The patient’s story began in January 2023, when a chest CT scan revealed a lung mass that surgery confirmed as moderately differentiated invasive adenocarcinoma, with spread to lymph nodes in three stations. Genetic testing identified an L858R mutation in exon 21 of the EGFR gene, the classic sensitive alteration that makes tumors highly responsive to EGFR inhibitors. Within weeks, a suspected brain metastasis pushed her disease to stage IV, and she began four cycles of pemetrexed combined with platinum chemotherapy, switching from cisplatin to nedaplatin after her kidneys showed signs of strain. In May 2023 she received 50 Gy of radiation across 25 fractions to the supraclavicular and mediastinal lymph node drainage regions, and on May 22 she started oral icotinib at 125 milligrams three times daily. Her blood counts before the drug began were largely unremarkable.

For more than two years, the strategy appeared to work. Then in November 2025 she was admitted with chest tightness, shortness of breath, and profound fatigue. Her hemoglobin had crashed to 41 grams per liter, a level that leaves the blood barely able to deliver oxygen. Symptoms improved with supportive care, her lung disease was assessed as stable, and icotinib continued. A month later she returned with the same complaints, her hemoglobin down to 39 grams per liter and her white cell count climbing. After a transfusion lifted her hemoglobin to 81 grams per liter, endoscopy ruled out gastrointestinal bleeding, and her physicians suspected a drug-related blood disorder and stopped the drug. By January 2026, the picture had transformed: her white cell count had exploded to 85.6 billion cells per liter while platelets collapsed, and a bone marrow examination revealed 21 percent blasts, abnormal myeloid progenitor cells flooding a marrow that was more than 95 percent cellular.

The diagnostic workup read like a molecular detective story. Flow cytometry showed leukemic cells carrying CD34, CD117, CD13, HLA-DR, and partially CD38, with a small fraction expressing CD33. Cytogenetic analysis revealed loss of one copy of chromosome 7, a karyotype of 45,XX,−7 in all 20 cells examined. Gene sequencing uncovered mutations in PPM1D at a variant allele frequency of 48.8 percent, compound heterozygous TET2 mutations, a DNMT3A mutation, and a RUNX1 insertion. Under the 2016 World Health Organization classification, which defines acute myeloid leukemia by 20 percent or more blasts in marrow or blood, the diagnosis was unambiguous. What remained ambiguous was the cause, because the patient had been exposed to cytotoxic chemotherapy, radiation, and a targeted drug in sequence.

To weigh the evidence, the team applied the Naranjo Assessment Scale, a standard pharmacovigilance tool for judging whether a drug plausibly caused an adverse event. The case scored 4 points, landing in the category of a possible association rather than a probable or definite one. The temporal sequence fit, the diagnosis was objectively confirmed, and no alternative single cause could be established with certainty, but the drug was never rechallenged, therapeutic drug monitoring was not performed, and the withdrawal-improvement criterion is fundamentally inapplicable to a clonal cancer that takes months or years to develop. The authors were careful to state that with multiple cytotoxic exposures in the patient’s history, further evidence would be needed to confirm whether the leukemia was induced solely by the EGFR inhibitor.

To place the case in context, the researchers searched PubMed, Web of Science, Embase, and two Chinese databases up to June 30, 2026, screening 849 records after removing duplicates and ultimately including eight articles describing 13 cases. Adding their own patient brought the pooled analysis to 14 cases of myeloid neoplasms arising during or after EGFR-TKI therapy. The primary tumors were predominantly L858R-mutated lung adenocarcinomas, and the implicated drugs included gefitinib in five cases, erlotinib in five, osimertinib in two, and icotinib in two. The interval from starting the inhibitor to the leukemia diagnosis ranged from 4 to 32 months, with a median of roughly 10 months, strikingly shorter than the approximately 5-year average latency associated with alkylating agents and radiation therapy.

The disease spectrum was telling. Treatment-related acute myeloid leukemia was the most common diagnosis with eight cases, and more than half of those were acute promyelocytic leukemia, a distinctive subtype driven by a specific chromosome 15 to 17 translocation. Four cases were myelodysplastic syndromes and two were chronic myeloid leukemias. Karyotypes spanned balanced translocations, chromosomal deletions, and normal results, with no single signature abnormality. Most patients presented with cytopenias, particularly anemia and severe thrombocytopenia. Crucially, the authors noted, two of the reviewed cases had received no chemotherapy and no radiation before the leukemia emerged, meaning the EGFR inhibitor was the only cytotoxic exposure, a detail that keeps the question of drug causation alive even though most cases cannot be cleanly disentangled from prior treatments.

The biology remains unresolved. Clonal hematopoiesis, the expansion of mutated blood stem cells that the WHO’s fifth edition now recognizes as a precursor state to myeloid cancers, is thought to set the stage, but whether EGFR inhibitors push these clones toward malignancy is unknown. There are hints: neutrophils express the EGFR receptor, and EGFR signaling can enhance their tumor necrosis factor-driven phagocytic activity in ways that EGFR inhibitors block in the laboratory. Exogenous growth factors can also drive bone marrow stromal cell proliferation through the same receptor. The authors are explicit that these mechanisms are speculative and that confounding by prior chemotherapy or radiotherapy cannot be excluded. What the genetics of this case do show is a pattern consistent with therapy-related disease: monosomy 7 is significantly enriched in treatment-related AML, and PPM1D mutations, found in roughly 20 percent of treatment-related myeloid neoplasms but only 0.4 percent of matched de novo cases, are tied to platinum exposure. Yet TET2 and DNMT3A mutations occur in primary disease too, and the patient’s age of 75 is not itself discriminating, since de novo AML still predominates in the elderly.

The outcome was grim. Treated with lisaftoclax, a next-generation BCL-2 inhibitor, combined with the hypomethylating agent azacitidine, the patient achieved partial remission after one cycle, though she suffered severe marrow suppression with her platelets falling to 12 billion per liter. The leukemia relapsed and progressed, and she died four months after diagnosis. That trajectory mirrors the broader statistics for treatment-related myeloid neoplasms, which carry a median overall survival of only 8 to 14 months and fare worse than their primary counterparts. Allogeneic stem cell transplantation remains the only potentially curative option, with a five-year survival of 22 percent, while the widely used combination of venetoclax with hypomethylating agents achieves complete remission in just 39.1 percent of cases with a median survival of seven months. Lisaftoclax itself has shown remission rates of 55.6 to 62.5 percent after one cycle in early studies of AML and higher-risk myelodysplastic syndrome, though its role in treatment-related disease remains to be defined.

The practical message from the Wuhan team is straightforward even if the science is unsettled: patients with EGFR-mutated lung cancer who take these targeted drugs for years should have their complete blood counts monitored regularly over the long term, and clinicians should treat persistent or progressive cytopenia as a warning sign that warrants bone marrow evaluation rather than a routine side effect to be watched passively. Hematologic toxicity is not a typical feature of EGFR inhibitors, which is precisely why an evolving anemia in a patient doing well on targeted therapy can be mistaken for something benign until it is anything but. With no conclusive evidence yet that EGFR inhibitors are independent risk factors for treatment-related myeloid neoplasms, this single case, scored only as a possible drug reaction, cannot change prescribing practice on its own. But as targeted agents extend survival for patients with EGFR-driven lung cancer, the case stands as a reminder that precision medicine’s clean safety profile deserves continued scrutiny, one carefully documented patient at a time.

Subject of Research: A reported case of treatment-related acute myeloid leukemia during icotinib therapy for EGFR-mutated non-small cell lung cancer, with a review of 14 similar cases

Article Title: Acute myeloid leukemia following icotinib therapy for non-small cell lung cancer: a case report and literature review

Article References: Huang, H., Song, L., Tong, F., & Li, L. (2026). Acute myeloid leukemia following icotinib therapy for non-small cell lung cancer: a case report and literature review. Frontiers in Oncology, 16, Article 1967553. https://doi.org/10.3389/fonc.2026.1967553

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1967553

Keywords: acute myeloid leukemia, icotinib, EGFR inhibitors, non-small cell lung cancer, treatment-related myeloid neoplasm, targeted therapy, case report, clonal hematopoiesis, PPM1D mutation, monosomy 7, azacitidine, lisaftoclax