Semantic Deep Learning: Prior Knowledge and a Type of Four-Term Embedding Analogy to Acquire Treatments for Well-Known Diseases.

Background: How to treat a disease remains to be the most common type of clinical question. Obtaining evidence-based answers from biomedical literature is difficult. Analogical reasoning with embeddings from deep learning (embedding analogies) may extract such biomedical facts, although the state-of-the-art focuses on pair-based proportional (pairwise) analogies such as man:woman::king:queen ("queen = -man +king +woman").

Rhegmatogenous Retinal Detachment and Solar Radiation in Northwestern Spain.

Objective: The aim of this study was to investigate the association between rhegmatogenous retinal detachment (RRD) and solar radiation in northwestern Spain.

Methods: All RRD cases in Pontevedra from 2008 and 2014 were retrospectively analyzed. Climatological data from 4 weather stations in the area were collected.

Spiradenocarcinoma of the eyelid: a case report.

A 84-year-old-woman presented a painless eyelid mass in her right eyelid. A biopsy was made and the anatomopathologic study showed a spiradenoma with malignant changes. The patient rejected any kind of treatment in spite of the prognosis of the lesion.

Primary orbital liposarcoma.

A 45-year-old-man presented a slightly painful proptosis and diplopia for 7 months. He had been kept elsewhere on oral steroids without evidence of any clinical response over an 8-week period with suspected diagnosis of an inflammatory pseudotumor upon referral to our clinic. An intraoperative biopsy positive for primary liposarcoma was followed by debulking surgery.