Traditional clinical diagnosis is based on the evaluation of analytical data by a medical staff. This approach may require bulky or unpractical analysis equipment, as well as trained personnel able to handle it and evaluate the data. Thus, it has a limited throughput that precludes the systematic screening of diseases. This could be bypassed thanks to novel photonic devices combining miniaturized optics and electronics, data post processing, and artificial intelligence. The potential of such devices is nicely illustrated by Thiago C. Cavalcanti and coauthors, who built an intelligent smartphone-based multimode imaging otoscope for the mobile diagnosis of otitis media. The optics of the otoscope, which include several UV and visible LEDs and light collection elements, are embedded in an earplug, the back of which is connected to the smartphone camera. This enables imaging in a practical way the spectral reflectance and fluorescence of the eardrum. The evaluation of the corresponding multispectral images provides information about the ear status: normal, adhesive otitis medium, or otitis medium with effusion. Machine learning algorithms are implemented to automatize the evaluation and enable a reliable diagnosis. The device has been tested on 69 patients. These findings open the way to the systematic screening of otitis, and thus to more efficiently preventing complications that may arise and lead to hearing loss.

---

---

 Add Comment

You must log in to add comments.