Explore the vast range of titles available.

The accelerated adoption of digital pathology and advances in deep learning have enabled the development of robust models for various pathology tasks across a diverse array of diseases and patient cohorts. However, model training is often difficult due to label scarcity in the medical domain, and a model’s usage is limited by the specific task and disease for which it is trained. Additionally, most models in histopathology leverage only image data, a stark contrast to how humans teach each other and reason about histopathologic entities. We introduce CONtrastive learning from Captions for Histopathology (CONCH), a visual-language foundation model developed using diverse sources of histopathology images, biomedical text and, notably, over 1.17 million image–caption pairs through task-agnostic pretraining. Evaluated on a suite of 14 diverse benchmarks, CONCH can be transferred to a wide range of downstream tasks involving histopathology images and/or text, achieving state-of-the-art performance on histology image classification, segmentation, captioning, and text-to-image and image-to-text retrieval. CONCH represents a substantial leap over concurrent visual-language pretrained systems for histopathology, with the potential to directly facilitate a wide array of machine learning-based workflows requiring minimal or no further supervised fine-tuning. Developed using diverse sources of histopathology images, biomedical text and over 1.17 million image–caption pairs, evaluated on a suite of 14 diverse benchmarks, a visual-language foundation model achieves state-of-the-art performance on a wide array of clinically relevant pathology tasks.

Tyrosine kinase domains dynamically fluctuate between two main structural forms that are referred to as type I (DFG-in) or type II (DFG-out) conformations. Comprehensive data comparing type I and type II inhibitors are currently lacking for NTRK fusion-driven cancers. Here we used a type II NTRK inhibitor, altiratinib, as a model compound to investigate its inhibitory potential for larotrectinib (type I inhibitor)-resistant mutations in NTRK. Our study shows that a subset of larotrectinib-resistant NTRK1 mutations (V573M, F589L and G667C) retains sensitivity to altiratinib, while the NTRK1 V573M and xDFG motif NTRK1 G667C mutations are highly sensitive to type II inhibitors, including altiratinib, cabozantinib and foretinib. Moreover, molecular modeling suggests that the introduction of a sulfur moiety in the binding pocket, via methionine or cysteine substitutions, specifically renders the mutant kinase hypersensitive to type II inhibitors. Future precision treatment strategies may benefit from selective targeting of these kinase mutants based on our findings. Romel Somwar et al. find that cancer-causing NTRK gene fusions resistant to one form of inhibitor therapy can be resistant to other inhibitor types. Using molecular simulations, they show that some NTRK1 mutations resistant to the type I inhibitor larotrectinib are hypersensitive to the type II inhibitor altiratinib, potentially due to the introduction of a sulfur moiety in the kinase binding pocket.

The central nervous system (CNS) represents a site of sanctuary for many metastatic tumors when systemic therapies that control the primary tumor cannot effectively penetrate intracranial lesions. Non-small cell lung cancers (NSCLCs) are the most likely of all neoplasms to metastasize to the brain, with up to 60% of patients developing CNS metastases during the disease process. Targeted therapies such as tyrosine kinase inhibitors (TKIs) have helped reduce lung cancer mortality but vary considerably in their capacity to control CNS metastases. The ability of these therapies to effectively target lesions in the CNS depends on several of their pharmacokinetic properties, including blood–brain barrier permeability, affinity for efflux transporters, and binding affinity for both plasma and brain tissue. Despite the existence of numerous preclinical models with which to characterize these properties, many targeted therapies have not been rigorously tested for CNS penetration during the discovery process, whereas some made it through preclinical testing despite poor brain penetration kinetics. Several TKIs have now been engineered with the characteristics of CNS-penetrant drugs, with clinical trials proving these efforts fruitful. This Review outlines the extent and variability of preclinical evidence for the efficacy of NSCLC-targeted therapies, which have been approved by the US Food and Drug Administration (FDA) or are in development, for treating CNS metastases, and how these data correlate with clinical outcomes.

Desmoplastic small round cell tumor (DSRCT) is characterized by the t(11;22)(p13;q12) translocation, which fuses the transcriptional regulatory domain of EWSR1 with the DNA-binding domain of WT1, resulting in the oncogenic EWSR1-WT1 fusion protein. The paucity of DSRCT disease models has hampered preclinical therapeutic studies on this aggressive cancer. Here, we developed preclinical disease models and mined DSRCT expression profiles to identify genetic vulnerabilities that could be leveraged for new therapies. We describe four DSRCT cell lines and one patient-derived xenograft model. Transcriptomic, proteomic and biochemical profiling showed evidence of activation of the ERBB pathway. Ectopic expression of EWSR1-WT1 resulted in upregulation of ERRB family ligands. Treatment of DSRCT cell lines with ERBB ligands resulted in activation of EGFR, ERBB2, ERK1/2 and AKT, and stimulation of cell growth. Antagonizing EGFR function with shRNAs, small-molecule inhibitors (afatinib, neratinib) or an anti-EGFR antibody (cetuximab) inhibited proliferation of DSRCT cells. Finally, treatment of mice bearing DSRCT xenografts with a combination of cetuximab and afatinib significantly reduced tumor growth. These data provide a rationale for evaluating EGFR antagonists in patients with DSRCT. This article has an associated First Person interview with the first two authors of the paper.

Multi-kinase RET inhibitors, such as cabozantinib and RXDX-105, are active in lung cancer patients with RET fusions; however, the overall response rates to these two drugs are unsatisfactory compared to other targeted therapy paradigms. Moreover, these inhibitors may have different efficacies against RET rearrangements depending on the upstream fusion partner. A comprehensive preclinical analysis of the efficacy of RET inhibitors is lacking due to a paucity of disease models harboring RET rearrangements. Here, we generated two new patient-derived xenograft (PDX) models, one new patient-derived cell line, one PDX-derived cell line, and several isogenic cell lines with RET fusions. Using these models, we re-examined the efficacy and mechanism of action of cabozantinib and found that this RET inhibitor was effective at blocking growth of cell lines, activating caspase 3/7 and inhibiting activation of ERK and AKT. Cabozantinib treatment of mice bearing RET fusion-positive cell line xenografts and two PDXs significantly reduced tumor proliferation without adverse toxicity. Moreover, cabozantinib was effective at reducing growth of a lung cancer PDX that was not responsive to RXDX-105. Transcriptomic analysis of lung tumors and cell lines with RET alterations showed activation of a MYC signature and this was suppressed by treatment of cell lines with cabozantinib. MYC protein levels were rapidly depleted following cabozantinib treatment. Taken together, our results demonstrate that cabozantinib is an effective agent in preclinical models harboring RET rearrangements with three different 5′ fusion partners (CCDC6, KIF5B and TRIM33). Notably, we identify MYC as a protein that is upregulated by RET expression and downregulated by treatment with cabozantinib, opening up potentially new therapeutic avenues for the combinatorial targetin of RET fusion- driven lung cancers. The novel RET fusion-dependent preclinical models described here represent valuable tools for further refinement of current therapies and the evaluation of novel therapeutic strategies.

First Person is a series of interviews with the first authors of a selection of papers published in Disease Models & Mechanisms, helping early-career researchers promote themselves alongside their papers. Roger Smith and Igor Odintsov are co-first authors on ‘ Novel patient-derived models of desmoplastic small round cell tumor confirm a targetable dependency on ERBB signaling’, published in DMM. Roger completed the research described in this article while a research technician in the lab of Marc Ladanyi at Memorial Sloan Kettering Cancer Center, New York, NY, USA. He is now an MD/PhD student in the lab of Marc Mendillo at Northwestern University, Chicago, IL, USA, investigating stress response pathways in cancer. Igor completed the research described in this article while a postdoctoral fellow in the lab of Marc Ladanyi at Memorial Sloan Kettering Cancer Center. He is now a resident anatomic pathologist at Brigham and Women's Hospital, Boston, MA, USA.

We sought to develop a pharmacologically advanced inhibitor of the oncoprotein RET. Vepafestinib potently inhibited on-target mutants that confer resistance to approved RET inhibitors while exerting superior efficacy against intracranial disease due to enhanced penetration and retention in the brain.

The molecular pathogenesis of thymoma remains largely unknown. It has been recently demonstrated, that activation of Wnt signaling pathway leads to increased incidence of thymoma in murine models. The present study investigated the activation of molecules of the Wnt signaling pathway in human thymoma. A total of 112 thymoma cases with complete clinical and follow-up data and 8 controls were included in the present study. Patients with thymoma and controls were examined immunohistochemically for β-catenin and E-cadherin. The mRNA expression levels of CTNNB1, CCND1, MYC, AXIN2 and CDH1 were analyzed by reverse transcription-quantitative PCR. Immunohistochemically, β-catenin and E-cadherin were overexpressed in neoplastic cells of all thymomas. In type A, B1 and non-invasive type B2 thymoma, both molecules were located in the cytoplasm, in contrast to invasive type B2 and B3 thymoma, where membranous immunopositivities were observed. mRNA expression levels of genes involved in the Wnt pathway and of E-cadherin were significantly increased in both type A and B thymoma compared with controls; increasing gradually from type B1 to B3, and with higher stage of disease. In recurrent type B thymoma, the mRNA expression of the molecules was significantly higher. Despite the activation of Wnt pathway in indolent type A thymoma, the negative feedback of the pathway was preserved by overexpression of inhibitory molecule axin2, which was not overexpressed in type B thymoma. In summary, the Wnt pathway was activated in human thymoma and may contribute to oncogenesis. Detection of molecules of the Wnt pathway may be of diagnostic and prognostic value.

RET receptor tyrosine kinase is activated in various cancers (lung, thyroid, colon and pancreatic, among others) through oncogenic fusions or gain-of-function single-nucleotide variants. Small-molecule RET kinase inhibitors became standard-of-care therapy for advanced malignancies driven by RET. The therapeutic benefit of RET inhibitors is limited, however, by acquired mutations in the drug target as well as brain metastasis, presumably due to inadequate brain penetration. Here, we perform preclinical characterization of vepafestinib (TAS0953/HM06), a next-generation RET inhibitor with a unique binding mode. We demonstrate that vepafestinib has best-in-class selectivity against RET, while exerting activity against commonly reported on-target resistance mutations (variants in RET L730 , RET V804 and RET G810 ), and shows superior pharmacokinetic properties in the brain when compared to currently approved RET drugs. We further show that these properties translate into improved tumor control in an intracranial model of RET-driven cancer. Our results underscore the clinical potential of vepafestinib in treating RET-driven cancers.

Disclosure: Y. Kang: None. D.A. Rubinson: None. I. Odintsov: None. A. Vaidya: None. J.S. Williams: None. G.K. Adler: None. M.E. McDonnell: None. M. Patti: Advisory Board Member; Self; Fractyl Laboratories. Consulting Fee; Self; AstraZeneca, Hanmi Pharmaceuticals, MBX. Grant Recipient; Self; Dexcom. N.E. Palermo: None. Background: Nonfunctioning malignant pancreatic neuroendocrine tumors (PNETs) can transform to functional PNETs. Transformation into insulin-secreting PNET (ISPNET) is associated with poor overall survival and refractory hypoglycemia, often requiring ICU care. We present a case of refractory hypoglycemia due to ISPNET, which was successfully treated with alpelisib, an inhibitor of phosphatidylinositol 3-kinase (PI3K) currently approved for breast cancer. Clinical Case: A 79-year-old man with a 5-year history of type 2 diabetes and 10-year history of grade II, well-differentiated, non-functioning PNET (Ki67 15%) with hepatic metastasis presented with worsening symptomatic hypoglycemia. He discontinued his antidiabetic agent 3 months ago due to hypoglycemia and was on lanreotide and everolimus as PNET treatment. Evaluation revealed plasma glucose 29 mg/dL, proinsulin >700 pmol/L (reference: 3.6-22), insulin 116 uIU/mL (2.6-24.9), C-peptide 12.0 ng/mL (1.1-4.4), beta-hydroxybutyrate 0.1 mmol/L (<0.3), and negative sulfonylurea panel. Imaging showed extensive progression of hepatic metastases, and debulking hepatectomy was performed. Pathology showed grade III PNET with positive insulin staining and mib-1 index of 22%. After 2 months of euglycemia post-surgery, he was readmitted to the ICU with refractory hypoglycemia. Evaluation showed plasma glucose 91 mg/dL, proinsulin, and insulin levels above-assay (>700 pmol/L and >930.0 uIU/mL, respectively) and C-peptide 71.8 ng/mL; imaging revealed hepatic tumor progression. Severe hypoglycemia as low as 34 mg/dL was noted despite therapy with IV dextrose (30%, up to 100 cc/hr), glucagon infusion (0.6 mg/hr), dexamethasone 20 mg QD, and diazoxide 30 mg TID. Debilitating anasarca developed despite aggressive diuresis, attributed to high-volume IV dextrose, diazoxide, and renal dysfunction. Two hours after 300 mg alpelisib administration, patient exhibited hyperglycemia, allowing discontinuation of dextrose infusion 8 hours later. Glucagon infusion was subsequently discontinued, and he was discharged to home. Within 20 days, the patient was able to discontinue all diuretics, dexamethasone, and cornstarch. Patient had no side effects related to alpelisib and performance status returned to baseline. Glucose ranged between 160-300 mg/dL without any hypoglycemia. On week 4 of alpelisib, plasma glucose was 340 mg/dL, proinsulin >700 pmol/L, insulin 181 uIU/mL, and C-peptide 20.2 ng/mL. Conclusion: To our knowledge, this is the first case demonstrating alpelisib as a safe and effective option in refractory hypoglycemia associated with insulinoma. While it is unclear whether this is mediated by anti-tumor effects or solely by blockade of insulin action, PI3-K inhibition should be considered as a treatment option in patients with ISPNET in whom hypoglycemia has failed to improve with conventional approaches. Presentation: Friday, June 16, 2023