A new study shows how routine blood tests and machine learning can help predict the risk of threatened miscarriage early in pregnancy.
As laboratory diagnostics continue to evolve alongside artificial intelligence (AI) and machine learning (ML), a new study out of China highlights the growing potential of routine blood testing as a predictive tool in obstetric care. Researchers have developed a machine learning model capable of identifying early signs of threatened miscarriage using only standard blood test results.
For lab leaders, this is a key opportunity to use existing lab data and infrastructure to support early pregnancy risk detection. Properly standardized blood data can help shift diagnostics upstream, potentially identifying issues before symptoms appear.
A threatened miscarriage (also called threatened abortion) is a medical condition in early pregnancy where a woman experiences vaginal bleeding before 20 weeks gestation, but the cervix remains closed and the pregnancy is still viable—meaning the fetus is alive with a detectable heartbeat.
To perform the research, scientists from the Second Affiliated Hospital of Shaanxi University of Chinese Medicine, collected medical records and analyzed data from 1,764 patients with threatened miscarriage and 1,489 healthy control subjects. Blood test data of the study participants were collected and a data preprocessing tactic called the Z-score normalization technique, also known as standardization, was applied to routine blood indicators. The high-level, general-purpose programming language called Python was used to facilitate the data transformation of eight different ML algorithms. The performance of those ML models were evaluated by calculating their area under the curve (AUC) values, which represents the overall performance of the various models.
Machine Learning Unlocks Predictive Power of Routine Blood Tests
The ML algorithm known as Deep Neural Network (DNN) achieved the best predictive performance with an AUC value of 96.76% That model also had the top metrics for accuracy at 91.88%, specificity (91.62%) and sensitivity (92.11%).
According to the study, approximately 25% of all pregnancies involve a threatened miscarriage and women who experience this condition are 2.5 times more likely to experience a miscarriage compared to healthy pregnancies.
“We’re not sure why this happens in some pregnancies but not in others,’ said Lisa Jackson, MD, assistant clinical professor in the Raquel and Jaime Gilinski Department of Obstetrics, Gynecology and Reproductive Science at Icahn School of Medicine at Mount Sinai, in an interview with The Bump.
Lisa Jackson, MD, assistant clinical professor in the Raquel and Jaime Gilinski Department of Obstetrics, Gynecology and Reproductive Science at Icahn School of Medicine at Mount Sinai noted, “The good news is that the majority of the time, bleeding in early pregnancy with the presence of fetal heart tones and a closed cervix doesn’t result in miscarriage.” (Photo credit: Icahn School of Medicine at Mount Sinai)
“For most women, the outlook is good, and they go on to have normal pregnancies,” added women’s health expert Jennifer Wider, MD.
Neither Jackson nor Wider were affiliated with the study out of China.
The study links incidents of threatened miscarriage with various factors, including chromosomes, genes, hormonal imbalances, immune factors, and maternal health as well as environmental influences. The researchers also stated in their study, “Women experiencing threatened abortion often encounter significant stress and require multiple clinical assessments to accurately determine their pregnancy status. Threatened abortion not only poses a threat to the health of pregnant women but may also affect their future fertility.”
The researchers believe their ML algorithm has immense clinical value and can assist doctors in accurately identifying high-risk patients. They suggest exploring the application of the DNN model in different clinical settings and patient populations and incorporating more diverse data sources to optimize its predictive capabilities.
A New Role for Clinical Labs in Reproductive Health
The researchers concluded, “Our research on constructing a prediction model for threatened abortion through routine blood tests has revealed the great potential of ML algorithms in detecting threatened abortion. This algorithm is expected to analyze routine blood data to identify at-risk pregnancies at an early stage, significantly improving the early detection of this common pregnancy complication. It will assist healthcare providers in intervening earlier and reducing the incidence of (threatened) abortion.”
More research and trials are needed before any ML model may be used for the early detection of threatened miscarriage incidents. However, this latest study does suggest the potential of machine learning algorithms could be beneficial in timely predictions of the medical condition, which would improve patient outcomes.
For laboratories, this research underscores the untapped value in the routine data already being collected daily. As machine learning models like the deep neural network in this study continue to demonstrate clinical relevance, lab leaders are uniquely positioned to facilitate their integration—by ensuring high-quality, normalized data flows and supporting interdisciplinary collaboration between data scientists and clinicians.
Moving forward, blood testing won’t just confirm conditions, it may soon help predict and prevent them. With strategic investment in data analytics and model validation, labs can transform from diagnostic endpoints into proactive partners in patient care.
The new Plasma Immuno Prediction Score achieves 96% accuracy in forecasting TNBC outcomes, offering laboratories a powerful tool for precision oncology.
For laboratory professionals, the latest findings in plasma proteomics underscore how the clinical laboratory is becoming central to guiding cancer treatment decisions. The discovery of blood-based protein signatures that can predict immunotherapy outcomes in triple-negative breast cancer (TNBC) demonstrates how lab-developed tests and biomarker assays can directly influence patient care, moving precision oncology forward.
A news release explained that a team of researchers in China has identified a set of plasma proteins that can reliably predict whether patients with TNBC will respond to immunotherapy, potentially transforming treatment strategies for one of the most aggressive forms of breast cancer.
They found that three proteins in particular—arginase 1 (ARG1), nitric oxide synthase 3 (NOS3), and CD28—were strongly linked to treatment outcomes. From this, the team developed a predictive model called the Plasma Immuno Prediction Score (PIPscore), which achieved nearly 86% accuracy in forecasting responses.
“This study transforms how we approach TNBC immunotherapy,” said Yizhou Jiang, MD, co-corresponding author. “By translating complex plasma proteomics into a practical score, we’ve bridged the gap between research and clinical utility.”
Breaking the Bottleneck in TNBC
Triple-negative breast cancer accounts for about 15% of breast cancer cases worldwide and is notoriously difficult to treat because it lacks the hormonal and HER2 targets used in other subtypes. Immunotherapy has emerged as a promising option, but predicting which patients will benefit remains a challenge.
Currently, clinicians rely on biomarkers like PD-L1 expression or tumor mutational burden. However, these markers often fail to capture the complexity of immune responses, leaving doctors without reliable tools to guide decisions. Tumor biopsies, another option, are invasive and impractical for frequent monitoring.
Yizhou Jiang, MD, Fudan University Shanghai Cancer Center, Fudan University, said “Plasma proteomics provides a non-invasive window into systemic immunity. Our work shows that the blood can tell us as much, if not more, than the tumor itself about how a patient will respond.”
How the Study Worked
The research team analyzed dynamic changes in plasma proteins over the course of immunotherapy. Patients who responded to treatment showed sharp rises in immune-activating proteins such as CXCL9 and interferon-gamma (IFN-γ). Those who achieved a pathologic complete response (pCR)—meaning no detectable cancer remained after treatment—had higher levels of ARG1 and CD28, but lower levels of NOS3.
According to the study, these proteins appear to regulate critical aspects of immune activation and tumor suppression. Elevated NOS3, for example, correlated with fewer CD8+ T cells in tumors, suggesting an immunosuppressive role. In contrast, ARG1’s role in arginine metabolism may boost T-cell function and strengthen immune attack on tumors.
To integrate these findings, the researchers developed the PIPscore, a composite of six proteins including ARG1, NOS3, and IL-18. This model stratified patients into high- and low-response groups with impressive precision. The area under the curve (AUC)—a common measure of predictive performance—was 0.858, indicating strong accuracy.
Perhaps most strikingly, the PIPscore predicted 12-month progression-free survival with 96% accuracy, highlighting its potential clinical value.
Linking Blood to Tumor Biology
To strengthen their conclusions, the team also used single-cell RNA sequencing to link blood protein signatures with changes in the tumor microenvironment. For example, patients with higher NOS3 levels showed reduced infiltration of CD8+ T cells into tumors, aligning blood-based findings with tissue-level biology.
“This dual approach—measuring proteins in the blood and validating them against the tumor microenvironment—offers a holistic view of how immunotherapy works,” Jiang said. “It underscores that systemic immunity, not just local tumor factors, dictates treatment success.”
Clinical Implications
The potential benefits of this approach are wide-ranging. Oncologists could use the PIPscore to determine upfront whether a TNBC patient is likely to respond to immunotherapy, sparing non-responders from ineffective treatments, unnecessary side effects, and high costs. Because the test is blood-based, it could be repeated over time, allowing clinicians to adjust treatment plans in real time.
“The PIPscore not only predicts response but also opens doors to targeting metabolic pathways like arginine deprivation to overcome resistance,” Jiang noted. “These findings underscore the importance of systemic immunity.”
Beyond TNBC, the researchers believe the method could be applied to other cancers where immunotherapy outcomes are highly variable.
In addition to plasma proteomics, the field of pharmacogenomics offers another layer of precision in cancer care by examining how genetic variations influence drug response.
When combined with tools like the PIPscore, pharmacogenomic profiling could help oncologists tailor both immunotherapy and supportive treatments to individual patients. For laboratory professionals, this integration underscores the expanding role of molecular diagnostics in personalizing therapy—ensuring patients not only receive the right drug but also the right dosage based on their genetic and immune profiles.
Next Steps
The study’s authors acknowledge that further validation is needed before the PIPscore can enter routine clinical practice. Larger, multi-center trials will be necessary to confirm its reliability across diverse patient populations. Still, experts view the findings as a major step toward more precise cancer care.
As immunotherapy adoption grows, laboratory professionals will be essential in validating, standardizing, and implementing predictive tools like the PIPscore in clinical practice. Their expertise in assay development, quality control, and biomarker interpretation ensures that discoveries at the research level can be reliably translated into real-world diagnostics, ultimately improving outcomes for patients with aggressive cancers like TNBC.
The World Health Organization has released updated Target Product Profiles to guide manufacturers in developing faster, more affordable, and accurate tuberculosis screening tools.
Since 2014, the World Health Organization (WHO) has regularly published Target Product Profiles (TPP) for tuberculosis (TB) screening tests. Aimed at manufacturers of TB tests and equipment, they are akin to a wish list for addressing global TB screening needs. The WHO recently published updated TPPs.
The WHO indicates that in 2023, a total of 1.25 million deaths occurred from TB globally. In that year, about 10.8 million people were estimated to have TB, with official diagnoses of about 8.2 million and another 2.6 million believed to have gone undetected. Of the 10.8 million, 1.3 million were children.
TB is an airborne infectious disease caused by the bacterium Mycobacterium tuberculosis. Primarily affecting the lungs, it can impact the entire body. It is typically diagnosed via tuberculin skin tests, blood tests, chest X-rays, and sputum analysis.
TB has recently been moved to the top of the list—over COVID-19—as the leading cause of death in the world from a single infectious agent. TB is also the top killer of people with HIV. About two out of five TB deaths are related to antibiotic-resistant strains of TB.
The WHO stated in the TPPs that, “If TB screening is to be implemented at the levels needed to significantly reduce the global TB burden, there is an urgent need for new tools for TB screening that can be readily and rapidly deployed and brought to scale.”
Tereza Kasaeva, MD, PhD, Director of the WHO Department for HIV, noted “The new WHO TPPs will inform manufacturers about the features that these new tests should have if they are to address the needs of the TB community and achieve the ambitious global commitments for TB detection and reduction of TB incidence and deaths.”
There have been many recent advancements, including mobile and portable digital radiographic equipment, as well as artificial intelligence (AI) in computer-aided detection software. Yet, the WHO notes, there are still challenges to scaling technology to close “the case detection gap.” Some of those challenges include availability, low cost, and high accuracy.
The TPP defined three distinct types of screening tools:
High-sensitivity, high-specificity screening test using a single screening step before referral for diagnostic testing or ruling out TB.
High-sensitivity screening as part of a two-step screening process before referral for diagnostic testing.
A moderate-sensitivity screening test with high specificity that can be used in settings that have limited access to health systems and can’t be reached via existing screening technology.
Current screening methods have accuracy limitations and difficulty in screening asymptomatic TB. The WHO TPPs noted that, “CXR (chest X-ray) can be used to screen most populations and has the best demonstrated sensitivity of available screening tools. CAD software for automated interpretation of CXR imaging can be used along CXR for screening all adults aged 15 years and older.”
The WHO also recommends rapid molecular diagnostic tests for screening all adults, and C-reactive protein (CRP) to screen adults living with HIV.
The latest TPP emphasizes effective community-based screening on a larger scale in settings with lower TB prevalence. This is a challenge to TB assay developers, especially in finding cost-effective methods with high accuracy. Ideally, they would have a small per-test cost and very little or no equipment or capital investments. If they do require centralized equipment for processing, they should have low per-test costs and implementation requirements.
For each category of testing, the TPP sets pricing goals. For high-accuracy one-step tests, the minimum is ≤ US$3; for high-sensitivity two-step tests, the minimum is ≤ US$2 with an optimal of ≤ US$1; and for a high-specificity test, the minimum is ≤ US$2.
The TPP states, “Ideally, a new test should be priced according to the cost of goods and estimated volumes, and should include a reasonable profit margin. Ensuring access to tests while maintaining business interests can be achieved through fair pricing, which requires transparency about the cost of goods and estimated volumes, and a reasonable profit margin.”
The expert panel that created the TPP understands the challenges, noting that “manufacturers may have to consider trade-offs, whereby a novel test has some clear advantages that may offset other downsides.”
Ultimately, the WHO states that the TPPs were issued to “guide and encourage product development for new screening tools in these settings.”
Danaher’s chief medical officer, Maximilian Schmid, says ‘infrastructure is there’ to push forward whole slide imaging.
Based on discussions at the recent Association for Diagnostics & Laboratory Medicine’s (ADLM) 2025 conference and observations from other sources, digital pathology trends continue to show sluggish adoption by clinical laboratories and anatomic pathology practices in the US. However, proponents of whole slide imaging say now is a prime opportunity to integrate the technology with a patient-centric care approach.
At least 65 vendors at the ADLM 2025 exhibitor hall indicated that their products touched whole slide imaging or digital pathology, showing this area is a hot focus for sellers.
Based on its customers, Leica has one of the largest installation bases of digital pathology scanners in the world. From that perspective, the technology is available, yet obstacles remain.
“The infrastructure is there,” said Maximilian Schmid, MD, chief medical officer at Danaher Diagnostics. “How do we bring it to the patients?”
Danaher Corporation, which held a press briefing at ADLM 2025, is parent company to Leica.
ADLM 2025 took place in Chicago July 27-31at McCormick Place. (Photo credit: Scott Wallask.)
Labcorp Report Notes Costs as a Barrier
Anyone who has followed the slow progression of digital pathology knows adoption from the diagnostic lab industry has been lukewarm. In 2024, Labcorp released a report about clinical laboratory trends that indicated cost remained a hurdle to further use of digital pathology scanners and software. Based on a survey of 115 US-based pathologists, lab managers, and lab directors, the report concluded that just 33% of respondents had started orplanned to implementdigital pathologyin lab workflows.
“Industry adoption of digital pathology has been slower than expected, largely due to high initial costs,” Deborah Sesok-Pizzini, MD, MBA, chief medical officer at Labcorp, told Today’s Clinical Lab at the time. That publication is a partner brand to Dark Daily.
Digital Pathology Trends are Rosier in Europe
Leica has been able to convey a message to its customers that despite the initial costs, the return on investment for digital pathology is high in terms of more accurate diagnoses and quicker processes, Schmid said.
“When I look ahead 10 years, digital pathology will be as normal to labs as H&E staining,” he predicted, referring to common hematoxylin and eosin stains.
He added that while Labcorp’s study reflected US trends, digital pathology adoption is stronger elsewhere, based on what Danaher and Leica see with customers.
“Europe seems to be a little bit ahead in terms digitization,” including with whole-slide imaging, he noted. However, even in the US, “academic medical centers are moving very fast in this direction.”
Schmid’s assertion about Europe is supported by other sources. For example, a business case published in July 2024 by the UK’s National Health Service (NHS) for Wales indicated that Northern Ireland and Scotland had near-fully digitized cellular pathology programs for the NHS, and England was building up its network. Wales was seen as trailing behind these countries.
“The national move towards scanning of histological material for primary diagnosis and more recently, the adoption of artificial intelligence (AI)/computational pathology to improve the accuracy, reliability and quality of reports, means that most pathologists, especially new trainees who are already using digital technology, will, in the future, choose to work in departments where digital technology will enhance and underpin their diagnosis thus benefiting the quality of patient care,” the NHS Wales business case paper stated.
Computational Pathology’s Growing Role
Computational pathology—in other words, the use of data science, information, and digital technologies for laboratory medicine—is a key to moving precision medicine forward via digital pathology, said Nicole Selenko-Gebauer, MD, MBA, group vice president and chief innovation officer at Danaher Diagnostics.
“We need to complement technology-driven focuses” with a patient-centric approach, Selenko-Gebauer added.
Even in 2022, The Dark Report had alerted its readers to the promise of computational pathology, noting Mayo Clinic Laboratories’ early success in launching related clinical assessment goals based on digital pathology and artificial intelligence. (If you’re not a Dark Report subscriber, check out our 14-day free trial.)
Taken at that viewpoint, digital pathology trends related to patient care will be an important milestone for the technology.
“A pivotal moment will be the clinical utilization of digital pathology—that it works and is accurate,” Schmid said.
Lab leaders can now align with WHO’s new product profile to develop innovative diagnostics that address gaps in newborn infection detection and antimicrobial resistance prevention.
The World Health Organization (WHO) has unveiled a new target product profile aimed at guiding the development of in vitro diagnostic tests for detecting serious bacterial infections in newborns and young infants, including neonatal sepsis—a major cause of infant mortality worldwide. This initiative comes in response to alarming statistics: 2.3 million newborns die every year, with around 15% of these deaths linked to sepsis. The burden is heaviest in low- and middle-income countries, where access to timely diagnosis and treatment remains limited.
Current diagnostic methods such as blood cultures and molecular diagnostics are often inaccessible, expensive, or unreliable, particularly in resource-limited settings. They suffer from low sensitivity, long turnaround times, and high infrastructure demands, making them poorly suited for the urgent needs of frontline healthcare facilities.
Lab Leaders See Shift in Global Standards
For laboratory leaders, WHO’s newly released target product profile for diagnosing serious bacterial infections in newborns and young infants could represent a pivotal shift in global diagnostic standards. As the demand for timely, accurate, and affordable diagnostic tools grows—particularly in low-resource settings—lab leaders have a unique opportunity to play a central role in shaping the next generation of in vitro diagnostics.
The new profile to is designed to define the essential and desirable characteristics of diagnostic tools needed to improve early detection of infections in infants aged 0–59 days. The profile covers two major clinical scenarios: use in primary health care settings and in higher-level hospitals. It’s intended to assist developers, regulators, and public health officials in designing effective tools tailored to diverse healthcare environments.
Yvan Hutin, MD, PhD, Director of the Department of Antimicrobial Resistance at WHO, emphasized the urgency of the effort, stating, “Timely and accurate diagnosis tests for serious bacterial infection is critical to reducing newborn mortality.”
Hutin’s areas of expertise include epidemiology, prevention, care and treatment of viral hepatitis, public health training, economic analyses, and financing. He has co-authored more than 120 articles in peer-reviewed journals. (Photo credit: WHO)
Silvia Bertagnolio, MD, Head of the Antimicrobial Resistance Surveillance, Evidence and Laboratory Strengthening Unit at WHO noted, “This new target product profile outlines the essential features needed in diagnostic tools to improve clinical decision-making, reduce unnecessary antibiotic use and prevent antimicrobial resistance, especially in low-resource settings where the burden of neonatal sepsis remains critical.”
Targeted Care
In many communities, healthcare workers must make critical decisions about antibiotic treatment or hospital referrals based on clinical judgment alone, without reliable diagnostics. The new tools envisioned by WHO would help identify which infants truly need antibiotics or hospitalization, allowing for more targeted and effective care.
For laboratory leaders, the release of this target product profile is more than just a technical guideline, it’s a call to action. By aligning their diagnostic development and evaluation efforts with WHO’s outlined priorities, lab leaders can help fill a critical gap in newborn and infant care, particularly in underserved regions. Their expertise will be essential in translating this profile into practical, scalable solutions that improve early detection, guide appropriate treatment, and ultimately save lives. As stewards of innovation and quality in diagnostics, lab leaders are uniquely positioned to drive meaningful progress in the global fight against neonatal infections and antimicrobial resistance.
The NHS plans to map the DNA of all newborns in England over the next 10 years to evaluate their risk for hundreds of diseases
Clinical laboratories in the UK will soon see a new increase for specialized testing. Under a new 10-year plan by England’s National Health Service (NHS), every newborn baby in the country will have their complete DNA mapped using tiny blood samples taken from their umbilical cords shortly after birth. The blood samples will assess an infant’s risk for hundreds of diseases with the intent of predicting and preventing those illnesses while mitigating demand for services and saving money. The plan itself is set to be revealed in the near future.
In October, the NHS announced it would be analyzing the genetic code of up to 100,000 newborns via a small drop of blood collected from the heel. Currently, testing usually occurs when a baby is five days old and looks for nine rare, but serious gene disorders that develop in early childhood and have effective treatments available.
The new umbilical cord testing utilizes genomics, artificial intelligence (AI), predictive analytics, and other technologies to provide faster diagnoses and treatments for approximately 7,000 single-gene disorders.
“With the power of this new technology, patients will be able to receive personalized healthcare to prevent ill-health before symptoms begin, reducing the pressure on NHS services and helping people live longer, healthier lives,” stated Wes Streeting, NHS Secretary of State for Health and Social Care in a statement.
Wes Streeting, NHS Secretary of State for Health and Social Care said, “”Genomics presents us with the opportunity to leapfrog disease, so we’re in front of it rather than reacting to it.” (Photo credit: GOV.UK)
This new, whole genome sequencing procedure for infants is part of a 10-year plan by the NHS to establish major shifts in how healthcare is delivered in the UK to improve the quality of care and increase transparency. The intent is to move delivery from hospital to community, which includes the implementation of “neighborhood health teams” to aggregate services. The plan also includes transitioning from analog to digital methods and from treating illnesses to preventing them.
“Our 10-year plan will build on the founding promise of the NHS, so that it provides health care free at the point of risk, not just need,” said Streeting.
He also stated technological advances will help alleviate pressures on the NHS and contribute to its future success, thus improving the overall health of the population.
“As we deliver the transformational shifts in our 10-year plan, from hospital to community, analog to digital, and sickness to prevention, it will have radical implications for services,” he said. “Much of what’s done in a hospital today will be done on the high street, over the phone, or through the app in a decade’s time.”
While many scientists, doctors and patient advocacy groups applaud this testing on infants, there are some concerns it may spark an ethical debate. Parents will have the ability to give consent, but the testing may result in information they may not want to know, which could have a negative psychological effect on children and parents who are aware they have a higher risk for certain diseases.
There are also concerns regarding the security of crucial patient data and how such information can be prone to security breaches. The DNA and health records of infants are stored on secure systems and require strict authorization to access, but this type of data is very valuable to cyber hackers. The NHS has stressed it is prioritizing digital security measures, including vigorous cybersecurity, data governance and the implementation of ethical guidelines for AI development.
Robin Lovell-Badge, PhD, principal group leader and head of the Laboratory of Stem Cell Biology and Developmental Genetics at the Francis Crick Institute also noted some reservations he has regarding how the collected data is provided to patients.
“You need people to have conversations with individuals who might be affected by genetic disease,” he said. “One of the things that worries me was an insufficient number of genetic counselors. It’s not just having the information, it’s conveying the information in an appropriate, helpful way.”
As technological advances become increasingly prevalent in medicine, clinical labs will be at the forefront of new initiatives such as the 10-year plan by the NHS.