Image Reconstruction Protocols Research Articles

Sensitivity of an AI method for [18F]FDG PET/CT outcome prediction of diffuse large B-cell lymphoma patients to image reconstruction protocols

BackgroundConvolutional neural networks (CNNs), applied to baseline [18F]-FDG PET/CT maximum intensity projections (MIPs), show potential for treatment outcome prediction in diffuse large B-cell lymphoma (DLBCL). The aim of this study is to investigate the robustness of CNN predictions to different image reconstruction protocols. Baseline [18F]FDG PET/CT scans were collected from 20 DLBCL patients. EARL1, EARL2 and high-resolution (HR) protocols were applied per scan, generating three images with different image qualities. Image-based transformation was applied by blurring EARL2 and HR images to generate EARL1 compliant images using a Gaussian filter of 5 and 7 mm, respectively. MIPs were generated for each of the reconstructions, before and after image transformation. An in-house developed CNN predicted the probability of tumor progression within 2 years for each MIP. The difference in probabilities per patient was then calculated between both EARL2 and HR with respect to EARL1 (delta probabilities or ΔP). We compared these to the probabilities obtained after aligning the data with ComBat using the difference in median and interquartile range (IQR).ResultsCNN probabilities were found to be sensitive to different reconstruction protocols (EARL2 ΔP: median = 0.09, interquartile range (IQR) = [0.06, 0.10] and HR ΔP: median = 0.1, IQR = [0.08, 0.16]). Moreover, higher resolution images (EARL2 and HR) led to higher probability values. After image-based and ComBat transformation, an improved agreement of CNN probabilities among reconstructions was found for all patients. This agreement was slightly better after image-based transformation (transformed EARL2 ΔP: median = 0.022, IQR = [0.01, 0.02] and transformed HR ΔP: median = 0.029, IQR = [0.01, 0.03]).ConclusionOur CNN-based outcome predictions are affected by the applied reconstruction protocols, yet in a predictable manner. Image-based harmonization is a suitable approach to harmonize CNN predictions across image reconstruction protocols.

Influence of the resolution mode on mean and maximum SUV for PET images acquired by a LabPET SOLO 4 scanner

In preclinical PET, SUV (standardized uptake value) is a robust quantifier that can be used to analyze PET images. Several factors – biological or technical – can affect SUV determination. Among technical factors, it is possible to cite the reconstruction protocols of PET images. This work evaluated the influence of two resolution modes – standard and high – on mean and maximum SUVs. The PET image of a mouse with a tumor in left flank was chosen from a PET image bank and reconstructed using two different protocols varying the resolution mode. The post-processing of images was performed using AMIDE software and eight volumes of interest (VOIs) were defined. Qualitatively, there was light improvement in structures definition in high-resolution image compared to standard resolution image. At the semiquantitative analysis, image reconstruction protocols using high-resolution mode did not significantly improve the recuperation of radiopharmaceutical uptake into analyzed tissues.

Aiming for [18F]FDG-PET acquisition time reduction in clinical practice for neurological patients

PurposePositron emission tomography (PET) imaging with [18F]FDG provides valuable information regarding the underlying pathological processes in neurodegenerative disorders. PET imaging for these populations should be as short as possible to limit head movements and improve comfort. This study aimed to validate an optimized [18F]FDG-PET image reconstruction protocol aiming to reduce acquisition time while maintaining adequate quantification accuracy and image quality. MethodsA time-reduced reconstruction protocol (5 min) was evaluated in [18F]FDG-PET retrospective data from healthy individuals and Alzheimer’s disease (AD) patients. Standard (8 min) and time-reduced protocols were compared by means of image quality and quantification accuracy metrics, as well as standardized uptake value ratio (SUVR) and Z-scores (pons was used as reference). Images were randomly and blindly presented to experienced physicians and scored in terms of image quality. ResultsNo differences between protocols were identified during the visual assessment. Small differences (p < 0.01) in the pons SUVR were observed between the standard and time-reduced protocols for healthy individuals (−0.002 ± 0.011) and AD patients (−0.007 ± 0.013). Likewise, incorporating the PSF correction in the reconstruction algorithm resulted in small differences (p < 0.01) in SUVR between protocols (healthy individuals: −0.003 ± 0.011; AD patients: −0.007 ± 0.014). ConclusionQuality metrics were similar between time-reduced and standard protocols. In the visual assessment of the images, the physicians did not consider the use of PSF adequate, as it degraded the quality image. Shortening the acquisition time is possible by optimizing the image reconstruction parameters while maintaining adequate quantification accuracy and image quality.

Aortic valve imaging using 18F-sodium fluoride: impact of triple motion correction

BackgroundCurrent 18F-NaF assessments of aortic valve microcalcification using 18F-NaF PET/CT are based on evaluations of end-diastolic or cardiac motion-corrected (ECG-MC) images, which are affected by both patient and respiratory motion. We aimed to test the impact of employing a triple motion correction technique (3 × MC), including cardiorespiratory and gross patient motion, on quantitative and qualitative measurements.Materials and methodsFourteen patients with aortic stenosis underwent two repeat 30-min PET aortic valve scans within (29 ± 24) days. We considered three different image reconstruction protocols; an end-diastolic reconstruction protocol (standard) utilizing 25% of the acquired data, an ECG-gated (four ECG gates) reconstruction (ECG-MC), and a triple motion-corrected (3 × MC) dataset which corrects for both cardiorespiratory and patient motion. All datasets were compared to aortic valve calcification scores (AVCS), using the Agatston method, obtained from CT scans using correlation plots. We report SUVmax values measured in the aortic valve and maximum target-to-background ratios (TBRmax) values after correcting for blood pool activity.ResultsCompared to standard and ECG-MC reconstructions, increases in both SUVmax and TBRmax were observed following 3 × MC (SUVmax: Standard = 2.8 ± 0.7, ECG-MC = 2.6 ± 0.6, and 3 × MC = 3.3 ± 0.9; TBRmax: Standard = 2.7 ± 0.7, ECG-MC = 2.5 ± 0.6, and 3 × MC = 3.3 ± 1.2, all p values ≤ 0.05). 3 × MC had improved correlations (R2 value) to the AVCS when compared to the standard methods (SUVmax: Standard = 0.10, ECG-MC = 0.10, and 3 × MC = 0.20; TBRmax: Standard = 0.20, ECG-MC = 0.28, and 3 × MC = 0.46).Conclusion3 × MC improves the correlation between the AVCS and SUVmax and TBRmax and should be considered in PET studies of aortic valves using 18F-NaF.

Influence of image reconstruction protocols on the image quality of a small animal PET scanner using 18F and 11C

The small animal positron emission tomography (PET) scanner from Molecular Imaging Laboratory (LIM/CDTN) is dedicated to pre-clinical studies on new 18F and 11C-based radiopharmaceuticals and to development of novel applications for well-known radiopharmaceuticals. Thus, quality control tests recommended by the publication NEMA NU 4-2008 are routinely carried out to ensure the proper performance of PET scanner. The aim of this work was to evaluate the influence of image reconstruction protocols on the image quality, accuracy of attenuation and scatter corrections parameters for 18F and 11C PET images. PET images of the image quality phantom filed with 18F-FDG or 11C-PK1122 were acquired and then reconstructed using different reconstruction protocols. The reconstruction variables evaluated were the algorithms (FBP, MLEM-3D, OSEM-3D), the resolution mode (high/standard) and the number of iterations (10 to 150). Uniformity, spill-over ratio (SOR) and recovery coefficients (RC) tests were performed for each reconstructed image according NEMA NU 4-2008. PMOD software was used for image analysis. FBP based protocol generated noisier images compared to iterative algorithms (MLEM-3D or OSEM-3D) based protocols. The increase in the number of iterations resulted in higher standard deviation of the analyzed parameters for all reconstructed images. MLEM-3D and OSEM-3D based protocols generates similar results when number of iterations and resolution mode were identical. SOR and RC mean values remained stable when the number of iterations ranged from 40 to 150. This study allowed the evaluation of different image reconstruction protocols on important parameters of 18F and 11C PET image quality. Additionally, standard image reconstruction protocols to be adopted in LIM/CDTN laboratorial routine for 18F and 11C images reconstruction in preclinical studies were defined.

Intra-scanner repeatability of quantitative imaging features in a 3D printed semi-anthropomorphic CT phantom

ObjectivesRadiomics has shown to provide novel diagnostic and predictive disease information based on quantitative image features in study settings. However, limited data yielded contradictory results and important questions regarding the validity of the methods remain to be answered. The purpose of this study was to evaluate how clinical imaging techniques affect the stability of radiomics features by using 3D printed anthropomorphic CT phantom to test for repeatability and reproducibility of quantitative parameters. Methods48 PET/CT validated lymph nodes of prostate cancer patients (24 metastatic, 24 non-metastatic) were used as a template to create a customized 3D printed anthropomorphic phantom. We subsequently scanned the phantom five times with a routine abdominal CT protocol. Images were reconstructed using iterative reconstruction and two soft tissue kernels and one bone kernel. Radiomics features were extracted and assessed for repeatability and susceptibility towards image reconstruction settings using concordance correlation coefficients. ResultsOur analysis revealed 19 of 86 features (22 %) as highly repeatable (CCC ≥ 0.85) with low susceptibility towards image reconstruction protocols. Most features analyzed depicted critical non-repeatability with CCC’s < 0.75 even under entirely consistent imaging acquisition settings. Edge enhancing kernels result in higher variances between the scans and differences in repeatability and reproducibility were detected between PSMA-positive and negative lymph nodes with overall more stable features seen in tumor positive lymph nodes. ConclusionsBoth, repeatability and reproducibility play a crucial role in the validation process of radiomics features in clinical routine. This phantom study shows that most radiomics features in contrast to previous studies, including phantom and clinical, do not depict sufficient intra-scanner repeatability to serve as reliable diagnostic tools.

A core laboratory approach to large-scale radiomics and machine-learning prediction of DLBCL outcomes after first-line treatment using results from the phase III GOYA study.

e19042 Background: We are using the phase 3 GOYA study to retrospectively investigate using quantitative image texture features (i.e. radiomics, evaluating tumor heterogeneity) to improve the prognostic value of baseline FDG-PET. We report on the considerations for analysis, which is ongoing, and solutions to methodological challenges. Methods: GOYA, an open-label, randomized, phase 3 study, enrolled 1418 patients from 207 centers; 1334 had a baseline FDG-PET scan with detectable lesions. Image data with annotated regions of interest (ROIs) defined by qualified physicians were transferred to the Cuneo imaging core laboratory. Images were interpolated to a common 4x4x4 mm voxel size and ROI masks using 4 different definition methodologies were used to define image sub-volumes that contained the lesions. Large ROIs in normal liver were also defined. The 42 Haralick radiomic image texture features were computed using the PET oncology radiomics test suite. Results: A total of 9286 lesions were analyzed. Interpolation of images to a fixed size is challenging in this cohort due the large number of features to be computed (~1.5M). As a test, the lesions were differentiated from liver ROIs using 5 classifiers (e.g. support vector machine, random forest, etc) with an accuracy of 0.995–0.999. The mean (±SD) of the metabolically active tumor volume was 21.2±22.2 mL (range 0.2–146 mL) with a preponderance of smaller lesions as shown by the median of 15.0 mL. The skewed distribution had downstream effects on many of the quantitative (e.g. total lesion glycolysis) and radiomic image texture features. The majority of texture features were computable; 5 (Coarseness, Contrast, Busyness, Complexity, Texture Strength) required interpolation to a 2x2x2 mm voxel size, requiring extended computation time. FDG-PET images came from a variety of PET/CT scanners; thus, a unified image reconstruction protocol was not possible. Conclusions: Radiomics analysis of large phase 3 imaging trials is feasible. However, careful attention should be paid to the entire processing chain, including tests for non-computable features values due to highly skewed distributions and lesion size. Clinical trial information: NCT01287741.

Accurate, robust and harmonized implementation of morpho-functional imaging in treatment planning for personalized radiotherapy.

In this work we present a methodology able to use harmonized PET/CT imaging in dose painting by number (DPBN) approach by means of a robust and accurate treatment planning system. Image processing and treatment planning were performed by using a Matlab-based platform, called CARMEN, in which a full Monte Carlo simulation is included. Linear programming formulation was developed for a voxel-by-voxel robust optimization and a specific direct aperture optimization was designed for an efficient adaptive radiotherapy implementation. DPBN approach with our methodology was tested to reduce the uncertainties associated with both, the absolute value and the relative value of the information in the functional image. For the same H&N case, a single robust treatment was planned for dose prescription maps corresponding to standardized uptake value distributions from two different image reconstruction protocols: One to fulfill EARL accreditation for harmonization of [18F]FDG PET/CT image, and the other one to use the highest available spatial resolution. Also, a robust treatment was planned to fulfill dose prescription maps corresponding to both approaches, the dose painting by contour based on volumes and our voxel-by-voxel DPBN. Adaptive planning was also carried out to check the suitability of our proposal.Different plans showed robustness to cover a range of scenarios for implementation of harmonizing strategies by using the highest available resolution. Also, robustness associated to discretization level of dose prescription according to the use of contours or numbers was achieved. All plans showed excellent quality index histogram and quality factors below 2%. Efficient solution for adaptive radiotherapy based directly on changes in functional image was obtained. We proved that by using voxel-by-voxel DPBN approach it is possible to overcome typical drawbacks linked to PET/CT images, providing to the clinical specialist confidence enough for routinely implementation of functional imaging for personalized radiotherapy.

322. Evaluation of the performance parameters on the Ingenuity TF PET/CT scanner

Purpose To evaluate the performance characteristic of Philips Ingenuity TF PET/CT system with the new Pixelar detector design. A comparison of different image reconstruction protocols for EARL accreditation programme was performed. Methods Performance measurements on PET scanner were performed according to the NEMA NU2-2012 standard. Image quality was extended by accounting for different reconstruction parameters (frame timing, PSF iteration number, PSF regularization). A NEMA Image quality phantom was used for suitable image reconstruction protocol evaluation, according to EARL guideline. Performance measurements on CT scanner were performed according to IAEA No.19 publication. Geometrical efficiency was evaluated by exposure of various beam collimations on radiochromic films. Image quality parameters (MTF, Noise) using all reconstruction filters and iterative reconstruction (IDose) were obtained by Catphan®600 acquisitions. Influence of current modulation on patient dose (DoseRight) was studied using an anthropomorphic phantom. Results Spatial resolutions ranged from 4.6 mm at 1 cm to 6.0 mm at 20 cm. Sensitivity measured in the centre and at 10 cm were 7.58 and 7.43 cps/kBq, respectively. The measured NECR peak was 122.7 kcps at 20.3 kBq/cm3. The scatter fraction was 36.6%. Results of image quality test are reported in table. Manufacturer’s EARL dedicated reconstruction protocol did not perfectly meet accreditation requirements as others reconstructions with PSF. CT acceptance parameters met manufacturers and international guidelines specifications, except from geometrical efficiencies for 5 mm, 2.5 mm and 1 mm beam collimations. Decreasing DoseRight index from 26 to 24 and 16, dose reductions of 19.8% and 67.3% were observed. Conclusions Philips Ingenuity TF PET/CT system has overall good performance characteristics, comparable with similar scanners from other manufacturers. The CT part performance meets manufacturer’s specifications with some discrepancies, which do not have a significant impact on clinical use. Regarding EARL accreditation programme, a better adjustment of reconstruction protocol is needed.

Ptychotomography at DLS Coherence Beamline I13

We describe the implementation and execution of ptychotomography at I13-1, the coherence branchline at Diamond Light Source. The data collection and image reconstruction protocol is demonstrated with the three dimensional reconstruction of a nanoporous gold sample.

Classification and evaluation strategies of auto-segmentation approaches for PET: Report of AAPM task group No. 211.

The purpose of this educational report is to provide an overview of the present state-of-the-art PET auto-segmentation (PET-AS) algorithms and their respective validation, with an emphasis on providing the user with help in understanding the challenges and pitfalls associated with selecting and implementing a PET-AS algorithm for a particular application. A brief description of the different types of PET-AS algorithms is provided using a classification based on method complexity and type. The advantages and the limitations of the current PET-AS algorithms are highlighted based on current publications and existing comparison studies. A review of the available image datasets and contour evaluation metrics in terms of their applicability for establishing a standardized evaluation of PET-AS algorithms is provided. The performance requirements for the algorithms and their dependence on the application, the radiotracer used and the evaluation criteria are described and discussed. Finally, a procedure for algorithm acceptance and implementation, as well as the complementary role of manual and auto-segmentation are addressed. A large number of PET-AS algorithms have been developed within the last 20years. Many of the proposed algorithms are based on either fixed or adaptively selected thresholds. More recently, numerous papers have proposed the use of more advanced image analysis paradigms to perform semi-automated delineation of the PET images. However, the level of algorithm validation is variable and for most published algorithms is either insufficient or inconsistent which prevents recommending a single algorithm. This is compounded by the fact that realistic image configurations with low signal-to-noise ratios (SNR) and heterogeneous tracer distributions have rarely been used. Large variations in the evaluation methods used in the literature point to the need for a standardized evaluation protocol. Available comparison studies suggest that PET-AS algorithms relying on advanced image analysis paradigms provide generally more accurate segmentation than approaches based on PET activity thresholds, particularly for realistic configurations. However, this may not be the case for simple shape lesions insituations with a narrower range of parameters, where simpler methods may also perform well. Recent algorithms which employ some type of consensus or automatic selection between several PET-AS methods have potential to overcome the limitations of the individual methods when appropriately trained. In either case, accuracy evaluation is required for each different PET scanner and scanning and image reconstruction protocol. For the simpler, less robust approaches, adaptation to scanning conditions, tumor type, and tumor location by optimization of parameters is necessary. The results from the method evaluation stage can be used to estimate the contouring uncertainty. All PET-AS contours should be critically verified by a physician. A standard test, i.e., a benchmark dedicated to evaluating both existing and future PET-AS algorithms needs to be designed, to aid clinicians in evaluating and selecting PET-AS algorithms and to establish performance limits for their acceptance for clinical use. The initial steps toward designing and building such a standard are undertaken by the task group members.

P3. The influence of reconstruction parameters on PET image quality

Introduction The PET image quality is directly related to the reconstruction algorithm. The statistical reconstruction methods can improve the image quality by improving the spatial resolution. This may change the image contrast and the quantitative measurement of SUV (standardized uptake value) in small structures. The aim of this study is to determine the changes in SUV, image contrast, recovery coefficient and spatial resolution within the reconstruction parameters such as iterations and subsets number and also the applied corrections. Material and method The measurements of SUV (average and maximum), contrast, recovery coefficient and spatial resolution were obtained through a Jaszczak Deluxe phantom image, acquired on a PET-CT GE scanner Discovery 710®. The phantom was filled with a 18F-FDG solution of 22 MBq. These parameters were studied based on the corrections applied during the reconstruction, including the Point Spread Function (PSF) and the Time-Of-Flight (TOF) option. Comparisons have also been made in terms of number of iterations and subsets: OSEM (2.4 and 10 iterations, 1.12 and 24 subsets). Results Reconstruction results for 4 iterations and 24 subsets showed that the average SUV value, with TOF function is reduced by 5.2%, 2.1% and 4.2% compared to non-TOF, PSF and TOF-PSF reconstructions respectively. With PSF, the average SUV value is reduced by 3% and 2% compared to the non-PSF and TOF-PSF reconstructions respectively. However the average SUV value is increased by 2% compared to the reconstruction integrating the TOF function. The contrast measurement is reduced by 80%, 75% and 45% compared to the expected absolute value for the reconstructions with TOF, PSF and TOF-PSF respectively. The recovery coefficient is improved with the correction of TOF and PSF however improvements become less important when the diameter of the sphere is large. The spatial resolution with the full width at half maximum (FWHM) measured on the largest sphere is underestimated by 5.5%, 0.6% and 5.6% for TOF, PSF and TOF-PSF reconstructions respectively. Conclusion The reconstruction parameters can improve image quality, however any image reconstruction protocol should be validated before clinical routine or research use.

CT head-scan dosimetry in an anthropomorphic phantom and associated measurement of ACR accreditation-phantom imaging metrics under clinically representative scan conditions

To measure radiation absorbed dose and its distribution in an anthropomorphic head phantom under clinically representative scan conditions in three widely used computed tomography (CT) scanners, and to relate those dose values to metrics such as high-contrast resolution, noise, and contrast-to-noise ratio (CNR) in the American College of Radiology CT accreditation phantom. By inserting optically stimulated luminescence dosimeters (OSLDs) in the head of an anthropomorphic phantom specially developed for CT dosimetry (University of Florida, Gainesville), we measured dose with three commonly used scanners (GE Discovery CT750 HD, Siemens Definition, Philips Brilliance 64) at two different clinical sites (Walter Reed National Military Medical Center, National Institutes of Health). The scanners were set to operate with the same data-acquisition and image-reconstruction protocols as used clinically for typical head scans, respective of the practices of each facility for each scanner. We also analyzed images of the ACR CT accreditation phantom with the corresponding protocols. While the Siemens Definition and the Philips Brilliance protocols utilized only conventional, filtered back-projection (FBP) image-reconstruction methods, the GE Discovery also employed its particular version of an adaptive statistical iterative reconstruction (ASIR) algorithm that can be blended in desired proportions with the FBP algorithm. We did an objective image-metrics analysis evaluating the modulation transfer function (MTF), noise power spectrum (NPS), and CNR for images reconstructed with FBP. For images reconstructed with ASIR, we only analyzed the CNR, since MTF and NPS results are expected to depend on the object for iterative reconstruction algorithms. The OSLD measurements showed that the Siemens Definition and the Philips Brilliance scanners (located at two different clinical facilities) yield average absorbed doses in tissue of 42.6 and 43.1 mGy, respectively. The GE Discovery delivers about the same amount of dose (43.7 mGy) when run under similar operating and image-reconstruction conditions, i.e., without tube current modulation and ASIR. The image-metrics analysis likewise showed that the MTF, NPS, and CNR associated with the reconstructed images are mutually comparable when the three scanners are run with similar settings, and differences can be attributed to different edge-enhancement properties of the applied reconstruction filters. Moreover, when the GE scanner was operated with the facility's scanner settings for routine head exams, which apply 50% ASIR and use only approximately half of the 100%-FBP dose, the CNR of the images showed no significant change. Even though the CNR alone is not sufficient to characterize the image quality and justify any dose reduction claims, it can be useful as a constancy test metric. This work presents a straightforward method to connect direct measurements of CT dose with objective image metrics such as high-contrast resolution, noise, and CNR. It demonstrates that OSLD measurements in an anthropomorphic head phantom allow a realistic and locally precise estimation of magnitude and spatial distribution of dose in tissue delivered during a typical CT head scan. Additional objective analysis of the images of the ACR accreditation phantom can be used to relate the measured doses to high contrast resolution, noise, and CNR.

Iterative reconstruction in image space (IRIS) in cardiac computed tomography: initial experience

Improvements in image quality in cardiac computed tomography may be achieved through iterative image reconstruction techniques. We evaluated the ability of "Iterative Reconstruction in Image Space" (IRIS) reconstruction to reduce image noise and improve subjective image quality. 55 consecutive patients undergoing coronary CT angiography to rule out coronary artery stenosis were included. A dual source CT system and standard protocols were used. Images were reconstructed using standard filtered back projection and IRIS. Image noise, attenuation within the coronary arteries, contrast, signal to noise and contrast to noise parameters as well as subjective classification of image quality (using a scale with four categories) were evaluated and compared between the two image reconstruction protocols. Subjective image quality (2.8±0.4 in filtered back projection and 2.8±0.4 in iterative reconstruction) and the number of "evaluable" segments per patient 14.0±1.2 in filtered back projection and 14.1±1.1 in iterative reconstruction) were not significant different between the two methods. However iterative reconstruction had a lower image noise (22.6±4.5HU vs. 28.6±5.1HU) and higher signal to noise and image to noise ratios in the proximal coronary arteries. IRIS reduces image noise and contrast-to-noise ratio in coronary CT angiography, thus providing potential for reducing radiation exposure.

Quantitative Analysis of a Three-Dimensional Computed Tomography Image Rendering Protocol for Skull Base Applications

Objective: Three-dimensional (3D) volume-rendering image protocols are being utilized at our institution in the development of virtual surgical simulators for improving surgical planning and predicting operative exposures in complex skull base surgery. Image reconstruction protocol validation is essential to verify the accuracy and precision needed for clinical applications. Methods: Three cadaver heads were examined using fine-cut CT. Archived data were transferred to a workstation and, using a 3D visualization and modeling system, high-resolution 3D-CT volumetric reconstructions of the skull base were generated using a specific voxel threshold protocol. Based on conventional craniometric and skull base anatomical landmarks (n = 20), linear anthropomorphic distances (n = 20) were measured on 3D-CT reconstructions. Subsequently, soft tissues were removed from the cadaver heads and linear measurements were repeated manually using digital calipers. All measurements were taken three times, independently, yielding 120 total measurements. Results: No statistically significant difference (p > 0.01) was found between 3D-CT reconstruction and physical cadaver measurements for 19 of 20 linear distances (95%). 3D-CT volumetric reconstructions overestimated one anthropomorphic measurement (glabella to left porion; p = 0.0095) repeatedly. Conclusions: 3D-CT volume-rendered images of the skull base generated using our voxel threshold protocol is both accurate and precise. Virtual 3D skull base models generated using this protocol represent true skull base anatomy. This study represents an essential first step toward the development of a validated and accurate 3D surgical approach assessment model and skull base surgical simulator currently under development at our institution.