Study of pharmacogenomic information in FDA-approved drug labeling to facilitate application of precision medicine

Mehta, Darshan; Uber, Ryley; Ingle, Taylor; Li, Catherine; Liu, Zhichao; Thakkar, Shraddha; Ning, Baitang; Wu, Leihong; Yang, Junshuang; Harris, Steve; Zhou, Guangxu; Xu, Joshua; Lesko, Lawrence J.; Fang, Hong

doi:10.1016/j.drudis.2020.01.023

Cited by 34 publications

(22 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Greater than half of all biomarker-drug pairs identified in our study were associated with clinically actionable measures of PGx information. Consistent with previous studies, this progress has continued to be most prominent in cancer therapies, which comprise the majority of new PGx drug approvals and account for the greatest proportion of biomarker-drug pairs with testing requirements [17][18][19].…”

Section: Discussionsupporting

confidence: 76%

Pharmacogenomic Biomarkers in US FDA-Approved Drug Labels (2000–2020)

Kim

Ceccarelli

2021

JPM

View full text Add to dashboard Cite

Pharmacogenomics (PGx) is a key subset of precision medicine that relates genomic variation to individual response to pharmacotherapy. We assessed longitudinal trends in US FDA approval of new drugs labeled with PGx information. Drug labels containing PGx information were obtained from Drugs@FDA and guidelines from PharmGKB were used to compare the actionability of PGx information in drug labels across therapeutic areas. The annual proportion of new drug approvals with PGx labeling has increased by nearly threefold from 10.3% (n = 3) in 2000 to 28.2% (n = 11) in 2020. Inclusion of PGx information in drug labels has increased for all clinical areas over the last two decades but most prominently for cancer therapies, which comprise the largest proportion (75.5%) of biomarker–drug pairs for which PGx testing is required. Clinically actionable information was more frequently observed in biomarker–drug pairs associated with cancer drugs compared to those for other therapeutic areas (n = 92 (59.7%) vs. n = 62 (40.3%), p < 0.0051). These results suggest that further evidence is needed to support the clinical adoption of pharmacogenomics in non-cancer therapeutic areas.

show abstract

Section: Discussionsupporting

confidence: 76%

Pharmacogenomic Biomarkers in US FDA-Approved Drug Labels (2000–2020)

Kim

Ceccarelli

2021

JPM

View full text Add to dashboard Cite

show abstract

“…Specifically, we extracted a list of FDA approved drug labels by using a search query “human Rx” under labeling type in the FDALabel databases (version 2.5, https://nctr-crs.fda.gov/fdalabel/ui/search ) ( Mehta et al, 2020 ). Consequently, we obtained queried results with summary information of human prescription (Rx) drug labeling.…”

Section: Methodsmentioning

confidence: 99%

DICE: A Drug Indication Classification and Encyclopedia for AI-Based Indication Extraction

Bhatt

Roberts

Chen

et al. 2021

Front. Artif. Intell.

Self Cite

View full text Add to dashboard Cite

Drug labeling contains an ‘INDICATIONS AND USAGE’ that provides vital information to support clinical decision making and regulatory management. Effective extraction of drug indication information from free-text based resources could facilitate drug repositioning projects and help collect real-world evidence in support of secondary use of approved medicines. To enable AI-powered language models for the extraction of drug indication information, we used manual reading and curation to develop a Drug Indication Classification and Encyclopedia (DICE) based on FDA approved human prescription drug labeling. A DICE scheme with 7,231 sentences categorized into five classes (indications, contradictions, side effects, usage instructions, and clinical observations) was developed. To further elucidate the utility of the DICE, we developed nine different AI-based classifiers for the prediction of indications based on the developed DICE to comprehensively assess their performance. We found that the transformer-based language models yielded an average MCC of 0.887, outperforming the word embedding-based Bidirectional long short-term memory (BiLSTM) models (0.862) with a 2.82% improvement on the test set. The best classifiers were also used to extract drug indication information in DrugBank and achieved a high enrichment rate (>0.930) for this task. We found that domain-specific training could provide more explainable models without performance sacrifices and better generalization for external validation datasets. Altogether, the proposed DICE could be a standard resource for the development and evaluation of task-specific AI-powered, natural language processing (NLP) models.

show abstract

“…Drug labeling information has been used to address a broad range of scientific questions, such as finding up-to-date information on drug safety arising from pharmacovigilance 5,6 , providing context for drug repositioning 7 , and precision medicine 2,[8][9][10] . We are implementing a variety of functions to enhance such a research endeavor.…”

Section: Fdalabel For Drug Repurposing Studies and Beyondmentioning

confidence: 99%