2022 Theses Doctoral
Development and optimization of a clinical harmonic motion imaging system for breast tumor characterization and neoadjuvant chemotherapy response assessment
Breast cancer is the most common cancer in women, accounting for almost one-thirdof new cancer diagnoses in the United States. The mortality rate has decreased by 42% since 1989 due to early diagnosis, improvements in imaging techniques and treatment regimens. Despite all the advances in imaging modalities, there is still a need for a non-invasive, nonionizing, and low-cost diagnosis technique with high sensitivity and specificity to reduce the rate of invasive biopsies. For individuals diagnosed with locally advanced breast cancer and early-stage breast cancer, neoadjuvant chemotherapy (NACT) has become the standard of care. Pathologic complete response (pCR) is the ideal outcome of NACT, which is correlated with the prognosis and overall survival of the patients. The pCR is achieved in only about 15-20% of patients determined at the time of surgery; therefore, most patients receive a treatment that is not beneficial for them and has considerable side effects. Thus, early detection and monitoring of breast tumor response to NACT is critical for treatment planning and improving overall survival.
Ultrasound-based elasticity imaging techniques have gained interest in the clinic due to their potential to provide qualitative and/or quantitative information about tissue stiffness, which is presently not unachievable with standard ultrasonography. These techniques rely on the fact that a breast tumor’s stiffness or Young’s modulus is higher than that of the surrounding normal tissues. In this dissertation, the clinical feasibility of a technique called harmonic motion imaging (HMI) for breast tumor classification, as well as for NACT response prediction and monitoring of solid tumors is investigated. HMI is an ultrasound-based elasticity imaging technique that evaluates the mechanical properties of the underlying tissues by inducing amplitude modulated (AM) displacements at a specific frequency.
First, we investigated whether HMI can characterize and differentiate human breast tumors based on their relative stiffness. We enrolled female patients with benign and malignant tumors and imaged them with a clinical HMI system. The malignant tumors were found to be associated with lower HMI displacements or higher stiffness than the benign tumors. Then, in order to verify our clinical findings, we estimated HMI displacements in the postsurgical breast specimens from the same subjects and compared them against the in-vivo estimations. Our findings indicated that HMI successfully differentiated tumors from the surrounding tissue in both ex-vivo and in-vivo conditions, with an excellent correlation between the results in the two different settings.
Second, we introduced and characterized a new HMI setup consisted of a multi-element focused ultrasound transducer (FUS) with electronic beam steering capability. Therefore, instead of mechanical translation of the HMI setup, the acoustic force could be electronically steered in the volumetric space to accelerate the data acquisition. A pulse sequence was developed to drive the HMI transducers assembly, the FUS and imaging transducer, using a single ultrasound data acquisition system to have a compact setup that is more applicable for clinical settings. The data acquisition was further improved by investigating the effect of AM frequencies on the quality of the HMI images and tumor detection. We found that higher AM frequencies are needed in order to improve the detection and characterization of small and stiff inclusions. On the contrary, soft and large inclusions are better resolved at lower AM frequencies.
Lastly, we investigated the feasibility of using HMI for early prediction of response to neoadjuvant chemotherapy in cancer mouse models and breast cancer patients. We acquired longitudinal HMI images from pancreatic and breast cancer murine tumors during treatment with chemotherapeutic drugs and monitored the changes in the mechanical properties of the tumors. The tumors were found to soften when responsive to treatment, followed by the stiffness increase in the case of drug resistance. However, the untreated mice underwent steady stiffening of the tumors. Next, we imaged breast cancer patients at different timepoints during their chemotherapy treatment. We found that tumors in the patients who achieved pCR had higher pre-treatment stiffness and higher softening from pre-treatment to a short-interval follow-up on treatment compared to the ones in patients with residual cancer cells at the completion of treatment. These findings indicate the promising potential of HMI in the early prediction of solid tumor response to chemotherapy interventions.
This item is currently under embargo. It will be available starting 2024-10-28.
More About This Work
- Academic Units
- Biomedical Engineering
- Thesis Advisors
- Konofagou, Elisa E.
- Ph.D., Columbia University
- Published Here
- November 2, 2022