By Budi Oetomo Narendra 
Researchers at the Johns Hopkins Kimmel Cancer Center have developed an artificial intelligence (AI)-driven liquid biopsy that analyzes genome-wide patterns of cell-free DNA (cfDNA) fragments circulating in the blood. This groundbreaking test examines how these DNA pieces break apart and where they appear across the entire genome. Leveraging this intricate information, the system can identify early signs of liver fibrosis and cirrhosis and holds significant promise for detecting broader indicators of chronic disease, marking a substantial leap forward in diagnostic medicine.
Expanding the Diagnostic Frontier: Beyond Cancer
The study, partly funded by the National Institutes of Health, was officially published on March 4 in the esteemed journal Science Translational Medicine. This publication is particularly noteworthy as it signifies the first systematic application of this type of DNA fragmentation analysis, widely known as fragmentome technology, to the detection of chronic diseases unrelated to cancer. Historically, the fragmentome approach has predominantly been investigated and refined as a method for identifying various forms of cancer, demonstrating considerable promise in that domain. The pivot to non-cancerous chronic conditions opens up an entirely new realm of diagnostic possibilities, potentially transforming how a wide array of illnesses are detected and managed. This expansion underscores the versatility and profound potential of analyzing the "fragmentome" – the complete collection of cfDNA fragments in an individual’s bloodstream – as a comprehensive biomarker for physiological health.
The Science Behind the Breakthrough: Unpacking the Fragmentome
Cell-free DNA (cfDNA) consists of short DNA fragments released into the bloodstream, primarily from dying cells through processes like apoptosis (programmed cell death) and necrosis (uncontrolled cell death). While cfDNA has long been known to circulate in the blood, its diagnostic potential has surged with advancements in genomic sequencing and bioinformatics. Its utility stems from the fact that cells from various tissues, including diseased ones, contribute to the cfDNA pool, offering a dynamic and accessible "snapshot" of the body’s physiological state.
The fragmentome approach capitalizes on the intricate ways DNA is packaged within cells. DNA is wound around proteins called histones to form nucleosomes, which in turn form higher-order structures called chromatin. When cells die, nucleases (enzymes that break down DNA) cleave the DNA preferentially in regions not protected by nucleosomes. This precise enzymatic activity results in cfDNA fragments of characteristic sizes and genomic positions. Disease states, particularly those involving inflammation, tissue damage, or altered gene expression, can modify chromatin structure, leading to distinct fragmentation patterns. By meticulously analyzing these subtle changes across the entire genome, the Johns Hopkins team can infer the presence and nature of underlying pathologies.
In this new research, investigators performed whole-genome sequencing on cfDNA samples collected from an extensive cohort of 1,576 individuals, many of whom were diagnosed with liver disease and additional medical conditions. By meticulously examining DNA fragments across the entire genome, they searched for patterns that might signal disease. The team analyzed both the precise size of DNA fragments and their specific distribution throughout the genome, crucially including repetitive DNA regions that have often been overlooked or rarely studied in previous analyses. Each individual analysis was incredibly data-rich, encompassing approximately 40 million fragments spanning thousands of genomic regions. This generated an enormous dataset, far surpassing the complexity and scale of information typically processed by most conventional liquid biopsy tests, highlighting the advanced nature of this methodology.
The integration of artificial intelligence, specifically sophisticated machine learning algorithms, was paramount to this breakthrough. Given the sheer volume of data generated—where human analysis alone would be impractical—AI excels at identifying complex, subtle patterns within massive datasets that might be invisible to the human eye. In this context, machine learning models were trained on a vast array of cfDNA fragmentation profiles from individuals with known liver conditions, allowing them to "learn" the specific signatures associated with early fibrosis, advanced fibrosis, and cirrhosis. Using these identified patterns, researchers successfully created a classification system that demonstrated high sensitivity in detecting early liver disease, advanced fibrosis, and full-blown cirrhosis.
"This builds directly on our earlier fragmentome work in cancer, but now using AI and genome-wide fragmentation profiles of cell-free DNA to focus on chronic diseases," explained Victor Velculescu, M.D., Ph.D., co-director of the cancer genetics and epigenetics program at the Johns Hopkins Kimmel Cancer Center and co-senior author of the study. "For many of these illnesses, early detection could make a profound difference, and liver fibrosis and cirrhosis are important examples. Liver fibrosis is reversible in its early stages, but if left undetected, it can progress to cirrhosis and ultimately significantly increase the risk of liver cancer."
Addressing a Critical Public Health Challenge: Liver Disease
The importance of this diagnostic advancement cannot be overstated, particularly in the context of liver disease, a growing global health concern. According to the Centers for Disease Control and Prevention (CDC), approximately 4.5 million adults in the U.S. are diagnosed with liver disease, and millions more may have undiagnosed conditions like non-alcoholic fatty liver disease (NAFLD), which can silently progress to fibrosis and cirrhosis. Cirrhosis, the end stage of chronic liver disease, is the 12th leading cause of death in the United States, accounting for over 50,000 deaths annually. The economic burden associated with liver disease is substantial, with healthcare costs estimated in the tens of billions annually, underscoring the urgent need for more effective early detection and intervention strategies.
Current diagnostic methods for liver fibrosis often present significant limitations. Standard blood-based tests, which measure liver enzymes such as AST (aspartate aminotransferase) and ALT (alanine aminotransferase), frequently lack the necessary sensitivity, especially in the crucial early stages of disease. These traditional blood markers typically fail to detect early fibrosis and identify cirrhosis only about half the time, leading to delayed diagnoses and missed opportunities for intervention. Imaging techniques, such as specialized ultrasound or magnetic resonance elastography (MRE) scans, can provide more accurate assessments of liver stiffness, a proxy for fibrosis. However, these tools require specialized equipment and expertise that are not always readily available, particularly in primary care settings or underserved regions, creating accessibility barriers for many individuals at risk.
Velculescu notes that roughly 100 million people in the United States have liver conditions that elevate their risk of progressing to cirrhosis and liver cancer. "Many individuals at risk don’t know they have liver disease," Velculescu points out. "If we can intervene earlier—before fibrosis progresses to cirrhosis or cancer—the impact could be substantial." He further emphasizes that identifying these precursor conditions at an early stage may enable doctors to treat underlying diseases sooner, such as metabolic syndrome or viral hepatitis, and potentially prevent the devastating progression to cirrhosis and the subsequent development of liver cancer.
Why DNA Fragment Analysis Is Different: A Holistic View
Unlike many conventional liquid biopsy methods that primarily search for specific cancer-related gene mutations, the fragmentome approach adopts a broader, more holistic view. It focuses not on individual genetic alterations but on how DNA fragments are cut, packaged, and distributed throughout the entire genome. According to the researchers, this comprehensive perspective is precisely what makes the method applicable to a much wider range of conditions beyond cancer, crucially including diseases that can eventually raise cancer risk. The study was also co-led by Robert Scharpf, Ph.D., professor of oncology, and Jill Phallen, Ph.D., assistant professor of oncology, both instrumental in developing this innovative diagnostic platform.
"The fact that we are not looking for individual mutations is what makes this study so powerful," stated first author Akshaya Annapragada, an M.D./Ph.D. student in the Velculescu lab. "We are analyzing the entire fragmentome, which contains a tremendous amount of information about a person’s physiologic state. The sheer scale of these data, coupled with the power of machine learning, enables the development of specific classifiers for many different health conditions." This distinction highlights the fragmentome’s ability to capture subtle, systemic changes in the body that might not be reflected in specific gene mutations, offering a more comprehensive and nuanced picture of health and disease.
Chronology of Innovation: From Cancer to Chronic Illnesses
This groundbreaking research did not emerge in isolation but rather grew organically out of prior foundational work. The origins of this study can be traced back to a 2023 Cancer Discovery study, also led by Velculescu, which focused intently on the fragmentome of liver cancer. During the course of that investigation, while studying patients with confirmed liver tumors, the scientists made a critical observation: some individuals who presented with fibrosis or cirrhosis, but not yet overt cancer, showed mostly normal fragmentation profiles. However, upon closer scrutiny, these individuals also contained subtle, yet distinct, DNA signals that were unequivocally linked to their underlying liver disease. This pivotal observation served as the catalyst, prompting the Johns Hopkins team to shift their focus and systematically examine the fragmentome patterns specifically associated with liver fibrosis and cirrhosis, thus laying the groundwork for the current breakthrough. This demonstrates a clear scientific progression, where insights gained from cancer research were skillfully translated and expanded to address other significant chronic diseases.
In another significant analysis within the current study, involving 570 people with suspected serious illness, researchers developed a novel "fragmentation comorbidity index." This sophisticated measure was designed to distinguish individuals with high and low Charlson Comorbidity Index (CCI) scores. The Charlson Comorbidity Index is a widely used and validated metric in clinical practice that estimates how a person’s additional health conditions (comorbidities) affect their risk of death. Strikingly, the fragmentome-based index proved capable of predicting overall survival independently, and in some cases, it even proved more specific and accurate than traditional inflammatory markers, which are commonly used in clinical assessments. Furthermore, certain fragmentation signatures were observed to be consistently associated with poorer clinical outcomes, reinforcing the prognostic power of this new index.
"The fragmentome can serve as a foundation for building different classifiers for different diseases, and importantly, these classifiers are disease-specific and do not cross-react," Annapragada clarified. "A liver fibrosis classifier is distinct from a cancer classifier. This is a unique, disease-specific test built from the same underlying platform." This capability to generate specific, non-overlapping classifiers for various conditions from a single test platform represents a significant advantage, ensuring diagnostic precision and avoiding misinterpretations.
Broader Implications and Future Horizons
The implications of this research extend far beyond liver disease. The study also included individuals identified as being at elevated risk for a diverse range of other medical conditions. Within this broader cohort, researchers observed intriguing fragmentome signals that appeared to be linked to cardiovascular, inflammatory, and neurodegenerative disorders. While the current study population did not include a sufficient number of cases to build separate, robust disease classifiers for each of these additional conditions, these preliminary findings are profoundly suggestive. They strongly indicate that the fragmentome technology may eventually have much wider medical applications, potentially serving as a universal screening tool for various chronic illnesses. Researchers have explicitly stated their plans to investigate these broader applications in future work, paving the way for a more comprehensive diagnostic paradigm.
This technology opens doors to a new era of proactive healthcare. Imagine a future where a single, non-invasive blood test could regularly screen for early markers of not just liver disease, but also nascent cardiovascular issues, inflammatory conditions like rheumatoid arthritis, or even early neurodegenerative changes, long before symptoms manifest or irreversible damage occurs. Such a capability could fundamentally shift the paradigm from reactive treatment—where interventions begin after significant disease progression—to preventive intervention. This proactive approach could potentially avert countless cases of severe illness, improve patient outcomes dramatically, and significantly reduce the immense burden on global healthcare systems by allowing for earlier, more effective, and less costly treatments.
It is important to note that the liver fibrosis assay described in the study remains a prototype and has not yet been introduced as a clinical test. The team’s immediate next steps involve meticulously refining and rigorously validating the classifier specifically for liver disease in larger, more diverse patient cohorts. Concurrently, they plan to embark on extensive research to explore and characterize fragmentome signatures connected to other chronic illnesses, building towards the broader vision of a multi-disease diagnostic platform.
Leadership and Collaborative Effort
The success of this complex research project is a testament to extensive collaboration and dedicated leadership. Along with Drs. Velculescu, Annapragada, Scharpf, and Phallen, the expansive research team included Zachariah Foda, Hope Orjuela, Carter Norton, Shashikant Koul, Noushin Niknafs, Sarah Short, Keerti Boyapati, Adrianna Bartolomucci, Dimitrios Mathios, Michael Noe, Chris Cherry, Jacob Carey, Alessandro Leal, Bryan Chesnick, Nic Dracopoli, Jamie Medina, Nicholas Vulpescu, Daniel Bruhm, Sarah Bacus, Vilmos Adleff, Amy Kim, Stephen Baylin, Gregory Kirk, Andrei Sorop, Razvan Iacob, Speranta Iacob, Liana Gheorghe, Simona Dima, Katherine McGlynn, Manuel Ramirez-Zea, Claus Feltoft, Julia Johansen and John Groopman.
Funding for this pivotal research came from a diverse array of sources, underscoring its broad scientific and clinical interest. Key contributors included the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation, SU2C in-Time Lung Cancer Interception Dream Team Grant, Stand Up to Cancer-Dutch Cancer Society International Translational Cancer Research Dream Team Grant, the Gray Foundation, The Honorable Tina Brozman Foundation, the Commonwealth Foundation, the Mark Foundation for Cancer Research, the Danaher Foundation and ARCS Metro Washington Chapter, the Family of Dan Y. Zhang AACR Scholar in Training Award, the Cole Foundation and National Institutes of Health grants CA121113, CA006973, CA233259, CA062924, CA271896, T32GM136577, T32GM148383 and DA036297. This broad financial backing highlights the significant investment in developing cutting-edge diagnostic tools that promise to reshape public health.
In conclusion, the Johns Hopkins Kimmel Cancer Center’s AI-driven liquid biopsy represents a monumental stride in early disease detection. By harnessing the power of genome-wide cfDNA fragmentation patterns and advanced machine learning, this technology offers an unprecedented ability to identify liver fibrosis and cirrhosis at their earliest, most treatable stages. More importantly, its demonstrated potential to detect a spectrum of other chronic conditions heralds a new era in preventive medicine, promising to revolutionize how diseases are diagnosed, managed, and ultimately, prevented, thereby improving countless lives globally.