Breast Cancer is a second deadly cancer disease and more than 2 million new cases are being reported annually in recent years. Early diagnosis and mass screening is validated as an effective tool for better disease management. The development of non-invasive affordable first-level screening techniques is required for efficient mass screening. The proposed system consists of a near infra-red optical transmitter connected with a signal modulator and receiver coupled with a demodulator mechanism. A modulated optical waveform provides deep tissue details by high tissue penetration. The constant signal waveform generator and signal oscilloscope were connected for wave generation and measurement. The optical waveform is digitally modulated by light modulation techniques at the transmitter side and demodulated at the receiver side. The normal spot and cancerous spot of the breast phantom model are placed in the path of the optical transmitter and the reflected signal is received by the receiver prior to demodulation. Phase Jitter, Time Jitter, and 3dB noise level were measured from the received signal for the normal as well as a cancerous spot of the breast phantom. The obtained data is given as input to a backpropagation based Artificial Intelligence system for training, validation, and classification. The results indicate that AI based classification of breast tumor from optically transilluminated data will provide substantial improvement of efficiency of the AI system. Also the proposed system can be utilized along with ultrasonography as a hybrid imaging modality.

Citation Format: P. Ponram. Artificial intelligence based detection of breast cancer from transilluminated optical data [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-083.