Author ORCID Identifier

0000-0002-1896-1446

Date of Graduation

8-2017

Document Type

Thesis (MS)

Program Affiliation

Medical Physics

Degree Name

Masters of Science (MS)

Advisor/Committee Chair

Steven H. Lin

Committee Member

Jingfei Ma

Committee Member

Ken-Pin Hwang

Committee Member

R. Jason Stafford

Committee Member

Brian P. Hobbs

Abstract

For patients diagnosed with stages IIa-IIb esophageal cancer, the current standard of care treatment is tri-modality therapy (TMT), where neoadjuvant chemoradiation (nCRT) is followed by surgical resection. Histopathology of resected tumors reveals that pathological complete response (pCR) is achieved in 20-30% of patients through nCRT alone. Because of the high mortality and morbidity associated with esophagectomy, it may be advantageous for patients exhibiting pCR from nCRT alone to be placed under observation rather than completing their TMT. Therefore, a method for predicting response at an early time-point during nCRT is highly desirable. Conventional methods such as endoscopic ultrasound, re-biopsy, and morphologic imaging are insufficient for this purpose. During nCRT, morphologic changes in tumors are often preceded by changes in the tumor biology. Diffusion Weighed Imaging (DWI) is an MRI modality which is sensitive to microscopic motion of water molecules in tissue. Quantitative DWI provides a measure of the cellular microenvironment which is impacted by cellularity, extra-cellular volume fraction, structure of the extracellular matrix, and cellular membranes. This work sought to investigate if changes in quantitative DWI may be used as an early imaging biomarker for the prediction of response to nCRT in esophageal cancer.

DWI scans were performed on a small group of esophageal cancer patients (stages IIa to IIIb) before, at interim, and after completion of their nCRT. Quantitative diffusion parameter maps were estimated for DWI scans using the following models of diffusion: mono-exponential, intra-voxel incoherent motion (IVIM), and kurtosis. Summary measures of quantitative diffusion parameters were extracted from tumor voxels through volumetric contouring. These summary measures were retrospectively compared between histopathologically confirmed groupings of patients as pCR and non-pCR. The study found that the relative change in mean ADC could completely separate groupings of pCR and non-pCR patients (AUC=1) at a cutoff of 27.7%. Measurement by volume contouring was shown to be highly reproducible between readers. This pilot study demonstrates the promise of using DWI for organ sparing approaches after nCRT in esophageal cancer.

Keywords

DWI, therapy response, ADC, diffusion, esophageal cancer, oesophageal cancer, IVIM, DKI, inter-reader reliability, image contour