Surgeons tend to perform less than optimally when scanning magnetic resonance imaging (MRI) of breast cancer patients, primarily due to a lack of training in radiology. A new virtual reality tool has improved this ability, however, researchers reported in JCO Cancer Clinical Informatics.
Mohamed El Beheiry, PhD, of the Institut Pasteur in Paris, and colleagues conducted a study of 18 breast surgeons, half of whom were residents and the other half practicing surgeons. The surgeons analyzed the MRIs of 25 patients using a standard slice-based visualization tool and a 3D virtual reality tool called DIVAwhich instantly generated 3D patient reconstructions from original MRI scans in their native version DICOM file format.
Surgeons analyzed scans much faster and more accurately with DIVA. The median time it took them to determine the number of lesions and their location was 145 seconds with slice visualization and 37 seconds with DIVA, almost four times faster (P<0.001). Improvements were also made to determine which breast and which quadrant contained lesions.
“In DIVA, we demonstrate a method to potentially significantly improve surgeons’ skills in reading MRI exams without the need for massive retraining,” El Beheiry and colleagues wrote. “The reported increase in the efficiency of medical image reading has the potential to improve surgical planning and significantly improve communication between surgeons and radiologists.”
In the following interview, El Beheiry, an engineer and physicist, elaborated on the technique and its potential benefits.
Can you describe the 3D virtual environment created by DIVA? What do doctors see and how can they interact with it?
El Beheiry: Physicians using DIVA are introduced to an immersive 3D representation of their patient’s MRI in virtual reality, which can be viewed as an “avatar” of the patient. The avatar is depicted in exact physical scale and can be fully explored using a handheld virtual reality controller. For example, the doctor can virtually “cut” the avatar to better visualize the internal organs and also perform precise 3D measurements.
What special equipment is needed?
El Beheiry: DIVA simply requires a computer powered by a VR-capable graphics card (e.g. NVIDIA 10 series equivalent or better) running Windows and a commercial PC VR headset (our favorite headset is the HTC VIVE).
One of the most notable findings from your study was how quickly surgeons were able to analyze breast MRIs with DIVA. Please tell us more about it.
El Beheiry: This finding, while remarkable, was not necessarily surprising. The immersive nature of patient reconstructions generated by DIVA provides a level of abstraction from the medical image. Surgeons don’t look at grayscale x-ray sections; they look at a realistic virtual representation of their patient. Interactions with these representations are designed to be simple, fast and very intuitive.
The implications are potentially very significant. For a high-volume cancer center like the Institut Curie – more than 3,000 breast surgeries per year – the time saved per surgeon is considerable (several man-days per year). What is further remarkable is that the accuracy of image analysis is also improved with these time savings.
You have noted additional medical benefits resulting from the use of DIVA. What were they?
El Beheiry: The departments of the Institut Curie which deployed DIVA report a better understanding and culture of the medical image in the broad sense. DIVA is sort of a shortcut to understanding medical images for non-radiologists. Other medical benefits include:
- We demonstrate the general skill level of surgeons in reading MRI sequences
- Our results show a considerable reduction in MRI analysis time per patient thanks to DIVA, which is a time saver for doctor-patient interactions
- DIVA opens up the possibility that medical imaging information can be easily shared and understood by patients to further involve them in treatment decisions.
What future technical developments will further improve DIVA?
El Beheiry: DIVA technology is currently commercialized by AVATAR MEDICAL SAS, a startup founded in 2020. The latest iterations of the technology include, perhaps ironically, an X-ray interface fully synchronized with the virtual reality context featured in our article from JCO. The idea here is to wrap surgeons in the culture of the medical image. The wealth of clinical information contained in these data must be made accessible to all clinicians, and future developments will be conducted in this spirit.
Read the study here.
The study was supported by the Institut Pasteur, Gilead Science and the National Research Agency.
El Beheiry is an employee of Avatar Medical. The Institut Pasteur and Avatar Medical have an R&D agreement relating to the development of certain software components based on DIVA technology.