← Issue №5/ week of Aug 2, 2026/Hepatology

Machine Learning Predicts Treatment Response and Prognostic Pathways From Whole-Blood Transcriptome in Primary Biliary Cholangitis.

From GI Signals issue №5: what this paper found, what it changes, and where it sits against the current standard of care, reviewed by Simon Mathews, MD.

Hepatology prospective cohort · n=247 · Aug 1, 2026 · Liver International · IF 6.7

Machine Learning Predicts Treatment Response and Prognostic Pathways From Whole-Blood Transcriptome in Primary Biliary Cholangitis.

Diagnosticartificial intelligencebiomarkerbasic sciencetranslational
Clinical takeawayNo clinical action yet: a preclinical biomarker study in PBC requiring validation before clinical use.
What it foundML model predicted OCA response (AUROC 0.93) and disease progression (AUROC 0.94) in PBC using a 105-gene panel from whole-blood transcriptomics.
ContextFirst demonstration of a blood-based transcriptomic signature for PBC prognosis and treatment response, linking pathways (FXR, stress response) to outcomes.
Measured against the standard of care· AGA/AASLD 2026 Clinical Practice Guideline (Gastroenterology 2026;171(5):817-839; Hepatology 2026), complementing AASLD 2018/2021 Practice Guidance

Decision at stakepredicting treatment response and disease progression in PBC

No single passage of this standard matched the paper closely enough to quote, so none is shown. The standard is cited above.

Our full summary of this standard

Adults with PBC: ursodeoxycholic acid (UDCA) is first-line for every patient, started at diagnosis (a strong recommendation); the 13-15 mg/kg/d dose comes from the 2018/2021 AASLD guidance, which the 2026 guideline adds to rather than replaces. If antimitochondrial antibodies (AMA) are negative and PBC is suspected, test for the PBC-specific antinuclear antibodies SP100 and GP210, which may avoid a liver biopsy. Measure liver stiffness (elastography) to estimate the risk of progression; the guideline says the best interval for repeat elastography is unknown. If the response to UDCA is inadequate, consider adding a PPAR agonist; if UDCA is not tolerated, consider a PPAR agonist instead. The patient summary names seladelpar and elafibranor, which have FDA accelerated approval as second-line drugs and are not recommended in decompensated cirrhosis. Off-label fibrates (bezafibrate, fenofibrate) are discouraged in decompensated cirrhosis. Obeticholic acid is no longer a US option: it was withdrawn from the market in late 2025, and the 2026 guideline makes no recommendation on it. For moderate-to-severe cholestatic pruritus, consider the IBAT inhibitor linerixibat through shared decision-making that weighs its GI side effects (diarrhea, abdominal pain), cost and patient preference; the older bile-acid sequestrant, rifampin, naltrexone and sertraline ladder from the 2021 guidance still applies. After liver transplant for PBC, give UDCA prophylaxis to reduce recurrence and graft failure. Certainty of evidence ranges from moderate to very low. The guideline found too little evidence to recommend anything for patients who do not respond to second-line therapy, and the long-term effect of PPAR agonists on mortality, transplant-free survival and decompensation is not established.

AGA/AASLD 2026 Clinical Practice Guideline (Gastroenterology 2026;171(5):817-839; Hepatology 2026), complementing AASLD 2018/2021 Practice Guidance · reviewed 2026-10-25 ↗
Syed H … Mason AL · Liver International : Official Journal of the International Association for the Study of the Liver · IF 6.7 · PubMed ↗Permalink
← Read the whole of issue №5 Every paper GI Signals surfaces gets a page like this one. All issues