The study design and protocol were reviewed and approved by the Institutional Review Board of Seoul National University Hospital (IRB No. H-2301-043-1393) and Seoul Metropolitan Government Seoul National University Boramae Medical Center (IRB No. 10-2023-34). Patients diagnosed with carcinosarcoma, spindle cell carcinoma, pleomorphic carcinoma, pulmonary blastoma, or those exhibiting focal sarcomatoid features in tissue biopsy results were identified from two institutions between January 2005 and December 2021. Inclusion criteria is patients receiving radical aim of RT for PSC. The exclusion criteria included patients who underwent salvage surgery on the pulmonary mass after radiotherapy and those diagnosed with distant metastasis or another carcinoma at the time of diagnosis. Out of 149 identified PSC patients, 62 had received radiotherapy targeted at lung lesions. Patients who underwent radiotherapy with palliative intent (n = 25) or those who had surgery either before or after radiotherapy (n = 16) were excluded. Consequently, 21 patients who received radical radiotherapy were selected for further analysis. The median follow-up duration was 7.4 months (range, 1.5 to 148.0 months).

Chest computed tomography (CT) scans were performed to assess the extent of the tumor, while fluorodeoxyglucose-positron emission tomography (FDG-PET) scan was used to evaluate the presence of distant metastases. Tumor pathology was confirmed through tissue biopsy. Pulmonary function tests (PFT) were conducted before treatment initiation. RT doses ranged from 42 Gy to 66 Gy, administered in various fractionation schedules from 4 to 33 fractions. The gross tumor volume (GTV) was delineated for the primary lung tumor and the involved mediastinal lymph nodes. The internal target volume (ITV) was delineated using four-dimensional CT (4D-CT) for patients with tumors in the lower lobe with significant horizontal movement, as well as for those with small tumors for stereotactic ablative radiation therapy. The clinical target volume (CTV) was delineated with an expansion of 3‒7 mm from the GTV, and adjustments were made as needed. The planning target volume was delineated from the CTV/ITV with a 5‒7 mm margin. Concurrent chemoradiotherapy (CCRT) was considered for patients with nodal invasion who had an acceptable performance status. For patients who underwent CCRT, taxane/carboplatin or taxane/cisplatin were mainly applied, and one patient received etoposide/cisplatin due to focal neuroendocrine feature. Programmed death ligand-1 (PD-L1) inhibitor was applied sequentially in selected two PD-L1 positive patients.

Follow-up chest CT scans were performed every three months. Tumor response was evaluated by measuring changes in the long axis of the primary lung lesion. A complete response was defined as the disappearance of the lesion on imaging, while a partial response was defined as a decrease of at least 30% in the initial tumor size. The eventual response is defined as the response evaluated at the time of the data analysis. Grade ≥3 toxicities, including respiratory, dermatological, and hematological events, were classified according to the Common Terminology Criteria for Adverse Events (version 5.0). Events occurring within 6 months were defined as acute toxicities, while those occurring after 6 months were classified as late toxicities.

Progression-free survival (PFS) was defined as the time interval from the initiation of curative radiotherapy to the time of tumor progression or death from any cause, with censoring applied to patients at their last follow-up. Survival rates were estimated using the Kaplan-Meier method. Overall survival (OS) was defined as the time interval from the initiation of curative RT to death from any cause. Univariate and multivariate analyses were conducted to identify prognostic factors for PFS using the Cox proportional hazards regression model. Backward stepwise selection was adopted for multivariate analyses. The prescribed radiation dose was calculated as the equivalent dose in 2-Gy fractions (Gy_EQD2, calculated by the Linear-Quadratic model, α/β = 10 Gy). A swimmer plot was adopted to visualize the longitudinal duration of individual overall systemic responses, excluding two patients who were lost to follow-up (n = 19). Each bar represents one patient, with annotation of the initiation of disease progression, partial response, and complete remission along the bar. A bar without any other annotation indicates stable disease. When there was no evidence of systemic progression at the last follow-up, the patient was regarded as having an ongoing response. Durable responses, defined as responses lasting more than 6 months, are annotated separately. A waterfall plot was adopted for visualizing the size change of the primary tumor after treatment, but five patients who were unable to measure the extent of disease due to radiation fibrosis were excluded when evaluating eventual response. All statistical analyses were performed using Stata 16.0 (StataCorp LLC, College Station, TX, USA).

From January 2008 to December 2021, a total of 21 patients with pathologically confirmed sarcomatoid carcinoma who received radical RT were included in this study. The majority of these patients were male (n = 18; 85.7%), with a median age of 71 years and generally tolerable performance status. Baseline PFT revealed median values of forced expiratory volume in 1 second (FEV₁) and diffusing capacity of the lung for carbon monoxide (DLCO) to be 89% and 92.5%, respectively. More than half of the patients (n = 12; 57.1%) had a history of smoking. As per the American Joint Committee on Cancer 8th edition staging system NSCLC, 17 patients (81.0%) were classified into stage III or IV. Four patients diagnosed with stage IV were treated with a radical aim after 2020, when the concept of oligometastasis is accepted. Among them, two patients had contralateral lung nodule, one patient had suspected malignant pleural effusion without pathological confirmation and one patient had oligometastasis in the brain. Approximately half of the patients (n = 11; 52.4%) were diagnosed with pure sarcomatoid carcinoma, while the rest (n = 10; 47.6%) presented focal sarcomatoid features in their pathology reports. PD-L1 expression was positive in nine patients (42.9%), with the status unknown for 10 patients (47.6%). CCRT was administered to nine patients (42.9%), while 12 patients (57.1%) underwent radiotherapy alone. The median prescribed radiation dose was 60 Gy_EQD2, with a range from 42 Gy_EQD2 to 125 Gy_EQD2.

Among the 19 patients available for response evaluation, five demonstrated a durable response to radical radiotherapy, which response lasted longer than 6 months. In the responder group, defined as those showing complete remission or partial response, all had a primary tumor size of less than 5 cm. The majority of responders had pure PCS (n = 4; 80.0%) as per their pathology reports, and no patient in this group showed negative PD-L1 expression status (range, 40% to 90%). Detailed patient characteristics are summarized in Table 1.

The median PFS and overall survival for the entire cohort (n = 21) were 4.6 months and 7.9 months, respectively (Fig. 1A). Among the 19 patients with follow-up data, 11 had experienced a good response to radiotherapy in the form of partial response or complete remission in their primary tumors. However, disease progression was observed in six of these patients after confirming their response. Durable response, defined as a response lasting more than 6 months, was observed in five patients, including three with complete tumor remission (Fig. 2). Three patients who achieved complete remissions had PFS of 148.0 months, 19.6 months, and 66.4 months, respectively. Patient #1, a 71-year-old male with stage III disease, had a 3.5-cm tumor and underwent CCRT, receiving a radiation dose of 66 Gy in 33 fractions with docetaxel and cisplatin. Patient #2, a 72-year-old male with stage II disease, presented with a 4.8-cm tumor and underwent radiotherapy alone, receiving 48 Gy in 4 fractions. Patient #3, a 51-year-old female with stage III disease, had a 3.3-cm tumor and underwent CCRT with docetaxel and cyclophosphamide. Ten patients exhibited a response within 1 month of RT, but this response subsided in six of them. The swimmer plot and the waterfall plot of the response is depicted in Figs. 2 and 3, respectively. At 1 month after treatment, follow-up chest CT imaging was available for 19 patients and none of them showed tumor growth after radiotherapy (Fig. 3A). One patient showed no change in the primary tumor while 11 patients (57.9%) displayed a 30% or greater reduction of their tumors. However, two patients experienced progression in distant metastasis without local progression; one patient had progression in both lungs and another in the retrocaval, cardiophrenic lymph node, and adrenal gland. In five patients, post-treatment tumor size became unmeasurable due to RT-induced fibrosis. Among the remaining 14 patients, five (35.7%) eventually showed a response (Fig. 3B).

Distant metastasis (n = 10; 47.6%) was the most common site of failure, often accompanied by regional recurrence (n = 5; 23.8%) and local recurrence (n = 1; 4.8%). Simultaneous regional recurrence and local recurrence occurred in one patient (4.8%), while only regional recurrence was observed in another patient (4.8%). The detailed pattern of failure is presented in Supplementary Table S1.

Univariate analysis of prognostic factors for PFS revealed that tumor size greater than 5 cm was associated with a significantly higher HR of 3.584 (95% confidence interval [CI], 1.030–12.468; p = 0.045). Smoking history, PD-L1 expression status, distant metastasis, pure PCS, and treatment modalities did not show statistical significance. In the multivariate analysis, administering a higher radiation dose of >58 Gy_EQD2 was significantly associated with a high PFS (HR = 0.204; 95% CI, 0.046–0.915; p = 0.038), suggesting a potential benefit of high-dose radiation in treating PCS (Table 2).

Overall, the incidence of acute adverse events was 38.1%, and the incidence of late adverse events was 23.8%. Table 3 presents the adverse events, with a focus on respiratory toxicity. Among the acute adverse events, grade 1-2 respiratory toxicity was reported in eight patients (38.1%). Late respiratory toxicity of grade 1-2 was reported in five patients, and one patient (4.8%) experienced grade 3 respiratory toxicity well managed with supportive care.