Gene Solutions study finds preterm-birth signals in routine NIPT data

25 Sep 2026 (ET): Gene Solutions said a study in AJOG Global Reports found cell-free DNA fragmentomic “end motif” patterns in routine first-trimester NIPT sequencing associated with spontaneous preterm birth, analyzed with its DNAsphere.AI platform. Held-out validation AUC was 0.970; the classifier remains investigational and needs prospective confirmation.

HEALTH desk — a first-trimester blood test that pregnancies already get is built to screen chromosomes, and an early delivery is a different risk. Gene Solutions says that same draw may carry an end-motif pattern tied to spontaneous preterm birth, and the honest label is investigational: a held-out AUC of 0.970 still needs a larger prospective check before anyone treats the score as a diagnosis.

The study is published in AJOG Global Reports. AJOG is the American Journal of Obstetrics & Gynecology. The paper’s title is “First-Trimester Non-Invasive Prediction of Preterm Birth Using Cell-Free DNA Fragmentomics.” The DOI the company filing points to is 10.1016/j.xagr.2026.100605.

The analysis used fragmentomics: the length, the positioning, and the molecular ends of cell-free DNA fragments. Cell-free DNA, or cfDNA, is DNA floating in the blood, outside of cells. Machine learning looked at end motifs, the short codes at those cut ends, that differed between pregnancies that ended in spontaneous preterm birth and pregnancies that reached term. Spontaneous means labor started early on its own. Those glosses are this desk’s.

The discovery work used DNAsphere.AI, Gene Solutions’ proprietary precision data and AI platform. Proprietary means the company owns the platform. This desk did not open it.

In the study’s held-out validation cohort, the end-motif model achieved an AUC of 0.970, with 94% sensitivity at a specificity above 90%. A held-out cohort is a slice of the study the model did not train on. AUC is a 0-to-1 score for how well a model separates two groups. 0.5 is a coin flip. 1.0 is a clean split on that set. 0.970 is a strong split on this set, and it is still one study’s number. Sensitivity is the share of preterm pregnancies the model flagged. Specificity above 90% means it stayed quiet on more than 90% of the term pregnancies. Those figures are the company’s report of the study. This desk did not rerun the model.

The AI analysis used existing first-trimester NIPT data. It did not require another blood draw or additional sequencing.

The findings are preliminary and require confirmation in larger, prospectively enrolled populations. Prospective means people are enrolled first and then followed, instead of being picked after the birth is already known. The preterm-birth classifier remains investigational. It is not intended to diagnose preterm labor, replace obstetric assessment, or direct treatment without further clinical validation. That limit is the company’s. It is the limit on this filing.

A prospective observational study registered in Vietnam is evaluating cfDNA and clinical information for prediction of spontaneous preterm birth, preeclampsia, and gestational diabetes. Estimated enrollment is 1,105. Estimated means a planned size, not a finished count. Preeclampsia is a pregnancy condition with high blood pressure. Gestational diabetes is high blood sugar that starts in pregnancy. Those glosses are this desk’s. Additional prospective collaborations are being established with clinical centers in India. Being established is not a completed trial.

Gene Solutions’ triSure NIPT platform has been expanding since 2018. The company describes itself as providing precision cell-free DNA solutions through multi-omics and artificial intelligence for pregnancy care, cancer detection, and precision medicine. Multi-omics means more than one kind of biological measurement, not only the DNA letters. That gloss is this desk’s.

What the card shows. The card is Gene Solutions’ study-design diagram from the company news post, placed on a maroon canvas. It walks a first-trimester NIPT blood sample through cell-free DNA, sequencing, fragmentomics, and a preterm-birth risk score, with DNAsphere.AI on the path. Those steps are the diagram’s. The AUC, the 94%, and the investigational limit are not a clearance stamp on the card. The card has no desk date laid over it. The catalog chip is HEALTH.

Plain English for the rest of the card. A common early-pregnancy blood test already sequences cfDNA. Gene Solutions says the cut-ends of those fragments differed, in this study, between spontaneous preterm birth and term birth. On a held-out slice the end-motif model scored an AUC of 0.970, and it flagged 94% of the preterm pregnancies while keeping specificity above 90%. The lab used data it already had. The company says the classifier is still investigational. It does not diagnose preterm labor, replace an obstetric exam, or tell a clinician how to treat. A Vietnam study with an estimated 1,105 participants, and centers in India that are still being set up, are the next checks. The card is the company diagram.

PRIMARY here: Gene Solutions’ 25 Sep 2026 announcement, the company scientific recap plus the PR Newswire APAC syndication the same day — Tier A PRIMARY. STATUS PRIMARY. The desk label on the chip is CONFIRMED because that is the catalog word for a primary page we can stand on. The NIPT use of routine first-trimester sequencing, the AJOG Global Reports title, fragmentomics of length, position, and molecular ends, end motifs, DNAsphere.AI, AUC 0.970, 94% sensitivity at specificity above 90%, no second blood draw, the preliminary and investigational limit, the Vietnam estimate of 1,105, the India collaborations, and triSure’s expansion since 2018 are that announcement’s. The paper DOI is 10.1016/j.xagr.2026.100605. NOT claimed: a cleared diagnostic, a diagnosis of preterm labor, a replacement for obstetric assessment, a treatment order, a finished enrollment of 1,105, a completed India trial, a model this desk reran, a login, a stock tip, or investment advice. The card is the company study-design diagram. Distinct from the already-filed veeva-study-builder-agent, bighat-biosciences-series-c, verana-health-agent-claire, and basalt-health-series-a.