The long-term goal of this project is to develop 1H MRS and MRI methods for prediction and detection of therapeutic response of non-Hodgkin's lymphoma (NHL), a prevalent malignancy whose incidence has increased by more than 90% since 1950 that ranks fourth in overall economic impact. A multi-institutional clinical study in which the PI is participating has already demonstrated that 31P MRS can predict about 2/3 of the tumors that fail to exhibit a complete clinical response, but this method is limited to relatively large superficial tumors. Our approach has been to develop proton imaging and spectroscopy methods that are capable of early detection of response in all patients. During the first funding cycle of this grant we examined a xenograft model, DLCL2, of diffuse large B-cell lymphoma (DLBCL), the most common form of NHL affecting about 1/3 of all patients. We have evaluated the utility of 1H MRS measurements of total choline and lactate and 1H MRI (diffusion-weighted and T2-weighted with dynamic contrast-enhanced and T1A-weighted) in this tumor. We have examined the response to multiple clinically relevant therapies including: combination chemotherapy with Cyclophosphamide, Hydroxydoxorubicin, Oncovin and Prednisone (CHOP), CHOP plus Bryostatin 1 (CHOPB), Rituximab alone, Rituximab plus CHOP and radiation therapy. The most exciting results from these studies are: 1) significant decreases in lactate detected within one cycle of CHOP that correlated with decreased tumor proliferation;2) regional changes in the apparent diffusion constant and in T2 corresponding to areas of response in the tumor;3) design and implementation of a novel hybrid pulse sequence for 1H imaging of lactate utilizing multi-slice Hadamard spatial encoding and chemical shift imaging via a selective multi-quantum coherence transfer. This sequence was developed on the animal scanner and has now been translated to a clinical scanner and applied to an NHL patient. The objectives of the current proposal are 1) to define the mechanisms underlying therapy associated decreases in lactic acid, 2) to define the mechanism leading to MRI-detected regional response of the tumor, and 3) to extend the study to more responsive DLBCL xenograft models and to follicular lymphoma, the next most common form of NHL. We will employ a number of sophisticated and innovative methods to achieve these goals including 1) use of a two- compartment model to distinguish between changes in steady-state lactate concentration due to changes in glycolysis and washout, 2) determination if FDG-PET imaging and DCE MRI, which are more readily available in the clinic than 13C MRS, can distinguish between changed in lactate levels due to alterations in glycolytic rate or due to a change in the rate of lactate elimination, 3) 17O imaging methods to image perfusion and oxygen consumption simultaneously and 4) validation of a metabolic network model for 13C NMR by predicted and experimental oxygen consumption rates. We will also explore in depth possible mechanisms that could contribute to regional therapeutic response in the tumor.
This project is directed towards developing NMR methods for prediction and early detection of therapeutic response of individual patients with non-Hodgkin's lymphoma (NHL), one of the prevalent forms of human cancer. A clinical program in which we participate has already developed a 31P NMR method for predicting about 2/3 of the tumors that will fail to respond to conventional treatment. This method, however, is limited to patients with large superficial tumors. Here we are developing much more sensitive 1H NMR methods that can be used for early detection of response in all NHL patients. The same methods should be applicable to other prevalent forms of human cancer.
Showing the most recent 10 out of 12 publications
