Proof-of-concept study ‘highlights that using AI to integrate different types of clinically informed data to predict disease outcomes is feasible’ researchers say
Artificial intelligence (AI) and machine learning are—in stepwise fashion—making progress in demonstrating value in the world of pathology diagnostics. But human anatomic pathologists are generally required for a prognosis. Now, in a proof-of-concept study, researchers at Brigham and Women’s Hospital in Boston have developed a method that uses AI models to integrate multiple types of data from disparate sources to accurately predict patient outcomes for 14 different types of cancer.
The process also uncovered “the predictive bases of features used to predict patient risk—a property that could be used to uncover new biomarkers,” according to Genetic Engineering and Biotechnology News (GEN).
Should these research findings become clinically viable, anatomic pathologists may gain powerful new AI tools specifically designed to help them predict what type of outcome a cancer patient can expect.
“Experts analyze many pieces of evidence to predict how well a patient may do. These early examinations become the basis of making decisions about enrolling in a clinical trial or specific treatment regimens,” said Faisal Mahmood, PhD (above) in a Brigham press release. “But that means that this multimodal prediction happens at the level of the expert. We’re trying to address the problem computationally,” he added. Should they be proven clinically-viable through additional studies, these findings could lead to useful tools that help anatomic pathologists and clinical laboratory scientists more accurately predict what type of outcomes cancer patient may experience. (Photo copyright: Harvard.)
AI-based Prognostics in Pathology and Clinical Laboratory Medicine
The team at Brigham constructed their AI model using The Cancer Genome Atlas (TCGA), a publicly available resource which contains data on many types of cancer. They then created a deep learning-based algorithm that examines information from different data sources.
Pathologists traditionally depend on several distinct sources of data, such as pathology images, genomic sequencing, and patient history to diagnose various cancers and help develop prognoses.
For their research, Mahmood and his colleagues trained and validated their AI algorithm on 6,592 H/E (hematoxylin and eosin) whole slide images (WSIs) from 5,720 cancer patients. Molecular profile features, which included mutation status, copy-number variation, and RNA sequencing expression, were also inputted into the model to measure and explain relative risk of cancer death.
The scientists “evaluated the model’s efficacy by feeding it data sets from 14 cancer types as well as patient histology and genomic data. Results demonstrated that the models yielded more accurate patient outcome predictions than those incorporating only single sources of information,” states a Brigham press release.
“This work sets the stage for larger healthcare AI studies that combine data from multiple sources,” said Faisal Mahmood, PhD, Associate Professor, Division of Computational Pathology, Brigham and Women’s Hospital; and Associate Member, Cancer Program, Broad Institute of MIT and Harvard, in the press release. “In a broader sense, our findings emphasize a need for building computational pathology prognostic models with much larger datasets and downstream clinical trials to establish utility.”
Future Prognostics Based on Multiple Data Sources
The Brigham researchers also generated a research tool they dubbed the Pathology-omics Research Platform for Integrative Survival Estimation (PORPOISE). This tool serves as an interactive platform that can yield prognostic markers detected by the algorithm for thousands of patients across various cancer types.
The researchers believe their algorithm reveals another role for AI technology in medical care, but that more research is needed before their model can be implemented clinically. Larger data sets will have to be examined and the researchers plan to use more types of patient information, such as radiology scans, family histories, and electronic medical records in future tests of their AI technology.
“Future work will focus on developing more focused prognostic models by curating larger multimodal datasets for individual disease models, adapting models to large independent multimodal test cohorts, and using multimodal deep learning for predicting response and resistance to treatment,” the Cancer Cell paper states.
“As research advances in sequencing technologies, such as single-cell RNA-seq, mass cytometry, and spatial transcriptomics, these technologies continue to mature and gain clinical penetrance, in combination with whole-slide imaging, and our approach to understanding molecular biology will become increasingly spatially resolved and multimodal,” the researchers concluded.
Anatomic pathologists may find the Brigham and Women’s Hospital research team’s findings intriguing. An AI tool that integrates data from disparate sources, analyzes that information, and provides useful insights, could one day help them provide more accurate cancer prognoses and improve the care of their patients.
Former Vice President received an exclusive tour of a completely fake medical testing laboratory within Theranos, which he found “most impressive”
One thing clinical laboratory leaders and pathologists may still be curious about concerning the whole Theranos affair is how the company founder Elizabeth Holmes could fool so many high-ranking individuals—including then Vice President Joe Biden—into endorsing a completely fraudulent medical laboratory test process.
But it was the lengths to which Holmes and Balwani went to “trick” Joe Biden into endorsing Theranos—and subsequently receive the positive press that followed—that MSN found most intriguing.
According to MSN, in July of 2015 Holmes and Balwani procured Biden’s endorsement by giving the VP a tour of a “completely fake, staged lab.”
“What’s most impressive to me is you’re not only making these lab tests more accessible, you’re charging historically low prices, which is a small fraction of what is charged now, while maintaining the highest standards, and empowering people whether they live in the barrio or a mansion, putting them in a position to help take control of their own health,” stated then VP Joe Biden (above with Elizabeth Holmes) in a Theranos press release. Sadly, many clinical laboratory leaders who were skeptical and outspoken about Theranos’ claims were ignored by the press. (Photo copyright: ABC News.)
Wall Street Journal Reporter Exposes Theranos Fraud
According to a 2018 article by John Carreyrou which was part of his expose´ of Theranos published in The Wall Street Journal, “Ms. Holmes and Mr. Balwani wanted to impress Vice President Biden with a vision of a cutting-edge, automated laboratory. Instead of showing him the actual lab with its commercial analyzers, they created a fake one, according to former employees who worked in Newark. They made the microbiology team vacate a room it occupied, had it repainted, and lined its walls with rows of [Theranos] miniLabs stacked up on metal shelves.”
And the ruse worked. A 2015 Theranos press release outlined the visit at the time and stated that Biden found the facility inspiring and was impressed by the work being done by the company.
“I just had a short tour and I’m glad because you can see first-hand what innovation is all about just walking through this facility. This is the laboratory of the future,” Biden said in the press release.
In 2015, then Vice President Joe Biden toured the Theranos facility with Elizabeth Holmes, observed their supposedly innovative finger stick test system, and met with several Theranos employees. Later reports exposing the fraud stated that Holmes and Balwani were desperate to obtain Biden’s approval as it would provide positive press for Theranos, a good reputation within the industry, and lure potential investors. Theranos later tweeted a photo (above) of the visit showing Biden and Holmes walking amongst numbered blood-testing machines with a huge Theranos logo banner in the background. (Photo copyright: Connor Radnovich/The Chronicle.)
Biden’s visit occurred just a few months before Carreyrou’s Wall Street Journal report questioned the efficacy of Theranos’ blood testing technology and alleged the lab testing company tried to cover up its failures and mislead investors and patients.
Prior to that hard-hitting exposé, Holmes was heralded by the media as a star in the field of medicine. She was even prominently featured on magazine covers of influential business periodicals such as Fortune, Forbes, and Inc.
Others Who Were Bamboozled by Holmes and Balwani
Biden was not the only high-profile individual who was fooled by Holmes, Balwani and their billion-dollar con job. Other high-profile people included:
Theranos ceased operations in September of 2018 amidst the exposing of the fraud and inability to locate a buyer for the company. The shutdown rendered all investments in the company worthless.
Holmes to Receive New Hearing in Federal Court
In January of this year, Holmes was found guilty of three counts of wire fraud and one count of conspiracy to commit wire fraud for lying to investors about Theranos products. She faces up to 20 years in prison and a fine of $250,000 plus restitution for each count.
And so, clinical laboratory leaders and pathologists now have a better idea as to how Joe Biden was hoodwinked and endorsed a completely fake blood testing laboratory at Theranos. Can he be blamed for his ignorance of clinical laboratory test technology? Probably not. But it makes for interesting reading.
Amid cost pressures, healthcare providers also plan to cut staff though some jobs are plentiful; adequate staffing at medical laboratories continues to be a challenge
Thanks to the COVID-19 pandemic and subsequent “Great Resignation,” masses of people have left the workforce and companies large and small in all industries are struggling to retain employees. Clinical laboratories have been particularly hard hit with no relief in sight.
Now comes the results of a PricewaterhouseCoopers (PwC) survey which shows 50% of US companies in various industries—including major healthcare providers—plan to lay off employees. And 83% of organizations intend to move forward with a “streamlined workforce,” according to the latest PwC Pulse: Managing Business Risks in 2022 report.
How this will affect the workload on remaining hospital and medical laboratory staff is clear. And healthcare consumers may not take well to healthcare provides running leaner and with fewer staff than they currently do.
Nevertheless, the PwC survey results “illustrate the contradictory nature of today’s labor market, where skilled workers can still largely name their terms amid talent shortages even as companies look to let people go elsewhere,” Bloomberg wrote on the CPA Practice Advisor website.
“Organizations are still walking a tightrope when it comes to talent as we begin to see the longer-term impacts of the ‘Great Resignation.’ Finding the proper balance between investing in specialized talent, managing headcount costs, and driving productivity and morale will remain a top focus,” said Bhushan Sethi (above), People and Organization Joint Global Leader at PwC and an adjunct professor at NYU Stern School of Business in a PwC news release. Clinical laboratories are finding it particularly challenging to fill staff positions across all areas of lab operations. (Photo copyright: PwC.)
Healthcare Has Biggest Challenges, says PwC
Clinical laboratory leaders and pathologist groups are well aware of the unique financial pressures on healthcare systems and medical labs, as well as shortages of pathologists, medical technologists, clinical laboratory scientists, information technology (IT) professionals, and other healthcare workers.
“Healthcare is seeing bigger talent challenges than other industries and is more focused on rehiring employees who have recently left,” the PwC report acknowledged. This is the second Pulse survey PwC conducted in 2022. The 722 respondents included leaders working in human capital and finance.
Finding Right Talent, Focusing on Growth, Automation
Finding the right employees is so important to companies that PwC ranks “talent acquisition” as the second highest risk (38%) behind cyber-attacks (40%).
“Finding the right talent continues to be a challenge for business leaders,” PwC said. “After a frenzy of hiring and a tight labor market over the past few years, executives see the distinction between having people and having people with the right skills.”
Unlike the high-touch and personal nature of healthcare, industries such as consumer technology, media, and telecommunications can turn to automation to alleviate staffing struggles. And that is what nearly two-thirds, or 63%, of companies in those sectors, aim to do, PwC said.
Other survey talent findings:
50% of companies plan layoffs.
46% are dropping or eliminating sign-on bonuses.
44% are rescinding job offers.
Conversely, the surveyed executives also told PwC they are “cautiously optimistic” and plan on growing and investing even as the economy gives mixed signals:
83% of companies are focused on growth.
70% plan an acquisition.
53% aim to invest in digital transformation, 52% in IT, 49% in cybersecurity and privacy, and 48% in customer experience.
“After more than two years dealing with uncertainty related to the pandemic, business leaders recognize the urgent need to focus on growth in order to compete, and they’re zeroing in on what they can control,” PwC said.
New Remote Work Programs, Reduction in Real Estate Investing, Big Tech
Although companies report having more than enough physical office space, many (42%) have launched remote work programs:
70% have expanded or plan to increase “permanent” remote work options as jobs permit.
22% are reducing real estate investment (financial services and healthcare industries lead the way with 30% and 29%, respectively, saying real estate buys are cooling off).
“While companies continue to invest in many areas of the business, they’re scaling back the most in real estate and capex ex [capital expenditure]. After two years of remote work, many companies simply need less space, and they’re allocating capital accordingly,” the PwC report noted.
In a somewhat parallel release to PwC’s findings, news sources are reporting reductions in real estate and staff at high-profile Big Tech companies.
Meta Platforms, Inc. in Menlo Park, Calif. (formerly Facebook Inc.), is closing one of its New York offices and cutting back on plans to expand two other locations in the city, the Observer reported.
Business Insider reported, “More than 32,000 tech workers have been laid off in the US till July, including at Big Tech companies like Microsoft and Meta (formerly Facebook), and the worst has not been over yet for the tech sector that has seen massive stock sell-off.”
According to Forbes, “San Francisco-based electronic signature company DocuSign will lay off 9% of its more than 7,400 employees (roughly 670 employees), the company announced in a Securities and Exchange filing Wednesday, saying the cuts are ‘necessary to ensure we are capitalizing on our long-term opportunity and setting up the company for future success.’”
And Bloomberg recently reported that Intel is planning to layoff thousands of people “around the same time as its third-quarter earnings report on Oct. 27.”
Healthcare Providers Plan Layoffs, Seek IT Pros
Meanwhile, major healthcare provider networks also are planning staff cuts amid service closures, rising costs, and other issues, according to Becker’s Hospital Review:
Ascension in St. Louis, Mo., plans to close an Indiana hospital and nine medical practices and lay off 133 employees.
“Our health system, like others around the nation, is facing significant financial pressures from historic inflation, rising pharmaceutical and labor costs, COVID-19, expiration of CARES Act funding, and reimbursement not proportional with expenses,” BHSH said in a statement shared with Becker’s.
Amidst these layoffs, however, IT jobs in healthcare seem to be growing. According to Becker’s Health IT, some healthcare providers have posted information technology openings:
Mayo Clinic in Rochester, Minn., has 43 IT job openings.
So, though it appears IT positions continue to expand, clinical laboratory leaders and pathology practice managers may want to prepare now for dealing with customers’ response to leaner healthcare systems overall.
Encouraging patients—even children—to be more directly involved in their own medical care may reduce the burden on healthcare workers and might even help those clinical laboratories struggling to hire enough phlebotomists to collect specimens
Researchers at Emory University School of Medicine have concluded a study which found that school-aged children can successfully use a nasal swab to obtain their own SARS-CoV-2 test specimens. This may come as a surprise to hospital and clinical laboratory personnel who have performed nasal swabbing for COVID-19 tests. Some people, adults included, find the procedure so uncomfortable it brings tears.
And yet, after being shown a 90-second how-to video and given a handout with written instructions and pictures, 197 Atlanta children who had COVID-19 symptoms between July and August of 2021 performed their own self-swabbing. A healthcare worker then collected a second swabbed sample. All samples were submitted to a clinical laboratory for PCR analysis.
The Emory study provides another example of how the healthcare system is engaging patients to be directly involved in their own medical care. Results of the study could positively impact clinical laboratories facing a shortage of personnel, as well as schools where children have to take repeated COVID-19 tests with the assistance of trained professionals.
In a study with 197 school-age children, researchers at Emory University School of Medicine found that children could self-swab themselves for COVID-19 testing after watching a 90-second instructional video. Clinical laboratory leaders who are short on personnel may find these results intriguing. (Photo copyright: Emory University.)
How Did the Children Do?
The self-collected swabs and those collected by a healthcare worker agreed 97.8% of the time for a positive result and 98.1% of the time for a negative result. The analysis showed that both collection methods identified the 44% of symptomatic kids who were positive for COVID-19.
“Seeing how closely the results line up between the children and trained healthcare workers is a strong indicator that these age groups are fully capable of swabbing themselves if given proper instruction,” said Jesse Waggoner, MD, an Assistant Professor of Infectious Diseases with the Emory University School of Medicine and one of the lead authors on the study, in an Emory University press release.
A higher percentage of children age eight and under needed assistance, such as more instruction before correctly completing self-collection—21.8% compared to 6.1% for children older—but SARS-CoV-2 detection among the two age groups did not differ.
Does FDA Approve of Self-Swabbing?
The US Food and Drug Administration (FDA) has not authorized COVID-19 tests that include self-swabbing by children under age 14. However, data from the Emory study, published in JAMA, is now available to test manufacturers seeking authorization for pediatric self-collection.
“Pediatric self-swabbing will support expanded testing access and should make it even easier to test school age populations with fewer resources,” said Tim Stenzel, MD, PhD, Director of the Office of In Vitro Diagnostics at the FDA, in the Emory statement. “This study furthers our knowledge of test accuracy with these types of samples and provides test manufacturers with data to support their EUA (Emergency Use Authorization) requests to the FDA.”
Self-swabbing versus Clinical Laboratory Worker
While it has been longstanding medical practice to have healthcare workers collect samples for respiratory tract infection testing, the Emory researchers suggest that allowing children to collect their own COVID-19 samples could be one way to reduce the burden of a shortage of healthcare workers.
The researchers also believe pediatric self-swabbing would expand access to diagnostic tests and make it easier to test school-age populations.
“Every minute of a healthcare worker’s time is at a premium,” said senior study author Wilbur Lam, MD, Professor of Pediatrics and Biomedical Engineering, Emory University and Georgia Tech, in a National Institutes of Health (NIH) press release. “Why not allow a kid to self-swab? It’s a win-win! They would rather do it themselves and it frees up the healthcare worker to do other things,” he added.
In 2020, a Stanford University School of Medicine study published in JAMA showed test samples collected by adults who swabbed their own nasal passages were as accurate as those collected by healthcare workers. This study involved 30 participants who had previously tested positive for COVID-19.
Though the Emory University and Stamford University studies were small, they agreed in their findings which is significant. Clinical laboratory executives and pathologists should expect this trend toward direct-to-consumer and other forms of self-testing to continue, even among young patients.
These findings may be useful to clinical laboratory professionals when physicians want guidance in effective treatments for COVID-19 patients, particularly when there are concerns about a rebound of the infection
Drug interactions are a major concern for physicians and clinical laboratories. That is especially true given the push for nearly universal COVID-19 vaccinations and boosters. Now, a study conducted in Denmark may show that the use of Paxlovid as an antiviral drug to treat early SARS-CoV-2 infection could trigger drug-drug interactions (DDI) in some patients.
For clinical laboratory managers, insights into the issues associated with Paxlovid may be useful in helping client physicians diagnose their patients and anticipate possible negative drug reactions where other anti-viral drugs are involved.
Also of interest to medical laboratory leaders is the fact that the federal Centers for Disease Control and Prevention (CDC) in May released a Health Alert Network (HAN) Health Advisory about the potential for COVID-19 rebound after Paxlovid treatment.
COVID-19 Rebound, according to the CDC, “has been reported to occur between two and eight days after initial recovery and is characterized by a recurrence of COVID-19 symptoms or a new positive viral test after having tested negative.”
In an article she penned for STAT, Joan Susan Bregstein, MD (above), a pediatric emergency medicine physician and professor of pediatrics at Columbia University Irving Medical Center in New York, wrote, “Is Paxlovid worth it? The CDC advisory states in black, bold, and no uncertain terms that, despite the risk of rebound COVID, ‘Paxlovid continues to be recommended for early-stage treatment of mild to moderate COVID-19 among persons at high risk for progression to severe disease.’ But the definition of ‘high risk’ in this situation has been a moving target since the first days of COVID-19.” Clinical laboratory leaders can attest to the accuracy of that statement. (Photo copyright: Columbia University.)
Do Anti-Viral Drugs Interact with Other Medications?
Paxlovid is the retail name for a combination of two anti-viral drugs: nirmatrelvir and ritonavir. The medication for COVID-19 was developed by American pharmaceutical company Pfizer (NYSE:PFE) and received Emergency Use Authorization from the US Food and Drug Administration in August of this year.
The drug is taken orally for five days by people who test positive for the SARS-CoV-2 coronavirus to head off disease progression as well as serious illness, according to the CDC advisory.
But a “sizeable proportion” of elderly people are on medications that could interact with Paxlovid, Reuters reported.
“Two oral antiviral drugs—nirmatrelvir/ritonavir (NMV/r) and molnupiravir—have been approved for early outpatient treatment of COVID-19 to prevent severe disease. Ritonavir, contained in NMV/r is known to have significant DDI with several drugs frequently used by the elderly. This communication puts the problem with DDI with oral antiviral COVID-19 treatment into perspective,” the study authors wrote.
Their analysis of prescription data from Denmark residents found “extensive use of drugs likely to interact with NMV/r” as follows:
Anticoagulants (blood thinners): used by 20% of people over age 65 and by 30% of people over 80.
Statins (cholesterol-lowering medications): taken by 15% to 18% of people over 65.
Analgesics (for pain), calcium channel blockers (used to decrease blood pressure in patients with hypertension), or digoxin (used to treat heart conditions): taken by 20% of those studied.
In their paper, the researchers offered guidance to physicians. “Before prescribing NMV/r, the patient’s full medical history, including herbals and over-the-counter and recreational drugs, must be known and co-treatment carefully managed by the treating physician or by a specialist to avoid detrimental effects.”
However, one infectious disease specialist told Scientific American it may just take the elderly who were taking Paxlovid more time to completely get over COVID-19.
“Being of an elderly age and then having other risk factors—like diabetes, heart disease, kidney disease, or some sort of cancer—does put you at higher risk of rebound,” Aditya Shah, MBBS, Mayo Clinic Infectious Disease Physician and Researcher, told Scientific American.
That study’s researchers retrospectively reviewed 92 million electronic health records (EHR) from US patients. They found most people (11,270) had been treated with Paxlovid. However, 2,374 patients took molnupiravir, which also was granted EUA status by the FDA and is marketed as Lagevrio.
That COVID-19-rebound study found:
After nirmatrelvir/ritonavir (Paxlovid) treatment: 3.53% had rebound infections, 2.3% with rebound symptoms, and .44% were hospitalized.
After molnupiravir (Lagevrio) treatment: 5.86% had rebound infections, 3.75% with rebound symptoms, and .84% were hospitalized.
“Patients who took molnupiravir were significantly older and had more comorbidities than those who took Paxlovid,” the researchers wrote. “Results further suggest that rebound was not unique to Paxlovid and may be associated with persistent viral infection in some patients treated with either of these two antivirals. There has been more attention to COVID-19 rebound following Paxlovid treatment than molnupiravir, which may be attributable to more people being treated with Paxlovid,” they concluded.
Clinical Laboratories Can Guide Doctors
In an article she penned for STAT, titled, “Paxlovid Rebound Happens, Though Why and to Whom Are Still a Mystery,” Joan Susan Bregstein, MD, a pediatric emergency medicine physician and professor of pediatrics at Columbia University Irving Medical Center in New York, wrote of COVID-19 rebound, “My emergency medicine physician colleagues are seeing tons of it. Although people tend to think of medical care as something that is certain, it is actually a real-time experiment. Paxlovid, like a lot of COVID-19 care, is a reminder of this.”
Similarly, Mayo Clinic’s Shah acknowledged difficulty in identifying a COVID-19 rebound case. “You need real documentation of three tests—a positive, a negative, a positive—and clear documentation of symptoms—all symptoms gone, symptoms come back,” Shah told Scientific American.
Thus, clinical laboratories play a vital role in diagnosing and treating COVID-19 rebound patients, because that is what clinical labs do: test, document, and report. And as the study of the Danish population pointed out, doctors need guidance as they prescribe oral antivirals to COVID-19 patients who are on other drugs and at possible risk of drug-drug interactions.
