Identifying Patient Populations in Texts Describing Drug Approvals Through Deep Learning-Based Information Extraction: Development of a Natural Language Processing Algorithm.

BACKGROUND: New drug treatments are regularly approved, and it is challenging to remain up-to-date in this rapidly changing environment. Fast and accurate visualization is important to allow a global understanding of the drug market. Automation of this information extraction provides a helpful starting point for the subject matter expert, helps to mitigate human errors, and saves time. OBJECTIVE: We aimed to semiautomate disease population extraction from the free text of oncology drug approval descriptions from the BioMedTracker database for 6 selected drug targets. More specifically, we intended to extract (1) line of therapy, (2) stage of cancer of the patient population described in the approval, and (3) the clinical trials that provide evidence for the approval. We aimed to use these results in downstream applications, aiding the searchability of relevant content against related drug project sources. METHODS: We fine-tuned a state-of-the-art deep learning model, Bidirectional Encoder Representations from Transformers, for each of the 3 desired outputs. We independently applied rule-based text mining approaches. We compared the performances of deep learning and rule-based approaches and selected the best method, which was then applied to new entries. The results were manually curated by a subject matter expert and then used to train new models. RESULTS: The training data set is currently small (433 entries) and will enlarge over time when new approval descriptions become available or if a choice is made to take another drug target into account. The deep learning models achieved 61% and 56% 5-fold cross-validated accuracies for line of therapy and stage of cancer, respectively, which were treated as classification tasks. Trial identification is treated as a named entity recognition task, and the 5-fold cross-validated F1-score is currently 87%. Although the scores of the classification tasks could seem low, the models comprise 5 classes each, and such scores are a marked improvement when compared to random classification. Moreover, we expect improved performance as the input data set grows, since deep learning models need to be trained on a large enough amount of data to be able to learn the task they are taught. The rule-based approach achieved 60% and 74% 5-fold cross-validated accuracies for line of therapy and stage of cancer, respectively. No attempt was made to define a rule-based approach for trial identification. CONCLUSIONS: We developed a natural language processing algorithm that is currently assisting subject matter experts in disease population extraction, which supports health authority approvals. This algorithm achieves semiautomation, enabling subject matter experts to leverage the results for deeper analysis and to accelerate information retrieval in a crowded clinical environment such as oncology.

Global mineralogical and aqueous mars history derived from OMEGA/Mars Express data.

Global mineralogical mapping of Mars by the Observatoire pour la Mineralogie, l'Eau, les Glaces et l'Activité (OMEGA) instrument on the European Space Agency's Mars Express spacecraft provides new information on Mars' geological and climatic history. Phyllosilicates formed by aqueous alteration very early in the planet's history (the "phyllocian" era) are found in the oldest terrains; sulfates were formed in a second era (the "theiikian" era) in an acidic environment. Beginning about 3.5 billion years ago, the last era (the "siderikian") is dominated by the formation of anhydrous ferric oxides in a slow superficial weathering, without liquid water playing a major role across the planet.

Sulfates in Martian layered terrains: the OMEGA/Mars Express view.

The OMEGA/Mars Express hyperspectral imager identified hydrated sulfates on light-toned layered terrains on Mars. Outcrops in Valles Marineris, Margaritifer Sinus, and Terra Meridiani show evidence for kieserite, gypsum, and polyhydrated sulfates. This identification has its basis in vibrational absorptions between 1.3 and 2.5 micrometers. These minerals constitute direct records of the past aqueous activity on Mars.

Mars surface diversity as revealed by the OMEGA/Mars Express observations.

The Observatoire pour la Minéralogie, l'Eau, les Glaces, et l'Activité (OMEGA) investigation, on board the European Space Agency Mars Express mission, is mapping the surface composition of Mars at a 0.3- to 5-kilometer resolution by means of visible-near-infrared hyperspectral reflectance imagery. The data acquired during the first 9 months of the mission already reveal a diverse and complex surface mineralogy, offering key insights into the evolution of Mars. OMEGA has identified and mapped mafic iron-bearing silicates of both the northern and southern crust, localized concentrations of hydrated phyllosilicates and sulfates but no carbonates, and ices and frosts with a water-ice composition of the north polar perennial cap, as for the south cap, covered by a thin carbon dioxide-ice veneer.

Perennial water ice identified in the south polar cap of Mars.

The inventory of water and carbon dioxide reservoirs on Mars are important clues for understanding the geological, climatic and potentially exobiological evolution of the planet. From the early mapping observation of the permanent ice caps on the martian poles, the northern cap was believed to be mainly composed of water ice, whereas the southern cap was thought to be constituted of carbon dioxide ice. However, recent missions (NASA missions Mars Global Surveyor and Odyssey) have revealed surface structures, altimetry profiles, underlying buried hydrogen, and temperatures of the south polar regions that are thermodynamically consistent with a mixture of surface water ice and carbon dioxide. Here we present the first direct identification and mapping of both carbon dioxide and water ice in the martian high southern latitudes, at a resolution of 2 km, during the local summer, when the extent of the polar ice is at its minimum. We observe that this south polar cap contains perennial water ice in extended areas: as a small admixture to carbon dioxide in the bright regions; associated with dust, without carbon dioxide, at the edges of this bright cap; and, unexpectedly, in large areas tens of kilometres away from the bright cap.