Introduction:Brain metastases (BrMs) are poorly studied and likely underreported in diffuse pleural mesothelioma (DPM), limiting understanding of risk factors associated with their development. We evaluated the genomic, histologic, and clinical landscapes of patients with DPM and BrMs. Methods:We retrospectively reviewed all patients with a mesothelioma diagnosis treated at the Memorial Sloan Kettering Cancer Center between January 1, 2010, and May 1, 2025, with cross-sectional brain imaging. Clinicopathologic data and treatment outcomes were annotated for patients with BrMs. Results:Among 194 patients with mesothelioma and brain imaging, 16 (8%) had BrMs. Half (n = 8) had tumors of epithelioid histology, 44% (n = 7) biphasic, and 6% (n = 1) sarcomatoid. Compared with our DPM cohort regardless of BrMs (n = 194) or a DPM cohort from The Cancer Genome Atlas (n = 74), patients with DPM and BrMs had a higher prevalence of LATS2 alterations (46% versus 7% versus 12%, respectively [p < 0.00001]). BrMs were typically (88%, n = 14) diagnosed after neurologic symptoms prompting imaging. Among patients with BrMs, the median overall survival from initial mesothelioma diagnosis was 35.4 (95% confidence interval, 19.3-not reached) months and 3.7 (95% confidence interval, 3.5-not reached) months from BrM diagnosis. Conclusions:Our findings suggest that BrM develops relatively late in DPM, is more common than previously reported, and may be enriched in patients with LATS2 alterations. Prospective, multi-institutional studies with standardized brain imaging are needed to further characterize the incidence of BrMs in DPM and associated risk factors. Routine brain surveillance at diagnosis and with symptoms should be considered for patients with DPM.

Histologic transformation to lung squamous cell carcinoma (LUSC) is an underrecognized mechanism of resistance in epidermal growth factor receptor (EGFR)-mutant lung adenocarcinoma (LUAD). Although AKT and MYC activation have been linked to LUSC features, the clinicogenomic determinants of this transformation remain undefined. In this study, we performed comprehensive clinical and multiomic profiling-including genomic, transcriptomic, methylation, and proteomic analyses-of EGFR-mutant tumors that were transforming, adenosquamous (LUAS), or de novo LUSC. Patients with EGFR-mutant LUSC or LUAS had shorter overall survival on first-line osimertinib compared with those with EGFR-mutant LUAD. Transforming tumors were enriched for alterations in the retinoblastoma (Rb) and AKT pathways, particularly cyclin-dependent kinase inhibitor 2A/B (CDKN2A/B) deletions. These alterations were also frequent in EGFR-wild-type LUSC and associated with shorter time-to-osimertinib discontinuation. In genetically engineered human in vivo models, Rb inactivation, in combination with AKT and MYC activation, enhanced the acquisition of LUSC features. Single-cell RNA profiling of such models recapitulated the molecular changes observed in the transforming clinical specimens and identified MET pathway up-regulation during transformation. Combined EGFR and MET inhibition suppressed tumor growth in patient-derived xenograft models of LUSC transformation. Together, these findings highlight Rb pathway inactivation as a promoter of LUSC transformation in EGFR-mutant lung cancer and identify MET signaling as a therapeutic vulnerability that may suppress plasticity in this setting and extend response to targeted therapy.

In this article, we summarize the progress made in lung cancer, mesothelioma, and thymic epithelial malignancy during the period 2005-2025. We enlisted multidisciplinary thoracic oncologic experts to tackle this task. The main focus of the article concerns how basic science with translational impact has improved the diagnosis, prognosis, and therapy of these cancers. During the past 20 years, we have come to the realization that "lung cancer" is a name that encompasses tumors with vast histologic, immune, and genomic differences that in turn influence prognosis and response to therapy. For example, programmed death-ligand 1 levels are being used as an immune signature which guides the use of immunotherapy. There is an 85% higher risk for developing lung cancer among first-degree relatives of patients with lung cancer. Accordingly, an increasing number of lung cancers are being identified in carriers of predisposing germline pathogenic inactivating mutations, suggesting that screening programs for early lung cancer detection may benefit family members. Underscoring the role of genetics, and the importance of germline testing, a different variant of mesothelioma has been identified developing in carriers of inactivating heterozygous germline mutations of BAP1 and of other tumor suppressor genes, including a new variant of mesothelioma caused by fusion genes. These variants of mesothelioma are characterized by specific histologic and molecular genetic alterations. These patients benefit from screening programs as they are at risk of multiple malignancies, their tumors are usually much less aggressive, and they are more responsive to therapy compared with sporadic, asbestos-induced mesotheliomas. Thus, the tailored therapeutic approach that is described here for lung cancer may extend to patients with mesothelioma, rather than the previous "one therapy fits all" approach. Progress in the rare thymic epithelial tumors has been less marked; however, recent insights into the biology of thymic tumors have resulted in the development of clinically relevant interventions.

PURPOSE:MYC, MYCN and MYCL amplifications are recurrent oncogenic events across solid tumors. Currently, no standardized selection biomarker is available to identify patients with MYC-dependent tumors. EXPERIMENTAL DESIGN:We analyzed copy number alterations of MYC family genes and their features in over 68,000 tumor-normal paired samples from pediatric and adult patients sequenced with MSK-IMPACT and annotated with FACETS. The relationship between amplification features and MYC mRNA expression levels were evaluated in over 10,000 samples from the TCGA. RESULTS:Across MSKCC samples, MYC amplifications were most common, found in 2,949 samples compared to 310 in MYCL and 217 in MYCN. While MYCN and MYCL amplifications were predominantly focal (<10Mb, 79% and 93% respectively), MYC amplifications were frequently broader (>10Mb, 62%). While most tumor types showed similar features between broad and focal amplifications of MYC, in select cancer types we identified differing co-occurrence and mutual-exclusivity patterns with other disease specific drivers. Furthermore, while MYC amplified TCGA samples showed higher mRNA expression than wild-type ones, the focality of MYC amplification was seen to have limited influence on expression levels. CONCLUSIONS:Our results suggest that MYC-dependency likely depends on many factors including, but not limited to, total copy number of the detected amplification, lineage specific factors, concomitant presence or absence of additional oncogenic alterations, and in some cases amplification focality.

BACKGROUND:Selective RET inhibitors (SRIs) have durable activity in RET fusion-positive lung cancers. Insidious side effects may emerge with long-term use. This is the first systematic analysis of intestinal lymphangiectasia (IL), an underrecognized disorder characterized by dilated intestinal lacteals and potential lymph leakage, in this population. METHODS:Patients with RET-altered lung cancers treated with selpercatinib or pralsetinib were eligible for this retrospective analysis. IL was classified as possible (radiologic findings), probable (radiologic and clinical findings), and definite (radiologic, clinical, and endoscopic or pathologic findings). RESULTS:Of 113 patients, 33 (29%) had IL-associated radiologic findings. Cumulative incidence at 1, 3, and 5 years was 11% (95% confidence interval [CI], 6%-18%), 27% (95% CI, 19%-36%), and 31% (95% CI: 22%-41%), respectively. Prior immune checkpoint inhibitor exposure was associated with IL risk (hazard ratio, 3.02; 95% CI: 1.53-5.96; p = 0.001). Among patients with IL-associated findings, IL was classified as possible, probable, and definite in 25%, 60%, and 15%, respectively. Median time from SRI initiation to radiologic findings (involving the ileum and duodenum in 67% and 24% of patients with IL, respectively) was 15 months. Radiologic findings were associated with additional clinical findings (e.g., gastrointestinal symptoms, hypoalbuminemia, hypocalcemia, third spacing) in 76% of patients. All patients with assessable endoscopic biopsies had mild lacteal dilatation. IL improved radiologically and clinically in 64% of patients after SRI dose reduction or discontinuation. CONCLUSION:Drug-induced IL occurs with selpercatinib or pralsetinib treatment. The frequency increases with longer drug exposure. Serial monitoring is essential and when necessary, dose modification. Most cases do not lead to treatment discontinuation.

PURPOSE:To evaluate the efficacy of ado-trastuzumab emtansine (T-DM1) among patients with advanced/metastatic HER2-amplified solid tumors. PATIENTS AND METHODS:This was a single-center, non-randomized phase 2 basket trial conducted at Memorial Sloan Kettering Cancer Center. All 95 patients had HER2-amplified metastatic/advanced disease and were enrolled in 1 of 5 cohorts. HER2 amplification was diagnosed either with next-generation sequencing or in-situ hybridization. All patients received intravenous T-DM1 3.6 mg/kg every 21 days. The primary endpoint was overall response rate (ORR). Overall survival (OS), progression-free survival (PFS), duration of response (DOR), and safety were secondary endpoints. RESULTS:We treated 95 patients, and 22 (23% [95% CI, 16-33%]) had a confirmed response by investigator assessment. The investigator-assessed confirmed ORRs by cohort were: salivary 11/19 (58% [95% CI, 36-77%]); lung 4/23 (17% [95% CI, 7-37%]); colorectal 0/7 (0% [95% CI, 0-35%]); endometrial 5/23 (22% [95% CI, 10-42%]); "other" 2/23 (9% [95% CI, 2-27%]). Median PFS was 3.6 months (95% CI, 2.6-5.4) and ranged from 1.4 months in the "other" cohort to 10.2 months in the salivary cohort. Median OS was 11.9 months (95% CI, 8.4-17.2) and ranged from 7.8 months in the "other" cohort to 29.2 months in the salivary cohort. Median DOR was 13.1 months (95% CI, 7.3-30.5), ranging from 6.4 months in the lung cohort to 18.4 months in the "other" cohort. CONCLUSIONS:T-DM1 demonstrated heterogenous efficacy among HER2-amplified tumor types, with promising response and outcomes among patients with salivary gland cancer.

PURPOSE:Oligometastatic disease (OMD) is an intermediate state of metastatic disease in which metastasis-directed therapy (MDT) may improve outcomes. The classification of OMD is inconsistent, typically defined by number of metastases without considering tumor biologic characteristics. To help optimize patient selection for MDT, we analyzed integrated genomic sequencing results from patients with metastatic non-small cell lung cancer (NSCLC). MATERIALS AND METHODS:Patients with metastatic NSCLC who had molecular sequencing through a validated institutional assay (MSK-IMPACT) were included. The number and location of metastases were manually annotated (assigned 1-10 and >10). Individual gene variant scores were analyzed by gene length and patient mutation burden. Analysis was completed in R and included maximally selected rank statistics, agglomerative hierarchical clustering, and the chi-square test for independence. RESULTS:In total, 844 patients had clinical data, tissue sequencing, and annotated imaging for analysis. Of these, 635 had >10 metastases (75.2%), and 209 had 1-10 metastases (24.8%). The cutpoint that maximized overall survival (OS) was four metastases, and six mutation signatures were identified. For patients with 1-4 metastases, TERT and KMT2D had inferior OS, while for those with ≥5 metastases, EGFR, ALK, and TBX3 had superior OS. CONCLUSION:The cutpoint that maximized difference in OS was four metastases, but incorporating genetic alteration information modified this criterion. These findings were proof of principle that integrating multimodal data beyond number of lesions can better identify patients with metastatic NSCLC who may be candidates for MDT.

Introduction Brain metastases (BrMs) are poorly studied and likely underreported in diffuse pleural mesothelioma (DPM), limiting understanding of risk factors associated with their development. We evaluated the genomic, histologic, and clinical landscapes of patients with DPM and BrMs. Methods We retrospectively reviewed all patients with a mesothelioma diagnosis treated at Memorial Sloan Kettering Cancer Center between January 1, 2010, and May 1, 2025, with cross sectional brain imaging. Clinicopathologic data and treatment outcomes were annotated for patients with BrMs. Results Among 194 patients with mesothelioma and brain imaging, 16 (8%) had BrMs. Half (n=8) had tumors of epithelioid histology, 44% (n=7) biphasic, and 6% (n=1) sarcomatoid. Compared with our DPM cohort regardless of BrMs (n=194) or a DPM cohort from The Cancer Genome Atlas (n=74), patients with DPM and BrMs had a higher prevalence of LATS2 alterations (46% vs. 7% vs. 12%, respectively [p < 0.00001]). BrMs were typically (88%, n=14) diagnosed after neurological symptoms prompting imaging. Among patients with BrMs, the median overall survival (OS) from initial mesothelioma diagnosis was 35.4 (95% CI, 19.3–NA) months and 3.7 (95% CI, 3.5–NA) months from BrM diagnosis. Conclusions Our findings suggest BrM develops relatively late in DPM, is more common than previously reported, and may be enriched in patients with LATS2 alterations. Prospective, multi-institutional studies with standardized brain imaging are needed to further characterize the incidence of BrMs in DPM and associated risk factors. Routine brain surveillance at diagnosis and with symptoms should be considered for patients with DPM.

INTRODUCTION:Machine learning algorithms may improve the efficiency and accuracy of pathologic response (PR) assessment in surgically resected lung cancers after neoadjuvant therapy. The aim of this study was to develop digital models for quantifying tumor bed (TB) area and residual viable tumor (VT) and to compare these results to previously published assessments of PR by pathologists from the International Association for the Study of Lung Cancer reproducibility study. METHODS:Manual pathologist annotations (N = 15,564) of regions including TB and VT were used to train a convolutional neural network model (digital artificial intelligence [AI]) and a convex hull algorithm (CHA). PR was determined by the percentage of VT in the TB area. The pathologists determined the average PR (APR) across slides (unweighted), which was compared with the weighted average for digital AI and CHA. The concordance between pathologist APR, digital AI, and CHA was calculated and correlated with outcomes. RESULTS:There was a strong correlation between approaches: APR versus digital AI (0.97), APR versus CHA (0.97), and digital AI versus CHA (0.99). Digital AI and CHA demonstrated 100% agreement for MPR. The κ concordance for MPR was 0.82 (95% confidence interval [CI]: 0.69, 0.96) for APR versus digital AI/CHA with six discordant cases. The concordance was higher for squamous cell carcinoma (κ = 0.92, 95% CI: 0.76, 1.0) than for nonsquamous carcinoma (κ = 0.77, 95% CI: 0.59, 0.96). APR and digital AI demonstrated similar relapse-free survival and overall survival. CONCLUSION:The overall high level of agreement supports the utility of machine learning approaches for evaluation of PR in patients with NSCLC.

PURPOSE:Concurrent platinum-based chemotherapy and radiation therapy (cCRT) followed by consolidative durvalumab is the standard of care (SoC) for fit patients with inoperable, locally advanced non-small cell lung cancer (LA-NSCLC). However, no SoC exists for patients who are cCRT-ineligible because of age, comorbidities, or frailty. Here, we investigated the efficacy and adverse events (AEs) of concurrent and consolidative durvalumab with definitive radiation therapy (RT) without chemotherapy. METHODS:In this multicenter, single-arm, prospective phase II study, patients received conventionally fractionated RT plus concurrent and consolidative durvalumab (1,500 mg fixed dose once every 4 weeks) for up to 12 months. The primary end point was 2-year progression-free survival (PFS) of 36% compared with historical results of 20% with sequential CRT (86% power). Additional end points included overall survival (OS) and cancer-specific survival (CSS). RESULTS:Fifty-eight patients (median age 82 years [IQR, 76-86]; 16 [28%] PD-L1-negative; eight [14%] Eastern Cooperative Oncology Group [ECOG] score 2; 46 [79%] ECOG score 1; four [7%] ECOG score 0) were treated per protocol. The study met its primary end point with a 2-year PFS of 39% (one-sided CI, 29 to 100) and a 2-year OS of 54%. Better performance status and PD-L1 positivity were associated with improved PFS; better ECOG was associated with improved OS; PD-L1 positivity was associated with better CSS. Grade 3/4 treatment-related AEs occurred in 12 (21%) patients. Grade 5 AEs occurred in four (7%) patients (radiation pneumonitis (n = 2) and cardiac arrest (n = 2)). Durvalumab was discontinued early because of AEs in 18 (31%) patients. CONCLUSION:Thoracic RT with concurrent and consolidative durvalumab is a promising treatment option for cCRT-ineligible patients with LA-NSCLC, demonstrating better PFS with a favorable safety profile compared with historical controls.

CHORUS (NCT04905316) was an open-label, non-randomized, single-arm phase I/II trial of the anti-IL-1beta monoclonal antibody canakinumab given with concurrent chemoradiotherapy and continued with anti-PD-L1 durvalumab consolidation in unresectable stage III non-small cell lung cancer. The primary endpoint was 24-month progression-free survival (PFS) from radiotherapy initiation; secondary endpoints included overall survival, objective response rate, and safety. Forty-one patients were enrolled, 37 received canakinumab, and 32 comprised the prespecified efficacy population. At a median follow-up of 25 months, 24-month PFS was 67% (95% confidence interval, 52–88); the lower bound of the prespecified one-sided 90% confidence interval was 56.5%, exceeding the 46% historical benchmark, and the primary endpoint was met. Twelve- and 24-month overall survival were 93% and 73%, respectively, and the confirmed objective response rate was 81%. Canakinumab-related grade 3–4 events included neutropenia (n = 5), febrile neutropenia (n = 1), lymphopenia (n = 1), lung infection (n = 1), fatigue (n = 1), and hypertension (n = 1); no canakinumab-related deaths occurred. Pneumonitis of any attribution occurred in 9 of 37 patients (24%; grade 2, n = 7; grade 3, n = 2). Peripheral immune cell profiling showed attenuated expansion of myeloid-derived suppressor cells and CTLA-4-expressing CD8 + T-cell subsets among responders. Circulating tumor DNA cleared by week 6 in 20 of 30 evaluable patients (66%). These findings support randomized evaluation of interleukin-1β blockade in this setting. IL-1β targeting may attenuate myeloid-suppressive pathways and support antitumor immunity. Here, the authors report a phase I/II trial of concurrent canakinumab, an anti-IL-1β antibody, with chemoradiotherapy, followed by consolidation canakinumab plus the anti-PD-L1 antibody durvalumab, in patients with unresectable stage III non-small cell lung cancer.

8087 Background: Tarlatamab is a DLL3-targeted bispecific T-cell engager used to treat small cell lung cancer (SCLC). SCLC subtypes can be defined by relative expression of master transcription factors (TFs). Whether subtype and DLL3 expression are associated with outcomes on tarlatamab is unknown. Methods: We analyzed all patients (pts) with SCLC, including SCLC transformed from lung adenocarcinoma, treated with tarlatamab at Memorial Sloan Kettering prior to October 2025. DLL3 expression and subtyping were determined by IHC by dedicated thoracic pathologists. Expression of subtype TFs ASCL1, NEUROD1, POU2F3, and YAP1 was assessed. “Neuroendocrine (NE)-high” was defined as any ASCL1 or NEUROD1 expression. Pts with any “neuroendocrine (NE)-low” subtype expression—POU2F3 or YAP1—were grouped for analysis. DLL3 fractional expression was dichotomized as “high” (≥50%) or “low” (0-49%). Progression-free and overall survival (PFS, OS) were analyzed with Kaplan-Meier curves and compared with the Cox proportional hazards model. Results: We identified 39 pts with SCLC treated with tarlatamab with DLL3 and/or subtyping data available. Median age was 65, 49% were women, and median prior lines of therapy was 1. DLL3 expression was available in 34 pts. Median fractional DLL3 expression was 80% and ranged from 0 to 100%. 7 pts had low expression and 27 had high expression. There were more pts with transformed SCLC in the DLL3-low group (3/7 [43%] vs 1/27 [4%], p=0.02). Among pts with response data available, response was inferior in the DLL3-low group: complete or partial response (CR or PR) in 0/7 (0%; [95% CI, 0 to 35]) vs 12/25 (48%; [95% CI, 30 to 67]); stable disease (SD) in 2/7 (29%) vs 5/25 (20%); progressive disease (PD) in 5/7 (71%) vs 8/25 (32%); p=0.049. Median PFS was 1.4 mos in DLL3-low vs 4.5 mos in DLL3-high disease (p<0.001); median OS was 4.7 mos vs not reached (p=0.001). Incidence of CRS and ICANS did not differ (5/7 [71%] in DLL3-low vs 11/29 [38%] in DLL3-high with CRS, p=0.2; 0/9 vs 3/29 [10%] with ICANS, p>0.9). 30/39 pts had subtyping data available. 28/30 pts had NE-high tumors (15/28 ASCL1+/NEUROD1+; 11/28 ASCL1+/NEUROD1-; 2/28 ASCL1-/NEUROD1+). 5/30 had subtype biomarkers of NE-low status: 4 YAP1+ and 1 POU2F3+; the 4 YAP1+ tumors also expressed ASCL1. Response rates were similar between pts with and without NE-low expression: 1/5 (20%; [95% CI, 1 to 62]) vs 11/25 (44%; [95% CI, 27 to 63]), p=0.6. PFS was also similar (mPFS 1.1 vs 2.7 mos, p=0.3). Conclusions: Low DLL3 expression was potentially associated with inferior response to tarlatamab. A contributing factor may be enrichment of transformed SCLC, which is associated with poor outcomes. We did not detect an impact of POU2F3/YAP1 expression. The majority of pts in this cohort had DLL3-high, NE-high SCLC—in line with prior studies—and they had a response rate of 48%. These data emphasize the need for more discerning predictive biomarkers for tarlatamab in an evolving landscape of treatment options.

List of 301 genes included in all versions of MSK-IMPACT and DFCI OncoPanel NGS platforms used for all analyses.

Kaplan-Meier curves of overall survival of patients with long-term response (LTR), short-term response (STR), and progressive disease (PD) from individual sites.

8081 Background: MTAP ( methylthioadenosine phosphorylase ) is located on chromosome 9p21 and often co-deleted with CDKN2A across a variety of cancers. MTAP deletions (del) are found in about 30% of diffuse pleural mesotheliomas (DPM). MTAP del has been associated with resistance to immunotherapy (IO) treatment (tx) in multiple tumor types. In patients (pts) with DPM, objective response rate (ORR) on ipilimumab/nivolumab is 40%, disease control rate (DCR) 77%, and median progression-free survival (PFS) 6.8 months (mos): the implications of MTAP status on IO outcomes is unclear but represents a potential predictive biomarker. With multiple targeted therapies underway for MTAP del tumors, such as PRMT5 inhibitors, this alteration is also of therapeutic importance. Methods: We prospectively identified pts with pathologically confirmed DPM whose tumors were sequenced with MSK-IMPACT version 7, a 505-gene next generation sequencing panel that includes MTAP and CDKN2A. IO regimens included anti-PD(L)1 monotherapy, dual checkpoint blockade with additional anti-CTLA4 tx, and anti-PD(L)1 + chemotherapy. MTAP del was defined as low read count and confirmed by FACETS copy number when able. Radiologists reviewed imaging to determine best response and PFS on IO using mRECIST or, when not applicable, RECIST. Overall survival (OS) was compared between MTAP del and MTAP wildtype (WT) cohorts using Kaplan-Meier curves and log-rank tests. Baseline demographics were compared using Fisher’s exact test. Results: We examined 156 pts with DPM: 39 had CDKN2A del (25%) and 32 had MTAP del (21%). 18/32 were treated with IO and available for analysis. 79 pts with MTAP WT DPM treated with IO were analyzed as a control. There were more pts treated with dual checkpoint blockade in the MTAP del vs MTAP WT group (83% vs 56%, single-agent IO 6% vs 39%, single-agent IO + chemo 11% vs 5%, p=0.03) and more men (94% vs 70%, p=0.04); there was no statistically significant difference in age (median 72 vs 69, p=0.5), histology (72% epithelioid vs 79%, p=0.5), or smoking status (current/former 61% vs 57%, p=0.9). All tumors with MTAP del also harbored a CDKN2A del: 5/79 tumors in the MTAP WT cohort had a CDKN2A del (p<0.001). Among the MTAP del cohort, 13 patients had (m)RECIST-evaluable disease. ORR on IO was 15% (2/13), DCR 38% (5/13), and median PFS 2.5 mos. OS was similar between the MTAP del and MTAP WT cohorts: median 25.4 vs 27.6 mos (HR 0.84, 95% CI 0.36 – 1.94, p=0.98). Conclusions: MTAP del (co-occurring with CDKN2A del ) was identified in both epithelioid and non-epithelioid DPM and was associated with a low ORR and short PFS on IO, but OS was similar compared to MTAP WT. Larger, multi-institution cohorts are needed to validate this finding. If confirmed, this could have implications for tx selection, particularly among pts with epithelioid DPM, in which the optimal choice between 3 FDA-approved first-line regimens is uncertain.

Supplemental Figure 3. A. Median progression-free survival and B. Overall survival for patients enrolled in the initial therapy study

PURPOSE:Delta-like ligand 3 (DLL3) and seizure-related homolog protein 6 (SEZ6) are appealing drug targets in neuroendocrine carcinomas (NECs) given their preferential expression on the tumor cell surface compared with normal cells. We aimed to describe the landscape of these proteins across NECs from eight different primary sites. PATIENTS AND METHODS:We used immunohistochemistry to assess 124 NEC tumor samples from any primary site for DLL3 and 53 for SEZ6 and defined positivity as ≥1% staining. RESULTS:DLL3 and SEZ6 were commonly expressed in our cohort (97 of 124, 78% and 43 of 53, 81% positivity rates, respectively) and frequently coexpressed when both tested (35 of 53, 66%). NECs of the breast, prostate, and GI system had the highest rates of DLL3 positivity (2 of 2, 100%; 15 of 16, 94%; and 14 of 17, 82%, respectively); all primary sites except lung exhibited 100% positivity rates for SEZ6. DLL3 expression and SEZ6 expression were seen in transformed NECs (12 of 17, 71% and 3 of 4, 75%, respectively) and in brain metastases (5 of 7, 71% and 1 of 2, 50%, respectively). Expression of both proteins tended to remain stable among 10 patients with serial biopsies. DLL3 expression did not affect progression-free survival (PFS) on first-line platinum/etoposide with or without immunotherapy among patients with metastatic lung NEC (median PFS 5.3 v 5.7 months in DLL3+ v DLL3-, P = .9) but was associated with longer overall survival (median 12.5 v 2.6 months, P = .03). CONCLUSION:We describe the landscape of DLL3 and SEZ6 coexpression across NECs, establishing a broad-based cohort of patients who might derive benefit from therapeutics in development targeting these cell surface determinants.

Supplementary Data from Germline Pathogenic Variants Impact Clinicopathology of Advanced Lung Cancer

PURPOSE:Diagnostic delays are common for patients with NUT carcinoma (NC), a cancer driven by NUT fusion oncoprotein. Rapid diagnosis is crucial for the best outcomes. We investigated NC's presenting features and associations with diagnostic delays. METHODS:We manually reviewed medical records from US participants in the NC Registry (2007-2024). Baseline features were analyzed descriptively, and multivariable logistic regression was used to examine odds ratios for delays. RESULTS:We analyzed 132 patients (median age 37 years). At presentation, 55% had metastatic disease and 64% had a thoracic primary. The median interval from symptom onset to diagnosis was 10 weeks (range, 1-165 weeks). The initial histopathologic diagnosis was poorly differentiated/squamous cell cancer in 52%, insufficient malignant cells in 14%, carcinoma in 12%, and NC in 22%. Immunohistochemistry (IHC) testing revealed NUT in 100%; 93% were keratin+, 88% p63+, 89% p40+, and 72% had Ki-67 ≥50%. IHC was used to diagnose NC in 79%. Nonthoracic primaries were associated with a longer time between symptom onset and NC diagnosis (odds ratio, 3.25). Of 32 patients who started treatment before diagnosis, 78% were on agents appropriate for NC. After a diagnosis of NC, 33% participated in clinical trials. CONCLUSION:NC almost always presents as squamous/poorly differentiated non-small cell lung or head and neck cancer. Overall, 78% of patients were initially diagnosed as non-NC. Efforts to raise awareness, recognition, and rapid diagnosis of NC by both oncologists and pathologists are critical if we are to improve outcomes.