are increasingly interoperable with PACS, including RIS, Sharing, Archival, Post Processing, and CAD. Over 99% systems are on-premises, cloud medical imaging services is projected to reach $400 M in 2018, growing at 27% CAGR.
Growth of scan data places ever increasing burdens on software and network efficiency in storing, retrieving or manipulating images. Physicians need powerful PCs or expensive custom workstations to manipulate the data on existing medical imaging systems.
Ultimately, the cost of supporting the electronic storage and distribution of medical scans is borne by healthcare providers. The Affordable Care Act is adding to the burden; providers need to cope with reduced reimbursement, yet make medical scans accessible electronically, while improving patient care.
Conventional cloud technology acts as a central storage. It requires data to be moved to end user before consumption which is impractical for medical images. For example, a typical breast tomosynthesis study is at least GB in size per patient scan, making it near impossible to transmit over Internet in real-time. Therefore, remote visualization is the key.
However, diagnostic remote visualization over Internet is challenging because of bandwidth and latency issues. Furthermore, the processing of these images for remote visualization needs to be offloaded to the cloud.
Conventional cloud is not apt to scale when processing immense data. One will need enormous processing capabilities built into the cloud. The readily available supercomputing processing capabilities exist on GPU, which can be innovatively integrated in the cloud. GPUs are great for artificial intelligence algorithms enabling them to execute in real-time. Machine Learning through AI algorithms is key to providing clinicians with decision support and other analytics involving imaging data. As competition begin to appreciate these technology hurdles for cloud medical imaging, they will encounter LifeVoxel.AI’s patents.
- Cloud access to patient records and history improves administration, billing, and clinical treatment
- Computers will need to assist clinicians to ensure patient safety and support clinical decisions
- Cloud will only reach widespread adoption when technological and economic issues are addressed
- Advanced visualization (3D/VR/AR) and Artificial Intelligence shall bring clarify to any device
GPU Computing for Latency and Efficiency
- GPU Computing is a powerful piece of technology and can process information at a rare of 3000x more than conventional CPU computing.
- GPU processing power has been surpassing CPU exponentially enabling class of algorithms that couldn’t be considered before to solve challenging problems.
- GPUs are considered as the heart of AI and can be put to great use in improving the quality and latency sensitivity in medical imaging. GPUs occupy a very small amount of space unlike the monstrous hardware space taken by CPUs.
- GPUs can perform highly powerful functions when they are required and GPUs also help take medical images to the cloud efficiently.
- AI algorithms has a distributed pattern. Deep neural network and AI in machine learning can be categorized as parallel problems which means parallel super computing solutions like GPUs can speed up 90%.
- GPUs are best suited for speeding up distributed pattern of AI algorithms where each unit in distributed system works independent of the other units. A neural network will learn several times faster on a GPU than a CPU.
- Artificial Intelligence AI and machine learning play a vital role in continuing competitive advantage and delivering fantastic user experience.
- GPU is very precious as it accelerates the tensor processing necessary for deep learning applications. A GPU has its own memory that keeps the whole graphics image as a matrix.
- GPU calculates change in the image using tensor math, whenever any change is made to the image like adding color to the pixel, GPU performs this process much faster instead of redrawing the entire screen every time the image changes. These deep learning approaches have shown impressive performances in resembling humans in various fields, including medical imaging.
Unique Solution Architecture
LifeVoxel.AI is a leap in medical imaging software solution. Existing technology download raw DICOM scans before viewing. This require immense bandwidth and processing capabilities of the end user device. Instead, LifeVoxel integrated Artificial Intelligence (AI), Cloud Computing and Graphical Processing Unit technology to deliver server side visualization over Internet on any browser and any device. This helped improve visualization and allowed for the transmission of high-quality images while eliminating the need to transmit the raw scan. Advanced 2D and 3D diagnostic visualization of medical images are delivered instantaneously adapted to physicians’ and their devices requirements using AI algorithms such as evolutionary and deep learning. This high-end technology of intuitive AI and adaptive imaging can help minimize and or potentially even eliminate the incidence of human error.
LifeVoxel.AI Cloud, a cloud visualization platform offers a unique solution that overcome the limitations:
- Utilizes cost effective yet super computing power using Graphics Processing Unit (GPU)
- Proprietary artificial intelligence based patented algorithms running on GPU supercomputing
- Fast interactive viewing over Internet allows data manipulation remotely
- Predictive and machine learning unleashes knowledge stored within enormous data
- Innovation by others is fostered through secure web based API for developers
Patented AI Based Cloud Visualization (12 International Patents)
- The platform uses virtual views to deliver high quality viewing over Internet without transmitting large data to end users.
- Eliminating up to $26 billion a year in unnecessary radiology spending and reducing misdiagnosis rate is a national priority.
- Lifevoxel.AI Cloud can pioneer game-changing delivery to physicians of immediate diagnostic-quality access to prior imaging, and this will curb unnecessary use.
- Bandwidth and latency limitations of the internet are overcome by using predictive buffering.
- The platform can transform best practices, assure quality, and improve outcomes by delivering a platform for artificial intelligence on Big Data analytics.
- Immediate Diagnostic-Quality Access, and Big Data Analytics – can be performed fastest, cheapest, and most effectively using the Lifevoxel.AI Cloud platform, without the inherent limitations of competing approaches based on legacy technology.
- Streams visualization real-time using patented server-side technology. LifeVoxel.AI provide instant access to time critical stroke patient images to remote physicians and patients in 3D
- LifeVoxel.AI with its twelve patents is disrupting medical imaging market by offering technological advancement and economic advantage using GPU Cloud.
- The patented platform fosters innovation by others, to bring the “best of breed” solutions to customers’ needs.