BACKGROUND: Several studies have indicated that intrapleural infusion of bevacizumab is an effective treatment for non-small cell lung cancer (NSCLC) with malignant pleural effusion (MPE). However, the impact of bevacizumab administered through an indwelling pleural catheter (IPC) on the prognosis of these patients is unknown.
METHODS: Consecutive advanced NSCLC patients with symptomatic MPE receiving an IPC alone or bevacizumab through an IPC were identified in a tertiary hospital. The patient characteristics and clinical outcomes were collected.
RESULTS: A total of 149 patients were included, and the median age was 60.3 years. Males and nonsmokers accounted for 48.3% and 65.8%, respectively. A total of 69.8% (104/149) of patients harbored actionable mutations, including 92 EGFR-activating mutations, 11 ALK fusions, and 1 ROS1 fusion. A total of 81.9% (122/149) of patients received IPC alone, and 18.1% (27/149) received bevacizumab through an IPC. The incidence of spontaneous pleurodesis during the first 6 months was greater in the bevacizumab-treated group than in the IPC-treated group in the subgroup with actionable mutations (64.3% vs. 46.9%, P = 0.28). The median overall survival (OS) in patients with actionable mutations treated with bevacizumab through an IPC was 42.2 months, which was significantly longer than the 26.7 months in patients who received an IPC alone (P = 0.045). However, the median OS did not differ between the two arms in the subgroup without actionable mutations (10.8 vs. 41.0 months, P = 0.24). No significant difference between the bevacizumab through an IPC group and the IPC group was detected in the number of participants who had adverse events, either in patients with actionable mutations (14.3% vs. 8.4%; P = 0.42) or in patients without actionable mutations (16.7% vs. 12.8%; P = 1.00).
CONCLUSIONS: Bevacizumab through an IPC resulted in a significantly improved prognosis for NSCLC patients with MPE and actionable mutations. However, patients without actionable mutations do not benefit from bevacizumab through IPC.

BACKGROUND: Accurate prediction of tumor molecular alterations is vital for optimizing cancer treatment. Traditional tissue-based approaches encounter limitations due to invasiveness, heterogeneity, and molecular dynamic changes. We aim to develop and validate a deep learning radiomics framework to obtain imaging features that reflect various molecular changes, aiding first-line treatment decisions for cancer patients.
METHODS: We conducted a retrospective study involving 508 NSCLC patients from three institutions, incorporating CT images and clinicopathologic data. Two radiomic scores and a deep network feature were constructed on three data sources in the 3D tumor region. Using these features, we developed and validated the \'Deep-RadScore,\' a deep learning radiomics model to predict prognostic factors, gene mutations, and immune molecule expression levels.
RESULTS: The Deep-RadScore exhibits strong discrimination for tumor molecular features. In the independent test cohort, it achieved impressive AUCs: 0.889 for lymphovascular invasion, 0.903 for pleural invasion, 0.894 for T staging; 0.884 for EGFR and ALK, 0.896 for KRAS and PIK3CA, 0.889 for TP53, 0.895 for ROS1; and 0.893 for PD-1/PD-L1. Fusing features yielded optimal predictive power, surpassing any single imaging feature. Correlation and interpretability analyses confirmed the effectiveness of customized deep network features in capturing additional imaging phenotypes beyond known radiomic features.
CONCLUSIONS: This proof-of-concept framework demonstrates that new biomarkers across imaging features and molecular phenotypes can be provided by fusing radiomic features and deep network features from multiple data sources. This holds the potential to offer valuable insights for radiological phenotyping in characterizing diverse tumor molecular alterations, thereby advancing the pursuit of non-invasive personalized treatment for NSCLC patients.

BACKGROUND: Selecting therapeutic strategies for cancer patients is typically based on key target-molecule biomarkers that play an important role in cancer onset, progression, and prognosis. Thus, there is a pressing need for novel biomarkers that can be utilized longitudinally to guide treatment selection.
METHODS: Using data from 508 non-small cell lung cancer (NSCLC) patients across three institutions, we developed and validated a comprehensive predictive biomarker that distinguishes six genotypes and infiltrative immune phenotypes. These features were analyzed to establish the association between radiological phenotypes and tumor genotypes/immune phenotypes and to create a radiological interpretation of molecular features. In addition, we assessed the sensitivity of the models by evaluating their performance at five different voxel intervals, resulting in improved generalizability of the proposed approach.
RESULTS: The radiomics model we developed, which integrates clinical factors and multi-regional features, outperformed the conventional model that only uses clinical and intratumoral features. Our combined model showed significant performance for EGFR, KRAS, ALK, TP53, PIK3CA, and ROS1 mutation status with AUCs of 0.866, 0.874, 0.902, 0.850, 0.860, and 0.900, respectively. Additionally, the predictive performance for PD-1/PD-L1 was 0.852. Although the performance of all models decreased to different degrees at five different voxel space resolutions, the performance advantage of the combined model did not change.
CONCLUSIONS: We validated multiscale radiomic signatures across tumor genotypes and immunophenotypes in a multi-institutional cohort. This imaging-based biomarker offers a non-invasive approach to select patients with NSCLC who are sensitive to targeted therapies or immunotherapy, which is promising for developing personalized treatment strategies during therapy.

OBJECTIVE: Personalized treatments such as targeted therapy and immunotherapy have revolutionized the predominantly therapeutic paradigm for non-small cell lung cancer (NSCLC). However, these treatment decisions require the determination of targetable genomic and molecular alterations through invasive genetic or immunohistochemistry (IHC) tests. Numerous previous studies have demonstrated that artificial intelligence can accurately predict the single-gene status of tumors based on radiologic imaging, but few studies have achieved the simultaneous evaluation of multiple genes to reflect more realistic clinical scenarios.
METHODS: We proposed a multi-label multi-task deep learning (MMDL) system for non-invasively predicting actionable NSCLC mutations and PD-L1 expression utilizing routinely acquired computed tomography (CT) images. This radiogenomic system integrated transformer-based deep learning features and radiomic features of CT volumes from 1096 NSCLC patients based on next-generation sequencing (NGS) and IHC tests.
RESULTS: For each task cohort, we randomly split the corresponding dataset into training (80%), validation (10%), and testing (10%) subsets. The area under the receiver operating characteristic curves (AUCs) of the MMDL system achieved 0.862 (95% confidence interval (CI), 0.758-0.969) for discrimination of a panel of 8 mutated genes, including EGFR, ALK, ERBB2, BRAF, MET, ROS1, RET and KRAS, 0.856 (95% CI, 0.663-0.948) for identification of a 10-molecular status panel (previous 8 genes plus TP53 and PD-L1); and 0.868 (95% CI, 0.641-0.972) for classifying EGFR / PD-L1 subtype, respectively.
CONCLUSIONS: To the best of our knowledge, this study is the first deep learning system to simultaneously analyze 10 molecular expressions, which might be utilized as an assistive tool in conjunction with or in lieu of ancillary testing to support precision treatment options.

UNASSIGNED: Aberrant mesenchymal-epithelial transition/hepatocyte growth factor (MET/HGF) regulation presented in a wide variety of human cancers. MET exon 14 skipping, copy number gain (CNG), and kinase domain mutations/arrangements were associated with increased MET activity, and considered to be oncogenic drivers of non-small cell lung cancers (NSCLCs).
UNASSIGNED: We retrospectively analyzed 564 patients with MET alterations. MET alterations were classified into structural mutations or small mutations. MET CNG, exon 14 skipping, gain of function (GOF) mutations, and kinase domain rearrangement were defined as actionable mutations.
UNASSIGNED: Six hundred thirty-two MET mutations were identified including 199 CNG, 117 exon 14 skipping, 12 GOF mutations, and 2 actionable fusions. Higher percentage of MET structural alterations (CNG + fusion) were detected in advanced NSCLC patients. Moreover, MET CNG was enriched while exon 14 skipping was rare in epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKI)-treated advanced NSCLC patients. Ten of the 12 MET GOF mutations were also in EGFR-TKI-treated patients. Fifteen (68.1%) of the 22 patients treated with crizotinib or savolitinib had a partial response. Interestingly, one patient had a great response to savolitinib with a novel MET exon 14 skipping mutation identified after failure of immune-checkpoint inhibitor.
UNASSIGNED: Half of the MET alterations were actionable mutations. MET CNG, exon 14 skipping and GOF mutations had different distribution in different clinical scenario but all defined a molecular subgroup of NSCLCs for which MET inhibition was active.

WHAT IS KNOWN AND OBJECTIVE?: Leptomeningeal metastasis (LM) is a serious complication of advanced non-small cell lung cancer (NSCLC) that is diagnosed in approximately 3%-5% of patients. LM occurs more frequently in patients with NSCLC harbouring epidermal growth factor receptor (EGFR) mutations or anaplastic lymphoma kinase (ALK) rearrangements and is usually accompanied by a poor prognosis, with a median overall survival (OS) of several months if patients receive conventional treatments. However, tyrosine kinase inhibitor (TKI) therapy after LM diagnosis is an independent predictive factor for extended survival. Here, we aim to summarize the latest advances in targeted therapy for LM and provide patients with better treatment options. METHODS: By reviewing the recent progress of targeted therapy in NSCLC with LM, especially the efficacy of newer generation TKIs, we aim to provide clinicians with a reference to further optimize patient treatment plans. RESULTS AND DISCUSSION: Osimertinib was confirmed to have a several-fold higher CNS permeability than other EGFR-TKIs and was recommended as the preferred choice for patients with EGFR-positive LM whether or not they harboured the T790M mutation. Second-generation ALK-TKIs have a higher rate of intracranial response and can be positioned as front-line drugs in NSCLC with LM. However, the sequence in which ALK-TKIs are administered for effective disease control requires further evaluation. In addition, targeted therapy revealed a potential choice in patients with LM and rare mutations, such as ROS1 and BRAF. WHAT IS NEW AND CONCLUSIONS?: The development of therapeutic agents with greater CNS penetration is vital for the management of CNS metastasis from NSCLC, particularly in the EGFR-mutant and ALK-rearranged subtypes. Systemic therapy with newer generation TKIs is preferred as the initial intervention. This is because newer generation TKIs are designed to penetrate the blood-brain barrier and possess significantly higher intracranial activities. However, their further effectiveness is limited by inadequate blood-brain barrier penetration and acquired drug resistance. Further studies are needed to further understand the mechanisms underlying resistance to treatment.

UNASSIGNED: Thymic epithelial tumors (TETs) are a heterogeneous group of rare malignancies which may be devastating, difficult to treat, and for which treatment options are limited. Herein, we investigated the comprehensive genomic alterations of TETs in a Chinese population for providing clinical management, especially targeted therapy.
UNASSIGNED: Comprehensive genomic profiling (CGP) was performed with DNA targeted sequencing of cancer-associated genes (CSYS) from a cohort of 40 Chinese TET patients. TMB was measured by an in-house algorithm. MSI status was inferred based on the MANTIS (Microsatellite Analysis for Normal-Tumor InStability) score. The expression status of PD-L1 was estimated by immunohistochemistry.
UNASSIGNED: The mutational profiling of thymomas (Ts) and thymic neuroendocrine tumors (TNETs) showed scattered mutation distributions with no recurrently mutated genes. In contrast, thymic carcinomas (TCs) did show highly recurrent mutations including CDKN2A, CYLD, CDKN2B, and TP53. Among them, CDKN2A and CDKN2B mutations were the top potentially actionable alterations in TCs. PD-L1 expression was mainly present in Ts and TCs, and was predominant in males and smokers.
UNASSIGNED: Our study provided a comprehensive genetic alteration view on the largest Chinese cohort of TETs to date. The results identified different genomic mutational profiles of Ts, TCs, and TNETs, and analyzed potential druggable biomarkers with clinical implications in Chinese TET patients, which provided the evidence for precision medicine of rare TET patients.

UNASSIGNED: The ability to rapidly, inexpensively, and accurately identify cancer patients based on actionable genomic mutations in tumour specimens is becoming critically important in routine clinical diagnostics. Targeted panel sequencing is becoming popular because it enables comprehensive and cost-effective diagnosis. However, the implementation of a next-generation sequencing (NGS) assay in clinical settings requires careful analytical validation to demonstrate its ability to detect multiple genomic variants.
UNASSIGNED: Here, we developed a custom-targeted NGS panel to identify actionable variants, including single nucleotide variants, insertions, and deletions; copy number variants; and gene fusions, across 73 genes for targeted cancer therapy. We implemented a practical validation strategy with diluted samples and reference standard samples that modelled key determinants of accuracy, including mutant allele frequency, insertion/deletion length, amplitude of copy number, and hotspot gene fusions.
UNASSIGNED: The analytical validation results demonstrated that our panel can identify different types of genomic alterations in these test samples with high levels of accuracy, sensitivity, and reproducibility.
UNASSIGNED: Our panel could be deployed as a routine clinical test to comprehensively detect actionable variants in cancer patients to guide targeted therapy decisions.

OBJECTIVE: Analysis of circulating tumor DNA (ctDNA) offers an unbiased and noninvasive way to assess the genetic profiles of tumors. This study aimed to analyze mutations in ctDNA and their correlation with tissue mutations in patients with a variety of cancers.
METHODS: We included 21 cancer patients treated with surgical resection for whom we collected paired tissue and plasma samples. Next-generation sequencing (NGS) of all exons was performed in a targeted human comprehensive cancer panel consisting of 275 genes.
RESULTS: Six patients had at least one mutation that was concordant between tissue and ctDNA sequencing. Among all mutations (n = 35) detected by tissue and blood sequencing, 20% (n = 7) were concordant at the gene level. Tissue and ctDNA sequencing identified driver mutations in 66.67% and 47.62% of the tested samples, respectively. Tissue and ctDNA NGS detected actionable alterations in 57.14% and 33.33% of patients, respectively. When somatic alterations identified by each test were combined, the total proportion of patients with actionable mutations increased to 71.43%. Moreover, variants of unknown significance that were judged likely pathogenic had a higher percentage in ctDNA exclusively. Across six representative genes (PIK3CA, CTNNB1, AKT1, KRAS, TP53, and MET), the sensitivity and specificity of detection using mutations in tissue sample as a reference were 25 and 96.74%, respectively.
CONCLUSIONS: This study indicates that tissue NGS and ctDNA NGS are complementary rather than exclusive approaches; these data support the idea that ctDNA is a promising tool to interrogate cancer genetics.