Skip to main content
. Author manuscript; available in PMC: 2019 May 15.
Published in final edited form as: ACS Nano. 2018 Aug 20;12(9):9081–9090. doi: 10.1021/acsnano.8b03029

Fig 1. Artificial Intelligence Diffraction Analysis (AIDA) architecture for breast cancer diagnoses.

Fig 1.

(a) Schematic of the work-flow. (i) Cellular samples obtained from breast cancer patients via fine-needle aspiration are immunostained for triple markers: ER/PR in red and HER2 in blue. (ii) Diffraction patterns of stained and unstained cells are imaged by the AIDA device. (iii) Deep learning algorithms based on convolutional neural networks are applied to identify cancer cells and extract their color information directly from raw images. The ER/PR and HER2 expression levels are analyzed to classify breast cancer subtypes. (b) Disposable sample cartridge for cell capture and staining. A silicone gasket is placed between a glass slide and a plastic top for a watertight sealing. Cells are captured on the glass slide. (c) AIDA imaging system. The device is equipped with a high-resolution CMOS imager and a dual-LED (λ = 470 nm, 625 nm) light source. The dichromatic mirror (DIC, 550 nm cutoff) centers light inputs to an aperture with normal incidence angle. The system body was fabricated in a photopolymer resin via 3d-printing.