Tumor segmentation and size assessment are of great clinical value in cancer diagnosis and treatment monitoring. As an example, tumor size is used to monitor neoadjuvant chemotherapy (NAC) response. As the NAC changes the stiffness (i.e. elasticity) of the tumor, elasticity imaging can be used to monitor the NAC response. Single transducer harmonic motion imaging (ST-HMI) is an ultrasound elastography method to assess elasticity of tissue. In ST-HMI, a force from the ultrasound propagating wave is used to oscillate tissue at a particular frequency and another ultrasound beam is used to assess the oscillatory motion. Then, the elasticity of the tissue is inferred from the motion. Instead of collecting a single frequency oscillation at a time, ST-HMI is expanded to collected several frequencies of data in a single acquisition (i.e. multi-frequency ST-HMI). We have demonstrated that lesions with different sizes and stiffness can be detected by exploiting oscillation frequency without prior knowledge of the inclusion characteristics. However, the boundary delineation is performed manually. The objective of this work is to develop an artificial intelligence algorithm to delineate tumor boundaries in the multi-frequency ST-HMI images automatically. A large data set of multi-frequency ST-HMI images of phantom with and without the presence of noise and artifacts, in vivo, and ex vivo mouse tumors, and in vivo human breast tumors will be used to train and test the artificial intelligence algorithm. By learning these multi-frequency data with a variety of noise sources, the constructed network is expected to learn boundary effects and be able to separate tumor boundaries from the background.

This is an UNPAID research project.

Faculty Advisor

  • Professor: Elisa Konofagou
  • Center/Lab: Biomedical Engineering
  • Location: 630 West 168th Street, Physicians & Surgeons 19-418 · New York, NY 10032
  • Prof. Konofagou’s main interests are in the development of novel elasticity imaging techniques and applications, such as breast elastography, electromechanical wave imaging (EWI), myocardial elastography, harmonic motion imaging (HMI), pulse wave imaging (PWI), and focused ultrasound therapy, in particular research on the blood-brain barrier opening and neuromodulation.

Project Timeline

  • Earliest starting date: 10/15/2022
  • End date: 12/31/2022
  • Number of hours per week of research expected during Fall 2022: ~10

Candidate requirements

  • Skill sets: Python, Matlab, experience with ultrasound imaging
  • Student eligibility: freshman, sophomore, junior, senior, master’s
  • International students on F1 or J1 visa: eligible
  • Academic Credit Possible: Yes
  • Additional comments: N/A