Makeup lovers might only know to contour as a way of highlighting facial features, but a technique by the same name is used to determine extremely precise radiation therapy treatment. Once the doctors have identified the size and shape of the tumor in diagrammatic detail, contouring helps them design specific target volumes for radiation therapy.
This idea of deep learning to improve the extremely time-taking, labor-inducing is developed and improvised by Researchers from the University of Texas, MD Anderson Cancer Center, in Houston. Contouring is done by examining the medical images of the tumour and accordingly device particular target volumes of rumours and of the surrounding tissues which help to determine the exact amount of radiation to be used. This results in an increased rate of perfection, however, with a considerable scope for human error. If a tumor is located in a very vulnerable place, a little bit of extra radiation would affect the neighboring healthy tissues. If the amount of radiation isn’t sufficient to kill off a tumor and it may continue growing.
Contouring is also an extremely subjective affair, that is it varies from doctor to doctor. The MD Anderson researchers have developed a deep learning process for an objective approach to the tasks of contouring. Their AI-based approach would control the variability by different medical professionals and prevent bias, decreasing the chance of missing the target tumor or over treating normal tissues.
The researchers used NVIDIA Tesla GPUs on the Maverick supercomputer at the Texas Advanced Computing Center (TACC) and the cuDNN-accelerated Tensorflow deep learning library to analyze data from 52 oropharyngeal cancer patients.
“If we were to do it on our local GPU, it would have taken two months,” said Carlos Cardenas, Ph.D., a researcher working with principal investigator Laurence Court, both at MD Anderson, in an interview with TACC. “But we were able to parallelize the process and do the optimization on each patient by sending those paths to TACC, and that’s where we found a lot of advantages by using the TACC system.”
This algorithm based network to identify and recreate physical contouring pattern is efficient and neutral. This development will be significant in cancer treatment by making radiation therapy for target-specific and efficient and it will reduce uncertainty and damaging healthy tissues. Automating time-consuming processes like contouring and radiation treatment planning could greatly benefit patient care in clinics with low resources, many of which are found in low and middle-income countries because of its cost-effectiveness.
The researchers’ next steps are to translate this process into clinical use, ultimately providing physicians with better information that can lead to improved patient treatments and outcomes.
This is the end that they are aiming at and for that, they’re developing a radiation planning assistant — a fully automated radiation therapy planning tool, which hopefully would launch to partner cancer centers in low-income countries by early next year.
Source: Deep Learning on Medium