By Diagnostics World Staff
November 18, 2025 | The Association for Molecular Pathology hosted its 2025 Annual Meeting & Expo last week. Highlights from some of the research presented at the event include faster molecular tests for dangerous regional fungal infections, new Lyme disease tests, AI tools for improving molecular diagnosis, new use for old DNA samples to understand the evolution of colorectal cancer, and more.
Ancient DNA Technique Tracks Disease Changes Over Time
Researchers at the University of Chicago adapted techniques originally used to study ancient DNA from archaeological specimens to recover genetic information from old medical samples. They were able to analyze DNA samples stored nearly 100 years ago and prove the adaptability of archaeological genomics techniques. To test their technique, the researchers used samples of tissue from colorectal cancer specimens collected between 1932 and 2023. That work is ongoing, but Alexander Guzzetta, M.D., Ph.D., who oversaw the project alongside ancient DNA researcher Maanasa Raghavan, Ph.D., noted, “There is massive potential here for other groups to use these approaches to unravel the root causes underlying the shifting landscape of modern diseases. This approach could unlock the ability to study how diseases evolve over decades and shed light on how their biology has changed through time.”
Faster Molecular Test For Dangerous Regional Fungal Infections
Researchers at Indiana University Health have created a test that rapidly detects three serious U.S. fungal pathogens often misdiagnosed as other respiratory illnesses: Histoplasma, Blastomyces, and Coccidioides. Current diagnostics of these pathogens require cultures that take weeks, but the researchers report that multiplex real-time PCR test achieved 100% sensitivity and specificity directly from samples, eliminating a DNA extraction step to reduce lab time and biosafety risk.
Detecting Lyme Disease Earlier
Researchers at Dartmouth Hitchcock Medical Center have developed a droplet digital PCR (ddPCR) test that may detect Lyme disease more accurately and earlier than traditional methods — potentially avoiding years-long diagnostic delays that can lead to severe complications. The ddPCR assay detects Borrelia burgdorferi DNA directly with a detection threshold as low as 5–10 bacterial cells. The team was able to successfully diagnose a patient with a four-year diagnostic delay who had negative antibody results but responded to antibiotics.
Multiple Studies In Acute Myeloid Leukemia
Several teams studying acute myeloid leukemia shared promising results.
A UC San Diego study that tracked mutations before and after stem cell transplant in 74 patients. Persistent mutations in two genes strongly predicted relapse. The work could improve post-transplant monitoring and follow-up treatment decisions.
University of Michigan scientists evaluated 600 AML cases by integrating RNA fusion analysis into standard next-generation sequencing, finding gene fusions in 15% of patients and identifying 23 cryptic fusions missed by cytogenetics, including some that were clinically actionable.
Moffitt Cancer Center researchers validated a deep-sequencing test for FLT3 mutations, which are common in AML and linked to higher relapse rate. It could be a powerful tool for detecting relapse earlier and guiding transplant decisions.
AI Tools Show Major Promise In Improving Molecular Diagnosis
A similarly rich area of work, AI tools to improve diagnostics were showcased by many groups.
Researchers at The Hospital for Sick Children created a web-based AI platform to classify cancers using RNA-sequencing data—even when samples vary in preparation and storage. The model achieved 93% diagnostic accuracy on subtypes covered by the platform. The system can also adapt and incorporate new subtypes, increasing its accuracy with each new sample. The goal for the platform is to cover new subtypes with only five reference samples.
A team at Soonchunhyang University (South Korea) trained two AI models to diagnose tumors using cerebrospinal fluid circulating tumor DNA and MRI imaging — offering a potential alternative to invasive biopsies. The approach may allow earlier diagnosis and better surgical planning without requiring repeated brain or spinal biopsies.
At Wake Forest University School of Medicine, researchers used an AI-based karyotyping tool to analyze chromosomal abnormalities in GATA2 deficiency syndrome, a rare condition linked to blood cancers. This approach may provide clearer insight into disease trajectory and relapse risk in myeloid malignancies.
Researchers at Augusta University developed a computational pipeline comparing AI models that predict genetic and transcriptomic tumor features from routine slide images, aiming to reduce reliance on expensive sequencing tests. Long-term, this approach may help deliver precision oncology insights directly from pathology slides, accelerating and broadening access to personalized cancer care.
New Genetic Test Finds Hidden Cases Of Rare Movement Disorder
A Harvard/Brigham and Women’s Hospital team developed a targeted DNA test for X-linked dystonia-parkinsonism, a rare inherited movement disorder primarily affecting men of Filipino descent. Symptoms of the disorder mimic Parkinson's disease, leading to frequent misdiagnosis. The new DNA test correctly identified all known carriers and three previously misdiagnosed cases. It also correctly alerted two patients to false-negative results from standard genetic tests.
Technique Helps Doctors Identify Possible Causes Of Pregnancy Loss
Two groups of researchers — one at Dartmouth–Hitchcock Medical Center and one at Queens University’s Kingston Health Sciences Centre and the University of Ottawa — found that optical genome mapping could help uncover causes of recurrent pregnancy loss. The teams found structural chromosome changes missed by traditional testing methods and identified a possible cause of recurrent pregnancy loss that was previously overlooked