Identification and Characterization of Immune Checkpoint Inhibitor-Induced Toxicities From Electronic Health Records Using Natural Language Processing.

PURPOSE: Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment, yet their use is associated with immune-related adverse events (irAEs). Estimating the prevalence and patient impact of these irAEs in the real-world data setting is critical for characterizing the benefit/risk profile of ICI therapies beyond the clinical trial population. Diagnosis codes, such as International Classification of Diseases codes, do not comprehensively illustrate a patient's care journey and offer no insight into drug-irAE causality. This study aims to capture the relationship between ICIs and irAEs more accurately by using augmented curation (AC), a natural language processing-based innovation, on unstructured data in electronic health records. METHODS: In a cohort of 9,290 patients treated with ICIs at Mayo Clinic from 2005 to 2021, we compared the prevalence of irAEs using diagnosis codes and AC models, which classify drug-irAE pairs in clinical notes with implied textual causality. Four illustrative irAEs with high patient impact-myocarditis, encephalitis, pneumonitis, and severe cutaneous adverse reactions, abbreviated as MEPS-were analyzed using corticosteroid administration and ICI discontinuation as proxies of severity. RESULTS: For MEPS, only 70% (n = 118) of patients found by AC were also identified by diagnosis codes. Using AC models, patients with MEPS received corticosteroids for their respective irAE 82% of the time and permanently discontinued the ICI because of the irAE 35.9% (n = 115) of the time. CONCLUSION: Overall, AC models enabled more accurate identification and assessment of patient impact of ICI-induced irAEs not found using diagnosis codes, demonstrating a novel and more efficient strategy to assess real-world clinical outcomes in patients treated with ICIs.

On the Origins of Omicron's Unique Spike Gene Insertion.

The emergence of a heavily mutated SARS-CoV-2 variant (Omicron; Pango lineage B.1.1.529 and BA sublineages) and its rapid spread to over 75 countries raised a global public health alarm. Characterizing the mutational profile of Omicron is necessary to interpret its clinical phenotypes which are shared with or distinctive from those of other SARS-CoV-2 variants. We compared the mutations of the initially circulating Omicron variant (now known as BA.1) with prior variants of concern (Alpha, Beta, Gamma, and Delta), variants of interest (Lambda, Mu, Eta, Iota, and Kappa), and ~1500 SARS-CoV-2 lineages constituting ~5.8 million SARS-CoV-2 genomes. Omicron's Spike protein harbors 26 amino acid mutations (23 substitutions, 2 deletions, and 1 insertion) that are distinct compared to other variants of concern. While the substitution and deletion mutations appeared in previous SARS-CoV-2 lineages, the insertion mutation (ins214EPE) was not previously observed in any other SARS-CoV-2 lineage. Here, we consider and discuss various mechanisms through which the nucleotide sequence encoding for ins214EPE could have been acquired, including local duplication, polymerase slippage, and template switching. Although we are not able to definitively determine the mechanism, we highlight the plausibility of template switching. Analysis of the homology of the inserted nucleotide sequence and flanking regions suggests that this template-switching event could have involved the genomes of SARS-CoV-2 variants (e.g., the B.1.1 strain), other human coronaviruses that infect the same host cells as SARS-CoV-2 (e.g., HCoV-OC43 or HCoV-229E), or a human transcript expressed in a host cell that was infected by the Omicron precursor.

Continuous genomic diversification of long polynucleotide fragments drives the emergence of new SARS-CoV-2 variants of concern.

Highly transmissible or immuno-evasive SARS-CoV-2 variants have intermittently emerged, resulting in repeated COVID-19 surges. With over 6 million SARS-CoV-2 genomes sequenced, there is unprecedented data to decipher the evolution of fitter SARS-CoV-2 variants. Much attention has been directed to studying the functional importance of specific mutations in the Spike protein, but there is limited knowledge of genomic signatures shared by dominant variants. Here, we introduce a method to quantify the genome-wide distinctiveness of polynucleotide fragments (3- to 240-mers) that constitute SARS-CoV-2 sequences. Compared to standard phylogenetic metrics and mutational load, the new metric provides improved separation between Variants of Concern (VOCs; Reference = 89, IQR: 65-108; Alpha = 166, IQR: 149-181; Beta 131, IQR: 114-149; Gamma = 164, IQR: 150-178; Delta = 235, IQR: 217-255; and Omicron = 459, IQR: 395-521). Omicron's high genomic distinctiveness may confer an advantage over prior VOCs and the recently emerged and highly mutated B.1.640.2 (IHU) lineage. Evaluation of 883 lineages highlights that genomic distinctiveness has increased over time (R 2 = 0.37) and that VOCs score significantly higher than contemporary non-VOC lineages, with Omicron among the most distinctive lineages observed. This study demonstrates the value of characterizing SARS-CoV-2 variants by genome-wide polynucleotide distinctiveness and emphasizes the need to go beyond a narrow set of mutations at known sites on the Spike protein. The consistently higher distinctiveness of each emerging VOC compared to prior VOCs suggests that monitoring of genomic distinctiveness would facilitate rapid assessment of viral fitness.

Knowledge synthesis of 100 million biomedical documents augments the deep expression profiling of coronavirus receptors.

The COVID-19 pandemic demands assimilation of all biomedical knowledge to decode mechanisms of pathogenesis. Despite the recent renaissance in neural networks, a platform for the real-time synthesis of the exponentially growing biomedical literature and deep omics insights is unavailable. Here, we present the nferX platform for dynamic inference from over 45 quadrillion possible conceptual associations from unstructured text, and triangulation with insights from single-cell RNA-sequencing, bulk RNA-seq and proteomics from diverse tissue types. A hypothesis-free profiling of ACE2 suggests tongue keratinocytes, olfactory epithelial cells, airway club cells and respiratory ciliated cells as potential reservoirs of the SARS-CoV-2 receptor. We find the gut as the putative hotspot of COVID-19, where a maturation correlated transcriptional signature is shared in small intestine enterocytes among coronavirus receptors (ACE2, DPP4, ANPEP). A holistic data science platform triangulating insights from structured and unstructured data holds potential for accelerating the generation of impactful biological insights and hypotheses.