Histologic transformation is one of the more unsettling resistance mechanisms in EGFR-mutant non-small cell lung cancer. EGFR tyrosine kinase inhibitors, including osimertinib, work well initially, but acquired resistance typically emerges after roughly 10 to 14 months of treatment. Known escape routes include the EGFR T790M mutation, amplification of MET or HER2, epithelial-to-mesenchymal transition, and, less commonly, the tumor abandoning its original lineage altogether. When researchers pooled data from patients who underwent repeat biopsy after EGFR inhibitor resistance and had a confirmed histologic diagnosis, transformation occurred in approximately 2.8 percent of cases, or 74 of 2,624 patients. The most frequent form was conversion to small-cell lung cancer, which accounted for 65 percent of transformation cases, while adenocarcinoma-to-squamous cell carcinoma transdifferentiation, abbreviated AST, made up 15 percent, or 11 of the 74 cases.

Rarity does not mean clinical irrelevance. Among patients with EGFR-mutant lung adenocarcinoma resistant to osimertinib, the reported incidence of AST rises to between 5.5 and 9 percent. One retrospective study found numerically shorter overall survival in patients with histologic transformation compared with those without it, although the difference did not reach statistical significance: 30.8 versus 41.2 months from the start of tyrosine kinase inhibitor therapy, and 6.6 versus 12.9 months measured from the moment of transformation. Because AST signals a profound biological rewiring of the tumor and complicates every subsequent treatment decision, understanding its molecular underpinnings has become an active frontier in thoracic oncology.

The clinical course in this case began conventionally enough. Surgery in February 2019 removed an acinar-predominant invasive adenocarcinoma of the right lower lobe, staged pT1cN0M0, with immunohistochemistry positive for the adenocarcinoma markers TTF-1 and Napsin A. A 50-gene sequencing panel identified the EGFR exon 19 deletion alongside a TP53 p.R175H mutation. Pleural metastases appeared in October 2021, prompting first-line osimertinib at 80 milligrams daily. Progression followed in October 2024, and the team switched to second-line aumolertinib. By January 2025, PET/CT showed diffuse right pleural thickening and hypermetabolic hilar lymph nodes, while plasma-based sequencing on a 680-gene panel detected no clinically actionable variants. Third-line treatment with pemetrexed, nedaplatin, bevacizumab, and sintilimab began in February 2025.

The pivotal moment arrived in May 2026, when imaging revealed metastases in the pleura, peritoneum, lymph nodes, and bone. A percutaneous lung biopsy delivered the surprise: squamous cell carcinoma. Immunohistochemistry told a coherent story of lineage switching, with strong positivity for the squamous markers CK5/6 and p40 and complete loss of TTF-1 expression. Diffuse strong nuclear p53 staining matched the known TP53 missense mutation. Tissue sequencing, with matched peripheral blood as a germline control, retained the original EGFR 19del and TP53 p.R175H alterations but added two new findings: SMAD4 c.1052A>T, p.D351V, at a variant allele frequency of 15.79 percent, and ATM c.3415G>A, p.E1139K, at 4.97 percent. Both were classified as class III variants of uncertain significance under American College of Medical Genetics and Genomics criteria. Notably, both genes had been included in the original 50-gene panel, and neither variant was detected at initial diagnosis.

Three explanations exist for finding squamous carcinoma on a repeat biopsy in this setting: outgrowth of a pre-existing squamous subclone, a second primary tumor, or true lineage transdifferentiation. The authors could not definitively distinguish among them because the original surgical slides, held at another hospital, were unavailable for re-review, and a minor unsampled squamous component can never be fully excluded even in a large resection specimen. Still, the shared retention of EGFR 19del and TP53 p.R175H points to a common clonal origin, and the clinical and pathological arc from pure adenocarcinoma to pure squamous carcinoma is consistent with AST, even if selective expansion of an undetected subclone remains a formal alternative.

The mechanistic discussion centers on the newly detected variants. Established AST-associated mechanisms include loss of LKB1/STK11, which links metabolic and redox changes to suppression of adenocarcinoma lineage factors and activation of squamous programs, as well as aberrant PI3K/AKT/mTOR and MYC signaling, and dysregulation of transcription factors such as loss of NKX2-1 with SOX2 activation. None of these alterations appeared in the tested samples from this patient. Instead, attention turned to SMAD4. Decades of work, largely through conditional deletion in animal models, show that SMAD4 loss helps initiate and progress squamous cell carcinoma by letting epithelial cells escape TGF-β-mediated growth inhibition, expanding cancer stem-cell populations, and increasing genomic instability. SMAD4-deficient epithelia also release TGF-β into the tumor microenvironment, fostering inflammation and immune evasion. In Smad4fl/fl;KrasG12D mice, ERK1/2 inhibition produced widespread lung squamous carcinoma, and isolated SMAD4 knockout can drive spontaneous squamous tumors in mouse skin, oral cavity, and stomach.

The specific alteration here, however, is not deletion but a missense change. A recent functional study showed that SMAD4 p.D351V can impair SMAD-dependent BMP/TGF-β signal transduction and disrupt normal interactions between SMAD4 and SMAD1/SMAD3, providing a theoretical basis for a biological association with AST in this patient. Whether the variant also affects SMAD4’s noncanonical role in three-dimensional genome regulation, in which SMAD4 sequesters the chromatin regulator EP300 away from SOX2 enhancers and thereby suppresses SOX2 transcription, remains unknown. The ATM p.E1139K variant is even less certain. ATM governs the DNA damage response and cellular redox homeostasis, and experimental loss of ATM function increases reactive oxygen species accumulation and mitochondrial dysfunction. Separately, studies of AST suggest that excessive ROS can suppress Wnt signaling through FOXO3A and disturb the balance between adenocarcinoma and squamous lineage programs. But these threads do not connect into a demonstrated causal chain, and ATM p.E1139K has not been shown to impair ATM activity at all.

Treatment after transformation is equally unsettled. For patients whose tumors acquire EGFR T790M, osimertinib remains an option supported by the AURA3 trial, which showed progression-free survival benefit over chemotherapy, though evidence specific to the AST setting is limited to case reports. For patients without T790M, expert consensus generally favors chemotherapy regimens designed for squamous cell carcinoma, with immunotherapy ideally evaluated in clinical trials. Immunotherapy results in transformed tumors have been mixed: in one study of 21 patients with EGFR-mutant lung cancer who developed histologic transformation, median progression-free survival on immunotherapy was just 1.6 months, though that analysis did not stratify by transformation subtype or PD-L1 expression. A contrasting case report described a patient whose PD-L1 expression jumped from 1 percent to 90 percent after squamous transformation and who achieved a durable 47-month response to pembrolizumab monotherapy. In the present case, PD-L1 tumor proportion score was below 1 percent, yet the combination of nab-paclitaxel, carboplatin, and tislelizumab produced a 40.3 percent reduction in the pleural lesion by October 2026, meeting RECIST version 1.1 criteria for partial response, with no clinically significant adverse or immune-related events. The authors caution that the contribution of each treatment component cannot be disentangled in a single patient.

The case also carries a practical diagnostic lesson. When plasma sequencing failed to yield actionable information, the authors point to evidence that cell-free DNA from pleural effusion supernatant is a reliable substrate for targeted sequencing in non-small cell lung cancer and can complement or outperform blood-based testing, and that closed pleural brushing with cell-block preparation can improve diagnostic yield in malignant effusion. For patients progressing with pleural disease, effusion may therefore serve as a valuable alternative source for molecular profiling. The authors are careful about the limits of their findings: the two variants are candidate molecular events, not established drivers, and their role in lineage plasticity requires functional validation and larger longitudinal cohorts. What the case delivers is a template, a tumor’s genome sampled before and after a radical change of identity, showing how longitudinal sequencing can surface candidate mechanisms in one of lung cancer’s rarest and most consequential escape maneuvers.

Subject of Research: Histologic transformation of EGFR-mutant lung adenocarcinoma to squamous cell carcinoma with newly detected SMAD4 and ATM variants

Article Title: Case report and literature review: histologic transformation from EGFR-mutant lung adenocarcinoma to squamous cell carcinoma with newly detected SMAD4 and ATM variants

Article References: Wang, C., Wang, S., Wu, D., Lou, Y., Liu, H., Tang, Y., Su, F., & Jia, L. (2026). Case report and literature review: histologic transformation from EGFR-mutant lung adenocarcinoma to squamous cell carcinoma with newly detected SMAD4 and ATM variants. Frontiers in Oncology, 16, Article 1982968. https://doi.org/10.3389/fonc.2026.1982968

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1982968

Keywords: lung adenocarcinoma, squamous cell carcinoma, histologic transformation, EGFR-TKI resistance, SMAD4, ATM, lineage plasticity, next-generation sequencing, immunotherapy, osimertinib, case report, thoracic oncology