Explore the vast range of titles available.

Murine double minute 2 (MDM2) inhibitors have shown promising activity in TP53 -wild type tumors and are under active investigation across a spectrum of malignancies. Herein, we report a 51-year-old female with MDM2 -amplified, TP53 -wild type adenoid cystic carcinoma who was treated with a MDM2 inhibitor and developed persistent pancytopenia despite drug discontinuation. Her pancytopenia was associated with 20 distinct pathogenic TP53 mutations in peripheral blood and bone marrow not present in drug-resistant tumor tissue. Plasma TP53 mutations were similarly detected among 4 other patients treated at our institution, with the number of mutations correlating strongly with duration of treatment. This case suggests that MDM2 inhibitors are associated with TP53 clonal hematopoiesis, which may confer a risk of subsequent myeloid malignancy. As multiple MDM2 inhibitor trials are ongoing, our findings underscore the need for further investigation into the potential long-term deleterious effects of these inhibitors in the hematopoietic stem and progenitor compartment.

Essential thrombocythemia (ET) is an indolent myeloproliferative neoplasm characterized by megakaryocyte hyperplasia, thrombocytosis, thrombotic and hemorrhagic complications, and potential transformation into myelofibrosis and acute myeloid leukemia. The vast majority of cases are driven by a somatic mutation in JAK2, CALR, or MPL. CALR, a gene that codes for the calcium-binding chaperone calreticulin, is the predominant mutation in patients with non-mutated JAK2 essential thrombocythemia, accounting for 20–25% of the overall somatic mutation frequency in ET. In this brief review of ET, we introduce a rare CALR mutation through a case presentation of a 58-year-old man with diffuse pulmonary emboli in the setting of thrombocytosis. We subsequently characterize the main types of CALR mutations and their value in diagnosis and prognosis of disease course, and lastly discuss the current clinical approach to ET.

Large language models (LLMs) are routinely used by physicians and patients for medical advice, yet their clinical safety profiles remain poorly characterized. We present NOHARM (Numerous Options Harm Assessment for Risk in Medicine), a benchmark using 100 real primary care-to-specialist consultation cases to measure frequency and severity of harm from LLM-generated medical recommendations. NOHARM covers 10 specialties, with 12,747 expert annotations for 4,249 clinical management options. Across 31 LLMs, potential for severe harm from LLM recommendations occurs in up to 22.2% (95% CI 21.6-22.8%) of cases, with harm of omission accounting for 76.6% (95% CI 76.4-76.8%) of errors. Safety performance is only moderately correlated (r = 0.61-0.64) with existing AI and medical knowledge benchmarks. The best models outperform generalist physicians on safety (mean difference 9.7%, 95% CI 7.0-12.5%), and a diverse multi-agent approach improves safety compared to solo models (mean difference 8.0%, 95% CI 4.0-12.1%). Therefore, despite strong performance on existing evaluations, widely used AI models can produce severely harmful medical advice at nontrivial rates, underscoring clinical safety as a distinct performance dimension necessitating explicit measurement.

Large language models (LLMs) and medical AI tools are routinely used by physicians and patients for medical advice, yet their clinical safety profiles remain poorly characterized. We present NOHARM (Numerous Options Harm Assessment for Risk in Medicine), a 1,100-task benchmark of primary care-to-specialist consultation cases to measure the frequency and severity of potentially harmful errors from LLM-generated medical consultation recommendations. NOHARM covers 10 specialties, with 12,747 expert annotations for 4,249 clinical management options. Across 20 notable LLMs and 4 widely used retrieval-augmented generation (RAG) clinical AI tools, direct application of recommendations carried potential for severe harm in up to 24.6% of cases, with errors of omission accounting for more than 80% of severe errors. Harm potential was not uniform across systems, with clinical AI tools outperforming generalist LLMs, and multi-agent AI teaming further improving performance in generalist models. In a randomized study of 101 U.S.-licensed generalist physicians, AI assistance improved physician performance compared to conventional resources. However, AI-assisted physicians frequently omitted valuable AI-generated recommendations and still scored lower than many AI systems alone. Had those recommendations been incorporated, combined human-AI responses would have outperformed both the human and AI system as used, suggesting complementary strengths and unrealized potential in human-AI teaming. Collectively, these results show that despite strong performance on medical knowledge benchmarks, widely used AI tools can produce medical consultation advice with the potential for severe harm, and highlight the need for explicit measurement of clinical safety. The benchmark and leaderboard are publicly available to support ongoing evaluation and improvement of AI systems used for clinical care.

Large language models (LLMs) are routinely used by physicians and patients for medical advice, yet their clinical safety profiles remain poorly characterized. We present NOHARM (Numerous Options Harm Assessment for Risk in Medicine), a benchmark using 100 real primary care-to-specialist consultation cases to measure frequency and severity of harm from LLM-generated medical recommendations. NOHARM covers 10 specialties, with 12,747 expert annotations for 4,249 clinical management options. Across 31 LLMs, potential for severe harm from LLM recommendations occurs in up to 22.2% (95% CI 21.6-22.8%) of cases, with harm of omission accounting for 76.6% (95% CI 76.4-76.8%) of errors. Safety performance is only moderately correlated (r = 0.61-0.64) with existing AI and medical knowledge benchmarks. The best models outperform generalist physicians on safety (mean difference 9.7%, 95% CI 7.0-12.5%), and a diverse multi-agent approach improves safety compared to solo models (mean difference 8.0%, 95% CI 4.0-12.1%). Therefore, despite strong performance on existing evaluations, widely used AI models can produce severely harmful medical advice at nontrivial rates, underscoring clinical safety as a distinct performance dimension necessitating explicit measurement.

Background Altered signaling pathways typify breast cancer and serve as direct inputs to steroid hormone receptor sensors . We previously reported that phospho-Ser134-GR (pS134-GR) species are elevated in triple-negative breast cancer (TNBC) and cooperate with hypoxia-inducible factors, providing a novel avenue for activation of GR in response to local or cellular stress. Methods We probed GR regulation by factors (cytokines, growth factors) that are rich within the tumor microenvironment (TME). TNBC cells harboring endogenous wild-type (wt) or S134A-GR species were created by CRISPR/Cas knock-in and subjected to transwell migration, invasion, soft-agar colony formation, and tumorsphere assays. RNA-seq was employed to identify pS134-GR target genes that are regulated both basally (intrinsic) or by TGFβ1 in the absence of exogenously added GR ligands. Regulation of selected basal and TGFβ1-induced pS134-GR target genes was validated by qRT-PCR and chromatin immunoprecipitation assays. Bioinformatics tools were used to probe public data sets for expression of pS134-GR 24-gene signatures. Results In the absence of GR ligands, GR is transcriptionally activated via p38-dependent phosphorylation of Ser134 as a mechanism of homeostatic stress-sensing and regulated upon exposure of TNBC cells to TME-derived agents. The ligand-independent pS134-GR transcriptome encompasses TGFβ1 and MAPK signaling gene sets associated with TNBC cell survival and migration/invasion. Accordingly, pS134-GR was essential for TNBC cell anchorage-independent growth in soft-agar, migration, invasion, and tumorsphere formation, an in vitro readout of cancer stemness properties. Both pS134-GR and expression of the MAPK-scaffolding molecule 14-3-3ζ were essential for a functionally intact p38 MAPK signaling pathway downstream of MAP3K5/ASK1, indicative of a feedforward signaling loop wherein self-perpetuated GR phosphorylation enables cancer cell autonomy. A 24-gene pS134-GR-dependent signature induced by TGFβ1 predicts shortened overall survival in breast cancer patients. Conclusions Phospho-S134-GR is a critical downstream effector of p38 MAPK signaling and TNBC migration/invasion, survival, and stemness properties. Our studies define a ligand-independent role for GR as a homeostatic “sensor” of intrinsic stimuli as well as extrinsic factors rich within the TME (TGFβ1) that enable potent activation of the p38 MAPK stress-sensing pathway and nominate pS134-GR as a therapeutic target in aggressive TNBC.

Although the extremely thermophilic archaea (Topt ≥ 70°C) may be the most primitive extant forms of life, they have been studied to a limited extent relative to mesophilic microorganisms. Many of these organisms have unique biochemical and physiological characteristics with important biotechnological implications. These include methanogens that generate methane, fermentative anaerobes that produce hydrogen gas with high efficiency, and acidophiles that can mobilize base, precious and strategic metals from mineral ores. Extremely thermophilic archaea have also been a valuable source of thermoactive, thermostable biocatalysts, but their use as cellular systems has been limited because of the general lack of facile genetics tools. This situation has changed recently, however, thereby providing an important avenue for understanding their metabolic and physiological details and also opening up opportunities for metabolic engineering efforts. Along these lines, extremely thermophilic archaea have recently been engineered to produce a variety of alcohols and industrial chemicals, in some cases incorporating CO2 into the final product. There are barriers and challenges to these organisms reaching their full potential as industrial microorganisms but, if these can be overcome, a new dimension for biotechnology will be forthcoming that strategically exploits biology at high temperatures.