Date of Award

Summer 2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical & Aerospace Engineering

Committee Director

Julie Zhili Hao

Committee Member

Stacie I. Ringleb

Committee Member

Siqi Guo

Committee Member

Krishnanand Kaipa

Abstract

This dissertation presents a stepwise compression-relaxation (SCR) testing method built upon a two-dimensional (2D) tactile sensor for mechanical characterization of soft tissues and tumor detection. The core of the 2D sensor entails one whole polydimethylsiloxane (PDMS) microstructure embedded with a 3×3 sensing-plate/transducer array. A soft sample was compressed by the 2D sensor with a step incremental depth at a ramp speed, and then relaxed for certain hold time. When a soft sample was compressed by the 2D sensor, the sensing-plates translated the sample response at different tissue sites to the sensor deflections, which were registered as resistance changes by the transducer array.

Instant elasticity (Einstant) and loss factor (tan δ) extracted from the measured data were used to quantify the sample elasticity and viscoelasticity, respectively. First, a three-way ANOVA analysis was conducted on the data of soft materials (PDMS/silicone rubbers) to evaluate the influence of testing parameters (incremental depth, hold time, and ramp speed) on the measured results. The results revealed that both Einstant and tan δ were significantly dependent on testing parameters. Next, the measured results on the soft tissues showed different elasticity and viscoelasticity between muscle tissues and fat/skin tissues. The measured results on the tumor tissues indicated different elasticity and viscoelasticity among the five breast tumor (BT) tissues, and between the two pancreatic tumor (PT) tissues before and after treatment. Due to the larger sample size of the BT tissues, the elasticity distribution among the measure BT tissue sites was used to determine the location, shape and size of the tumor in a BT tissue.

The correlation of stress drop (Δσ) (obtained from the difference between the instant and relaxed sensor deflections at each step incremental depth) with the applied strain (ε) was used for tumor detection. Pearson correlation analysis was conducted to quantitatively analyze the measured Δσ-ε relation as slope of stress drop versus applied strain (m=Δσ/ε) and coefficient of determination (R2) as a measure of the goodness of fit of the linear regression for distinguishing tumor tissue from normal tissue. The measured results on soft materials showed that m was significantly dependent on testing parameters, but R2 showed no significant dependency on testing parameters. The measured results on the tumor tissues indicated R2 was significantly varied among the center, edge and outside sites of the BT tissues. However, no difference was found between the BT outside sites and the normal tissues. R2 also revealed significant difference between before and after treatment of the PT tissues, while no difference between the PT tissues after treatment and the normal tissues. R2 of the PT tissues before treatment was significantly different from that of the BT center sites, but m failed to capture their difference. Furthermore, dummy tumors made of silicone rubbers were found to behave differently from the native tumors.

In summary, the feasibility of the SCR testing method for tissue characterization and tumor detection was experimentally validated on the measured soft samples, including PDMS, silicone rubbers, porcine and bovine normal tissues, mouse BT and PT tissues. Future work will investigate the feasibility of the SCR testing method for differentiation between benign tumors and malignant tumors.

DOI

10.25777/j9av-4579

Share

COinS