Despite advancements in non-small cell lung cancer (NSCLC) management through the use of molecular biomarkers, the recently introduced 9 th edition of the TNM staging system remains based exclusively on anatomic descriptors, with no consistently demonstrated improvement in risk stratification for early-stage disease. This study explores the integration of a molecular prognostic classifier into the conventional TNM staging system. We analyzed 502 patients with stage I–III lung adenocarcinoma (LUAD) who underwent surgical resection with tumor-based gene expression profiling at the Quebec Heart and Lung Institute. A molecular prognostic classifier was developed and integrated into the 9 th edition TNM staging system to generate a novel model (TNMEx). Prognostic performance was compared with the 8 th and 9 th TNM editions using prognostic discrimination and reclassification metrics. External validation of the molecular classifier was performed in 271 LUAD cases from The Cancer Genome Atlas (TCGA). An independent cohort of 606 resected LUAD patients from the National Cancer Center Hospital (Tokyo) was used to externally compare the prognostic performance of the 8 th and 9 th TNM staging systems in the absence of molecular data. The molecular prognostic classifier was developed based on the expression levels of 26 prognosis-associated genes, weighted by their corresponding coefficients. The classifier was subsequently integrated into the 9 th edition TNM staging to generate the TNMEx model. The TNMEx system demonstrated superior prognostic performance, achieving a higher concordance index (C-index = 0.72) compared to the 9 th edition TNM (C-index = 0.65, p=0.006). Moreover, TNMEx significantly improved patient risk reclassification compared to both the 8 th (net reclassification improvement [NRI] = 0.27, integrated discrimination improvement [IDI] = 0.04) and 9 th editions (NRI = 0.40, IDI = 0.05), underscoring its superior ability to stratify outcomes. The 8 th and 9 th editions showed only limited improvement in overall prognostic accuracy and risk stratification, as reflected by their relatively modest C-index values (0.62 and 0.65, respectively) and minimal reclassification gains (NRI = −0.06, IDI = 0.003). Incorporating a molecular-based prognostic model significantly enhanced the ability to recognize patients at high risk and to predict their survival outcomes more accurately than traditional TNM staging systems.

Abstract Salivary duct carcinoma (SDC) is primarily categorized as de novo (SDCDN) or ex pleomorphic adenoma (SDCXPA). Previous reports indicate a higher frequency of HMGA2 and PLAG1 fusion events in SDCXPA. Surgical resection remains the main intervention due to limited guidance on new treatment strategies. However, frequent recurrence and challenging management of metastasis highlight the necessity for innovative treatments. This study aimed to investigate SDC characteristics, including perineural invasion (PNI), and elucidate its carcinogenic mechanisms and adverse prognostic factors. We analyzed 52 patients with SDC diagnosed in the National Cancer Center Hospital from 2014 to 2023. To compare gene expression profiles, we performed immunohistochemical staining, including Her2, androgen receptor (AR), PLAG1, and HMGA2, followed by human epidermal growth factor receptor 2 (HER2) in situ hybridization and RNA sequencing of 30 cases. Differential analysis identified genes subjected to immunohistochemical staining and statistical analysis. Based on histologic classification using PLAG1 and HMGA2, 52 cases were classified as 26 cases (50%) SDCDN and 26 cases (50%) SDCXPA. Compared with SDCXPA, SDCDN showed higher perineural, venous, and lymphatic invasion rates (P = .0005.0294, and .0044, respectively). Genetic expression profiling revealed clustering tendencies between these subtypes. Focusing on PNI, gene expression was decreased in early growth response 1 in tumor portions infiltrating perineural tissues, indicating a negative correlation (P < .0001). Similarly, ARID1A expression was elevated in tumor regions of cases with perineural invasion; however, immunohistochemical analysis showed no difference between perineural and non-perineural areas. Furthermore, RNA sequencing identified three novel fusion genes in SDC. In conclusion, clinical disparities between SDCDN and SDCXPA based on molecular and pathological features were observed. We found early growth response 1-brain-derived neurotrophic factor-linking crosstalk between cancer cells and nerves for PNI in SDC, offering insights into future treatment and prognostic factors. Citation Format: Airi Sakyo, Eijitsu Ryo, Seiichi Yoshimoto, Go Omura, Chihiro Fushimi, Toshihiko Sakai, Yoshifumi Matsumoto, Azusa Sakai, Kohtaro Eguchi, Yo Suzuki, Kazuki Yokoyama, Yoshitaka Honma, Yasushi Yatabe, Fumihiko Matsumoto, Taisuke Mori. EGR1 Expression and Perineural Invasion in Salivary Duct Carcinoma: A comparison of SDCDN and SDCXPA [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 5948.

Abstract Metastasis remains the leading cause of cancer mortality, yet effective strategies to eliminate metastasis-initiating cells are lacking. Here, we dissected intratumoral heterogeneity in colon cancer using single-cell analyses and identified a KRT17 + slow-cycling cancer cell population that exhibits features of metastasis-initiating cells ( L1CAM ) and senescent-like cells ( CDKN2A, BCL2L1 ). Spatial transcriptomics and immunostaining revealed that these cells localize at tumor-stroma interfaces, where they are closely associated with TGF-β1-producing subset of cancer-associated fibroblasts (CAFs) and exhibit SMAD3 activation. Mechanistically, TGF-β1 induces KRT17 expression in patient-derived cancer cells, while co-culture with CAFs drives the emergence of KRT17 migratory cells in a KRT17-dependent manner. Functionally, genetic ablation of KRT17 or senolytic targeting of BCL2L1 suppresses peritumoral invasion and liver metastasis in xenograft models. Clinically, KRT17 cells co-localize with TGF-β1 CAFs, and their co-expression with L1CAM correlates with advanced disease stage. These findings support a model in which stromal TGF-β signaling promotes the emergence of a KRT17 invasive cancer cell state with senescence-associated features and suggest that senolytic strategies may represent a potential approach to limit metastatic progression in colon cancer.

OBJECTIVE:Pathological "invasion" influences patient prognosis in lung adenocarcinoma (LUAD); however, diagnosis is often associated with high interobserver variability because non-lepidic adenocarcinoma (NLA) and cancer-related fibrosis (CRF) are intricately mixed, and CRF contains both invasive and non-invasive components. In this study, to enable quantitative and reproducible diagnosis of areas showing high intratumoral heterogeneity, we proposed an artificial intelligence (AI)-based analysis distinguishing between NLA and CRF. This approach could reduce interobserver variability in prognostic prediction and help examine the different biological meanings of NLA and CRF. METHODS:To clarify the physiological structure of lung parenchyma, we used elastin staining specimens. CRF and NLA were separately annotated in the first cohort (n = 35) and used for supervised learning. The AI then analyzed whole non-lepidic areas in the second cohort (n = 188); we then examined the relationship between clinicopathological features and the AI analysis. RESULTS:For the first cohort, the accuracy was 89.2% on average. For the second cohort, groups with high CRF ratios (>50%) in the AI analysis showed a statistical correlation with types B-C in Noguchi's classification (p < 0.001), grades 1-2 in the WHO grading system (p = 0.018), and a higher 5-year disease-free survival (DFS) rate (85.7% vs. 69.4%, p = 0.006). In groups with a low CRF ratio (n = 111), a central distribution of CRF was associated with a lower DFS rate (41.2% vs. 74.5%, p = 0.001). CONCLUSION:The AI engine for distinguishing between NLA and CRF enabled prognostic prediction of LUAD with reduced interobserver variability. It may also be useful for examining the biological significance of NLA and CRF.

Invasive lobular carcinoma (ILC) is a special type of breast cancer. The histological subtypes of ILC exhibit diverse morphological features, and the prognosis differs accordingly. Compared with patients with classic-ILC (C-ILC), patients with pleomorphic-ILC (P-ILC) have a worse prognosis, owing to high-grade nuclear atypia and mitotic cells. However, the molecular differences between C-ILC and P-ILC remain unclear. To address this gap, we performed spatial transcriptomic profiling on four fresh-frozen C-ILC samples and four fresh-frozen P-ILC samples, followed by confirmation of reproducibility in bulk RNA-seq datasets. We identified significant enrichment of cellular response to heat stress in P-ILC. Furthermore, molecular clustering analysis using genes differentially expressed across ILC samples in spatial transcriptome revealed that ILC has three molecular subtypes: proliferative (PR), immunoreactive (IM), and stroma-rich (ST), associated with distinct prognostic outcomes. Although these molecular subtypes did not completely correspond to C-ILC or P-ILC, PR tended to include P-ILC, and ST tended to include C-ILC. These molecular clusters exhibited features comparable to previously reported subtypes. Despite these phenotypic features, ILC is generally treated similarly to invasive ductal carcinoma (IDC), with limited consideration of molecular subtype classification. Our findings suggest that molecular profiling may more accurately reflect prognosis than conventional histological classification and may provide potential diagnostic markers and therapeutic targets.

INTRODUCTION:Histologic descriptors such as lymphovascular invasion (LVI), visceral pleural invasion (VPI), spread through air spaces (STAS), and histologic grade have each been associated with adverse outcomes in lung adenocarcinoma. However, with the exception of VPI, these features are not formally incorporated into the TNM staging system. We evaluated the prognostic value and incremental contribution of these histologic descriptors within the framework of the ninth edition TNM staging system. METHODS:In total, 1745 individuals diagnosed with stages I to III invasive nonmucinous lung adenocarcinoma were included in this study, comprising 1139 French-Canadian patients who underwent surgical resection at the IUCPQ-Université Laval (discovery cohort) and 606 patients from the National Cancer Center Hospital in Tokyo, Japan (validation cohort). The objective of this study was to assess the prognostic contribution of histologic descriptors, including histologic grade, STAS, and LVI, as complements to conventional ninth edition TNM staging. RESULTS:Grade 3 tumors, LVI, and STAS were identified in 880 (50.4%), 809 (46.4%), and 775 (44.4%) of 1745 cases, respectively. Histologic grade and LVI demonstrated the strongest associations, particularly in early stage disease, whereas STAS exhibited a stage-dependent effect, being more impactful in stages II to III. VPI had less consistent prognostic value. Incorporating these histologic descriptors into TNM staging improved prognostic model performance, with the largest gains driven by histologic grade and LVI, whereas STAS provided modest complementary prognostic refinement. CONCLUSION:These findings demonstrate that key histologic descriptors-including histologic grade, LVI, and STAS-represent robust and consistent prognostic parameters. Importantly, these descriptors provide complementary, stage-dependent information that may enhance risk stratification and inform refinement of future TNM staging frameworks, including the forthcoming tenth edition.

8033 Background: SMARCA4 mutations occur in 5–10% of non-small cell lung cancers (NSCLCs) and are associated with poor overall survival, especially class I SMARCA4 alterations (truncating mutations, fusions, and homozygous deletion), which often result in SMARCA4 loss. However, the biological and clinical significance of SMARCA4 alteration classes and their distribution among pan-cancers in Asians remain unclear. Methods: We analyzed pan-cancer across solid tumors patients enrolled in Center for Cancer Genomics and Advanced Therapeutics (C-CAT) database (approval number CDU2021-001N) between June 2019 and October 2025 and patients with surgically resected NSCLCs at our hospital between 1999 and 2023 (approval number 2005-109) were performed by whole exome sequencing. Somatic mutations were annotated using Annovar and OncoKB. SMARCA4 mutations were classified into two groups: class I or class II (missense or other). Copy number alterations were detected using facets and GISTIC2. Transcriptomic differences were assessed using RNA sequencing with DESeq2 and gene set enrichment analysis (GSEA). Tumor immune cell composition was assessed using CIBERSORTx. Results: In the C-CAT pan-cancer dataset (N=10,155), SMARCA4 mutations were frequently identified in NSCLC (NSCLC vs. others; 12.3% vs. 4.0%). In NSCLC, TP53 was the most frequent co-altered gene in patients with class I and class II mutation (73.0% vs. 63.0%). Other major genes did not show clear class-associated differences. In our NSCLC cohort (N=1,166), the frequency of SMARCA4 class I and II mutation was 2.3% (all truncating) and 2.3% (all missense), respectively. The class I SMARCA4 alterations were associated with significantly poorer survival outcomes compared with wild- type tumors (OS: HR = 2.04, 95% CI = 1.24–3.12, p = 0.007; RFS: HR = 2.09, 95% CI = 1.28–3.20, p = 0.005), whereas class II alterations were not significantly different in survival (OS: HR =0.50, 95% CI =0.40-0.61, p =0.17; RFS: HR = 0.46, 95% CI = 0.42–0.76, p =0.12). Regarding transcriptomic analysis and genomic profiling, class I tumors showed activation of ASCL1 -related transcriptional programs, suppression of squamous/basal lineage differentiation, increased CD8⁺ T-cell infiltration, and recurrent copy number loss of SMARCA2 and CDKN2A/B compared with wild type tumors. In contrast, class II tumors showed limited transcriptomic divergence from wild-type tumors, downregulating SPRR2E , KRT5 , and KRT13 without a significant CD8⁺ T-cell difference. Conclusions: In NSCLC, class I SMARCA4 altered tumors are associated with poor survival outcomes, and an immune-active molecular phenotype with recurrent copy number loss of SMARCA2 and CDKN2A/B , whereas class II tumors lack these features and do not show adverse survival compared to wild-type tumors. This highlights the distinct clinical implications of SMARCA4 alteration class.

INTRODUCTION:Minimal residual disease (MRD) detection using liquid biopsy is an emerging tool for risk stratification and monitoring for recurrence in resected early stage NSCLC. There is increasing need for clear guidance on its optimal clinical implementation. METHODS:The Asian Thoracic Oncology Research Group (ATORG) convened a multidisciplinary panel of 27 experts to develop a consensus statement on the clinical application of circulating tumor DNA-based MRD testing in early stage resected NSCLC, using a structured Delphi methodology. Statements were organized into the following broad thematic domains: assay validity and standardization; harmonization in research and trials; clinical application; challenges in implementation; consensus recommendations; infrastructure for regional MRD adoption; and roadmap for pragmatic trials. RESULTS:A total of 23 position statements were developed, of which all except one achieved strong consensus. The consensus highlighted the need to define minimum analytical performance thresholds for MRD assays, improve standardization of reporting metrics, and clear guidelines for pre-analytical handling. Harmonization of blood sampling time points and terminology across clinical trials is also essential to confirm the prognostic value of MRD assays. Although current MRD assays demonstrate high specificity and positive predictive value, variable sensitivity precludes routine use for adjuvant therapy de-escalation outside clinical trials. Broader access, sustainable funding, ongoing consensus building, and collaborative real-world data generation are also critical to support clinical implementation and adoption. Future clinical trials must account for the distinct biology and changing standards of care associated with different driver genes. CONCLUSION:These consensus recommendations provide a pragmatic framework to guide the responsible integration of MRD testing into clinical research and practice.

PURPOSE Rare cancers are molecularly heterogeneous and lack robust clinical evidence to guide treatment. Whether cancer driver gene expression is governed by tumor lineage or specific somatic alterations—and whether this distinction carries clinical relevance—remains poorly characterized, particularly in Asian populations. This study aimed to determine the relative contributions of tumor lineage and DNA mutations to RNA expression of cancer driver genes and to evaluate the clinical implications of genomic alteration status in Asian patients with rare cancers. METHODS DNA and RNA were extracted from formalin-fixed paraffin-embedded tumor samples obtained through the MASTER KEY Asia research network. Next-generation sequencing was performed to evaluate genomic alterations and gene expression profiles. RNA expression of cancer driver genes was quantified as TPM Z scores. Elastic net regression with 10-fold cross-validation assessed the relative contributions of tumor lineage and DNA mutations to RNA expression, and findings were independently replicated in another cohort. RESULTS Integrated DNA and RNA data were available for 128 patients. Tumor lineage was the predominant determinant of RNA expression in 17 of 27 cancer driver genes (63%), whereas ERBB2 and MDM2 demonstrated genomic-predominant regulation, replicated across both cohorts. Alterations in ERBB2 or ARID1A were associated with a lower frequency of disease progression compared with patients without either alteration ( P = .01), independently of ERBB2 RNA expression. Functional subclassification of TP53 revealed that gain-of-function mutations were associated with higher rates of disease progression ( P = .007) and shorter progression-free survival than loss-of-function variants (hazard ratio, 2.355 [95% CI, 1.086 to 5.105]; P = .029). Tumor mutational burden was not associated with treatment response. CONCLUSION These findings demonstrate that genomic alteration status, including TP53 functional subclassification, provides clinically meaningful information beyond transcriptional profiling alone, supporting integrative genomics and transcriptomics analysis for precision oncology in rare cancers.

INTRODUCTION:Tumor spread through air spaces (STAS) and histologic tumor grade have been reported as adverse prognostic factors in non-small cell lung cancer (NSCLC), yet their impact on the extent of surgical resection remains insufficiently defined. This study aimed to evaluate the prognostic impacts of STAS and tumor grade in stage IA patients (pT1a/bN0) from JCOG0802/WJOG4607L comparing segmentectomy and lobectomy. METHODS:STAS and tumor grading were assessed in 593 tumors by 32 expert pulmonary pathologists across the world. The prognostic impact of STAS in lobectomy (n=294) versus segmentectomy (299) in relation to relapse-free survival (RFS) and overall survival (OS) was evaluated, as well as the association of STAS with histologic grading. RESULTS:STAS was identified in 227 cases (38.3%) and was significantly associated with shorter relapse-free survival (RFS) and overall survival (OS) (RFS HR=2.204; OS HR=1.917; both p<0.001). Multivariable analysis confirmed STAS as an independent adverse factor for RFS (p<0.001) and OS (p=0.009). STAS was associated with higher local recurrence in both lobectomy and segmentectomy approaches (p=0.012 and p=0.002, respectively). STAS independently predicted shorter RFS in both lobectomy and segmentectomy subgroups, but only OS in the lobectomy subgroup. Among invasive non-mucinous adenocarcinomas, histologic grade 3 was strongly associated with STAS (p<0.001) and poor outcomes. Both STAS and grade 3 independently predicted shorter RFS, whereas only grade 3 predicted OS. CONCLUSIONS:These findings reinforce STAS and high-grade adenocarcinoma as clinically significant prognostic factors and support their incorporation into risk stratification for stage IA (pT1a/bN0) NSCLC patients.

Polatuzumab vedotin (PV), a CD79B-directed antibody–drug conjugate, is approved for treating diffuse large B-cell lymphoma (DLBCL). However, the clinical relevance of CD79B expression and the frequency of loss or decrease after PV-containing therapy remain unclear. We evaluated immunohistochemical CD79B expression using H-score (0–300) in 91 patients with DLBCL before PV-containing therapy and assessed post-treatment changes. Low CD79B expression was observed in 15 (16

BACKGROUND:Dual inhibition of epidermal growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF) pathways may improve outcomes in metastatic EGFR-mutant NSCLC, but VEGF inhibitors are not universally effective. We evaluated the impact of VEGFA and VEGFR2 expressions on EGFR-TKI outcomes. METHODS:EGFR-mutant NSCLC patients from the National Cancer Center Hospital, were retrospectively analysed. The early-stage cohort comprised stage I-IIIA patients (1997-2019). The advanced-stage cohort included metastatic patients (2018-2022). VEGFA/VEGFR2 expressions were dichotomised by median transcripts per million. RESULTS:Among 447 early-stage patients (median age 66), high VEGFA was associated with smoking, TP53 co-mutation, higher Brinkman Index, and higher tumour mutation burden (all p < 0.01). High VEGFA predicted shorter relapse-free survival (HR 2.10, 95% CI 1.63-2.71, p < 0.01) and overall survival (HR 2.07, 95% CI 1.46-2.95, p < 0.01). VEGFR2 expression showed no prognostic impact. In 146 relapsed patients receiving first-line EGFR-TKIs, high VEGFA was linked to shorter progression-free survival (PFS) overall (HR 1.70,95%CI 1.13-2.56, p = 0.009), particularly for first/second-generation EGFR-TKIs (HR 1.66,95%CI 1.08-2.54 p = 0.023), but not for osimertinib (p = 0.491). In 60 advanced-stage patients on osimertinib, PFS was unaffected by VEGFA (p = 0.102). CONCLUSIONS:High VEGFA is associated with aggressive biology and inferior outcomes, correlating with shorter PFS for first/second-generation EGFR-TKIs but not for osimertinib.

Prognostic performance of mitotic index, DL score, and MJ risk models in cohorts C1 to C3.

Kaplan-Meier curves for recurrence-free survival (RFS) and overall survival (OS) in the subgroup of C2 and C3 patients with high risk score according to pathological Miettienen-Joensuu scoring system and Imatinib sensitive mutations depending on the deep Miettinen-Joensuu models. Kaplan-Meier curves for RFS (A) and OS (B) depending on the deep Miettinen-Joensuu model employing the C2 DL Score in this subgroup of the C2 subcohort. Kaplan-Meier curves for RFS (C) and OS (D) depending on the deep Miettinen-Joensuu model employing the C3 DL Score in this subgroup of the C3 subcohort. *: p < 0.05; ***: p < 0.001. Tests are log-rank tests.

Background:Gastrointestinal stromal tumor (GIST) is the most common gastrointestinal mesenchymal tumor, driven by tyrosine-protein kinase KIT and platelet-derived growth factor receptor A (PDGFRA) mutations. Specific variants, such as KIT exon 11 deletions, carry prognostic and therapeutic implications, whereas wild-type (WT) variants derive limited benefit from tyrosine kinase inhibitors (TKIs). Given the limited reproducibility of established clinicopathological risk models, deep learning (DL) applied to whole-slide images (WSIs) emerged as a promising tool for molecular classification and prognostic assessment. Patients and methods:We analyzed 8398 GIST cases from 21 centers in 7 countries, including 7238 with molecular data and 2638 with clinical follow-up. DL models were trained on WSIs to predict mutations, treatment sensitivity, and recurrence-free survival (RFS). Results:DL predicted mutational status in GIST from WSIs, with area under the curve (AUC) of 0.87 for KIT, 0.96 for PDGFRA. High performance was observed for subtypes, including KIT exon 11 del-inss 557-558 (0.67) and PDGFRA exon 18 D842V (0.93). For therapeutic categories, performance reached 0.84 for avapritinib sensitivity, 0.81 for imatinib sensitivity. DL models predicted RFS, with hazard-ratios (HR) of 8.44 (95%CI 6.14-11.61) in the overall cohort and 4.74 (95%CI 3.34-6.74) in patients receiving adjuvant therapy. Prognostic performance was comparable to pathology-based scores, with highest discrimination in the overall cohort and in patients without adjuvant therapy (9.44, 95%CI (5.87-15.20)). Conclusion:DL applied to WSIs enables prediction of molecular alterations, treatment sensitivity, and RFS in GIST, performing comparably to established risk scores across international cohorts, providing a baseline for future multimodal predictors.

DL predicts treatment sensitivity and supports clinical decision-making in GIST. A, Standard clinical workflow from initial diagnosis to treatment selection in patients with GIST. B, Proposed integration of the DL model into the diagnostic–therapeutic pathway, enabling early triage of patients for further molecular analysis or tailored treatment selection. C, AUC with 95% CI for DL-based prediction of treatment sensitivity, reported for the internal validation cohort (pink), external validation cohort (green), and an external biopsy-only validation cohort (light green). D, Confusion matrices for avapritinib and imatinib sensitivity, shown at the optimal threshold by Youden’s index (green) and at the threshold yielding the highest F1 score (pink).