Skip Navigation

Radiation Physics
Functional Imaging Group Projects

PROJECTS

 

Develop algorithms and models to extract the subvolume of the tumor for outcome prediction

imaging studies considered a tumor as a uniform entity by using a mean value of a physiological imaging parameter to depict the whole tumor. This approach ignores heterogeneity of the tumor pathophysiology and biology. It is plausible that heterogeneity within a tumor is responsible for observed heterogeneous tumor responses to treatment. This motivates investigations of tumor voxel-level changes longitudinally, such as from pre to after starting therapy. However, reliably determining voxel-to-voxel changes from a pair of image volumes acquired over a period depends on the voxel-level accuracy of image registration, which is a very challenging task when a tumor volume grows or shrinks over the period of follow-up. Errors in deformable image registration have a great impact on the analysis of voxel-to-voxel changes due to therapy or tumor growth. We have been developing an alternative approach that extracts an aggressive subvolume of the tumor from its heterogeneously-distributed physiological imaging parameters and assess its change after starting therapy for prediction of treatment response and outcomes (Figure 1). Our model extracts the “feature” subvolumes for therapy assessment by analyzing the imaging parameters in the feature space. Finally, to spatially guide adaptive RT, the prognostic or predictive subvolume of the tumor is required to be highly associated with the site for tumor recurrence/failure. We have been developing and validating our models in three cancer types, namely, head-and-neck cancer, brain metastasis, and intrahepatic cancer.

Figure 1.

 
 
 

Publications
Y. Cao.  Development of Image Software Tools for Radiation Therapy Assessment.  Medical Physics 32:2136, 2005.

Y. Cao. Automated Classification of Substructures of white mater via anisotropic water diffusion.  Medical Physics 34(6):2548, 2007.

Y. Cao and L. Chen.  MRI in Radiation Treatment Planning and Assessment.  In: Integrating New Technologies into the Clinic: Monte Carlo and Image-Guided Radiation Therapy (ed. Bruce H. Curran, James M. Balter, and Indrin J. Chetty).  Medical Physics Publishing (Madison, WI), 2006, pp401-424.

J.M. Balter, Y. Cao.  Advanced technologies in image guided radiation therapy.  Sem Radiat Oncol 7:293-297, 2007.

Back to top

 

Functional Imaging as a Biomarker for Neurotoxicity after Brain Irradiation
RT is a major treatment modality for malignant and benign brain tumors.  The major limiting factor is radiation-induced neurotoxicity.   This neurotoxicity manifests as late neurological sequelae and neurocognitive dysfunction with or without gross tissue necrosis. Late neurocognitive dysfunction presents as diminishing mental capacity for working memory, learning ability, executive function, and attention.  Recent multi-center studies of patients with low-grade gliomas who are without clinical signs of tumor recurrence after radiation treatment show that both a high total dose as well as a high dose per fraction is associated with neurocognitive deterioration, especially memory functions. The potential effect of RT on neurocognitive outcomes is an important factor in the determination of the risks versus benefits of treatment, which should be an integral part of clinical decision-making.  Our primary goal of the project is to identify functional imaging biomarkers for early assessment of individual sensitivity to radiation and prediction of late neurotoxicity. Such biomarkers might provide an opportunity to individualize the dose of RT or to begin neuroprotective therapy.Our on-going imaging studies of patients undergoing partial brain RT have showed that early changes in vascular and WM properties assessed by DCE and DT MRI correlated with delayed changes in memory and learning functions.  Also, we will test and validate the clinical value of these quantitative imaging metrics for assessing cerebral vascular and tissue injury and predicting neurocognitve function declines in patients with brain metastases undergoing WBRT.

Figure 2. Hippocampal White Matter Change.

 
Fig_3
Cingulum
Pre-RT
1 month after WBRT
   
The diffusivity perpendicular to the long axis WM increased by 63% one month after WBRT


Publications
V. Nagesh, C. I. Tsien, T. L. Chenevert, B. D. Ross, T. S. Lawrence, L. Junck, and Y. Cao.  Quantitative characterization of radiation dose dependent changes in normal appearing white matter of cerebral tumor patients using diffusion tensor imaging.  Int J Rad Onc Biol Phys 70(4):1002-10, 2008.

P.C. Sundgren, V. Nagesh, A. Elias, C. Tsien, L. Junck, D.M. Gomez-Hassan, T.S. Lawrence, T.L. Chenevert, L. Rogers, P. McKeever, Y. Cao.  Metabolic alterations: a biomarker for radiation induced normal brain injury. A MR Spectroscopy study. JMRI, 29(2):291-297, 2009.

Y. Cao, C.I. Tsien, P. Sundgren, V. Nagesh, Daniel Normolle, H. Buchtel, L. Junck, and  T.S. Lawrence.  DCE MRI as a biomarker for delayed radiation-induced neurocognitive dysfunctions.  Clinical Cancer Research, 15(5): 1747-1754, 2009.

V. Nagesh, C.I. Tsien, P.C. Sundgren, H. Buchtel, T.L. Chenevert, L. Junck, L. Rogers, T.S. Lawrence and Y. Cao.  Diffusion Tensor Imaging as a Biomarker for Neurotoxicity of Radiation.  Int J Rad Onc Biol Phys 72(1):S529, 2008.

Back to top

 

Hepatic Perfusion Imaging as a Biomarker for Residual Liver Function after Irradiation
Approximately two thirds of patients with intrahepatic cancer present with unresectable disease.  Our recent studies show that high dose conformal radiation combined with chemotherapy appears to prolong the survival of patients with unresectable intrahepatic cancers.  However, attempts to increase radiation dose still further have been limited by the development of radiation-induced liver disease (RILD).  Symptoms generally occur 2 weeks-2 months following completion of radiotherapy and include tender hepatomegaly and weight gain secondary to ascites.  Laboratory findings include elevations of alkaline phosphatase, transaminases, and bilirubin.  The clinical outcome ranges from mild, reversible damage to death.  The pathology of RILD is veno-occlusive disease (VOD), which is characterized by thrombosis within the central veins of the liver producing “post” hepatic congestion.  The pathological changes may be present in the entire liver (rarely), a lobe, or a fraction of it.  In the past, efforts to develop models to estimate the likelihood of developing radiation-induced liver disease (RILD) have been based primarily on the planned radiation dose distribution for the normal liver, expressed as a dose volume histogram.  While these models have permitted the safe delivery of far higher doses of radiation than have previously been possible, they also suggest that there is a broad range of individual patient sensitivity that is not reflected by predictions made solely based on the physical dose distribution or general clinical features.   If individual patient sensitivity could be better estimated before or during a course of treatment, it would permit higher doses of radiation to be delivered safely to the tumors of patients whose liver is relatively radiation resistant, thus prolonging survival without increasing complications.  We hypothesized that the regional perfusion one month after the completion of radiation therapy could be predicted by the local radiation dose and perfusion measured prior to the end of treatment.  Our studies have demonstrated that (1) the regional venous perfusion dysfunction after RT was correlated with the local accumulated dose at the end of RT; (2) the extent of the regional venous perfusion dysfunction after RT also depended upon each individual’s sensitivity to radiation; (3) the individual and regional sensitivity to radiation can be assessed by the individual portal venous perfusion-dose response function; and (4) the spatially-resolved portal venous perfusion could be a marker for overall liver function, knowledge of which is required for radiation treatment planning and plan adaptation.  We are testing and validating these imaging metrics for prediction of the maximal tolerable dose in patients who will undergo SBRT, thereby to individualize benefits and risks of therapy.

Figure 3. Early Assessment for Venous Perfusion Dysfunction

 
Fig_3
Prior to RT
After 45 Gy

 

Publications
Y. Cao, C. Pan, J. M. Balter, J. F. Platt,  I. R. Francis, J. A. Knol, D. Normolle, E. Ben-Josef, R. K. Ten Haken, and T. S. Lawrence. Liver function after irradiation based upon CT portal vein perfusion imaging.  Int J Rad Onc Biol Phys 70(1):154-160, 2008.

Y. Cao, J. F. Platt, I.R Francis, J. M. Balter, C. Pan, D. Normolle, E. Ben-Josef, R. K. Ten Haken, T. S. Lawrence.  The prediction of radiation-induced liver dysfunction using a local dose and regional venous perfusion model.  Med Phys 34(2):604-612, 2007.

Back to top

 

Predictive Value of DCE Imaging For Treatment Outcomes in Advanced Head and Neck Cancers
The current standard of care for advanced SCCOP involves aggressive concurrent chemo-RT.  Intensification of chemo-RT for SCCOP has resulted in improved control rates as well as increasing rates and severity of late treatment toxicity.   Despite the improvements in treatment regimens, there are still up to 20%-50% failure rates, including metastatic and locoregional failures, especially in patients who are smokers25 and with negative infection of the human papillomavirus (HPV).  It would be clinically important to have an earlier predictor of treatment outcomes, thus allowing for Intensification of definitive therapy for tumor likely to recur.  Our recent findings suggest that blood volumes (BV) in the primary gross tumor volume (GTV) 2 weeks after the start of 7-week definite chemo-RT was increased significantly in the patients with local control compared to the patients with local failure (p<0.03). ROC analysis showed that the change in BV of the primary GTV has better sensitivity and specificity for prediction of LF than the change in the primary GTV size (AUC 0.89 vs 0.69, respectively). Thus changes in tumor BV may serve as a predictor of outcome of HNC, and aid in stratifying patients for different intensity treatments.

Figure 4. Early Assessment for Venous Perfusion Dysfunction

 
Fig_3
BV pre-RT
BV 2 weeks after the start of RT

Publications
Y. Cao, J. Alspaugh, Z. Shen, J.M. Balter, T.S. Lawrence, R. K. Ten Haken.  A practical approach for quantitative estimates of voxel-by-voxel liver perfusion using DCE imaging and a compartmental model.  Med Phys 33(8):3057-3062, 2006.

Y. Cao, Z. Shen, T. Chenevert, and J. R. Ewing. An Estimate of Vascular Permeability and Cerebral Blood Volume Using Gd-DTPA Contrast Enhancement and Dynamic T2* Weighted MRI.   J Magn Reson Imag 24(2): 288-296, 2006.

Back to top

 

Early Assessment of Treatment Response in Patients with Brain Metastases

Development, Improvement and Validation of DCE Modeling
DCE imaging is emerging as a valuable biomarker for treatment assessment.  In order to enable multi-center clinical trials, improvement and validation of DCE models become important.  One of our on-going projects is to develop, improve, and validate software tools for DCE pharmacokinetic models of vascular parameters.

Publications
Y. Cao, J. Alspaugh, Z. Shen, J.M. Balter, T.S. Lawrence, R. K. Ten Haken.  A practical approach for quantitative estimates of voxel-by-voxel liver perfusion using DCE imaging and a compartmental model.  Med Phys 33(8):3057-3062, 2006.

Y. Cao, Z. Shen, T. Chenevert, and J. R. Ewing. An Estimate of Vascular Permeability and Cerebral Blood Volume Using Gd-DTPA Contrast Enhancement and Dynamic T2* Weighted MRI.   J Magn Reson Imag 24(2): 288-296, 2006.

Back to top