A new analysis shows why models fall short in practice, how liability and equity issues slow adoption, and what lab leaders should consider as AI becomes a growing part of diagnostic workflows.
Artificial intelligence (AI) has made notable advances in medical imaging, but radiologists are not being displaced. For laboratory and diagnostic leaders, a recent analysis in Works in Progress highlights why AI has not replaced human expertise in radiology—and what this means for managing technology adoption in labs and hospitals.
In 2016, AI pioneer Geoffrey Hinton declared that “people should stop training radiologists now.” Since then, more than 700 FDA-cleared radiology AI models have entered the market, covering everything from stroke detection to lung cancer screening.
Companies such as Annalise.ai, Lunit, Aidoc, and Qure.ai offer tools that can identify dozens of diseases across modalities, reorder worklists, or generate structured draft reports. “On paper, radiology looks like the perfect target for automation,” the article noted, citing its reliance on digital images, pattern recognition, and quantitative benchmarks. Yet demand for radiologists has never been higher. In 2025, US residency programs offered a record 1,208 positions, and vacancy rates remain high as well.
Why Hasn’t AI Taken Over?
For leaders overseeing diagnostic services, three key elements are why AI has not replaced radiologists.
First, models struggle in real-world deployment. “Performance can drop by as much as 20 percentage points” when systems trained on narrow datasets are applied across different scanners, imaging protocols, or patient populations, the article explained. What works in a benchmark test may falter in a hospital with diverse workflows.
Second, liability and regulatory hurdles remain high. Assistive models that require physician review face fewer barriers, but autonomous systems must self-abort on poor image quality, identify unfamiliar equipment, and withstand rigorous scrutiny. Insurers have also drawn hard lines: one malpractice policy states that “coverage applies solely to interpretations reviewed and authenticated by a licensed physician; no indemnity is afforded for diagnoses generated autonomously by software.” Another bluntly imposes an “Absolute AI Exclusion.” For labs, this underscores the importance of risk management before deploying AI tools.
Photo credit: “Artificial Intelligence – Resembling Human Brain” by deepakiqlect is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/?ref=openverse.
Photo credit: “Cancer” by davis.steve32 is licensed under CC BY 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/2.0/?ref=openverse.
Third, radiologists do much more than read scans. “Human radiologists spend a minority of their time on diagnostics and the majority on other activities, like talking to patients and fellow clinicians,” the commentary pointed out. Oversight of imaging protocols, interdisciplinary consultations, and patient communication all fall outside the reach of algorithms. Even as AI improves, demand for imaging may increase rather than decrease—a version of the Jevons paradox where greater efficiency leads to higher use. “The better the machines, the busier radiologists have become,” the article observed.
For laboratory leaders, the takeaway is not to fear replacement but to prepare for integration. AI tools are proving valuable in triaging urgent cases, flagging abnormalities, and drafting reports, but they remain narrow in scope—stroke, lung cancer, and breast lesions account for about 60% of models, yet represent only a fraction of total imaging work. As the article concluded, “Models can lift productivity, but their implementation depends on behavior, institutions and incentives.”
The challenge for labs is to create environments where AI augments human expertise rather than attempts to replace it. That means aligning technology adoption with clinical needs, providing training for staff, and working with insurers and regulators to ensure coverage and compliance.
For now, radiologists and the labs that support them are not going away. They are adapting, and AI will be a partner in that evolution.
Device provides physicians with quick insights into infant’s immune system capabilities without requiring clinical laboratory testing
International researchers have developed a revolutionary tool that rapidly assesses an infant’s immune system using a single drop of blood. The novel device provides healthcare professionals with real-time insights about a newborn’s immune response in less than 15 minutes.
Scientists from the Singapore-MIT Alliance for Research and Technology (SMART) in collaboration with colleagues from KK Women’s and Children’s Hospital (KKH) in Singapore created the Biophysical Immune Profiling for Infants (BLIPI) portable device to help alleviate potentially life-threatening illnesses in newborns. The device only uses 0.05 ml of blood.
The work was led by researchers from the Critical Analytics for Manufacturing Personalized Medicine (CAMP) and Antimicrobial Resistance (AMR) interdisciplinary research groups within SMART. SMART is a major research collaboration between the Massachusetts Institute of Technology (MIT) and the National Research Foundation of Singapore.
“BLIPI represents a major step forward by providing clinicians with fast, actionable immune health data using a noninvasive method, where it can make a real difference for newborns in critical care,” said Kerwin Kwek Zeming, PhD, research scientist at SMART CAMP and SMART AMR, and co-lead author of the study, in an MIT news release.
“Our goal was to create a diagnostic tool that works within the unique constraints of neonatal care—minimal blood volume, rapid turnaround, and high sensitivity,” said Kerwin Kwek Zeming, PhD, research scientist at SMART CAMP and SMART AMR, and co-lead author of the MIT study, in the news release. (Photo copyright: MIT.)
Bridging Gap between Science and Healthcare
To perform their study, the team used BLIPI to screen 19 infants—eight full-term and 11 preterm—and compared the differences in immune cells between the infants. The device uses microfluidic technology to measure immune cell characteristics, such as size and flexibility, to expose how the immune system is responding to changes within the cells. Traditional tests look only for the presence of germs, but BLIPI also looks at results such as C-reactive protein levels, white blood cell counts, and immature-to-total neutrophil ratios, to determine if an infant is fighting an infection.
“BLIPI exemplifies our vision to bridge the gap between scientific innovation and clinical need. By leveraging microfluidic technologies to extract real-time immune insights from whole blood, we are not only accelerating diagnostics but also redefining how we monitor immune health in fragile populations,” said Jongyoon Han, PhD, professor of electrical engineering and biological engineering at MIT and coauthor of the paper, in the news release. “Our work reflects a new paradigm in point-of-care diagnostics: rapid, precise, and patient-centric.”
Saving Infant Lives
BLIPI only needs one tiny drop of blood, which equals 1/20 of the blood volume typical lab tests require. The onsite tool removes the need for sending blood samples to clinical labs, which may enable clinicians to make earlier decisions regarding treatment options for critical situations like sepsis or necrotizing enterocolitis.
“KKH cares for about two-thirds of all babies born weighing less than 1,500 grams (52.91 ounces or 3.31 pounds) in Singapore. These premature babies often struggle to fight infections with their immature immune systems. With BLIPI, a single prick to the baby’s finger or heel can give us rapid insights into the infant’s immune response within minutes. This allows us to tailor treatments more precisely and respond faster to give these fragile babies the best chance at a healthy start not just in their early days, but throughout their lives,” said Yeo Kee Thai, MD, senior consultant in the department of neonatology at KKH, and senior author of the study, in the news release.
BLIPI also could be extremely beneficial to healthcare settings in remote areas or with limited resources.
Further research and clinical trials are needed to validate the diagnostic accuracy of BLIPI. In addition, the researchers plan to improve the design to render it usable for widespread distribution. They also hope BLIPI will someday be used by pharmaceutical companies and medical researchers to evaluate immune responses to neonatal therapies in real time.
New AI tool doubled efficiency in busy university radiology department
Creative artificial intelligence (AI) solutions are being developed to address critical staffing shortages in radiology that could help with similar shortages in overworked pathology and clinical laboratories as well.
In a recent clinical study at 11-hospital Northwestern Medicine, researchers developed a new generative AI radiology tool to assist radiologists that demonstrates high accuracy and efficiency rates when working with multiple types of imaging scans.
For the study, approximately 24,000 radiology reports were analyzed and then compared for clinical accuracy with and without the AI tool. The tool evaluates an entire scan and generates a report that is 95% complete and personalized to each patient. A template based on that report is then provided to radiologists for review, according to a Northwestern Medicine Feinberg School of Medicine news release.
The study reported an average 15.5% increase in radiograph efficiency without compromising accuracy. Some radiologists even produced gains as high as 40%. The radiology reports were scrutinized during a five-month period last year and enabled radiologists to improve the time it took to return a diagnosis.
“This is, to my knowledge, the first use of AI that demonstrably improves productivity, especially in healthcare. Even in other fields, I haven’t seen anything close to a 40% boost,” said the study’s senior author Mozziyar Etemadi, MD, PhD, assistant professor of anesthesiology and biomedical engineering at Northwestern University McCormick School of Engineering, in the news release. (Photo copyright: Northwestern University.)
Doubled Efficiency for One Radiology Team
“For me and my colleagues, it’s not an exaggeration to say that it doubled our efficiency. It’s such a tremendous advantage and force multiplier,” said study co-author Samir Abboud, MD, emergency radiology in the department of radiology at Northwestern Medicine, in the news release.
“Having a draft report available, even before it is viewed by the radiologist, offers a simple, actionable datapoint that can be quickly and efficiently acted upon” added study senior author Mozziyar Etemadi, MD, PhD, assistant professor of anesthesiology and biomedical engineering at Northwestern University McCormick School of Engineering, in the news release. “This is completely different than traditional triage systems, which need to meticulously be trained one by one on each and every diagnosis.”
The AI tool can also alert radiologists to life-threatening conditions.
“On any given day in the ER, we might have 100 images to review, and we don’t know which one holds a diagnosis that could save a life,” Abboud said. “This technology helps us triage faster—so we catch the most urgent cases sooner and get patients to treatment quicker.”
Relying on In-house Data
Engineers at Northwestern developed the AI model using clinical data within the university’s own network, emphasizing that such tools can be created without assistance from other organizations.
“Our study shows that building custom AI models is well within reach of a typical health system, without reliance on expensive and opaque third-party tools like ChatGPT,” Etemadi noted.
The Journal of the American College of Radiology states the supply of radiologists is expected to increase by approximately 26% over the next 30 years. However, the need for radiologists is expected to grow between 17% and 27% over the same period. Becker’s Hospital Review reports there will be a shortage of up to 42,000 radiologists in the US by 2033.
Some health organizations are using a mixed model of permanent employees and contracted radiologists to meet the increasing demand for services. Others are also looking at options such as internal training programs, better benefits for workers, teleradiology, and remote radiologists to fulfill radiology needs.
“You still need a radiologist as the gold standard,” Abboud said. “Medicine changes constantly—new drugs, new devices, new diagnoses—and we have to make sure the AI keeps up. Our role becomes ensuring every interpretation is right for the patient.”
Can pathology practices and clinical laboratories learn from radiology’s situation? Development of AI solutions for those fields would likely have similar effects on workloads and overworked personnel.
Exploring the benefits of AI may be one way of helping meet clinical laboratory and pathology practice staff shortages.
Lumipulse G pTau217/ß-Amyloid 1-42 Plasma Ratio test measures blood biomarkers associated with the disease
Clinical laboratories could soon find themselves playing a significant role in Alzheimer’s care as the US Food and Drug Administration (FDA) announced its first-ever clearance of a blood test to assist in diagnosing the cognitive disease.
The agency issued a 510(k) clearance for the Lumipulse G pTau217/ß-Amyloid 1-42 Plasma Ratio test from Fujirebio Diagnostics, Inc. The test is intended for “the early detection of amyloid plaques associated with Alzheimer’s disease in adult patients, aged 55 years and older, exhibiting signs and symptoms of the disease,” according to an FDA press release.
“Nearly seven million Americans are living with Alzheimer’s disease, and this number is projected to rise to nearly 13 million,” said FDA Center for Devices and Radiological Health Director Michelle Tarver, MD, PhD, in the press release. “Today’s clearance is an important step for Alzheimer’s disease diagnosis, making it easier and potentially more accessible for US patients earlier in the disease.”
“Alzheimer’s disease impacts too many people, more than breast cancer and prostate cancer combined,” said FDA Commissioner Martin A. Makary, MD, MPH, in the FDA press release.
“Knowing that 10% of people aged 65 and older have Alzheimer’s, and that by 2050 that number is expected to double, I am hopeful that new medical products such as this one will help patients,” said FDA Commissioner Martin A. Makary, MD, MPH, in the FDA press release. (Photo copyright: Johns Hopkins University.)
Blood Protein Test Details
Fujirebio’s test calculates the ratio of two blood proteins, pTau217 and β-amyloid 1-42. “This ratio is correlated to the presence or absence of amyloid plaques in the patient’s brain, reducing the need for a PET scan,” the press release states.
The FDA said it has already authorized or cleared similar tests that use cerebrospinal fluid (CSF) samples obtained through an invasive spinal tap. “This new Lumipulse test only requires a simple blood draw, making it less invasive and much easier for patients to access,” the FDA said.
In granting the clearance, the agency looked at data from a clinical study of 499 adults displaying signs of cognitive impairment. Close to 92% of participants with positive test results had amyloid plaques as determined by amyloid PET scan or CSF test results. Approximately 97% with negative results from the Lumipulse test were also shown to be negative in the PET scan or CSF test. Fewer than 20% of the patients received indeterminate results, which would require additional testing.
The primary risks of the test are false positive or false negative results, the federal agency noted.
“Importantly, the Lumipulse G pTau217/ß-Amyloid 1-42 Plasma Ratio is not intended as a screening or stand-alone diagnostic test and other clinical evaluations or additional tests should be used for determining treatment options,” the FDA said.
‘Wild West’ Marketplace
In their reporting of this story,MedPage Today and the Associated Press both noted that some laboratory-developed tests can already measure plasma biomarkers associated with Alzheimer’s.
“But those tests aren’t reviewed by the FDA and generally aren’t covered by insurance,” the AP reported. “Doctors have also had little data to judge which tests are reliable and accurate, leading to an unregulated marketplace that some have called a ‘wild west.’”
Neurologist Richard S. Isaacson, MD, told CNN that he’s already using the test for research purposes. “It can provide better clarity into whether a person experiencing memory loss may have Alzheimer’s disease,” he said.
However, he cautioned that more research is needed to allow doctors to make best use of the test.
“I think the next step as a field is, we need to advance education about what these tests mean and what they don’t and who they should be used for,” he told CNN. “Because they mean different things in different people depending on their risk factors and whether they have symptoms. So, we’re still early.”
“The results must be interpreted in conjunction with other patient clinical information,” the FDA acknowledged in their press release.
Other Alzheimer’s Tests
The FDA said it reviewed the Lumipulse test through the 510(k) premarket notification pathway, in which the federal agency determines if a device is “substantially equivalent to a legally marketed predicate device.”
In this case, the agency found that the test is substantially equivalent to Fujirebio’s Lumipulse G β-amyloid Ratio (1-42/1-40), which measures the same proteins in CSF samples. The FDA authorized that device in 2022, according to a Fujirebio press release. That test uses Fujirebio’s automated Lumipulse G1200 instrument system.
Other companies including Roche, Eli Lilly, and C2N Diagnostics are also developing Alzheimer’s tests they intend to submit for FDA review, the AP reported.
Diagnostic laboratories must dig into where their suppliers source their items to avoid surprises
Clinical laboratories in the US are under additional financial and operational pressure because of tariff fluctuations from the federal government among global trading partners. This brief overview explores clinical lab tariff strategies that may help organizations better prepare for difficult stretches.
Across the overall laboratory industry—including diagnostics settings—average duties on imported lab equipment and consumables hover around 23%, with personal protective equipment, lab instruments, and supply chains facing even higher tariffs depending on their origin, according to a report from Lab Manager, a sibling brand to Dark Daily.
“Don’t assume that just because something is made in the United States that it will be tariff free because many of the components are not,” Drew Kevorkian, founder and CEO at ARES Scientific, told Lab Manager. ARES Scientific maintains a useful online tariff guide.
These nuances can escalate costs, complicate budgeting, and squeeze margins, all of which pose a risk to financial viability for clinical labs.
When it comes to clinical lab tariff strategies, the starting point is clear. “Assume costs are going up one way or another. The best thing to do is be well informed,” Drew Kevorkian, founder and CEO at ARES Scientific, told Lab Manager. (Photo copyright: Ares Scientific.)
Tariff-Influenced Areas for Clinical Labs to Consider
Managers and others who oversee budgeting or product procurement should think about the following aspects as they plan out their clinical lab tariff strategies:
Component audits. Labs should undertake an audit of their consumables down to the stock-keeping unit (SKU), which is an alphanumeric descriptor that identifies a product. The goal is to identify the source of the products to determine if they come from tariff-affected countries, Kevorkian noted. This exercise allows clinical lab professionals to forecast cost scenarios and make informed procurement decisions.
Reused lab instruments. Tariffs introduce an opportunity to think about acquiring certified refurbished lab equipment. Such items, often already in the US, can offer meaningful savings and avoid import duties, according to Lab Manager.
Investigate in vitro diagnostics (IVD) sourcing. Clinical labs should look into where their IVD suppliers get their components, according to The Dark Report. Even if a lab buys from American-based suppliers, the IVD companies a laboratory works with might have ties overseas that aren’t immediately obvious. “All IVD companies get components from China,” Rob LaCroix, executive director of global strategy at LTC LLC, told attendees at the 2025 Executive War College on Diagnostics, Clinical Laboratory, and Pathology Management. “Just-in-time [purchasing] with tariffs is a problem,” he noted.
Clinical Lab Tariff Strategies Should Model Various Budget Scenarios
Lab Manager suggested that laboratories build tiered budget models that consider various tariff scenarios:
Baseline case assumes stable tariffs (around 23% as noted earlier).
Worst case simulates escalation or new restrictions.
Optimistic case explores tariff relief or exemptions.
These forecast ranges enable clinical labs to develop contingency plans, such as temporarily reducing discretionary spending or exploring ways to pass on incremental costs.
NPR reports that the shamed Theranos founder/CEO is providing advice to Evans, but the startup denies that claim
Prison bars can’t block Elizabeth Holmes from finding her way back into the news spotlight. The disgraced founder and former CEO of Theranos is reportedly advising her partner Billy Evans on his new artificial intelligence (AI) diagnostic startup company, named Haemanthus after the blood lily.
According to sources who spoke with NPR, Evans’ new company Haemanthus, Inc. is developing a blood testing device and has patented a process that uses Raman spectroscopy, which, according to NPR, “has been shown to help diagnose ALS, also called Lou Gehrig’s disease, as well as some forms of cancer. It has also been used to discover improvised explosive devices on battlefields.”
Evans has already raised millions of dollars for the fledgling startup, NPR reported, adding that a source claimed finances for the company have come from “mostly friends, family, and other supporters so far.”
According to Newsweek, Evans’ goal is to raise $50 million toward the development of a “medical testing product.”
The company will “do medical tests using bodily fluids,” Newsweek reported, adding, “An image of the alleged device published by The New York Times is eerily similar to Theranos’ ‘Edison’ testing machine.”
Elizabeth Holmes is currently housed in a federal facility in Bryan, Texas. Sources told NPR that she has been “providing advice” to Billy Evans, her partner, on his new AI/medical testing company Haemanthus, which denied those claims stating on X that Holmes “has no role, now or future.” (Photo copyright: Wikimedia Commons.)
Haemanthus Denies Holmes’ Involvement
Holmes has reportedly been providing insight to Evans throughout her prison term, though her role with his budding company is unclear, NPR noted.
As previously reported by Dark Daily, Holmes is “barred from receiving payments from federal health programs for services or products, which significantly restricts her ability to work in the healthcare sector.”
Haemanthus denied Holmes’ involvement with the company, claiming that she “has no formal role” and that “Haemanthus is not Theranos 2.0,” Fortune reported.
Previous lengthy posts by Haemanthus on social media platform X fully denied any involvement with Holmes but have since been deleted. The company now uses their platform to curtly retort the significance of Holmes’ involvement, leaning on their advancements and high standards. “Skepticism is rational. We must clear a higher bar,” they said. “When The NY Times contacted us, we invited them to see our lab, tech, and team. They declined. The headline was already written. Our reality inconvenient.”
Further posts on X showcase Haemanthus’ desire to have the same groundbreaking prowess Holmes clung to throughout her Theranos venture. The company claims to have developed “the world’s first AI-native sensors for health,” adding, “Our technology captures thousands of biomarkers simultaneously.”
And the Holmes Saga Continues
Haemanthus is comprised of about a dozen people, including individuals who “worked with Evans at Luminar Technologies, which develops sensor technology for autonomous vehicles, according to the company’s patent and Delaware incorporation paperwork,” NPR reported.
Holmes is currently serving an 11-year federal prison sentence for her role in fraud involving Silicon Valley startup Theranos, which boasted clinical laboratory blood-test breakthroughs that turned out to be riddled with faulty equipment and fraudulent results.
Though whistleblowers brought Holmes scheme to the light, she has never admitted wrongdoing for her actions and continues to claim her innocence. In May, the Ninth Circuit of Appeals denied her request for a rehearing of her case.
