UCLA

Radiotherapy is used to treat many cancers of the lung, liver, breast, and other anatomies. The goal of radiotherapy is to deliver high radiation doses to the target while sparing adjacent organs at risk. Accurate target definition is fundamental to ensuring a good outcome. Currently, only geometrical and anatomical information are acquired for purposes of target definition in radiotherapy treatment planning. Functional imaging techniques can provide additional biological information about tumors and the surrounding organs that is often complementary to the information provided by anatomic imaging. While useful, common functional imaging modalities are not generally employed within the radiotherapy workflow, which inhibits their integration into radiotherapy treatment planning.

Tissue biomechanical properties, such as elasticity, have been correlated to underlying tissue function. Towards this goal, a reliable elasticity estimation methodology, or elastography approach, was formulated for obtaining organ-specific tissue elasticity using images acquired within the current radiotherapy workflow. The feasibility and accuracy of this methodology was verified for breast, liver, and lung anatomies. Patient-specific elasticity distributions were estimated for lung and liver cancer radiotherapy patients, and functional correlation avoidance treatment plans were retrospectively developed. Brief descriptions of the major contributions are given below.

Using biomechanical, high-resolution breast geometry obtained from CT scans, a systematic assessment was performed to determine the feasibility of the elasticity estimation using clinically relevant displacement distributions. To simulate gravity induced breast deformation and mimic the chest-wall/breast tissue interface, the breast geometry was anchored at its base and allowed to deform between prone and supine patient positions. A forward iterative approach was used to solve the inverse elasticity problem. The feasibility and potential of the fast reconstruction of breast tissue elasticity using supine/prone patient postures was proven using these virtual phantom simulations.

The breast elastography methodology was extended to enable the performance of lung elastography using the internal breathing-induced deformation. For a systematic analysis, a physics-based virtual lung phantom with CT source geometry and a heterogeneous voxel-to-voxel elasticity distribution was employed. Our results showed that the elastography approach estimated the elasticity with 92% of voxels converging within 0.5 mm of the ground-truth displacement and 1 kPa of the ground-truth elasticity. This approach was extended to estimate and validate the elasticity distributions of 15 5DCT patient lung datasets and 11 patient 4DMR liver datasets. Overall, the results suggest that the lung elasticity can be measured with approximately 90% convergence using routinely-acquired clinical 4D scans, indicating the potential for a lung and liver elastography implementation within the radiotherapy clinical workflow.

Dose is a limiting factor for many imaging studies. As such, the feasibility of performing elastography using simulated lower dose CT scans was investigated. The results suggest that the original tube current-time product of 40 mAs scans can be reduced to 20 mAs without affecting the reliability of the elastography results. The relationship between elasticity and lung tissue function was investigated using 13 4DCT patients with known GOLD 2017 COPD status. Elasticity distributions proved to be an effective spatial indicator of lung pathophysiology and displayed the potential be translated into a binary indicator for diagnostic staging purposes. Finally, elastography results were used to retrospectively derive functional contours for 7 lung SBRT patients for treatment planning purposes. For consistency, both clinical and functional lung avoidance VMAT SBRT treatment plans were developed. The functional avoidance plans were able to successfully spare functional lung regions while maintaining target coverage and meeting all normal tissue constraints, suggesting that incorporating functional information into SBRT planning may facilitate improved patient outcome after radiotherapy.

The works presented here provide a way to use conventional imaging to perform elastography within the radiotherapy workflow. The focus of this work concentrated on the lung anatomy but have been preliminarily expanded to the breast and liver. In our future investigation, the tools provided here can be extended to other anatomical sites and can provide a foundation for performing elastography and obtaining functional information for treatment planning purposes within the clinical radiotherapy workflow.