Radonc/BME Medical Image and
Computational Analysis Laboratory
Lung tissue radiation dose response


Former students:
      Matt Wilhelm

In the lung, radiation side-effects on the surrounding normal tissue manifests as edema and infiltration of inflammatory cells, ultimately leading to fibrosis, but the natural history of long-term fibrosis and repair are not well understood. Our goal is to apply quantitative analysis of follow-up CT chest scans of patients after targeted radiation treatment to correlate directly the progression of fibrosis with the amount of radiation dose delivered during treatment. Our radiobiology colleagues are actively developing new agents to protect normal tissues, and accelerate recovery, in patients or victims of a radiological accident, such as occurred recently in Japan. Our ability to quantify tissue damage over time allows us to validate and compare the effectiveness of such agents, both in animal models and in human subjects.

A complication that we have recently overcome is that pulmonary blood vessels surrounding the regions of fibrosis are not well distinguished from the fibrotic areas themselves, leading to erroneous inclusion of such vessels into our radiation dose-response analysis. We worte software to segment and traverse the pulmonary vascular tree. By computing the vessel radius as we traverse each branch we are now able identify branches that extend into regions of fibrosis and subtract them from the analysis.

We are using retrospective and prospective data of patients treated at UF using both proton therapy and conventional X-ray-based therapy. A prospective study of lung radiation damage in breast cancer patients is on-going, funded through the Florida Department of Health . It is aimed at better modeling the radiation dose-response and compare the severity of toxicity between proton and X-ray therapies. A retrospective study of lung toxicity is being started using existing data sets of patients treated at the UF Proton Therapy Institute.

Figure 1
[A] CT image slice through the chest of a patient with a ~3cm lung lesion, and where is superposed the conformal radiation dose field used for his/her treatment.
[B] CT iamge slice at approximately the same location but taken 8 months after treatment, showing development of radiation fibrosis.

Related Publications
  1. O’Dell W, Gormaley A, Prida D.,
    Validation of the Gatortail Method for Accurate Sizing of Pulmonary Vessels from 3D Medical Images
    in Medical Physics, 2017;44(12). doi:10.1002/mp.12580.
  2. O'Dell W.G., Schell M.C., Reynolds D, Okunieff P.
    Dose broadening due to target position variability during fractionated breath-held radiation therapy
    Medical Physics, (July) 29(7):1430-37, 2002
  3. Govindarajan ST, Chandrasekharan S, O'Dell WG
    Automatic Segmentation of Blood Vessel in the presence of Fibrosis in Volumetric Lung CT Images
    BMES Scientific Meeting and Exhibition, Hartford Conn, October 2011
Related Presentations
  1. Siva Kumar S, Bradley J, Liang X, Lockney N, Mailhot R, Mendenhall NP, Pembroke M, Okunieff P, O’Dell W.
          Evaluation of radiographic pulmonary changes on serial chest CTs after radiation therapy for breast cancer: a comparison of proton vs. photon therapy.
          65th Annual Meeting of the Radiation Research Society. San Diego, CA:
          Nov. 2, 2019.