A I can help stratify COVID-19 risk on lung ultrasound

By Erik L. Ridley, AuntMinnie.com staff writer

June 1, 2021 -- Deep-learning algorithms can be used to automatically provide risk scores on lung ultrasound exams in COVID-19 patients, researchers from Italy reported in an article published online May 27 in the Journal of the Acoustical Society of America.

 A multi-institutional team of researchers trained deep-learning models to automatically score and provide semantic segmentations of lung ultrasound exams in COVID-19 patients. In testing, the deep-learning algorithms yielded a high level of agreement with lung ultrasound experts for stratifying patients as having either low- or high risk of clinical worsening, according to the group led by corresponding author Libertario Demi, PhD, of the University of Trento.

"These encouraging results demonstrate the potential of [deep-learning] models for the automatic scoring of [lung ultrasound] data, when applied to high-quality data acquired according to a standardized imaging protocol," the authors wrote.

Although lung ultrasound has been shown to be useful for evaluating COVID-19 patients, the modality is often limited to the visual interpretation of ultrasound data. As a result, there are concerns over reliability and reproducibility, according to the researchers.

Following up on prior work to develop standard lung ultrasound protocols and training deep-learning algorithms for COVID-19 patients, the researchers wanted to compare the performance of their models with that of expert clinicians. The first algorithm labels each video frame with a score, while the second provides semantic segmentation -- assigning one or more scores to each frame.

They used their algorithms to evaluate 314,879 ultrasound video frames from 1,488 lung ultrasound videos on 82 COVID-19 patients. The ultrasound exams were acquired using multiple ultrasound scanner types at the Agostino Gemelli University Hospital Foundation in Rome and the San Matteo Polyclinic Foundation in Pavia, Italy. None of the images were used in training the models.

Left: Examples of lung ultrasound images corresponding to score 3 (top), score 0 (middle), and score 2 (bottom). Right: the corresponding semantic segmentations obtained with the described deep-learning algorithms. Color maps are informative of the score level (red for score 3, blue for score 0, orange for score 2), as well as indicate the relevant part of the image that determines the score. These maps explain the decision progress of the algorithms. Images and caption courtesy of Libertario Demi, PhD.

The algorithm's scores for each frame are then used to generate an aggregate score for each video. The researchers tested two methods for producing an aggregate score: one based only on the labeled video frames and another that combined the labeled frames with the segmented frames.

These video-level scores were then compared with the risk scores provided by expert clinicians.

AI algorithm performance for risk stratifying COVID-19 patients
	AI risk scores based on only labeled frames	AI risk scores based on combination of labeled and segmented frames
Agreement with expert clinicians	83.3%	86%

Although the combined approach generally outperforms the first method, the improvement isn't significant. What's more, the first method seems more stable to threshold variations, according to the researchers.

"Therefore, the first approach can be considered as a self-standing strategy to classify [lung ultrasound] videos starting from a frame-based classification," the authors wrote. "Nevertheless, the semantic segmentation remains essential for clinicians, as it provides the explainability of the decision by highlighting the specific [lung ultrasound] patterns."

In future work, the researchers plan to expand their existing database and train deep-learning algorithms on video-labeled data instead of just frame-based labeled data. That approach would be more consistent with the process used by clinicians when evaluating lung ultrasound videos, they said.

How to implement POCUS at your hospital

By Amerigo Allegretto, AuntMinnie.com staff writer

June 1, 2021 -- So, you want to integrate point-of-care ultrasound (POCUS) into your hospital enterprise image network. But how? Breaking down what goes into a successful POCUS implementation was the subject of a May 26 talk at the Society for Imaging Informatics in Medicine meeting.

Successfully implementing POCUS requires swift strategy and teamwork, including taking advantage of existing imaging technology, according to registered nurse Laurie Perry from Cincinnati Children's Hospital in Ohio.

"Point-of-care ultrasound is only going to grow, and there will be divisions that continually need to be implemented. You'll need to do that quickly," she said.

POCUS has been looked at in recent years as a convenient clinical tool that can bring imaging to a patient's bedside. Advantages include portability and cost-effectiveness.

However, there are some limitations to POCUS, Perry said. One is that images remain on the ultrasound device and are not clinically available to anyone else taking care of the patient. POCUS images also can't be compared with other specialty images, and there are challenges to getting reimbursed.

Although many divisions at Cincinnati Children's Hospital wanted to use POCUS, Perry said they were unable to bill for the procedure until providers had been credentialed and images stored in the enterprise archive.

"If you're not storing images, you cannot bill for the ultrasound," Perry said.

To integrate POCUS with the rest of the enterprise, the implementation team needed to create a standard POCUS workflow within the electronic health record. Then, it needed to work with business managers to help them to understand the process for billing imaging studies.

The team that was assembled for this included clinical and business leaders from clinical divisions, imaging informatics professionals, electronic medical records analysts, a project manager, a trainer, and an enterprise imaging physician leader.

The group solved the existing challenges by working with clinical leaders and business managers in creating a process that fits into the existing workflow. Team members utilized ad-hoc order creation and automated as many steps as possible. They also worked with the radiology business director to identify and create appropriate procedure codes.

"All radiology and cardiology, including ultrasound, is sent to the hospital's enterprise archive. We've acquired and sent over 5,000 POCUS studies to the archive," Perry said. "We have billed more than $2.25 million of new services for our organization."

Children's Hospital began using POCUS for anesthesia in 2015 and has expanded its use in other care units and departments, such as emergency, pediatric intensive care, and cardiac intensive care among others. The hospital is currently working on implementing POCUS for physical medicine and rehab and its gastrointestinal division.

Perry said the hospital has used the following guiding principles over the years to successfully implement POCUS into its departments:

• Take advantage of DICOM to power workflow.
• Automate workflow for providers as much as possible.
• Provide a link to the images in the electronic medical record, preferably within the appropriate context for the division.
• The goal should be to create a standard, reproducible process.

Perry also said implementation teams should evaluate a division's current ultrasound machines for such things as WiFi capability, DICOM capability, and security requirements. Credentialing providers, as well as determining procedures and billing, developing the electronic medical record, and training users are also needed.