Identifying symptom etiologies using syntactic patterns and large language models.

Differential diagnosis is a crucial aspect of medical practice, as it guides clinicians to accurate diagnoses and effective treatment plans. Traditional resources, such as medical books and services like UpToDate, are constrained by manual curation, potentially missing out on novel or less common findings. This paper introduces and analyzes two novel methods to mine etiologies from scientific literature. The first method employs a traditional Natural Language Processing (NLP) approach based on syntactic patterns. By using a novel application of human-guided pattern bootstrapping patterns are derived quickly, and symptom etiologies are extracted with significant coverage. The second method utilizes generative models, specifically GPT-4, coupled with a fact verification pipeline, marking a pioneering application of generative techniques in etiology extraction. Analyzing this second method shows that while it is highly precise, it offers lesser coverage compared to the syntactic approach. Importantly, combining both methodologies yields synergistic outcomes, enhancing the depth and reliability of etiology mining.

Mutant PIK3CA is a targetable driver alteration in histiocytic neoplasms.

Langerhans cell histiocytosis (LCH) is an inflammatory myeloid neoplasm characterized by the accumulation of clonal mononuclear phagocyte system cells expressing CD1a and CD207. In the past decade, molecular profiling of LCH as well as other histiocytic neoplasms demonstrated that these diseases are driven by MAPK activating alterations, with somatic BRAFV600E mutations in >50% of patients with LCH, and clinical inhibition of MAPK signaling has demonstrated remarkable clinical efficacy. At the same time, activating alterations in kinase-encoding genes, such as PIK3CA, ALK, RET, and CSF1R, which can activate mitogenic pathways independent from the MAPK pathway, have been reported in a subset of histiocytic neoplasms with anecdotal evidence of successful targeted treatment of histiocytoses harboring driver alterations in RET, ALK, and CSF1R. However, evidence supporting the biological consequences of expression of PIK3CA mutations in hematopoietic cells has been lacking, and whether targeted inhibition of PI3K is clinically efficacious in histiocytic neoplasms is unknown. Here, we provide evidence that activating mutations in PIK3CA can drive histiocytic neoplasms in vivo using a conditional knockin mouse expressing mutant PIK3CAH1047R in monocyte/dendritic cell progenitors. In parallel, we demonstrate successful treatment of PIK3CA-mutated, multisystemic LCH using alpelisib, an inhibitor of the alpha catalytic subunit of PI3K. Alpelisib demonstrated a tolerable safety profile at a dose of 750 mg per week and clinical and metabolic complete remission in a patient with PIK3CA-mutated LCH. These data demonstrate PIK3CA as a targetable noncanonical driver of LCH and underscore the importance of mutational analysis-based personalized treatment in histiocytic neoplasms.