New artificial intelligence model agrees with interpretations of human medical technologists and microbiologists with extraordinary accuracy
Microbiology laboratories will be interested in news from Brescia University in Italy, where researchers reportedly have developed a deep learning model that can visually identify and analyze bacterial species in culture plates with a high level of agreement with interpretations made by medical technologists.
They initially trained and tested the system to digitally identify pathogens associated with urinary tract infections (UTIs). UTIs are the source for a large volume of clinical laboratory microbiological testing.
In their Nature paper, the researchers explained that microbiologists use conventional methods to visually examine culture plates that contain bacterial colonies. The scientists hypothesize which species of bacteria are present, after which they test their hypothesis “by regrowing samples from each colony separately and then employing mass spectroscopy techniques,” to confirm their hypotheses.
However, DeepColony—which was designed for use with clinical laboratory automation systems—looks at high-resolution digital scans of cultured plates and attempts to identify the bacterial strains and analyze them in much the same way a microbiologist would. For example, it can identify species based on their appearance and determine which colonies are suitable for analysis, the researchers explained.
“Working on a large stream of clinical data, and a complete set of 32 pathogens, the proposed system is capable of effectively assisting plate interpretation with a surprising degree of accuracy in the widespread and demanding framework of urinary tract infections,” the study authors wrote. “Moreover, thanks to the rich species-related generated information, DeepColony can be used for developing trustworthy clinical decision support services in laboratory automation ecosystems from local to global scale.”
“Compared to the most common solutions based on single convolutional neural networks (CNN), multi-network architectures are attractive in our case because of their ability to fit into contexts where decision-making processes are stratified into a complex structure,” wrote the study’s lead author Alberto Signoroni, PhD (above), Associate Professor of Computer Science, University of Brescia, and his researcher team in their Nature paper. “The system must be designed to generate useful and easily interpretable information and to support expert decisions according to safety-by-design and human-in-the-loop policies, aiming at achieving cost-effectiveness and skill-empowerment respectively.” Microbiologists and clinical laboratory managers will want to follow the further development of this technology. (Photo copyright: University of Brescia.)
How Hierarchical AI Works
Writing in LinkedIn, patent attorney and self-described technology expert David Cain, JD, of Hauptman Ham, LLP, explained that hierarchical AI systems “are structured in layers, each with its own distinct role yet interconnected in a way that forms a cohesive whole. These systems are significant because they mirror the complexity of human decision-making processes, incorporating multiple levels of analysis and action. This multi-tiered approach allows for nuanced problem-solving and decision-making, akin to a seasoned explorer deftly navigating through a multifaceted terrain.”
DeepColony, the researchers wrote, consists of multiple convolutional neural networks (CNNs) that exchange information and cooperate with one another. The system is structured into five levels—labeled 0 through 4—each handling a different part of the analysis:
At level 0, the system determines the number of bacterial colonies and their locations on the plate.
At level 1, the system identifies “good colonies,” meaning those suitable for further identification and analysis.
At level 2, the system assigns each good colony to a bacterial species “based on visual appearance and growth characteristics,” the researchers wrote, referring to the determination as being “pathogen aware, similarity agnostic.”
The CNN used at this stage was trained by using images of 26,213 isolated colonies comprising 32 bacterial species, the researchers wrote in their paper. Most came from clinical laboratories, but some were obtained from the American Type Culture Collection (ATCC), a repository of biological materials and information resources available to researchers.
At level 3, the system attempts to improve accuracy by looking at the larger context of the plate. The goal here is to “determine if observed colonies are similar (pure culture) or different (mixed cultures),” the researchers wrote, describing this step as “similarity aware, pathogen agnostic.” This enables the system to recognize variants of the same strain, the researchers noted, and has the effect of reducing the number of strains identified by the system.
At this level, the system uses two “Siamese CNNs,” which were trained with a dataset of 200,000 image pairs.
Then, at level 4, the system “assesses the clinical significance of the entire plate,” the researchers added. Each plate is labeled as:
“Positive” (significant bacterial growth),
“No significant growth” (negative), or
“Contaminated,” meaning it has three or more “different colony morphologies without a particular pathogen that is prevalent over the others,” the researchers wrote.
If a plate is labeled as “positive,” it can be “further evaluated for possible downstream steps,” using MALDI-TOF mass spectrometry or tests to determine susceptibility to antimicrobial measures, the researchers stated.
“This decision-making process takes into account not only the identification results but also adheres to the specific laboratory guidelines to ensure a proper supportive interpretation in the context of use,” the researchers wrote.
Nearly 100% Agreement with Medical Technologists
To gauge DeepColony’s accuracy, the researchers tested it on a dataset of more than 5,000 urine cultures from a US laboratory. They then compared its analyses with those of human medical technologists who had analyzed the same samples.
Agreement was 99.2% for no-growth cultures, 95.6% for positive cultures, and 77.1% for contaminated or mixed growth cultures, the researchers wrote.
The lower agreement for contaminated cultures was due to “a deliberately precautionary behavior, which is related to ‘safety by design’ criteria,” the researchers noted.
Lead study author Alberto Signoroni, PhD, Associate Professor of Computer Science, University of Brescia, wrote in Nature that many of the plates identified by medical technologists as “contaminated” were labeled as “positive” by DeepColony. “We maximized true negatives while allowing for some false positives, so that DeepColony [can] focus on the most relevant or critical cases,” he said.
Will DeepColony replace medical technologists in clinical laboratories any time soon? Not likely. But the Brescia University study indicates the direction AI in healthcare is headed, with high accuracy and increasing speed. The day may not be far off when pathologists and microbiologists regularly employ AI algorithms to diagnose disease.
As we noted, thousands of clinical laboratory tests and surgical pathology readings had to be delayed or cancelled due to the strikes.
An NHS worker in a Liverpool hospital told CNN that conditions felt like a “war zone” with patients being treated in the backs of ambulances, corridors, waiting rooms, cupboards, or not at all since hospitals are well over capacity.
“Those who can afford to get private insurance are,” Chris Thomas (above), told The Guardian. Thomas is Head of the Commission on Health and Prosperity for UK progressive policy think tank the Institute for Public Policy Research (IPPR). “People are not opting out of the NHS because they have stopped believing in it as the best and fairest model of healthcare,” he said. “Rather, those who can afford it are being forced to go private … and those without the funds are left to ‘put up or shut up.’” (Photo copyright: Institute for Public Policy Research.)
Two-Tier System Could Become UK’s Norm, Dividing Classes
The drive towards private insurance is leaving Britain on the brink of having a “two-tier” system where the NHS is overpowered by private healthcare. And it’s not an unwarranted fear. One in six people in Britain are prepared to use private healthcare instead of waiting for the NHS, The Guardian reported.
A report from the Institute for Public Policy Research (IPPR) claims a UK two-tiered system would not mimic what we have here in the US. Rather, if the trend continues in the private direction, it would more likely be comparable to dentistry in England, “… where poor NHS access exists for some and superior but expensive access exists for many. We stand at the precipice of a growing ‘opt-out’ by those who can,” according to the IPPR report, The Guardian noted.
More importantly, this could further divide classes. “Such a trend could threaten the deep and widespread public support for the NHS among voters and leave millions of patients vulnerable because of their ethnicity, postcode, income or job,” The Guardian noted the IPPR report as saying.
“It’s different when you see your everyday reality though naïve eyes. He saw the elderly patients on the jigsaw of trolleys crammed into the department, pushed against the wall, squeezed in the gap between the bed and nursing stations.
“He saw the fluids hanging from rails where we had no stands, lines running into the patient’s forearms. He saw the oxygen fed into their noses from cylinders propped along the bed, the cacophony of beeping machines and alarms.
“It doesn’t look like it does on the TV. It doesn’t even look like it does on reality TV,” she wrote.
The healthcare statistics are alarming. According to CNN:
There was a 20% increase in excess deaths the final week of December 2022, compared to the previous five years.
Half of patients waiting for emergency care that month waited for more than four hours, which was a record.
Also in December, 54,000 people waited more than 12 hours for emergency admission. The wait was “virtually zero” prior to the COVID-19 pandemic.
And “category 2” conditions, such as a stroke or heart attack, had a more than 90-minute wait time for ambulance attendance. The target response time is 18 minutes.
Dim Hopes for Improvement
Though the NHS has struggled in recent years, the challenges are seemingly worse now. “This time feels different. It’s never been as bad as this,” gastroenterologist Peter Neville, MD, a consultant physician who worked with the NHS since 1989, told CNN.
CNN noted that a perfect storm of challenges might have brought the NHS to where it is today. COVID-19, flu seasons paired with COVID, lack of financial support, lack of social support, staffing and morale issues are just some of the problems that the NHS must address.
Experts point out that as the NHS’ struggles increase so begins a loop where one problem feeds another. Patients who wait to be seen have treatments that take longer, then they get sicker, and the cycle continues.
Despite having one of the highest proportions of government healthcare spending on Earth, up to 40% of Britons report having accessed or plan to access private care, Breitbart reported.
Sadly, it’s unlikely enough cash will come in from the UK government to make significant improvements for the NHS. The budget announcement in November showed the NHS was to get an average 2% spending increase over the next two years, CNN reported.
Are there lessons here for US hospitals, clinical laboratories, and pathology groups? Perhaps. It’s always instructive to see how our fellow healthcare providers across the pond respond to public pressure for more access to quality care.
Coordinating at-home testing for monkeypox may provide opportunities for clinical laboratories to add value for their physician clients
Microbiologists and clinical laboratory managers who oversee medical laboratory tests for monkeypox (aka, mpox) will be interested to learn that, according to the US Centers of Disease Control (CDC), cases per day have dropped into the single digits.
The United States led the world in cases during the 2022-2023 outbreak, according to the most recent CDC statistics. As of February 15, the US has reported 30,193 cases of monkeypox with 32 deaths.
Nevertheless, January 31 was the day that the US public health emergency involving monkeypox officially expired. Data from the World Health Organization shows the number of daily monkeypox cases in most countries around the world is declining, although numbers of cases are still increasing in some South American countries.
The global monkeypox outbreak appears to have slowed considerably, but are we out of the woods?
“There were concerns that there would be ongoing transmission and that ongoing transmission would become endemic in the United States like other STIs: gonorrhea, chlamydia, syphilis. We have not seen that occur,” Jonathan Mermin, MD (above), Director of the National Center for HIV, Viral Hepatitis, STD, and TB Prevention at the CDC, told CNN.”We are now seeing three to four cases a day in the United States, and it continues to decline. And we see the possibility of getting to zero as real.” This decline in monkeypox test corresponds with a similar decline in COVID-19 clinical laboratory testing as well. (Photo Copyright: CDC.)
Untried Vaccine and At-home Testing for Monkeypox
When the monkeypox outbreak began in May of 2022, there were concerns about the US’ level of preparedness for dealing with a second pandemic while also battling COVID-19. But monkeypox was not entirely unknown to the scientific and medical communities.
Monkeypox first appeared in 1958 amongst a colony of monkeys being kept for research. The origin of the disease is not known. According to the CDC, the first reported human case of monkeypox was in 1970. Prior to the 2022 outbreak, most cases were found in central and western African countries. Cases outside of those areas could be traced back to travel from those specific countries.
When cases of monkeypox first appeared in the US, public health officials were concerned about the availability of testing, vaccines, and treatments. As CNN reported, though there was a new vaccine available, its effectiveness against monkeypox had never been tested on humans.
That treatment, known as TPOXX (Tecovirimat), was an antiviral drug approved by the FDA in 2018 to treat smallpox in adults and children, according to an FDA factsheet. The drug was difficult to obtain, and it took until August of 2022 for the federal government to declare monkeypox a public health emergency. That allowed it to deploy emergency funds towards fighting the outbreak.
The demographic found to be at the highest risk of monkeypox infection were men who have sex with other men. According to MedPage Today, “Daskalakis had both pandemic experience as former senior lead on equity in COVID-19 data and engagement for the New York City Department of Health and Mental Hygiene and an ‘in’ with the LGBTQ+ community from his work in HIV prevention and his transparency about being a gay man.”
When comparing monkeypox to HIV, Daskalakis said, “This one [monkeypox], you don’t have to change behaviors for generations; it’s for a few months. Once you build your force field of immunity with vaccines, people can make their own informed decisions about their risk.”
Opportunities for Clinical Laboratories
So, how should clinical laboratories respond if there’s another monkeypox flare up?
Daskalakis advocates for home testing. “People that are going to order home tests are going to be motivated to action in other ways. And so, thinking about HIV home testing, which was the grandparent of COVID-19 home testing, this really shows us how you reach people you’re not going to reach when you have lab-based, provider-only testing … When you look at the HIV home testing data from the CDC, 26% of the people that ordered a home test had never been tested before. That is way higher than what you would expect,” he told MedPage Today.
We are not out of the woods in regard to monkeypox, vigilance is still required. But with existing harm reduction measures in the most vulnerable community, at-home testing and advancements in vaccines could help us keep our numbers as low as possible.
Recent intrusions into the hospitals’ IT systems resulted in blocked medical records including medical laboratory data
Healthcare cyberattacks continue to be a threat that bring potentially costly business consequences for clinical laboratories. Just in the past month, two hospital systems had their health information technology (HIT) systems disrupted due to security incidents. In response, the hospitals’ medical laboratories were forced to switch from digital to paper documentation and, in at least one case, the organization reportedly had difficulty accessing electronic laboratory test results.
At Tallahassee Memorial, an “IT security issue” on Feb. 2 resulted in the organization shutting down its IT systems for 13 days, including at its clinical laboratory. The hospital’s computer network went back online on Feb. 15, according to a news release.
At Atlantic General Hospital, according to an AGH news release, IT personnel discovered a ransomware attack on Jan. 29 that affected the hospital’s central computer system. As a result, the walk-in outpatient laboratory was closed until Feb. 14.
These recent cyberattacks underscore the importance for clinical laboratory leaders to have plans and procedures already in place prior to a disruption in access to critical patient data.
Healthcare cyberattacks can be a “complete blindside for a lot of organizations that think they have protections in place because they bought a product or they developed a policy,” said Ben Denkers (above), Chief Innovation Officer at CynergisTek, an Austin, Texas-based cybersecurity company, in an exclusive interview with The Dark Report. Since clinical laboratory test results make up about 80% of a patient’s medical records, disruption of a hospital’s IT network can be life threatening. (Photo copyright: The Dark Report.)
Laboratory Staff Unable to View Digital Diagnostic Results at Tallahassee Memorial
Though the exact nature of the incident at Tallahassee Memorial HealthCare has not been divulged, hospital officials did report the incident to law enforcement, which suggests a cyberattack had occurred.
Electronic laboratory test results were among the casualties of the IT difficulties at TMH. “Staff have been unable to access digital patient records and lab results because of the shutdown,” a source told CNN.
Attempts by Dark Daily to reach a medical laboratory manager for comment at TMH were unsuccessful. However, in a news release posted online shortly after the cyberattack, the health system advised staff members on dealing with the IT outages.
“Patients and families may notice the switch to paper documentation during registration, admission, or during their care, as our providers will be using paper forms, prescription pads, handwritten notes, or other similar paper methods where they may usually use an electronic process,” the news release stated. “We apologize for any delays this may create. We practice for situations like this, and we are prepared to provide safe, high-quality care to our patients during computer system downtimes.”
Atlantic General Hospital Reports Ransomware Incident to the FBI
At Atlantic General Hospital, the outpatient walk-in laboratory and outpatient imaging department both temporarily closed because of the ransomware attack.
Staff members throughout the hospital were “forced to manually check patients in and out of appointments and record all other information by hand instead of online,” Ocean City Today reported.
The hospital immediately informed the FBI of the ransomware incident and continues to work with an incident response team to determine whether criminals accessed any sensitive data. It was not clear whether the organization ultimately paid a ransom to unlock its systems.
The hospital’s medical laboratory director did not respond to an email from Dark Daily seeking further comment.
Healthcare Cyberattacks Attempt to Gain Access to Data
Therefore, it is critical that clinical laboratory and hospital staff work with their IT counterparts to verify that technology and processes are in place to protect access to patient data.
In “Labs Must Audit Their Cybersecurity Measures,” Ben Denkers, who at that time was Chief Innovation Officer at CynergisTek, a cybersecurity firm based in Austin, Texas, told The Dark Report, “Testing, validating, and auditing whether measures are working as designed is a change of mentality for a lot of organizations.” (If you don’t subscribe to The Dark Report, try our free trial.)
An IT network attack is an attempt by a cybercriminal to gain unauthorized access to devices that contain and exchange data within an organization. Although this information may be on individual devices or on servers, network attacks are often only possible after a hacker enters a system through an endpoint, such as an individual’s email inbox.
“It’s important to understand that while the network server itself might have ultimately been the target, that doesn’t necessarily mean that it was compromised first,” Denkers told The Dark Report. “Phishing is a perfect example of a way an attacker could first gain access to a workstation, and then from there move laterally to a server.”
The final cost of a healthcare cyberattack often exceeds the ransom. Media coverage can lead to an organization’s diminished reputation within the community, and if protected health information (PHI) is accessed by the criminals, a hospital or health system may need to pay for identity theft monitoring for affected patients.
As part of the settlement, Banner Health paid a $1.25 million penalty and will carry out a corrective action plan to protect PHI in the future and resolve any alleged HIPAA violations, according to the HHS Office for Civil Rights.
This hefty penalty is a reminder to pathologists and clinical laboratory managers that—when it comes to cyberattacks—the classic adage “an ounce of prevention is worth a pound of cure” is appropriate advice.
‘Balwani is no Johnny Depp,’ says an expert on juror behavior, as prosecution and defense rest in fraud trial of the former executive of the now-defunct lab test company
Clinical Laboratory directors and pathologists continue to focus like a laser beam on the trials of former founders and executives of the now-defunct blood test company Theranos. But as the criminal fraud trial of ex-president and COO Ramesh “Sunny” Balwani comes to a close, legal experts maintain the 57-year-old businessman may face an uphill battle to win an acquittal.
Balwani faces 12 counts of wire fraud and conspiracy to commit wire fraud while serving as second in command at Theranos, the former Silicon Valley medical laboratory test startup. The fraud trials of Balwani and Theranos founder Elizabeth Holmes have made headlines for more than a year as the two once-high-flying executives face a reckoning for allegedly defrauding patients, investors, and physicians about their proprietary Edison blood-testing device, which they claimed could conduct hundreds of blood tests using a finger-prick of blood.
Before resting their case, Balwani’s defense team called only two witnesses: information-technology consultant Richard Sonnier III, and naturopathic physician Tracy Wooten, NMD, of Arizona, who sent more than 100 patients to Theranos.
Theranos LIS Not Accessible to Government Prosecutors
Sonnier was hired by Balwani’s legal team to assess the accessibility of data held in the LIS, which the defense believed would have provided evidence of Theranos test accuracy.
The WSJ noted that in 2018, the year Balwani and Holmes were indicted, the government subpoenaed a copy of the LIS, which Theranos provided. However, the LIS data was delivered on an encrypted hard drive.
“Not only was the hard drive itself encrypted, but the data it contained was also encrypted with a separate passcode required,” the WSJ wrote. “The government didn’t have the passcode to access the data, and a day or two after sending the hard drive to US attorneys, Theranos officials ordered the entire original database dismantled, according to court testimony.”
The WSJ reported that Sonnier testified he was unable to access the encrypted data on a backup hard drive despite having a list of possible passcodes found in Theranos documents. Sonnier also testified that it would have been “very straightforward” to reassemble the original LIS and “recover that data.” The missing password wouldn’t be an issue, Sonnier testified.
The Prosecution Rests
Federal prosecutors rested their case last month after calling more than 24 witnesses. The government alleges Balwani worked closely with Holmes and conspired with her to defraud investors and patients about the startup’s blood testing technology. They allege he knew about the accuracy and reliability problems that plagued Theranos’ Edison blood-testing device.
Holmes was convicted in January on three of the nine fraud counts and one of two conspiracy counts. She was acquitted on four counts related to defrauding patients, one charge of conspiracy to commit wire fraud and three charges of wire fraud.
While prosecutors failed to persuade jurors that Holmes intentionally sought to defraud patients, Bloomberg legal reporter Joel Rosenblatt told the Bloomberg Law Podcast he believes Balwani is “inherently more vulnerable” on the patient-related fraud counts because he “oversaw” the operation of Theranos’ clinical laboratories.
“As a result of that role, [Balwani] was more aware of not only the faulty Theranos blood test results, but all the problems that employees were pointing out about those results,” Rosenblatt added. “So, he was the first high-level executive to be dealing with those complaints.”
Rosenblatt noted that Balwani’s defense centers not only on trying to show that Theranos’ proprietary blood-testing machine worked, but that it “works maybe well enough or worked as well as other [medical] laboratories.” He said Balwani also maintains that Holmes, as CEO and founder, was in charge long before he joined Theranos as president.
“It’s a difficult argument to make because all the emails show how cooperative they were, how closely they worked together. They were intimately involved but they were working side by side for years and really during the years where all the money started coming in,” Rosenblatt said in the podcast.
“He has a lot of problems that [Elizabeth Holmes] didn’t have,” Taylor said. “He kind of fits the part from a juror’s standpoint. He’s got the power, the authority, he’s got the personal traits that make the allegations more credible from a perceptual standpoint for the jury.”
In contrast, Taylor says, “People don’t love Elizabeth Holmes, but I think what she had going for her was that she pitched herself as a true believer in the company. She was the voice and the face of Theranos.”
“Sunny Balwani is not Johnny Depp. He doesn’t have the halo that Johnny Depp has, or the fan base,” Taylor said. “He does not present as that type of person, so I don’t know that the jurors will have any sympathy towards him. And I think they would actually be more inclined to believe Holmes’ allegations.”
The Theranos fraud trials of Holmes and Balwani continue to capture the attention of clinical laboratory directors and pathologists who are now witnessing the final chapters in the downfall of the one-time Silicon Valley power couple.