Details

Autor(en) / Beteiligte

• Gertsenshteyn, Inna
• Epel, Boris
• Giurcanu, Mihai
• Leoni, Lara
• Barth, Eugene
• Fan, Xiaobing
• Markiewicz, Erica
• Zamora, Marta
• Tsai, Hsiu-Ming
• Friefelder, Richard
• Kucharski, Anna
• Bhuiyan, Mohammed
• Bodero, Darwin
• McVea, Andrew
• Holderman, Nathanial
• Karrison, Theodore
• Karczmar, Gregory
• Kao, Chien-Min
• Halpern, Howard
• Chen, Chin-Tu

Titel

Preliminary Investigation of Hypoxia within Tumor Using EPRI, DCE-MRI, and PET-CT with 18F-FMISO to Improve Radiotherapy

Ist Teil von

• The Journal of nuclear medicine (1978), 2019-05, Vol.60

Ort / Verlag

New York: Society of Nuclear Medicine

Erscheinungsjahr

2019

Link zum Volltext

Quelle

Free E-Journal (出版社公開部分のみ）

Beschreibungen/Notizen

• Objectives: Hypoxic cells are more resistant than normoxic cells to radiation treatment, so the identification of hypoxic regions within the tumor is necessary to predict or improve radiotherapeutic outcome. Preclinical studies in image-guided radiotherapy with electron paramagnetic resonance (EPR) pO2 images have shown that targeting hypoxic regions of tumors, defined as < 10 torr, significantly suppresses tumor regrowth. Though EPR is considered to be the gold standard for measuring pO2, this imaging modality is not readily available for clinical practice; both positron emission tomography (PET) and magnetic resonance imaging (MRI) are more readily available and used in the clinic routinely. In addition to EPR, 18F-fluoro-misonidazole (FMISO) PET has been used to target hypoxia in tumors. The purpose of this study is to measure hypoxia in a mouse tumor using EPR and PET to determine the efficacy of the FMISO tracer in accurately identifying hypoxia, and to consider applying a modification to the PET data using dynamic contrast enhanced (DCE)-MRI to better predict true hypoxia as identified by EPR. Methods: MCa4 tumors were grown in the leg of mice. Once the tumor grew in the range of 250-400 mm3, the mouse was anesthetized, and its leg was set in a custom-made bed and cast with soft vinylpolysiloxane. Under minimal anesthesia for a total of five hours, the mouse was imaged in EPR, followed by a T2-weighted and DCE MRI, a two-hour dynamic PET scan, and finally a CT scan. Data from all modalities were registered using fiducials and anatomic landmarks. The tumor to muscle ratio (TMR) above 1.4 defined hypoxia in the PET image in the last frame, and pO2 values below 10 torr defined hypoxia in the EPR image. Based on the Tofts model, ktrans and vemaps were obtained from DCE-MRI, as well as the relative signal increase (RSI) of contrast from time of injection. The Dice similarity coefficient was used to evaluate the overlap between the PET and EPR data. In addition, linear regression, ridge regression, least absolute shrinkage and selection (LASSO) regression, and generalized additive models were used to predict EPR at each voxel in terms of PET, ktrans, ve, and RSI. Results: Axial slices of the mouse shown in Fig. 1 qualitatively show that the region of uptake in FMISO PET more closely resembles increased activity in DCE-MRI measurements of ktrans and RSI, rather than the pO2 data from EPR, which we consider to be the gold standard of measuring hypoxia. Among the mice analyzed, the highest Dice coefficient between hypoxic regions in PET and EPR images was 0.23 with hypoxic fractions (HF) of 0.083 and 0.51, respectively. This follows the general trend of a higher HF in FMISO PET images than EPR images based on the set thresholds, with the exception of one mouse. Another mouse showed an HF of 0 in the EPR image, and an HF of 0.44 in the PET image, which suggests that FMISO does not target only hypoxia. The adjusted R2 of the regression models ranged from 0.20 to 0.24. Conclusions: Based on this preliminary dataset and low Dice coefficient values between PET and EPR images, we conclude that FMISO does not solely target hypoxia, and if it is to be used as a method of image-guided radiotherapy, it must be corrected by some other relevant physiological parameters, including but not limited to DCE-MRI imaging. Future analysis of this data includes adjusting hypoxia thresholds in PET data by tumor to lung ratio and by SUVmax, and developing spatial regression models for EPR in terms of PET, ktrans, ve, and RSI with better prediction properties.