Whitney J. Palmer

Chest X-ray can prospectively predict outcomes for a patient with COVID-19 in high pre-test probability settings or a quick turnaround of RT-PCR testing.

In a poster presented during the 2021 ARRS Virtual Annual Meeting, Russel Reeves, M.D., a radiology resident at Thomas Jefferson University in Philadelphia, described how this type of imaging can potentially help providers prepare for the type of care a patient will need.

For additional ARRS 2021 Virtual Annual Meeting conference coverage, click 

“We developed an accurate and reliable tool for classifying COVID-19 severity, which can be used both at the attending chest radiologist and junior resident level,” he said. “This study identifies the laboratory, clinical and radiographic data that predict important patient outcomes, such as death, intubation, and the need for chronic renal replacement therapy (CRRT).”

For their study, Reeves’ team identified 240 patients admitted to an urban multi-center health system between March 16, 2020, and April 13, 2020, who had RT-PCR-confirmed COVID-19, including 142 males and 98 females with an average age of 65.

Chest X-rays captured at the time of admission were evaluated by three cardiothoracic radiologists and three diagnostic radiology residents who scored the images based on extent and severity of COVID-19 pneumonia. In addition, they also collected information on demographics, clinical characteristics, as well as admission laboratory values and assessed the inter-rater reliability among attending providers, residents, and the combined group of readers.

Based on their analysis, inter-rater reliability for chest X-ray scoring was high among the attending, residents, and combination group (a 0.686 average). Their evaluation showed no difference in outcomes for gender, race, ethnicity, or among those who had lung cancer or chronic obstructive pulmonary disease histories.

The team did note that even though disease severity seen on chest X-ray was an independent predictor of death, it was a non-independent predictor of CRRT and intubation – but not extracorporeal membrane oxygenation.

These results, the team said, add to the existing knowledge base around the role chest X-ray can play in COVID-19 care.

“These findings are supported by prior work, validating the utility of [chest X-ray] for patient prognostication, while adding new insight into COVID-19 infection during the initial presentation,” they said.

For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-Newsletter 

Articles

In high pre-test probability and quick RT-PCR settings, chest X-ray scoring can be predictive. 

Chest X-ray Can Predict Outcomes for Patients With COVID-19

The push toward greater utilization of artificial intelligence in radiology (AI) shows no signs of slowing, so it would be a good idea to brush up on the legal aspects of using it, a group of industry experts said.

In an April 9 perspective article published in 

 legal and radiology experts from Harvard noted that the use of AI in diagnostic imaging is largely uncharted territory. Some risks to providers, practices, and technology developers have been identified, but more could be revealed over time.

To help you prepare, H. Benjamin Harvey, M.D., J.D., a neuroradiologist from Massachusetts General Hospital, and Vrushab Gowda, J.D., a Harvard Law School graduate who is also pursuing his medical degree from the University of North Carolina at Chapel Hill School of Medicine, discussed some specifics of the liabilities around AI use in musculoskeletal imaging.

These cases typically deal with medical malpractice, Harvey and Gowda said. While plaintiffs have the burden of showing four distinct occurrences of negligence, courts will also rely on professional society standards, as well as community practice and department protocols. Consequently, it is vital to create guidelines, complete sufficient user training, and correctly integration AI.

In addition, AI is extremely helpful, but if you rely on it alone – rather than using the tool and deciding to dismiss its findings and recommendations – your malpractice risk will be greater, they said.

When it comes to using AI with medical imaging, there are two main concerns, they said – do you let your patients know you’re using AI, and if you do, what do you need to tell them?

Given that AI is largely used for research investigations currently, this issue hasn’t crept up much. But, once the tools are implemented clinically, the story will be different, especially if the technology eventually becomes autonomous further downstream.

Ultimately, Harvey and Gowda said, proper handling of informed consent will likely require a patient-based or provider-based standard, and the difference between them can be unclear.

“The rub of the matter is that radiologists should be cognizant of their peers’ disclosures and anticipate the sorts of information patients would find important when deploying AI,” they said.

When mistakes happen, patients are going to look for someone to hold accountable, they said. If you’ve used an AI tool that results in a negative outcome, your hospital or radiology practice could end up as the defendant – not you.

This situation of “vicarious liability” can be sticky, they explained, because it puts your practice or facility leadership in the hot seat. Consequently, there is a need for clearly written protocols on how to handle these potential situations.

You aren’t the only one subjecting yourself to liability and risk when you use an AI tool. The developer also bears responsibility for when things go wrong, such as mislabeling findings or recommending a biopsy – or worse, not recommending one.

Under a strict liability framework, if a plaintiff can prove that the mistake is a result of a manufacturing problem – which could be tough to do – they could win their case. Conversely, negligence principles call for examining the features of the healthcare facility rather the product.

Either way, they said, radiologists need to be aware about these possibilities and stay current with what’s going on.

With a lack of clarity around liability and responsibility with the future use of artificial intelligence, Harvey and Gowda both said, as radiologists, you must make sure your voice is part of the discussion around developing stringent AI imaging standards.

“Looking forward, imaging departments should articulate clear protocols for their use, to include procedures in the event of human-AI discrepancy,” they said. “They can be aided by [American College of Radiology]- and [Society of Skeletal Radiology]-validated guidelines, training programs, and model practices for deploying AI technologies.”

Liability and responsibility concerns are not significant with AI use now, but as the tools enter clinical use, that will change. 



4 Must-Knows about Future Radiologist Liabilities with AI

Patients who have metastatic kidney cancer could benefit from undergoing baseline brain imaging, according to a recently published study.

Journal of the National Comprehensive Cancer Network

investigators from both Memorial Sloan Kettering (MSK) Cancer Center and Gustave Roussy Institute, a leading academic cancer center in France, showed that incidental brain metastases can occur in a clinically significant percentage of patients who have newly diagnosed metastatic renal cell carcinoma.

Consequently, they said, implementing brain imaging – even in asymptomatic patients – could positively impact overall survival rates.

For their study, they included 1,689 patients with metastatic renal cell carcinoma who had been considered to participate in a clinical trial at either facility between 2001 and 2019. These individuals had no clinical suspicion of brain involvement, but they had undergone brain CT or MRI.

When researchers evaluated the patients' scans, they found 72 patients (4.3 percent of the group) had asymptomatic brain metastases. Of that group, 22 were diagnosed via screening for first-line studies, and 50 via screening in the treatment-refractory setting. For those patients, the median 1-year overall survival rate was 48 percent, and the median overall survival was 10.3 months.

Brain imaging is routinely implemented for kidney cancer patients who show evidence of central nervous system involvement, but that is not the case for patients who are symptomatic, they said. Given these results, it could be useful to conduct brain imaging in this patient population.

“With 4 percent overall incidence in this cohort, one might conclude that baseline brain imaging should be considered in all patients with metastatic kidney cancer, particularly those with multi-organ involvement and/or pulmonary metastases,” said lead MSK researcher Ritesh R. Kotecha, M.D.

The team also identified other areas of cancer spread in patients who had brain involvement. At initial kidney cancer diagnosis, 43 patients (60 percent) had metastatic disease, but by the time of brain metastasis diagnosis, 62 patients (86 percent) had metastatic disease in two or more additional organ systems. The most common was lung (66 patients, 92 percent), followed by liver and bone (18 patients each, 25 percent).

“In current practice, chest, abdomen, and pelvis are routinely imaged from the time that metastatic disease is first detected, yet many oncologists do not image the brain,” said senior researcher Martin H. Voss, M.D., medical oncologist with MSK.

According to Eric Jonasch, M.D., professor of genitourinary medical oncology at The University of Texas MD Anderson Cancer Center, the findings of this study are important for two main reasons.

“First, they show that the overall prognosis of patients with brain metastases is consistently worse than the broader population of patients with metastatic renal cell carcinoma. We need to develop a deeper scientific understanding of why this patient population has a worse outcome, and we need to include them in future clinical trials,” said Jonasch, who is the vice chair of the NCCN Clinical Practice Guidelines in Oncology Panel for Kidney Cancer. “Second, they underscore the utility for MRI imaging of all patients with metastatic renal cell carcinoma both at initial diagnosis, and at regular intervals, to detect occult brain metastases, since specific treatment strategies are required for this patient population.”

The team did note that their study did not elucidate how frequently brain imaging in these patients should be repeated, pointing to the need for further study.

Study finds clinically significant percentage of patients with metastatic renal cell carcinoma develop asymptomatic brain metastases. 

Brain MRI and CT Could Identify Metastasis in Patients with Kidney Cancer

Low-dose CT (LDCT) isn’t just for lung cancer screening. Data gathered from the scans can also help predict a patient’s five-year risk of death from cardiovascular disease.

investigators from The Netherlands detailed how implementing a deep learning algorithm that is fueled with information captured during a low-dose CT exam for lung cancer screening can offer valuable details about calcification in the heart and the aorta.

Using LDCT in this way is a two-for-one opportunity to screen patients for both potentially deadly conditions. The scan is used to detect lung cancer in high-risk individuals, including heavy smokers, but for those individuals, cardiovascular disease actually surpasses lung cancer as the leading cause of death. By picking up these areas of calcification that are linked to plaque build-up, the scan can give providers a strong predictor of heart attacks, stroke, and mortality from cardiovascular disease.

Given the recent recommendation update from the U.S Preventive Services Task Force that opened up LDCT lung cancer screening to high-risk individuals ages 50-to-80 who have at least a 20-pack year smoking history, it is possible that providers will be able to pick up on cardiovascular disease mortality risk in an even wider population of patients, the team said.

In the past, data pulled from CT images has been used in conjunction with other risk factors, such as cholesterol levels, blood pressure, and self-reported history of illness. But, for this study, the team, led by Bob D. de Vos, Ph.D., from Amsterdam University Medical Center and the Image Sciences Institute at the University Medical Center Utrecht, drew upon deep learning to create an automated method that can quickly predict five-year cardiovascular disease mortality without putting a heavier work burden on providers.

“We have shown that five-year cardiovascular disease mortality can be predicted for lung screening participants in less than half a second, using only site-specific calcium scores automatically derived from lung screening low-dose CT,” the team said. “Hence, the proposed image-based analysis could aid in identification of lung screening participants at risk for cardiovascular disease mortality, without relying on self-reported participant data.”

Figure 2: Projections of all aligned chest CT scans show feasibility of slab-based quantification of calcium, resulting in an average image. For alignment, only translation, rotation, and scaling were allowed, resulting in a blurry image, because not all anatomy is exactly the same across participants. From left to right, the center axial, sagittal, and coronal sections are shown. Note that field of view is similar to cardiac CT, which is a consequence of image alignment by the used automatic calcium scoring method. Image alignment allowed the determination of calcification distributions into slabs as a proxy measure for proximal and distal calcifications, of which the borders are indicated by the horizontal lines.

For the study, they pulled data from LDCT scans conducted between August 2002 and April 2004 on 4,451 participants with an average age of 61 who were part of the National Lung Screening Trial. Using those details, the team trained their deep learning method to quantify six types of vascular calcification: thoracic aorta calcification (TAC), aortic and mitral valve calcification, coronary artery calcification (CAC) of the left main, the left anterior descending, and the right coronary artery. After training, they tested their method on data gleaned from 1,113 participants.

The team compared the methods with semi-automatic baseline prediction using self-reported participant characteristics, such as age, smoking history, and history of illness. According to their analysis, the prediction model using calcium scores performed better than the baseline model.

Specifically, the prediction model that used calcium scores reached a C statistic of 0.74 while the baseline model that relied on only participant characteristics achieved 0.69. The model had the best results, they said, when all variables were combined – a 0.76 C statistic.

“The analysis shows we found predictors that are typically not described in a literature, possibly because we performed analysis in lung cancer screening participants we are already at high risk of cardiovascular disease from a history of heaving smoking and the presence of extensive arterial calcification,” de Vos explained.

The method works in two stages, he said, first using deep learning to pinpoint the amount and location of arterial calcification in the coronary arteries and aorta. The second stage relies on a more conventional statistical approach to mortality prediction and identifies features that are most predictive of five-year mortality.

Incorporating the deep learning method into lung cancer screening would be simple, de Vos said, because it requires no special equipment and does not tack on any exam time. And, the benefits to providers and patients could be immense.

“The method uses only image information, it is fully automatic, and it is fast. The method obtains calcium scores in a complete chest CT in less than half a second. This means that the method should be easy to implement to routine patient work ups and screening,” he said. “Lung screening studies show that heavy smokers die from cardiovascular disease as much as from lung cancer. But, we also see that some people with very high calcium scores survive, while others with low scores do suffer from major cardiac events.”

Their results point to a need for additional research that can more specifically point to calcifications that are dangerous. To date, they have developed several automatic scoring methods that can use varied types of data, but they are continuing on with development. Now, de Vos said, they are concentrating on creating a calcium-scoring method that can use low-quality data, such as data affected by cardiac motion, low image resolution, and high noise level, to detect arterial calcification.

“We developed a method, for example, that can detect coronary calcification even when the lesions are below the clinically used threshold,” he said. “This way, we hope to increase the reproducibility of calcium scoring and enable more accurate prediction.”

But, for now, he said, their work has shown that it is possible to achieve real-time mortality prediction with lung screening chest CT by using only data from automatic deep learning-quantified arterial calcium.

A low-dose CT for lung cancer screening can also predict a five-year cardiovascular disease risk of death when using a deep learning algorithm.

Whitney J. Palmer

Articles

Chest X-ray Can Predict Outcomes for Patients With COVID-19

4 Must-Knows about Future Radiologist Liabilities with AI

Brain MRI and CT Could Identify Metastasis in Patients with Kidney Cancer

Add Predicting Heart Disease Mortality Risk to LDCT Capabilities