When he was admitted to hospital in May 2025, blood tests showed elevated levels of several tumor markers, including carcinoembryonic antigen at 16.29 nanograms per milliliter against a reference ceiling of 4.50. A baseline computed tomography scan showed a 56 by 35 millimeter soft-tissue mass in the right upper lobe with limited adjacent atelectasis, together with lymph node involvement on both sides of the chest. A separate lesion in the left upper lobe near the hilum was judged to be a solitary contralateral pulmonary metastasis, which established M1a disease. The clinical stage was therefore cT3N3M1a, stage IVA under the eighth edition of the AJCC system, corresponding to extensive-stage small cell lung cancer with a low metastatic burden. Scans of the abdomen, brain, and bones found no metastases outside the chest.

A CT-guided needle biopsy of the lung mass delivered the central diagnostic surprise: combined small cell lung cancer, a tumor containing both small cell and non-small cell components, with a focal adenocarcinoma element. Combined small cell lung cancer accounts for roughly 5 to 28 percent of all small cell lung cancers and is molecularly heterogeneous. Immunohistochemistry showed positivity for TTF-1, synaptophysin, CD56, chromogranin A, and INSM1, along with loss of the Rb protein and a Ki-67 proliferation index of approximately 45 percent. Next-generation sequencing of the composite tumor specimen identified an EGFR exon 21 L858R mutation with a mutant allele frequency of 60.78 percent, a driver alteration far more typical of lung adenocarcinoma than of small cell tumors.

Because component-level sequencing was not performed, the researchers could not determine which histologic component carried the mutation. No ALK, ROS1, or RET rearrangements and no mutations in KRAS, BRAF, or other common genes were found. The treatment team chose the standard first-line regimen for extensive-stage small cell lung cancer: four cycles of etoposide and carboplatin combined with the PD-L1 inhibitor atezolizumab. This approach rests on the phase III IMpower133 trial, in which adding atezolizumab to platinum-etoposide chemotherapy improved median overall survival to 12.3 months from 10.3 months, with a hazard ratio of 0.70 and a 95 percent confidence interval of 0.54 to 0.91, and improved median progression-free survival to 5.2 months from 4.3 months. The choice was not straightforward, because the optimal sequencing of chemotherapy, immunotherapy, and EGFR tyrosine kinase inhibitors in EGFR-mutant combined small cell lung cancer remains debated, and evidence for EGFR-targeted therapy in this rare tumor type is largely limited to individual case reports. The authors emphasized that their sequencing decision reflected the absence of prospective evidence rather than proven superiority of one approach.

The four cycles achieved a partial response, and no new distant metastatic lesions appeared during first-line treatment. But the liver chemistry told a troubling second story. Before therapy began, the patient’s alkaline phosphatase was already mildly elevated at 156 units per liter against an institutional upper limit of 125, and gamma-glutamyl transferase stood at 87 units per liter against an upper limit of 60, while transaminases and total bilirubin were normal. After four cycles, alkaline phosphatase had climbed to 532 units per liter and gamma-glutamyl transferase to 317, with only mild transaminase elevation and normal bilirubin. This cholestatic pattern, graded as CTCAE version 5.0 grade 2, was accompanied by a striking serological finding: antimitochondrial antibody M2, a hallmark of primary biliary cholangitis, was positive at 62.6 units per milliliter. No pre-treatment AMA testing had been done, so nobody could say whether the antibody had been present all along.

Imaging and elastography narrowed the possibilities without resolving them. Liver ultrasound showed hepatic cysts, the largest measuring 46 by 38 millimeters, but no intrahepatic or extrahepatic biliary dilatation. Transient elastography performed in September 2025 showed a controlled attenuation parameter of 160 decibels per meter and liver stiffness of 7.0 kilopascals, consistent with neither significant steatosis nor advanced fibrosis. Crucially, magnetic resonance cholangiopancreatography, endoscopic retrograde cholangiopancreatography, and liver biopsy were never performed, so the injury could not be confirmed histologically and its biliary subtype remained undefined. The patient also developed hypothyroidism with markedly elevated thyroid autoantibodies, with thyroglobulin antibody exceeding 4000 international units per milliliter and thyroid peroxidase antibody exceeding 600, and his thyroid function swung from a mildly elevated baseline TSH through thyrotoxicosis to overt hypothyroidism with a TSH of 90.1, a course consistent with immune checkpoint inhibitor-associated thyroiditis. Immunoglobulin G and A measured after the liver injury were elevated, findings consistent with broader immune activation but detected only after the damage appeared.

The clinical team stopped atezolizumab and continued ursodeoxycholic acid, deliberately withholding high-dose corticosteroids after multidisciplinary review, because the injury was cholestatic-predominant with normal bilirubin and PBC-related serology was present. That restraint was informed by the wider literature on immune checkpoint inhibitor-associated cholangiopathy, a rare and distinct cholestatic phenotype that differs from the more common immune-mediated hepatitis. Systematic reviews of immune-related cholangitis report a median alkaline phosphatase of 1328 units per liter with only mild transaminase elevation, a median onset of 5.7 months in French pharmacovigilance data, and limited glucocorticoid responsiveness: complete steroid response was achieved in only 11.5 percent of patients in one series, or 3 of 26, and median recovery to grade 1 took 102.5 days compared with 36 days for hepatitis. In a 530-patient cohort, imaging-defined sclerosing cholangitis occurred in 0.8 percent of patients versus 1.5 percent for hepatitis. Ursodeoxycholic acid monotherapy produced improvement in 81.5 percent of 27 treated patients, with a median of 39.3 days to improvement, though 23 percent relapsed after rechallenge. After atezolizumab was withdrawn, the cholestatic enzymes fell, a trajectory consistent with a contribution from drug-related biliary injury but not proof of it.

The authors laid out two competing explanations that the available data could not separate. One is immune-related biliary injury superimposed on a previously unrecognized primary biliary cholangitis, in which the mild pre-treatment cholestasis reflected an underlying autoimmune susceptibility that immunotherapy then aggravated. The other is de novo PBC-like injury triggered by checkpoint blockade, a phenomenon reported after immune checkpoint inhibitor exposure, sometimes with granulomatous hepatitis. As Ruini and colleagues have noted in the literature, without more complete baseline information it is not possible to clearly state whether immunotherapy was a mere trigger of a latent autoimmune disease or its primary cause. AMA-M2 positivity alone does not establish PBC, and the significance of PBC-specific antibodies with normal alkaline phosphatase remains uncertain, though the combination of AMA-M2 positivity with elevated alkaline phosphatase did fulfill the biochemical and serological criteria for the disease.

Oncologic control was maintained through a carefully sequenced continuation of therapy. Maintenance treatment with furmonertinib, 80 milligrams once daily, directed at the EGFR L858R alteration, plus oral etoposide, 50 milligrams once daily for 21 days of each four-week cycle, began on August 19, 2025, 26 days after the last atezolizumab dose. Dose-intensified consolidative thoracic radiotherapy, 60 gray in 30 fractions of 2 gray delivered by intensity-modulated radiotherapy to the right lung lesion and mediastinal regions, was completed from January 12 to February 24, 2026, concurrently with maintenance therapy and without interruption. The authors caution that this dose-intensified regimen should not be interpreted as a uniform standard or as reclassification to limited-stage disease. After a contrast-enhanced brain MRI on May 26, 2026 confirmed no brain metastases, prophylactic cranial irradiation of 25 gray in 10 fractions was delivered following an individualized benefit-risk assessment. Mild radiation-related esophagitis, pneumonitis, and fatigue were documented. At the last follow-up on July 28, 2026, partial response was maintained, adherence was good, and no grade 3 or higher adverse events occurred during maintenance.

The authors are careful about what a single case can and cannot show. The evidence rests on one patient and cannot be generalized; pre-treatment AMA-M2 testing was absent, so latent PBC could be neither established nor excluded; the missing imaging and biopsy left the histologic subtype undefined; and no rechallenge with atezolizumab was attempted, so the durability of the biochemical improvement is unknown. Retrospective data suggest that autoimmune or cholestatic liver disease may be tolerated under checkpoint inhibition with monitoring, including a review of 22 patients with autoimmune liver disease, 12 of whom had PBC, but such observations do not establish safety. The practical lesson the authors draw is procedural rather than prescriptive: documenting baseline autoimmunity before immunotherapy, and pursuing a biliary-specific diagnostic route including imaging and biopsy when cholestatic injury emerges during treatment, would allow future clinicians to distinguish a latent autoimmune biliary disease from a drug-driven cholangiopathy, a distinction that affects both hepatic management and the continuation of cancer therapy.

Subject of Research: Cholestatic liver injury during atezolizumab therapy in EGFR-mutant combined small cell lung cancer with possible underlying primary biliary cholangitis

Article Title: Case Report: Cholestatic liver injury during atezolizumab therapy in EGFR-mutant combined small cell lung cancer with possible underlying primary biliary cholangitis

Article References: Zhu, Y., Xiong, J., Tan, C., Tan, B., Huang, Y., & Zhao, G. (2026). Case Report: Cholestatic liver injury during atezolizumab therapy in EGFR-mutant combined small cell lung cancer with possible underlying primary biliary cholangitis. Frontiers in Oncology, 16, Article 1979503. https://doi.org/10.3389/fonc.2026.1979503

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1979503

Keywords: atezolizumab, cholestatic liver injury, combined small cell lung cancer, EGFR L858R, immune checkpoint inhibitor, cholangiopathy, primary biliary cholangitis, AMA-M2, ursodeoxycholic acid, immune-related adverse events, furmonertinib, IMpower133