Purpose In radiotherapy, PET images can be used to guide the delivery of selectively escalated doses to biologically relevant tumour subvolumes. Validation of PET for such applications requires demonstration of spatial coincidence between PET tracer uptake pattern and the histopathologically confirmed target. This study introduces a novel approach to histopathological validation of PET image segmentation for radiotherapy guidance. Methods and materials Sequential tissue sections from surgically excised whole-tumour specimens were used to acquire full 3D-sets of both histopathological images (microscopy) and PET tracer distribution images (autoradiography). After these datasets were accurately registered, a full 3D autoradiographic distribution of PET tracer was reconstructed and used to obtain synthetic PET images (sPET) by simulating the image deterioration induced by processes involved in PET image formation. To illustrate the method, sPET images were used in this study to investigate spatial coincidence between high FDG uptake areas and the distribution of viable tissue in two small animal tumour models. Results The reconstructed 3D autoradiographic distribution of the PET tracer was spatially coherent, as indicated by the high average value of the normalised pixel-by-pixel correlation of intensities between successive slices (0.84 ± 0.05 and 0.94 ± 0.02). The loss of detail in the sPET images versus the 3D autoradiography was significant as indicated by Dice coefficient values corresponding to the two tumours (0 and 0.1 at 70% threshold). The maximum overlap between the FDG segmented volumes and the extent of the viable tissue as indicated by Dice coefficient values, was 0.8 for one tumour (for the image thresholded at 22% of max intensity) and 0.88 for the other (threshold of 14% of max intensity). Conclusion It was demonstrated that the use of synthetic PET images for histopathological validation allows for bypassing a technically challenging and error-prone step of registering non-invasive PET images with histopathology.

Purpose: To develop methods for pre‐clinical validation of PET image‐guided selective dose escalation in IMRT treatment using the Small Animal Radiation Research Platform (SARRP, Xstrahl). Methods: Nude mice bearing subcutaneous FaDu human head and neck tumor xenografts were imaged with 18F‐FDG using a Siemens Inveon PET/CT. The PET image was used to create an image‐guided radiation treatment plan treating the entirety of the tumor with a 10Gy uniform radiation dose while sparing as much normal tissue as possible using a 15mm collimator. A dose escalation of 10 Gy was planned to treat the volume of highest FDG uptake within the tumor. To ensure localized dose distribution, the boost was to be delivered using dynamic arc techniques with a 5mm collimator. The isocenters for both fields were recorded on the planning PET image. The following day, the animals were anaesthetized, positioned in the SARRP irradiator, and imaged using on‐board cone‐beam to obtain an image in the treatment position. The planning CT was then deformably registered to the treatment CT using the BRAINSFit b‐spline algorithm in SLICER 3D. Finally, the transform was applied to the planning PET and isocenters for the treatment fields were transferred to the treatment CT while the animal was still under anesthesia. The animal was then irradiated according to the original PET/CT‐guided treatment plan. Results: SARRP allows for selective dose escalation treatments to small animal tumor models. In this study a subcutaneously grown tumor of 12mm in diameter was irradiated to 10Gy. The volume of the tumor characterized by increased FDG uptake received an additional 10Gy in the form of a 5mm‐diameter boost delivered with non‐coplanar arc field. Conclusion: The Small Animal Radiation Research Platform (SARRP) combined with a dedicated small animal PET/CT scanner enables pre‐clinical validation of PET image‐guided selective dose escalation in IMRT treatment.

Background and purposePET imaging with 18F-fluorothymidine (18F-FLT) can potentially be used to identify tumour subvolumes for selective dose escalation in radiation therapy. The purpose of this study is to analyse the co-localization of intratumoural patterns of cell proliferation with 18F-FLT tracer uptake.Materials and methodsMice bearing FaDu or SQ20B xenograft tumours were injected with 18F-FLT, and bromodeoxyuridine (proliferation marker). Ex vivo images of the spatial pattern of intratumoural 18F-FLT uptake and that of bromodeoxyuridine DNA incorporation were obtained from thin tumour tissue sections. These images were segmented by thresholding and Relative Operating Characteristic (ROC) curves and Dice similarity indices were evaluated.ResultsThe thresholds at which maximum overlap occurred between FLT-segmented areas and areas of active cell proliferation were significantly different for the two xenograft tumour models, whereas the median Dice values were not. However, ROC analysis indicated that segmented FLT images were more specific at detecting the proliferation pattern in FaDu tumours than in SQ20B tumours.ConclusionHighly dispersed patterns of cell proliferation observed in certain tumours can affect the perceived spatial concordance between the spatial pattern of 18F-FLT uptake and that of cell proliferation even when high-resolution ex vivo autoradiography imaging is used for 18F-FLT imaging.

PURPOSE:PET imaging allows for the visualization of tumor microenvironment and identification of aggressive or radioresistant tumor subvolumes that can be targeted with an escalated radiation dose. Multiple PET tracers have been developed for visualization of different aspects of tumor microenvironment; however, the spatial distribution of tracers in tumors is equally affected by tumor tissue viability and tracer delivery limitations. Given these issues and the low resolution associated with PET imaging, two different PET tracers can produce very similar images. Therefore, it is important to demonstrate that a novel PET tracer does provide additional useful information to that obtained with other tracers. This study investigates the added value of performing 18F-FLT PET imaging as well as 18F-FDG imaging.METHODS:Head and neck tumor xenografts grown in nude mice were used to study intratumoral tracerdistributions. 18F-FDG and 18F-FLT PET images were obtained on subsequent days using a small animal PET/CT. Pinnacle 9 was used to deformably register the CT image from the FLT PET/CT to the FDG PET/CT image set. The generated deformation was applied to the FLT PET image to achieve an unbiased FLT to FDG PET image registration. The Pearson correlation coefficient between FDG and FLT was calculated voxel- by-voxel within a tumor contour. Overlap analysis of thresholded tracer distributions was carried out by comparing Dice similarity coefficients.RESULTS:Both SQ20B and FaDu tumors showed a moderate voxel-by-voxel correlation between FDG and FLT intratumoral patterns of uptake with an average rho value of .56 and .63 respectively (range .37-.76) despite significant differences in tumor morphology. The average volumes under thedice coefficient surface for SQ20B and FaDu tumors were not significantly different.CONCLUSIONS:Despite being equally affected by the issues of tracer delivery, necrosis and PET resolution, FDG and FLT PET images displayed an observable difference at clinically relevant thresholds.

Purpose: It has been proposed that PET images can be used to guide the delivery of selectively escalated doses to biologically-relevant tumor subvolumes. Histopathological validation of PET imaging is challenging due to difficulties associated with precise registration of non-invasive in-vivo images to histopathological ex-vivo images. The aim of this study is to develop an alternative method of PET imaging validation for image-guidance applications. The method is applied to evaluation of the feasibility of FDG PET-based delineation of viable tissue in animal tumor models. Methods: Tumor-bearing mice were injected with 14C-FDG. Whole-tumor specimens were sectioned, obtaining 8μm thick sections every 120μm throughout the tumor. These sections were used to obtain 14C-FDG autoradiography and H&E microscopy images. Viable tumor tissue was delineated on each H&E image. Based on sequential digital photography images of the tissue block acquired during sectioning, the true 3D distributions of 14C-FDG and viable tissue were reconstituted. To simulate generation of a PET image, the 3D activity map was convolved with a 3D point-spread-function of Siemens Inveon small-animal PET scanner. Threshold-based analysis was used to evaluate the degree of coincidence between the areas of high FDG uptake in the simulated PET image and 3D distribution of viable tissue. Results: Averaging effects associated with PET imaging altered the true 3D spatial pattern of FDG intratumoral uptake. ROC analysis indicated good sensitivity of FDG PET image-segmentation for the detection of the viable tissue (AUC = 0.74). However, the specificity was low, as indicated by the low threshold value at which the maximum overlap occurred (22% of maximum uptake). Conclusion: A novel method of histopathological validation of PET imaging for image-guidance in radiotherapy was developed. Using this method, it was demonstrated that for the tumors with high viable tissue content, FDG-thresholding can be used for viable tissue detection.

Histopathologic validation of a PET tracer requires assessment of colocalization of the tracer with its intended biologic target. Using thin tissue section autoradiography, it is possible to visualize the spatial distribution of the PET tracer uptake and compare it with the distribution of the intended biologic target (as visualized with immunohistochemistry). The purpose of this study was to develop and evaluate an objective methodology for deformable coregistration of autoradiography and microscopy images acquired from a set of sequential tissue sections. Methods: Tumor-bearing animals were injected with 3′-deoxy-3′-18F-fluorothymidine (18F-FLT), 14C-FDG, and other markers of tumor microenvironment including Hoechst 33342 (blood-flow surrogate). After sacrifice, tumors were excised, frozen, and sectioned. Multiple stacks of sequential 8 μm sections were collected from each tumor. From each stack, the middle (reference) sections were used to obtain images of 18F-FLT and 14C-FDG uptake distributions using dual-tracer autoradiography. Sections adjacent to the reference were used to acquire all histopathologic data (e.g., images of cell proliferation, hematoxylin and eosin). Hoechst images were acquired from all sections. To correct for deformations and misalignments induced by tissue processing and image acquisition, the Hoechst image of each nonreference section was deformably registered to the reference Hoechst image. This transformation was then applied to all images acquired from the same tissue section. In this way, all microscopy images were registered to the reference Hoechst image. The Hoechst-to-autoradiography image registration was done using rigid point-set registration based on external markers visible in both images. Results: The mean error of Hoechst to 18F-FLT autoradiography registration (both images acquired from the same section) was 30.8 ± 20.1 μm. The error of Hoechst-based deformable registration of histopathologic images (acquired from sequential tissue sections) was 23.1 ± 17.9 μm. Total error of registration of autoradiography images to the histopathologic images acquired from adjacent sections was evaluated at 44.9 μm. This coregistration precision supersedes current rigid registration methods with reported errors of 100–200 μm. Conclusion: Deformable registration of autoradiography and histopathology images acquired from sequential sections is feasible and accurate when performed using corresponding Hoechst images.

Purpose: To utilize serial PET imaging it is important to minimize the effect of repositioning errors and anatomical changes. Compensating for these errors requires an objective and reliable method of deformable image registration. Here we report on the methodology used to deformably co‐register serial PET images using CT anatomy and compare intratumoral distributions of FDG and FLT as imaged in the same animal with PET/CT on two consecutive days. Methods: Nude mice bearing FaDu (human H&N) tumor xenografts were imaged with 18F‐FDG and 18F‐FLT on two consecutive days using a dedicated small animal PET/CT scanner (Siemens Inveon). Both data sets were reconstructed and loaded into Pinnacle 9.1. Despite careful repositioning of the animal using an animal‐specific pad with recorded landmarks, misalignment of FDG and FLT PET images hindered direct voxel‐by‐voxel analysis. To perform an objective co‐registration of the PET images, we relied on associated CT images. Animal bodies and tumors were contoured on both CT scans. A mesh was generated from the contours for visualization purposes. The CT images were deformably registered with a demons algorithm, and the resulting displacement vector field was applied to the FLT image, allowing for voxel‐by‐voxel analysis of co‐registered FLT and FDG PET images in Matlab. Results: Based off of visual inspection, the deformable image registration tools available in Pinnacle are adequate for co‐registration of the animal PET/CT images despite the deformations caused by repositioning. Voxel‐by‐voxel analysis of co‐registered FLT and FDG PET/CT images produced correlation coefficients ranging from .45 to .55, p<10–4. Conclusions: Using corresponding CT anatomy with the tools in Pinnacle 9.1 to generate deformation matrices is a viable approach to deformable PET image registration. The images produced facilitated a voxel‐by‐voxel comparison of FLT vs. FDG.

Purpose: To propose a new objective method for deformable coregistration of multimodality images acquired with digital autoradiography (DAR) and microscopy in the context of PET tracer histopathological validation. To analyze the spatial concordance between the uptake pattern of 18F-fluorothymidine (FLT) as imaged with DAR and the distribution of cell proliferation as revealed by immunofluorescence microscopy imaging.Methods: Tumor-bearing mice were injected with FLT and other markers including bromodeoxyuridine (cell proliferation). After sacrifice, tumors were excised, frozen and sectioned. Multiple stacks of sequential 8μm sections were collected from each tumor. Selected sections were used for DAR to image FLT uptake distribution. Adjacent sections were used to acquire histopathological data. To correct for imperfections of the tissue cutting and collection, all images were deformably coregistered to the FLT DAR image based on biological images of tumor blood flow (Hoechst) that was acquired from each tissue section used. For each FLT DAR — cell proliferation microscopy image pair, object-based analysis was conducted, including overlap and relative operating characteristics (ROC) analysis. Results: Total registration error of proposed coregistration method was 44.86μm. This supersedes current rigid registration methods with reported errors of 100–200μm. In tumors with well-compartmentalized functional aspects, area under the ROC curve (AUCroc= 0.7) indicated FLT DAR image thresholding as an accurate method of detecting cell proliferation. For these tumors, Dice overlap index indicated maximum detection rates at thresholds between 20% and 40% of the maximum DAR intensity. For the tumors characterized by more heterogeneous distribution of cell proliferation across the tumor section, FLT DAR image thresholding could not predict cell proliferation beyond random chance. Conclusions: We developed a comprehensive method of obtaining and analyzing coregistered images of cell proliferation markers and intratumoral uptake of FLT. Tumor microenvironment heterogeneity is a significant factor affecting the utility of FLT for imaging cell proliferation.