| Full text | |
| Author(s): |
Cosci, Alessandro
;
Takahama, Jr., Ademar
;
Correr, Wagner Rafael
;
Azevedo, Rebeca Souza
;
Fernandes da Costa Fontes, Karla Bianca
;
Kurachi, Cristina
Total Authors: 6
|
| Document type: | Journal article |
| Source: | JOURNAL OF MEDICAL IMAGING; v. 3, n. 4 OCT 2016. |
| Web of Science Citations: | 0 |
| Abstract | |
Actinic cheilitis (AC) is a disease caused by prolonged and cumulative sun exposure that mostly affects the lower lip, which can progress to a lip squamous cell carcinoma. Routine diagnosis relies on clinician experience and training. We investigated the diagnostic efficacy of wide-field fluorescence imaging coupled to an automated algorithm for AC recognition. Fluorescence images were acquired from 57 patients with confirmed AC and 46 normal volunteers. Three different algorithms were employed: two based on the emission characteristics of local heterogeneity, entropy and intensity range, and one based on the number of objects after K-mean clustering. A classification model was obtained using a fivefold cross correlation algorithm. Sensitivity and specificity rates were 86% and 89.1%, respectively. (C) 2016 Society of Photo-Optical Instrumentation Engineers (SPIE) (AU) | |
| FAPESP's process: | 98/14270-8 - Centro de ciencias opticas e fotonica (center for optical and photonics science). |
| Grantee: | Vanderlei Salvador Bagnato |
| Support Opportunities: | Research Grants - Research, Innovation and Dissemination Centers - RIDC |