Faculty Bibliography

Purpose: Small field dosimetry is challenging but critical in modern radiation treatment. Various publications have provided k(fclin,fmsr) to convert detector readings to dose in homogenous medium. However, validity of k values with inhomogeneity is unknown which is presented in this study. Methods: Plastic scintillator detector, PSD-W1 (Standard Imaging) is an ideal detector with k(fclin,fmsr) = 1.0 in water. An specifically designed lung phantom consisting of 5 cm thick commercial grade lung slab sandwiched between cork and solid water slabs was used. A custom made central hole was provided in the slab accommodating the W1 detector. The lung phantom was scanned on a CT simulator. The images were sent to the Eclipse TPS for planning. A 6 MV beam was planned on this phantom in a series of single fields ranging from 0.5 x 0.5 to 10 x 10 cm2 using AAA and Acuros algorithms with 100 MU delivered to the W1 center at 100 cm SAD with 0.1 x 0.1 cm2 calculation grid in both algorithms. Measurements with the W1 detector were performed in the same geometry and beam parameters on a Varian Edge machine. Results: Dose calculated by AAA and Acuros deviates from that by the W1 measurement as the field size decreases, from below 2% difference at 10 x 10 cm2 to above 10% difference at very small field (0.5 x 0.5 cm2) for 6 MV beam. The Acuros calculated dose approaches more closely to that of the W1 than the AAA at field sizes larger than 1 cm, beyond which point the deviation for Acuros continues to increase, while that for AAA is reduced. Conclusion: Our study validates the accuracy of PSDW1 in an inhomogeneous medium and indicates that, Acuros provides a superior inhomogeneity correction in small field down to 2 x 2 cm2 within 3.6% which is limit of accuracy in small field measurements

Purpose: TBI treatment at our institution has moved from traditional hand calculation to CT-based planning to incorporate dose heterogeneities and organs at risk dose limits. The main objective of this work is to report our institutional experience with CT-based TBI and to show a comparison with the traditional approach. Methods: Ten patients were CT simulated supine with arms immobilized for lung shielding. Legs are separated to achieve a width similar to umbilicus separation; rice bags were placed between the legs for compensation. Four plans (P1, P2, P3 and P4) were created for each patient, all prescribed at midplane-umbilicus. The first three plans use lateral 15X beams, with head compensation. P1 was planned using a hand calculation. P2 includes heterogeneity corrections and inferior subfield to improve coverage. P3 includes heterogeneity corrections, inferior subfield, and adjustment of field weights to maintain coverage while keeping mean lung doses below 10.5 Gy (prescription dose 12 Gy). P4 uses AP-PA 6X beams. Dose to target (mean, max, D98%, D95%, min), mean lung and liver doses are calculated for all plans; reported doses when unitless and normalized to prescription dose. Results: Coverage of the target (Body-2 cm), indicated by D98% was 84.1 +/- 2.8, 84.7 +/- 3.9, 81.0 +/- 1.8, and 92.2 +/- 1.9 whereas the maximum doses were 123 +/- 5, 135 +/- 4, 129 +/- 4, and 124 +/- 5 for P1, P2, P3, and P4 respectively. The mean relative lung and liver doses were lowest for P3 with values of 87.8 +/- 0.5 and 89.8 +/- 3.4. The largest mean lung dose (12.5 Gy) was observed for P4 plan as expected, showing the necessity of using lung shielding. Conclusion: We are able to achieve target coverage of D98% >80%, keeping the mean lung and liver doses <90% of prescription using optimal arm positioning and subfields. This approach is easy to implement without the complexity of introducing lung shielding required with the use of 6X AP-PAbeams

PURPOSE: The aim of this paper is to define the requirements and describe the design and implementation of a standard benchmark tool for evaluation and validation of PET-auto-segmentation (PET-AS) algorithms. This work follows the recommendations of Task Group 211 (TG211) appointed by the American Association of Physicists in Medicine (AAPM). METHODS: The recommendations published in the AAPM TG211 report were used to derive a set of required features and to guide the design and structure of a benchmarking software tool. These items included the selection of appropriate representative data and reference contours obtained from established approaches and the description of available metrics. The benchmark was designed in a way that it could be extendable and open to the implementation of bespoke segmentation methods, while maintaining its main purpose of being a standard testing platform for newly developed PET-AS methods. An example of implementation of the proposed framework, named PETASset, was built. In this work, a selection of PET-AS methods representing common approaches to PET image segmentation were evaluated within PETASset for the purpose of testing and demonstrating the capabilities of the software as a benchmark platform. RESULTS: A selection of clinical, physical and simulated phantom data, including 'best estimates' reference contours from macroscopic specimens, simulation template, and CT scans was built into the PETASset application database. Specific metrics such as Dice Similarity Coefficient (DSC), Positive Predictive Value (PPV), and Sensitivity (S), were included to allow the user to compare the results of any given PET-AS algorithm to the reference contours. In addition, a tool to generate structured reports on the evaluation of the performance of PET-AS algorithms against the reference contours was built. The variation of the metric agreement values with the reference contours across the PET-AS methods evaluated for demonstration were between 0.51 and 0.83, 0.44 and 0.86, and 0.61 and 1.00 for DSC, PPV, and the S metric, respectively. Examples of agreement limits were provided to show how the software could be used to evaluate a new algorithm against the existing state-of-the art. CONCLUSIONS: PETASset provides a platform that allows standardizing the evaluation and comparison of different PET-AS methods on a wide range of PET datasets. The developed platform will be available to users willing to evaluate their PET-AS methods and contribute more evaluation datasets

Purpose: Despite advances in the multimodality care in esophageal cancers, in particular the combination of chemotherapy, radiotherapy and surgery, the 5-year overall survival rate remains only 15-34%. The ability to predict treatment response is therefore of great interest, with the potential to personalize cancer treatment. This project performs a comprehensive texture feature (TF) analysis from esophageal tumor [18F]FDG PET/CT images to establish their predictive value when compared with the PET Response Criteria in Solid Tumors (PERCIST). Methods: Pre/Post chemo-radiation treatment [18F] FDG PET/CT images in sixty five patients were analyzed retrospectively. In all patients, the lesions were identified using nuclear medicine reports. The images were rigid-registered using CERR (a computational environment for clinical research), and segmented using thresholding method. Fifty features, based on the intensity histogram, second and high-order matrices, were extracted from the segmented regions from both image sets. The relative difference of SUVmax and of each TF were used as surrogates for treatment response. One-way ANOVA model of the intra-class correlation coefficient (ICC) and Spearman's rank correlation coefficient (SC) were used to establish correlations between SUVmax and TF treatment response. Results: Relative differences for fifty features were correlated with the corresponding SUVmax based on their ICC values, which were found in the range from 0.4 to 0.6. Two second-order and three high-order feature presented ICC = 0.6 (p < 0.05). Spearman's rank correlation coefficient ranged from -0.7 to 0.7. Two second-order and six high-order features presented 0.5 <= SC <= 0.6 (p < 0.05). Conclusion: Features with high ICC and SC values can be potentially used as additional metrics for treatment assessment. Hence, metabolic texture feature response provides a feasible approach for evaluating clinical outcomes in esophageal chemo-radiation

BACKGROUND: Characterization of textural features (spatial distributions of image intensity levels) has been considered as a tool for automatic tumor segmentation. The purpose of this work is to study the variability of the textural features in PET images due to different acquisition modes and reconstruction parameters. MATERIAL AND METHODS: Twenty patients with solid tumors underwent PET/CT scans on a GE Discovery VCT scanner, 45-60 minutes post-injection of 10 mCi of [(18)F]FDG. Scans were acquired in both 2D and 3D modes. For each acquisition the raw PET data was reconstructed using five different reconstruction parameters. Lesions were segmented on a default image using the threshold of 40% of maximum SUV. Fifty different texture features were calculated inside the tumors. The range of variations of the features were calculated with respect to the average value. RESULTS: Fifty textural features were classified based on the range of variation in three categories: small, intermediate and large variability. Features with small variability (range 30%). CONCLUSION: Textural features such as entropy-first order, energy, maximal correlation coefficient, and low-gray level run emphasis exhibited small variations due to different acquisition modes and reconstruction parameters. Features with low level of variations are better candidates for reproducible tumor segmentation. Even though features such as contrast-NGTD, coarseness, homogeneity, and busyness have been previously used, our data indicated that these features presented large variations, therefore they could not be considered as a good candidates for tumor segmentation.