Identifying the origin of brain metastases (BM) is essential for personalized treatment, particularly in patients with carcinoma of unknown primary or multiple primary neoplasms. We aimed to develop a multiparametric MRI-based three-dimensional (3D) deep learning (DL) fusion model for noninvasive differentiation of lung cancer (LC) versus non-lung cancer (NLC) origin BM. A total of 458 patients with 585 BM lesions from three centers were retrospectively included. A 3D DL model was developed using paired CE-T1 and T2-FLAIR sequences. The model employs a shared-weight Siamese encoder to extract cross-modal spatial features, combined with spatial attention mechanisms that highlight diagnostically relevant tumor regions. To capture intratumoral heterogeneity, Local Moran's I spatial autocorrelation analysis was used to partition each lesion into four biologically interpretable habitat subregions, generating subregion-level feature tokens. Cross-modal attention fusion was then applied to integrate complementary information from both sequences at both whole-lesion and subregion levels. Model robustness was evaluated using sensitivity analyses across histologic subtypes and at the patient level. Ablation studies were conducted to compare multimodal and single-modality performance and a multi-reader study to assess the model's clinical utility in assisting radiologists. The fusion model achieved an area under the receiver operating characteristic curve (AUC) of 0.925, significantly outperforming single-modality models using CE-T1 (AUC = 0.817) and T2-FLAIR (AUC = 0.802). Sensitivity analyses demonstrated robust performance across all subtypes (AUCs, 0.895-0.956). The patient-level analysis yielded an AUC of 0.942. Radiologists' performance improved with model assistance (accuracy trainee, 0.500-0.583; experienced, 0.595-0.679; expert, 0.690-0.845). The proposed 3D DL fusion model demonstrated strong performance in differentiating LC from NLC BM and improved radiologists' accuracy, supporting its potential as a noninvasive decision-support tool for BM origin assessment.

We report a case of ALK-rearranged pulmonary large-cell neuroendocrine carcinoma (LCNEC) with brain metastases and radiologically suspected leptomeningeal involvement, successfully treated with first-line lorlatinib. A 58-year-old woman was diagnosed with LCNEC, and molecular testing revealed an EML4-ALK fusion. Lorlatinib achieved a partial response, with regression of the primary lesion, brain metastases, and suspected leptomeningeal lesions. Our literature review showed that ALK-positive LCNEC occurs more often in younger, female, and never- or light-smoking patients. In such cases, molecular testing should be actively pursued. In LCNEC harboring actionable driver alterations, molecularly guided treatment, as used for NSCLC, may be a reasonable option.

Pineoblastoma (PB) is a rare malignant embryonal tumour arising from the pineal gland. It commonly affects the paediatric age group with a slight male predilection. The typical clinical features include symptoms and signs of raised intracranial pressure, but extracranial metastasis may be seen in poorly differentiated tumours. We present a 5-year-old male who presented with clinical features suggestive of raised intracranial pressure. Radiological investigations revealed a pineal region tumour, and histopathological evaluation confirmed the diagnosis of a PB. The patient was optimised, and he had a right occipital craniotomy and microscopic surgery decompression via interhemispheric transplenial with ventriculoperitoneal shunt inserted. Fever, as a post-operative complication, was noted which necessitated removal of the shunt. The patient did well and was followed up bi-weekly. Unfortunately, the patient presented 5 months later with a recurrence, and he succumbed to his illness.Highlights (Key Messages)• Pineoblastoma is a high-grade aggressive embryonal tumour arising from the pineal gland• It typically affects the paediatric population with a slight male preponderance• Prognosis is poor with high rates of mortality• Imaging studies are necessary but prompt histopathological evaluation is required for a definitive diagnosis of this tumour.

This study proposes a 2π solid angle-based heuristic framework for non-coplanar volumetric modulated arc therapy (VMAT) in patients with multiple brain metastases (MBM) to automate treatment angle selection. Approach: Fifty-two patients with MBM (2-8 targets) were retrospectively included. A hybrid heuristic framework was constructed: first, a volume-weighted cumulative tumor-skin distance (VW-CTSD) metric was introduced to quantify geometric requirements in the 2π solid angle space and to automatically identify non-coplanar trajectories minimizing normal brain exposure. Subsequently, a blocking factor (BF)-based sub-arc segmentation and collimator-angle selection strategy determined the sub-arc splitting and collimator angles. Using 30 Gy/5 fractions, three plans were generated for each patient: manual expert planning (Pref), VW-CTSD-guided trajectory selection alone (Ptsd), and the proposed combined heuristic strategy (Pcom). Dosimetric metrics and plan parameters were compared. Main results: All plans met clinical goals. Ptsd achieved automated trajectory selection with plan quality comparable to Pref. Values are presented as median (P25-P75). Compared with Pref, Pcom improved CI [0.86 (0.85-0.89) to 0.89 (0.85-0.91)] and reduced GI [3.37 (3.17-3.79) to 3.09 (2.74-3.48)]. Pcom also reduced normal brain tissue (NBT) V12Gy, MU, and complexity metric (CM) from 92.45 (62.54-141.46) to 83.57 (50.30-125.10) cc, from 2391 (2144-2527) to 1938 (1845-2085), and from 0.22 (0.21-0.24) to 0.20 (0.19-0.21), respectively. Control-point aperture analysis showed X-axis narrowing [129.75 (110.60-146.35) to 101.91 (83.34-120.20) mm] and Y-axis expansion [98.70 (80.13-105.86) to 119.17 (103.38-133.48) mm] in Pcom. All Pcom-versus-Pref comparisons for these metrics were statistically significant (P < 0.001). Significance: The proposed combined heuristic framework reduced NBT dose, GI, MU, and CM, suggesting its potential as a practical automated approach for non-coplanar VMAT angle selection in patients with MBM.

Primary adrenal high-grade undifferentiated pleomorphic sarcoma (UPS) is an exceptionally rare malignant mesenchymal neoplasm, and presentation with brain metastasis as the initial manifestation is even more uncommon. This report highlights the diagnostic evaluation and multidisciplinary management of a patient presenting with an undifferentiated brain metastasis accompanied by an adrenal mass. A 75-year-old woman presented with speech disturbance, cognitive slowing, memory decline, gait instability, and right-sided weakness. Brain computed tomography and contrast-enhanced magnetic resonance imaging revealed a left frontal lesion with multiple intracerebral nodules. Positron emission tomography/computed tomography further demonstrated a left adrenal lesion and additional metastatic foci. Owing to symptomatic mass effect, the brain lesion was surgically resected. Initial pathology suggested a poorly differentiated malignant tumor, and the tumor origin remained uncertain after the first immunohistochemical assessment. Subsequent multidisciplinary review, serial adrenal imaging, adrenal core needle biopsy, and comparative pathologic evaluation progressively redirected the diagnosis toward a high-grade undifferentiated sarcoma, morphologically consistent with UPS. The patient subsequently received postoperative brain radiotherapy, anthracycline-based chemotherapy, additional systemic therapy, and local treatment for progressive metastatic disease. Despite multimodal management, the tumor rapidly progressed with thigh metastasis, bilateral adrenal involvement, skeletal dissemination, and bilateral pathologic femoral fractures. The patient ultimately died of intracerebral hemorrhage. This case expands the clinical spectrum of primary adrenal UPS and underscores the importance of dynamic diagnostic reassessment, multisite tissue sampling, extended immunohistochemical evaluation, and multidisciplinary collaboration when evaluating undifferentiated brain metastases associated with an adrenal mass.

ROS1-targeted therapies have improved outcomes in ROS1-positive non-small cell lung cancer (NSCLC), but acquired resistance, CNS progression, and off-target toxicities remain important limitations. Zidesamtinib (Jideytro[®], NVL-520) is a next-generation, brain-penetrant, tropomyosin receptor kinase (TRK)-sparing ROS1 inhibitor designed to retain activity against resistance mutations, including G2032R. In preclinical models, it demonstrated potent, selective ROS1 inhibition, broad resistance coverage, and durable intracranial activity. Administered orally at 100 mg once daily, zidesamtinib exhibits a favorable pharmacokinetic profile. Data from the ARROS-1 phase 1/2 trial demonstrated durable clinical efficacy and manageable safety in previously treated ROS1-positive NSCLC, including patients with CNS metastases and G2032R resistance. These results led to its accelerated FDA approval on July 22, 2026, for patients previously treated with a ROS1-targeted tyrosine kinase inhibitor.

Brain metastasis (BM) is a common complication of advanced non-small cell lung cancer (NSCLC). Although immune checkpoint inhibitors (ICIs) have demonstrated efficacy in treating NSCLC-BM, their clinical benefits are often limited. The combination of radiotherapy (RT) and ICIs has emerged as a promising strategy for overcoming these challenges. This review summarizes the current knowledge regarding the immune landscape of NSCLC-BM as well as the clinical evidence supporting and challenging the RT-ICI combination, and it provides an expanded mechanistic framework detailing how RT reshapes the tumor immune microenvironment. RT can enhance the antitumor immune response by disrupting the blood-brain barrier, reprogramming the tumor microenvironment, and improving immune cell infiltration. Clinical studies have supported the synergistic potential of RT and ICIs against NSCLC-BM, showing improved survival and progression-free survival rates. However, challenges such as resistance and toxicity remain, and further investigation is required to refine treatment strategies. We further discuss optimal RT fractionation and sequencing strategies, toxicity mitigation, including practical approaches for managing radiation necrosis, and emerging biomarkers that may guide personalized treatment. A clearer understanding of these biological and clinical interactions is essential to refine treatment algorithms and maximize the therapeutic potential of radioimmunotherapy in NSCLC-BM.

Therapy-related chronic myeloid leukaemia (t-CML) is an extremely rare myeloid neoplasm. Here, we report a first case of CML developing during long-term treatment with alectinib, a highly selective tyrosine kinase inhibitor (TKI) targeting anaplastic lymphoma kinase (ALK). A 53-year-old never-smoking woman with ALK-positive non-small cell lung cancer (NSCLC) developed mediastinal lymph node and brain metastases after curative surgery. Alectinib therapy achieved sustained systemic and intracranial disease control for nearly 3 years without serious adverse events. However, 3 years and 1 month after initiating alectinib, progressive leukocytosis with marked basophilia and circulating immature myeloid cells was observed. Bone marrow analysis confirmed the diagnosis of CML with BCR-ABL1 fusion. Nilotinib was initiated while alectinib was continued, resulting in a rapid hematologic response without serious adverse events. Further accumulation of cases is needed to clarify the potential leukemogenic risk of ALK-TKIs and to establish the safety and clinical utility of combined ALK-TKI and BCR-ABL-TKI therapy.