Purpose— Pretreatment multimodality imaging can provide useful anatomical and functional data about tumors, including perfusion and possibly hypoxia status. The purpose of our study was to assess non-invasively the tumor microenvironment of neck nodal metastases in patients with head and neck (HN) cancer by investigating the relationship between tumor perfusion measured using Dynamic Contrast Enhanced MRI (DCE-MRI) and hypoxia measured by 18 F-fluoromisonidazole ( 18 F-FMISO) PET. Methods and Materials— Thirteen newly diagnosed HN cancer patients with metastatic neck nodes underwent DCE-MRI and 18 F-FMISO PET imaging prior to chemotherapy and radiation therapy. The matched regions of interests from both modalities were analyzed. To examine the correlations between DCE-MRI parameters and standard uptake value (SUV) measurements from 18 F-FMISO PET, the non-parametric Spearman correlation coefficient was calculated. Furthermore, DCE-MRI parameters were compared between nodes with 18 F-FMISO uptake and nodes with no 18 F-FMISO uptake using Mann-Whitney

Abstract It has been reported that the therapeutic effect of the proteasome inhibitor Velcade is due to its selective interference in the hypoxia pathway. We studied the effect of Velcade on tumor microenvironment and examined the underlying molecular mechanisms. We have generated a human colorectal cancer xenograft model (#C53) in which the hypoxia-inducible dual reporter fusion gene (HSV1-TK and eGFP) was under the control of hypoxia-response-element (HRE). In vitro, #C53 cells were treated with Velcade in normoxic and hypoxic conditions, and the following assays were performed in comparison with controls: eGFP (flow cytometry), CA9 (western blot), VEGF (ELISA), and TK activity (trapping of the marker substrate 14C-FIAU). In vivo, #C53 xenografts were treated with Velcade and various assays performed, including a) Dynamic contrast-enhanced (DCE) MRI pre and post treatment; b) dual hypoxia marker (pimonidazole and EF5) administration pre and post treatment; c) fluorescence microscopy of Hoechst 33342 (perfusion), eGFP, HIF-1α, and CA9; and d) plasma VEGF level (ELISA), Where applicable data from control and treated tumors were compared. Our results showed that in both in vitro and in vivo experiments Velcade treatment increased the level of HIF1α, but decreased those of hypoxia-induced eGFP, TK, CA9 and VEGF. Interestingly, in the dual hypoxia marker study there were significant EF5-positive regions that did not co-localize with pimonidazole-positive regions, suggesting de novo hypoxia and perhaps another novel effect of Velcade on tumor microvasculature. Consistent with these results, DCE MRI demonstrated decrease global tumor blood flow with Velcade treatment. Our data suggest that Velcade suppresses the hypoxia response by disrupting the HIF1 transcriptional activity. In addition, our results suggest a novel function of Velcade in modifying the tumor microenvironment and decreasing tumor perfusion as noninvasively detected by DCE MRI. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 1381.

A. Rizwan, X. Ni, R. O'Connor, S. Singer, J. Koutcher, and K. L. Zakian Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, Weill Cornell Medical College, New York, NY, United States, Sarcoma Biology Laboratory, Sarcoma Disease Management Program and Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY, United States

H-MRS, spectra were acquired on the tumor identified on T2-weighted images, and a volume of interest ( 2cc) was placed over the node. Single voxel spectroscopy data (PRESS, TR/TE=1600/136 and 256 averages) was obtained. Additionally, a spectrum (16 averages) was recorded of unsuppressed water. PET For 18 FDG PET imaging, F18-fluoride was produced by the cyclotron by proton irradiation of an enriched O-18 water target in a small-volume titanium chamber. 10 to 18 mCi of 18 FDG was administered by IV and image acquisition at the PET/CT scanner (Discovery ST) started after 2 hours of the injection. PET/CT images were reconstructed with the standard reconstruction array processor and corrected for attenuation. Analysis 1

Bladder cancer is one of the most common causes of death in industrialized countries. New tumor markers and therapeutic approaches are still needed to improve the management of bladder cancer patients. Choline kinase-α (ChoKα) is a metabolic enzyme that has a role in cell proliferation and transformation. Inhibitors of ChoKα show antitumoral activity and are expected to be introduced soon in clinical trials. This study aims to assess whether ChoKα plays a role in the aggressiveness of bladder tumors and constitutes a new approach for bladder cancer treatment. We show here that ChoKα is constitutively altered in human bladder tumor cells. Furthermore, in vivo murine models, including an orthotopic model to mimic as much as possible the physiological conditions, revealed that increased levels of ChoKα potentiate both tumor formation ( P ⩽0.0001) and aggressiveness of the disease on different end points ( P =0.011). Accordingly, increased levels of ChoKα significantly reduce survival of mice with bladder cancer ( P =0.05). Finally, treatment with a ChoKα-specific inhibitor resulted in a significant inhibition of tumor growth ( P =0.02) and in a relevant increase in survival ( P =0.03).

Introduction Diffusion-weighted MR imaging (DW MRI) is commonly performed to determine the apparent diffusion coefficient (ADC) of body tissues. Since DW images are reconstructed as magnitude images, they are contaminated with Rician noise [1, 2], which depends on the signal amplitude and the number of receiver channels N [3]. For a single receiver channel, it has been shown that the signal distortion caused by Rician noise leads to bias in ADC estimates [4, 5]. For multiple receiver channels, this noise-related bias is expected to be higher, as it was shown for T2 measurements with phased arrays [6]. We study the influence of Rician noise in DW images acquired with multichannel receivers on ADC values calculated using three methods: the noise-corrected maximum likelihood estimation (MLE) [7] and the uncorrected nonlinear least-squares fitting (NLSQ) and log-linear fitting (LL). We explore the accuracy and precision of these methods for a range of ADC and varying number of receiver channels using simulations and phantom and in vivo imaging of human prostate. Methods Simulations: All analyses were performed in Matlab (Mathworks; Natick, MA). Rician probability density for multichannel receivers [3] was incorporated into MLE; NLSQ and LL fitting was done without noise correction. Monte Carlo simulations (3000 trials) were performed with ideal data simulated using monoexponential model S(b)=S0exp(-b⋅ADC) for b=0-1600 s/mm in steps of Δb=100 s/mm. This signal was assumed to be real and contribute equally to each of N receiver channels, N=1-32. Gaussian noise with zero mean and standard deviation σ was added to both real and imaginary signal components in each channel. Signal-to-noise ratio (SNR) was varied by adjusting σ. The same SNR for different N was achieved by scaling σ as √ [6]. Simulations were first performed for a fixed ADC=1.0⋅10 mm/s and SNR=2-50. Another set of simulations was performed for SNR=10 and ADC=(0.2-2.0)⋅10 mm/s. Phantom imaging: A uniform copper sulfate solution phantom was imaged on a 3T whole-body unit (Signa HDx; GE, Milwakee, WI) with single shot spin echo echo-planar imaging sequence (TR/TE=3500/93.3 ms; matrix, 128x128 acquired, 256x256 interpolated; slice/gap, 3/3 mm; 17 uniformly spaced b-values 0 to 1000 s/mm). Endorectal coil (Medrad; Indianola, PA) (1 channel) combined with torso phased array (7 channels) was used for signal reception. To vary SNR, images were acquired at two different fields of view (FOV), FOV1=16x16 cm (1.25x1.25x3 mm voxel) and FOV2=32x32 cm (2.5x2.5x3 mm voxel). The noise parameter was determined as √ , where is the mean square intensity across voxels in an empty ROI [3]. ADC voxel maps were calculated by all three methods. In vivo imaging: After providing informed consent, a 59-year-old patient was imaged at 3T with the same coils and sequence (TR/TE=3500/104.9 ms; FOV, 16x16 cm; slice/gap, 3/3 mm; matrix, 96x96 acquired, 256x256 interpolated; b=01600 s/mm, Δb=100 s/mm). The noise parameter σ was determined as described above from an ROI in the center of endorectal coil. ADC voxel maps were calculated by each method. Results Simulations: NLSQ and LL progressively underestimate ADC as SNR decreases, while MLE is accurate within 10% for N=8 at SNR=5 and is bias-free at higher SNR (Fig. 1). However, MLE is less precise than NLSQ and LL up to SNR=10. MLE appears to be accurate across the entire range of ADC or N (Fig. 2a). NLSQ and LL underestimate higher ADCs more strongly than lower ones and this effect is exacerbated at higher N (Fig. 2b,c). Phantom imaging: As expected, for FOV2, ADCNLSQ and ADCMLE are equivalent (Pearson R=1.0, mean difference=1.12⋅10 mm/s), while for FOV1 ADCNLSQ<ADCMLE, especially at fitted ADC>2.5⋅10 mm/s. ADCLL shows large scatter vs ADCMLE at both FOVs. All methods provided similar average ADC across the phantom (2.2⋅10 mm/s). In vivo imaging: ADC maps showed a region of low ADC in transition zone (Fig. 3). The mean ADC in an ROI drawn in this region from MLE/NLSQ/LL analyses was (0.58/0.57/0.58)⋅10 mm/s, respectively. An ROI drawn in an adjacent area with higher ADC was (1.43/1.35/1.28)⋅10 mm/s and the ADC difference between the two ROIs was (0.85/0.78/0.70)⋅10 mm/s. As predicted by simulations, MLE yielded higher ADC estimates and higher ADC contrast between ROIs than NLSQ or LL.

S. B. Thakur, D. D. Dershaw, D. Giri, J. Zheng, C. Moskowitz, J. Ferrara, J. A. Koutcher, and E. A. Morris Medical Physics, Memorial Sloan Kettering Cancer Center, New York, NY, United States, Radiology, Memorial Sloan Kettering Cancer Center, New York, NY, United States, Pathology, Memorial Sloan Kettering Cancer Center, New York, NY, United States, Epidemiology-Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, United States, Epidemiology-Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY

6083 Background: 18F-FMISO PET/CT is a noninvasive hypoxia imaging modality. Hypoxia can develop within the tumor which typically corresponds to regions with poor blood perfusion. DCE-MRI can yield results on tumor perfusion. The present study compares hypoxic and perfusion status of HNC with nodal metastases using DCE-MRI and 18F-FMISO PET/CT imaging. Methods: 13 HNC (7 base of tongue, 5 tonsil, 1 larynx) patients with nodal metastases underwent both DCE-MRI and 18F-FMISO PET/CT scans prior to chemoradiotherapy. MRI was performed on a 1.5 Tesla GE Excite scanner. DCE-MRI studies were acquired using a fast multi-phase spoiled gradient echo sequence. Antecubital vein catheters delivered a bolus of 0.1 mmol/kg Gd-DTPA at 2 cc/s. For 18F-FMISO PET/CT,∼10.0 mCi of 18F-FMISO was administered IV and images were acquired ∼2 hours later. PET images were reconstructed; CT data were used for attenuation correction. Region of interests (ROIs) were manually drawn by a neuroradiologist. Quantitative DCE-MRI analyses were done using the 2-compartment Tofts model. The analyses calculated the pixel Ktrans(distribution rate constant), ve(extravascular-extra-cellular volume fraction), and kep(redistribution rate constant). 18F-FMISO PET/CT images were analyzed and the uptake by the tumor was scored as: no uptake (score 0); moderate-severe uptake (score 1). Semi-quantitative analysis included standardized uptake value (SUV) measurements. Statistical significance was calculated using a 2-sided Student's t-test, with p < 0.05. Results: A total of 17 nodes were analyzed. For the nodes that showed no hypoxia on PET imaging (n = 7), the mean (±SD) values were: 18F-MISO SUV (1.1 ± 03), Ktrans(0.33 ± 0.18), ve(0.53 ± 0.23), and kep(0.66 ± 0.25). For the nodes that showed moderate to severe 18F-MISO uptake (n = 10) the values were: 18F-FMISO SUV (2.8 ± 08), Ktrans(0.24 ± 0.07), ve(0.61 ± 0.13), and kep (0.43 ± 0.17). A significant lower kep for nodes with 18F-MISO uptake (p = 0.042). Conclusions: Preliminary result supports the hypothesis that the hypoxic nodes are poorly perfused nodes (lower kep values) versus nodes with no hypoxia. No significant financial relationships to disclose.

Background: Studies using phosphorus magnetic resonance spectroscopy (AIRS) have pointed to the significance of phospholipid metabolite alterations as biochemical markers for tumour progression or therapy response.Methods: Spectroscopic imaging was performed in colorectal flank tumours in nude mice. In vivo tumour doubling times for each cell line were measured. In vivo sensitivity of each tumour line to treatment with and NV1020 oncolytic viruses was assessed. Correlations between vital sensitivity, and tumour doubling time and phosphorus MRS were estimated.Results: For G207 virus, in vitro cytotoxicity tests showed cell viability at multiplicities of infection (ratio of viral particles per tumour cell) of 0.1 on day 6 as follows: C85, less than 1 per cent; HCT8, I per cent; LS174T, 9 per cent; HT29, 18 per cent; and C18, 92 per cent. Respective values for NV1020 were 1, 18, 4, 18 and 86 per cent. The phosphoethanolamine to phosphocholine ratio was significantly lower in virus-sensitive than -insensitive cells, and was dependent on tumour doubling time.Conclusion: Alterations in membrane phospholipid metabolites that relate to proliferation of cancer cells affect the efficacy of oncolytic viral therapy. MRS proved a highly sensitive non-invasive tool for predicting the efficacy of viruses.

Purpose: Dedicated small‐animal imaging devices are being used more frequently for translational molecular imaging studies. However, few studies have investigated the magnitude of animal motion during extended dynamic imaging studies or the precision of animal repositioning in multi‐modality and/or serial imaging protocols. The objective of this work was to determine the positional accuracy and precision with which tumors in situ can be reliably and reproducibly imaged on dedicated small‐animal imaging equipment. Method and Materials: A custom rodent animal cradle with a stereotactic template served to define a coordinate system and to facilitate rigid‐body image registration. Attached to the template were fiduciary markers containing PET tracer and MRI and CT contrast media for visualization on the respective scanners. To quantify animal tumor motion during imaging protocols, “gold standard” point markers were inserted into tumors grown on the hind limb of nude rats. Three types of imaging examination were performed with the animals continuously anesthetized and immobilized: (i) single‐modality imaging (microPET and MRI) in which the animals remained in the same scanner continuously for 2 hours, (ii) multi‐modality imaging studies in which the animals were transported from a microPET to an MR scanner located in another building, and (iii) serial microPET scans in which the animals were removed from the scanner, then re‐positioned and scanned. Results: The animal tumor moved by less than 0.2–0.3 mm over two‐hour microPET or MR imaging sessions. Transporting the animal between instruments introduced additional error of ∼0.2 mm. In serial animal imaging studies, in which the animal was returned to its cage and subsequently re‐positioned, reproducibility within ∼0.8 mm could be obtained. Conclusion: To our knowledge, this is the first study systematically and rigorously evaluating the accuracy and precision with which tumors can be repeatedly imaged in small‐animal imaging devices.

Purpose: To optimize the correlation between interstitial pO2 readings and PET‐derived concentrations in tumors using co‐registered Gd‐DTPA DCE‐MRI to distinguish viable and necrotic tumor cells. Method and Materials: Nude rats with Dunning R3327‐AT xenografts were positioned in a custom‐fabricated mold and imaged by (1) T2‐weighted MRI imaging; (2) DCE‐MRI following IV injection of gadopentetate dimeglumine; (3) microPET for 1/2 hr post‐IV injection of or . The MRI and PET images were used to direct intra‐tumoral pO2 measurements using an image‐guided robot system. PET‐ and MRI‐visible fiducial markers were used to register the respective image and robot coordinate systems. pO2 was measured at 0.5‐mm increments along different tracks within the tumor using an OxyLite™ 4000 Oxygen probe advanced by the robot, providing point‐to‐point correspondence between the pO2 measurement and image‐voxel intensity. Measurement points in the necrotic region and those outside tumor region, identified by DCE‐MRI and T2‐weighted MRI, were excluded from the correlation analysis. Necrotic tissue was identified as having at t=2 min after contrast injection, where I0 and I(t) are MR image intensities before injection and at time t. Results: The registration error between images and the robot is <0.3mm. For three FMISO studies with 539 measurements and 21 tracks, the correlation between interstitial pO2 readings and intensities was not improved — vs . Four animals were studied using with 423 measurements and 20 tracks. The negative correlation improved significantly (p =0.00118) — vs — if only viable‐tumor points were considered. Conclusion: The negative correlation between voxel intensities and intra‐tumoral pO2 was improved when DCE‐MRI was used to exclude necrotic tissue pO2s. Exclusion of necrotic points did not improve the correlation for , however.

Figure 4: ADC and Akep with increasing tumor volume (circle diameter proportional to tumor volume). Figure 3: The median ADC across the tumors increased slightly but not significantly with tumor size (left panel). However, median Akep across the tumor decreased significantly with tumor volume (p<0.05) (right panel). Field of view: 37.5 mm x 37.5 mm (128 pixels x 128 pixels), Slice thickness: 1 mm. (4 to 5 slices per tumor). Measuring Non-invasively Tumor Perfusion and Diffusion as A Function of Tumor Progression using Proton Magnetic Resonance Imaging (1H MRI)

Introduction Since visual inspection of large H-MRSI (magnetic resonance spectroscopic imaging) datasets of the prostate is time-consuming and requires spectroscopic expertise, introduction of a decision support system based on a pattern recognition (PR) model could help to promote the clinical use of MRSI. A PR approach does not require specific prior knowledge and can extract important spectral patterns from the in vivo training data automatically. It uses the complete information contained in the raw spectral data to address a diagnostic question – malignant vs. benign – directly. In this work, we have applied a Partial Least Squares (PLS) algorithm with Orthogonal Signal Correction (OSC) filtering which encompasses linear regression methods that have been employed successfully for prediction modelling in biological and biochemical applications [1]. Aim of the study To build a model for automated pattern recognition of H MRSI data indicating tumor-suspicious voxels in the prostate to assist clinicians in cancer diagnosis. Materials and methods 10 men (median age, 63 years; range, 43-68 years) who underwent 1.5-T endorectal MR imaging before radical prostatectomy and who fulfilled all inclusion criteria of no prior hormonal or radiation treatment and at least one tumor lesion at whole–mount pathologic examination were included. MRI and 3D H MRSI examinations were performed on a 1.5-T whole-body unit (Signa Horizon; GE Medical Systems) with an endorectal coil (Medrad) and an acquisition package (PROSE; GE Medical Systems). The spectroscopic acquisition parameters were as follows: PRESS voxel excitation, 1000/130 ms [TR/TE]; numbers of averages acquired, one; spectral width, 1250 Hz; number of points, 512; field of view, 11 x 5.5 x 5.5 cm; and 16 x 8 x 8 phase encoding steps. Spectral data were processed by using free software 3DiCSI v1.9.11 (http://mrs.cpmc.columbia.edu/3dicsi.html). The MRSI data were spatial zero filled to a 16 x 8 x 16 matrix and zero filled in the spectral dimension to 1024 points. The time-spectral dimension was apodized with a 4-Hz Gaussian function. The spectra were aligned and referenced to the water peak at 4.7 ppm. The range 3.6 0.6 ppm (198 points) was chosen. Magnitude spectra were exported to achieve better and reproducible results with the PR method [2, 3] as well as to fully automate and simplify preprocessing. All 2362 voxels within the prostate were labeled as healthy or tumor by an experienced spectroscopist according to established decision rules based on the resonances of total choline (Cho) at 3.2 ppm, creatine/phosphocreatine (Cr) at 3.0 ppm, polyamines (PA) at 3.1 ppm and citrate (Cit) at 2.6 ppm [4]. For all voxels visually marked as tumor, the correct lesion location was confirmed on the basis of histopathology maps with sextant precision. The labels healthy or tumor in the form of “0” or “1” in the Y prediction matrix were used to establish PR models by applying supervised multivariate statistical methods (OSC filtering and PLS analysis) on training sets. Using the “leave-one-patient-out” method, the generated models were applied to predict new data in an unknown test set and to evaluate classification accuracy. Prior to prediction, new samples in the test set were automatically pre-treated in the same way as the training set (scaling, centering, OSC filtering). Multivariate analysis was performed using SIMCA-P software v.11.5 (Umetrics, Sweden). Three different approaches to variable centering and auto scaling were compared (centered and scaled to Unit Variance (UV); centered but not scaled (Ctr); no centering or scaling (None)). The OSC algorithm was used to remove unwanted variation in the spectra that was irrelevant for the classification. Five orthogonal components were extracted removing over 70% of variation that did not contribute to discrimination. The classification accuracy for the model was computed as the ratio of the number of spectra predicted correctly to the total number of spectra in the test set. YPredPS values were provided by the SIMCA-P software. YPredPS is the Y value predicted by the model based upon the X block variables (resonance intensities at given ppm). A YPredPS value close to 1 would indicate that the object is likely to belong to the class. A YPredPS value close to 0 would indicate that the object is unlikely to belong to the class. Results The best results (highest Q2 value) were obtained using no centered or scaled spectra in comparison to UV and Ctr methods. In the computed models after OSC filtering, the first PLS component explained greater than 82.1% of the variation in the spectra between healthy and tumor voxels (R2Y=0.821). The overall predictive power of the training set calculated by cross-validation was greater than 80.4% (Q2 = 0.804). Using the models generated by the training set, the spectra in the test set were correctly predicted greater than 81% of the time. It must be noted that the results are strongly dependent on the choice of training datasets and peak position variation. Variables (ppm locations) with Variable Importance in the Projection values (VIP) larger than 1 are the most relevant for explaining differences between classes of spectra. Figure 1 contains a plot of the VIP values generated from the training set. The most important variable in differentiating tumor and healthy voxels was Cho. Cr, PA and Cit also had VIP values greater than one and thus were important for differentiating cancer voxels. Figure 2 contains an example of a spectrum from the test set which was classified correctly as tumor by the model. Discussion and conclusions The impact of data preprocessing is very important in carrying out this pattern recognition model. MRSI spectra have different ranges of intensities both within a given patient data set and between patients. Therefore, the interpretation can be distorted because these changes are not really responsible for the discrimination between the different classes. Auto-scalling (mean-centering and unit-variance scalling) has the limitation of giving the same weight to all the spectral variables because of their now equal variance. Our study showed that the best results were obtained by not applying scaled or centered spectra. One of the main sources of error in the computed models was applying an alignment method based on water peak referencing. The robustness of the algorithm is strongly dependent on peak alignment; and, since the water peak is partially suppressed, the true center frequency of the peak may not be accurately reflected. Another disadvantage is operating on magnitude spectra which have the advantage of invariance with regard to zero-order phase, but have increased linewidths. Magnitude spectra have been shown previously to yield better performance by pattern recognition methods including PLS [2]. The modelling results presented here are still in development. However, we have shown that the multivariate PLS method with OSC works well with the tested data sets and could help to automatically distinguish the tumor-suspicious voxels. The main advantage of this method is the much shorter time of analysis compared to visual inspection and the possibility of broad implementation in cancer centers not employing experienced spectroscopists. In terms of accuracy, the proposed method is still not comparable to the reference method (a visual inspection by an experienced spectroscopist). Ongoing improvements in preprocessing and refining the model as well as importing more datasets should result in better performance. References: 1. Trygg and Wold 16 (3): 119. (2002); 2. Kelm et al. 57(1): 150. (2007); 3. Devos et al. 170 (1): 164. (2004); 4. Shukla-Dave et al. 245 (2): 499. (2007)

The double suicide gene therapy strategy combining herpes simplex virus type-1 thymidine kinase (HSV-1 TK) and cytosine deaminase (CD) with ganciclovir (GCV) and 5-fluorocytosine (5-FC) had been carried out in a phase I clinical trail for prostate cancer. The objective of the present study was to ascertain whether co-expression of HSV-1 TK, CD, and uracil phosphoribosyltransferase (UPRT) in conjunction with GCV and 5-FC treatment, a triple suicide gene therapy strategy, would be more effective at killing and radiosensitizing prostate cancer cells than double suicide gene therapy. Rat prostate cancer R3327-AT cell lines stably expressing HSV1-TK and CD (TKCD), or HSV1-TK and CD/UPRT (TKCDUPRT) were selected and established by co-transfecting the cells with the plasmids encoding HSV1-TK and CD or CD/UPRT fusion gene respectively. The sensitivity of TKCD and TKCDUPRT cells to GCV and/or 5-FC treatment was assessed by colony formation assay. In experiments to test the radiosensitization effect, TKCD and TKCDUPRT cells were treated with GCV (0.2 μg/ml for TKCDUPRT cells or 1 μg/ml for TKCD cells) and/or 5-FC (0.2 μg/ml for TKCDUPRT cells or 10 μg/ml for TKCD cells) for 24 h before γ-irradiation (0-12 Gy), the surviving fractions was determined by the colony formation assay and normalized to those of non-irradiated cells treated with the prodrug only. The survival curves were fitted with LQ models, and sensitization enhancement ratios (SER) were calculated. TKCD and TKCDUPRT cells were sensitive to GCV or 5-FC alone in a dose dependent manner. Concurrent prodrug treatment with both GCV and 5-FC produced a more synergistic cytotoxic effect in TKCDUPRT-expressing cells than TKCD cells, even though the concentration of GCV and 5-FC used was much lower in TKCDUPRT cells than TKCD cells to achieve the same effect. A much lower concentration of GCV or 5-FC treatment to CDUPRT cells results in noticeable radiosensitization (SER 1.1-1.3). However, treatment with both 5-FC and GCV prior to radiation resulted in a marked increase of radiosensitivity in TKCDUPRT cells, yielding a SER of 2.1. On the other hand, a much higher dose of 5-FC and GCV is needed to radiosensitize TKCD cells (SER is 1.7). Our study showed that co-expression of HSV1-TK, CD, and UPRT suicide genes significantly enhanced the cytotoxicity and radiosensitization effect of combined 5-FC and GCV treatment in R3327-AT prostate cancer cells. The data suggest that this triple suicide gene therapy strategy may have potential as an adjuvant to improve the treatment outcome of radiotherapy for prostate cancer.

Purpose: To establish a criterion for distinguishing viable and necrotic tumor cells using the initial relative slope of the Gd‐DTPA DCE‐MRI and to apply this criterion to optimize the specificity of tumor hypoxia imaging based on PET. Method and Materials: Three nude rats with Dunning R3327‐AT prostate adenocarcinoma xenografts were imaged by dynamic MRI following tail vein injection of gadopentetate dimeglumine, with imaging parameters of 1‐mm slice thickness, 0.5‐mm spacing, and 0.13mm × 0.13mm voxel size. The time‐intensity curve of each voxel was obtained at 1‐min interval to 25‐min post‐injection. Histologic (standard H&E stain and high‐power microscopy) examination was performed of 8‐μm thick slices parallel to the DCE‐MRI imaging axis and the necrotic regions were identified. The initial relative slope, where I 0 and I(t) are MR image intensity before injection and at time t, was calculated for different times and compared with the pathologically defined necrotic region to determine the threshold that best distinguishes viable and necrotic cells. This criterion was then used to study the specificity of FMISO PET with the same xenografts scanned with DCE‐MRI and PET. Results: The optimal criterion for identifying viable and necrotic tumor regions was that a DCE‐MRI voxel was necrotic if at 2 minutes after contrast injection. When this criterion was applied to the PET images of three xenografts, necrotic region was found to have a wide range of image intensities, which is inconsistent with the hypothesis that high image intensity exclusively identifies hypoxic viable tumor. Among the 319 hypoxic voxels determined from the three animals' PET, only 75% corresponded to viable cells. Conclusion: A criterion was established to identify necrotic/viable tumor cells using DCE‐MRI. Using PET alone for hypoxia imaging is problematic because the specificity might be compromised by necrotic regions with high PET image intensity.

Introduction: Improved understanding and the ability to image specific important features of the tumor microenvironment in vivo will provide important prognostic information about tumors and factors that induce responses or resistance to treatment [1,2]. In this study, multimodality in vivo Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) imaging using stereotactic fiduciary markers [3] in the Dunning R3327-AT prostate tumor was performed, focusing on the relationship between Dynamic Contrast-Enhanced (DCE)-MRI using Magnevist (Gd-DTPA), and dynamic Ffluoromisonidazole (F-Fmiso) PET. The non-invasive measurements were verified using tumor tissue sections stained for haematoxylin/eosin (H&E), pimonidazole and F digital autoradiography.