No abstract

ObjectivesTo adapt and evaluate a deep learning language model for answering why-questions based on patient-specific clinical text. Materials and MethodsBidirectional encoder representations from transformers (BERT) models were trained with varying data sources to perform SQuAD 2.0 style why-question answering (why-QA) on clinical notes. The evaluation focused on: 1) comparing the merits from different training data, 2) error analysis. ResultsThe best model achieved an accuracy of 0.707 (or 0.760 by partial match). Training toward customization for the clinical language helped increase 6% in accuracy. DiscussionThe error analysis suggested that the model did not really perform deep reasoning and that clinical why-QA might warrant more sophisticated solutions. ConclusionThe BERT model achieved moderate accuracy in clinical why-QA and should benefit from the rapidly evolving technology. Despite the identified limitations, it could serve as a competent proxy for question-driven clinical information extraction.

To our knowledge, natural history has not been reported for cardiac sarcoidosis (CS) diagnosed by pathologic evaluation of the apical core at left ventricular assist device (LVAD) implantation or cardiac transplantation. We retrospectively identified 232 consecutive patients meeting CS criteria. Of these patients, 54 were diagnosed by pathologic confirmation of CS, 10 after evaluation of the apical core (LVAD implant) or explanted heart (transplant). We compared clinical characteristics at initial evaluation and outcomes for these 10 patients with those of 10 patients with known CS before LVAD implant/transplant. In the study group, five patients (50%) had confirmed extracardiac sarcoidosis before LVAD implant/transplant; five had not been diagnosed with sarcoidosis. Mean (standard deviation) left ventricular ejection fraction at initial evaluation was 23% (16%), and left ventricular end-diastolic dimension was 61 (10) mm. Four patients died during follow-up; however, no survival difference was found for the 10 patients diagnosed incidentally and the group with a previous diagnosis or institutional LVAD/transplant cohorts. Patients diagnosed with CS on pathological examination of the apical core/explanted heart may have severe dilated cardiomyopathy as the initial presentation. Outcomes for patients with CS after advanced heart failure therapies may be comparable with those of non-CS patients.

No abstract