Extracting relevant information for cancer diagnosis from dynamic full field OCT through image processing and learning

Clément Apelian; Clément Gastaud; A. Claude Boccara

doi:10.1117/12.2254824

17 February 2017 Extracting relevant information for cancer diagnosis from dynamic full field OCT through image processing and learning

Clément Apelian, Clément Gastaud, A. Claude Boccara

Author Affiliations +

Proceedings Volume 10053, Optical Coherence Tomography and Coherence Domain Optical Methods in Biomedicine XXI; 100531H (2017) https://doi.org/10.1117/12.2254824
Event: SPIE BiOS, 2017, San Francisco, California, United States

Abstract

For a large number of cancer surgeries, the lack of reliable intraoperative diagnosis leads to reoperations or bad outcomes for the patients. To deliver better diagnosis, we developed Dynamic Full Field OCT (D-FFOCT) as a complement to FFOCT. FFOCT already presents interesting results for cancer diagnosis e.g. Mohs surgery and reaching 96% accuracy on prostate cancer. D-FFOCT accesses the dynamic processes of metabolism and gives new tools to diagnose the state of a tissue at the cellular level to complement FFOCT contrast. We developed a processing framework that intends to maximize the information provided by the FFOCT technology as well as D-FFOCT and synthetize this as a meaningful image. We use different time processing to generate metrics (standard deviation of time signals, decorrelation times and more) and spatial processing to sort out structures and the corresponding imaging modality, which is the most appropriate. Sorting was achieved through quadratic discriminant analysis in a N-dimension parametric space corresponding to our metrics. Combining the best imaging modalities for each structure leads to a rich morphology image. This image displaying the morphology is then colored to represent the dynamic behavior of these structures (slow or fast) and to be quickly analyzed by doctors. Therefore, we achieved a micron resolved image, rich of both FFOCT ability of imaging fixed and highly backscattering structures as well as D-FFOCT ability of imaging low level scattering cellular level details. We believe that this morphological contrast close to histology and the dynamic behavior contrast will push forward the limits of intraoperative diagnosis further on.

Conference Presentation

Citation Download Citation

Clément Apelian, Clément Gastaud, and A. Claude Boccara "Extracting relevant information for cancer diagnosis from dynamic full field OCT through image processing and learning", Proc. SPIE 10053, Optical Coherence Tomography and Coherence Domain Optical Methods in Biomedicine XXI, 100531H (17 February 2017); https://doi.org/10.1117/12.2254824

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available