DLI-IT: a deep learning approach to drug label identification through image and text embedding.

BACKGROUND: Drug label, or packaging insert play a significant role in all the operations from production through drug distribution channels to the end consumer. Image of the label also called Display Panel or label could be used to identify illegal, illicit, unapproved and potentially dangerous drugs. Due to the time-consuming process and high labor cost of investigation, an artificial intelligence-based deep learning model is necessary for fast and accurate identification of the drugs. METHODS: In addition to image-based identification technology, we take advantages of rich text information on the pharmaceutical package insert of drug label images. In this study, we developed the Drug Label Identification through Image and Text embedding model (DLI-IT) to model text-based patterns of historical data for detection of suspicious drugs. In DLI-IT, we first trained a Connectionist Text Proposal Network (CTPN) to crop the raw image into sub-images based on the text. The texts from the cropped sub-images are recognized independently through the Tesseract OCR Engine and combined as one document for each raw image. Finally, we applied universal sentence embedding to transform these documents into vectors and find the most similar reference images to the test image through the cosine similarity. RESULTS: We trained the DLI-IT model on 1749 opioid and 2365 non-opioid drug label images. The model was then tested on 300 external opioid drug label images, the result demonstrated our model achieves up-to 88% of the precision in drug label identification, which outperforms previous image-based or text-based identification method by up-to 35% improvement. CONCLUSION: To conclude, by combining Image and Text embedding analysis under deep learning framework, our DLI-IT approach achieved a competitive performance in advancing drug label identification.

Therapeutic monoclonal antibodies and clinical laboratory tests: When, why, and what is expected?

BACKGROUND: We herein provide an overview of the clinical laboratory tests that should be performed before, during and after using therapeutic monoclonal antibodies (mAbs) and the clinical laboratory tests that may be affected by mAbs. METHODS: The labels of FDA-approved therapeutic mAbs were downloaded from DailyMed (the official website for drug labels) and were used as the sources of data for this review. RESULTS: It was found that most of the labels provided information relevant to the clinical laboratory tests, including the tests needed before mAbs treatment to check the patients' background status and to identify potentially sensitive patients, the tests needed during or after the treatment to evaluate the patients' response, and the mAbs that may lead to false positive or negative results for clinical laboratory tests. CONCLUSIONS: The present findings will be of interest to physicians, laboratory scientists, those involved in drug development and surveillance and individuals making health care policy.

FDA-approved drugs that interfere with laboratory tests: A systematic search of US drug labels.

Drug-related laboratory test interference or drug/laboratory test interactions (DLTI) are a major source of laboratory errors. DLTI is of concern with regard to both the clinical diagnosis and the monitoring of patients. Although there have been numerous reports about specific drugs that interfere with laboratory tests, there has not been a recent review on the topic. We herein provide a review of the known DLTI of US FDA-approved drugs based on a systematic search of DailyMed, a website containing the labels of US FDA-approved drugs. The labels for all human single-ingredient prescription drugs included in the database (1368) were searched using stemmed keywords and were manually reviewed for their relevance to DLTI. A total of 134 labels were positive, which indicated that the drug interferes with at least one clinical laboratory test. Antibacterial agents, psychotropic drugs and contrast media are the classes of drugs most likely to lead to DLTI. Urine was the clinical sample most frequently affected by DLTI. The FDA drug label is a source of information for studies of DLTI, although information is still lacking for most drugs, and additional improvements are needed for many of the existing records. Medical professionals, clinicians and laboratory staff should keep these possible interactions in mind when interpreting the results of laboratory tests, and should ensure that they obtain a complete and accurate record of all drugs being used by patients in order to anticipate potential DLTI. The development of a reporting system to address potential DLTI is warranted.

Inconsistent excipient listings in DailyMed: implications for drug safety.

Excipients, or inactive ingredients, are a frequent cause of medication intolerance and allergy. Patients and clinicians concerned about medication allergies and sensitivities rely on the U.S. National Library of Medicine's DailyMed for accurate lists of excipients. Based on our anecdotal discovery of several examples of excipient omissions, we wished to examine the accuracy of DailyMed's listings more systematically in a sample of commonly prescribed medications. The objective of the study is to identify the frequency of inconsistency of excipient reporting within the DailyMed website. We performed a database audit of the Structured Product Labeling XML file provided by the drug manufacturer to the Food and Drug Administration and DailyMed for two randomly selected formulations of each of 50 commonly prescribed medications. For each of the 100 formulations, we compared the excipients listed in the "Description" to those in the "Ingredients and Appearance" sections in DailyMed. The Structured Product Labeling data file provided by the drug manufacturer contained internal inconsistencies of excipients in 39% of the formulations examined. Despite the use of Structured Product Labeling, the drug manufacturer's medication labels provided to the FDA and reported by DailyMed often contain conflicting information about inactive ingredients. Patients with allergies and excipient sensitivity should be aware of these discrepancies and consult multiple sections of the label to identify potential allergy-inducing inactive ingredients.

The need for contraception in patients taking prescription drugs: a review of FDA warning labels, duration of effects, and mechanisms of action.

Introduction: This review provides a guide for the rational use of prescription drugs in patients of reproductive age. Areas covered: A comprehensive retrieval of the labels of FDA-approved drugs was performed to identify drugs where the label recommends contraceptive use during and/or after treatment. The acquired data were analyzed and organized into a table. Contraception was recommended or mandated for 268 single-ingredient drugs. These could be divided into four main categories, with many having effects across several categories: 177 drugs required contraception because they were associated with pregnancy loss or stillbirth, 177 drugs were associated with teratogenesis, 136 were associated with non-teratogenic adverse peri- or postnatal effects on the fetus (e.g. low birth weight), and 44 were associated with decreased efficacy of contraception or a change in ovulatory cycle. We also discuss the period of time contraception is required, as well as the known or hypothesized reasons for the reproductive toxicity of these agents. Expert opinion: We have provided a comprehensive overview of the FDA-approved drugs where the warning labels currently stipulate that contraception should be used. Although other references are available for clinicians, this review provides a useful source of information regarding the single-ingredient prescription drugs that may affect the outcome of pregnancy. This information is particularly relevant for researchers, as it provides an overview of the different drugs with reproductive toxicity, and because it highlights the specific needs for future research. In particular, more work (especially epidemiological studies) is needed to clarify the clinical relevance of these findings, most of which were obtained through animal studies.

FDA-approved medications that impair human spermatogenesis.

We herein provide an overview of the single-ingredient U.S. Food and Drug Administration (FDA)-approved drugs that affect human spermatogenesis, potentially resulting in a negative impact on male fertility. To provide this information, we performed an in-depth search of DailyMed, the official website for FDA-approved drug labels. Not surprisingly, hormone-based agents were found to be the drugs most likely to affect human spermatogenesis. The next category of drugs most likely to have effects on spermatogenesis was the antineoplastic agents. Interestingly, the DailyMed labels indicated that several anti-inflammatory drugs affect spermatogenesis, which is not supported by the peer-reviewed literature. Overall, there were a total of 65 labels for drugs of various classes that showed that they have the potential to affect human sperm production and maturation. We identified several drugs indicated to be spermatotoxic in the drug labels that were not reported in the peer-reviewed literature. However, the details about the effects of these drugs on human spermatogenesis are largely lacking, the mechanisms are often unknown, and the clinical impact of many of the findings is currently unclear. Therefore, additional work is needed at both the basic research level and during clinical trials and post-marketing surveillance to fill the gaps in the current knowledge. The present findings will be of interest to physicians and pharmacists, researchers, and those involved in drug development and health care policy.