| Question 1: Very small regions of high uptake: |
| Reference: | Fessler, J.A. and W.L. Rogers, Spatial resolution properties of penalized-likelihood image reconstruction: space-invariant tomographs. Image Processing, IEEE Transactions on, 1996. 5(9): p. 1346-1358
Li S, Zhang J, Krol A, Schmidtlein CR, Vogelsang L, Shen L, et al. Effective noise-suppressed and artifact-reduced reconstruction of SPECT data using a preconditioned alternating projection algorithm. Medical physics. 2015;42(8):4872-87.
|
| Choice A: | Are problematic due to tradeoffs between noise over-fitting during reconstruction and resolution recovery. |
| Choice B: | Are accurately measured with maximum SUV. |
| Choice C: | Require more iterations to resolve their uptake distribution. |
| Choice D: | Both a and c are correct. |
| Question 2: Depth of interaction (DOI) transaxial resolution loss: |
| Reference: | W. W. Moses, “Fundamental limits of spatial resolution in PET,” Nucl. Instrum. Methods Phys. Res. Sect. A 648, S236–S240 (2011).
Schmidtlein CR, Turner JN, Thompson MO, Mandal KC, Häggström I, Zhang J, et al. Initial performance studies of a wearable brain positron emission tomography camera based on autonomous thin-film digital Geiger avalanche photodiode arrays. Journal of Medical Imaging. 2017;4(1):011003-.
|
| Choice A: | Increases in the radial direction with distance from the central axis. |
| Choice B: | Has little effect near the central axis of the scanner. |
| Choice C: | Results from coincidence photons detections being mispositioned due to detector penetration. |
| Choice D: | All of the above. |
| Question 3: Roughly speaking, an otherwise identical Time-of -Flight PET scanner with half the coincidence timing resolution (CTR, e.g. CTR= 300 ps compared to 600 ps) will have: |
| Reference: | S. Surti, “Update on time-of-flight PET imaging,” J. Nucl. Med. 56(1), 98–105 (2015).
Schmidtlein CR, Turner JN, Thompson MO, Mandal KC, Häggström I, Zhang J, et al. Initial performance studies of a wearable brain positron emission tomography camera based on autonomous thin-film digital Geiger avalanche photodiode arrays. Journal of Medical Imaging. 2017;4(1):011003.
|
| Choice A: | Double the effective noise equivalent counts for an equivalent scan. |
| Choice B: | Twice the sensitivity of the scanner with the larger CTR. |
| Choice C: | Will have improved intrinsic resolution in comparison to the scanner with larger CTR. |
| Choice D: | Will have data with the same statistical properties with respect to the reconstructed images. |
| Question 4: The main limitation of the fixed thresholding approach in radiotherapy applications is: |
| Reference: | Lee JA. Segmentation of positron emission tomography images: some recommendations for target delineation in radiation oncology. Radiother Oncol. 2010;96:302–307.
- Biehl KJ, Kong FM, Dehdashti F, et al. 18F-FDG PET definition of
gross tumor volume for radiotherapy of non-small cell lung cancer: is a
single standardized uptake value threshold approach appropriate? J
Nucl Med. 2006;47:1808–1812.
|
| Choice A: | In segmenting spherical lesions in phantom. |
| Choice B: | Its inter and intra-user reproducibility. |
| Choice C: | Its accuracy over a large range of cases. |
| Choice D: | Not for therapy assessment since a fixed threshold will provide consistent results in longitudinal studies. |
| Question 5: The main limiting factor(s) for application of gradient-based methods (e.g. Watershed algorithm) are: |
| Reference: | Geets X, Lee JA, Bol A, Lonneux M, Gregoire V. A gradient-based
method for segmenting FDG-PET images: methodology and validation.
Eur J Nucl Med Mol Imaging. 2007;34:1427–1438.
|
| Choice A: | Image noise, which could be reduced by image denoising. |
| Choice B: | Image blur, which could be reduced by image deblurring. |
| Choice C: | Both a and b. |
| Choice D: | None of the above. |
| Question 6: Segmentation methods based on supervised learning typically require the following: |
| Reference: | Zaidi H, El Naqa I. PET-guided delineation of radiation therapy treatment
volumes: a survey of image segmentation techniques. Eur J Nucl
Med Mol Imaging. 2010;37:2165–2187.
|
| Choice A: | Intensity of PET data converted into SUV. |
| Choice B: | Training and testing data with labeled contours (ground truth). |
| Choice C: | Knowledge of PET acquisition parameters. |
| Choice D: | Known number of clusters. |
| Question 7: Which criteria can strongly affect the segmented volume, but is often neglected in the evaluation of PET auto-segmentation methods (PET-AS) methods? |
| Reference: | Hatt M, Lee J, Schmidtlein CR, El Naqa I, Caldwell C, De Bernardi E, et al. Classification and evaluation strategies of auto‐segmentation approaches for PET: Report of AAPM Task Group No. 211. In Press. Med Phys. 2017. |
| Choice A: | Accuracy for physical phantoms acquisitions (spheres). |
| Choice B: | Robustness for heterogeneous uptake distributions and across acquisitions from different scanners and different imaging protocols. |
| Choice C: | Repeatability between consecutive runs. |
| Choice D: | Reproducibility between different operators. |
| Question 8: Which of the following is true about evaluation of PET-AS methods? |
| Reference: | Hatt M, Lee J, Schmidtlein CR, El Naqa I, Caldwell C, De Bernardi E, et al. Classification and evaluation strategies of auto‐segmentation approaches for PET: Report of AAPM Task Group No. 211. In Press. Med Phys. 2017. |
| Choice A: | PET-AS methods should be evaluated only against resolution corrected PET images . |
| Choice B: | Threshold based methods are more accurate than decision tree and consensus based approaches. |
| Choice C: | In addition to vendor suggested acceptance tests, further evaluation using phantoms with different complexity and clinical images is recommended. |
| Choice D: | Adaptive threshold methods should be used exactly as published. |
| Choice E: | The segmentation requirements for treatment planning, dose painting and treatment assessment are the same. |
| Question 9: The following metrics can be used for evaluating a tumor segmentation contour:
i) Volume difference
ii) Barycenter difference
iii) Sensitivity
iv) Positive predictive value (PPV)
v) Jaccard similarity coefficient
vi) Dice Similarity Coefficient (DSC)
vii) Hausdorff distance (HD)
If the ground truth volume is known the following combination is recommended:
|
| Reference: | Hatt M, Lee J, Schmidtlein CR, El Naqa I, Caldwell C, De Bernardi E, et al. Classification and evaluation strategies of auto‐segmentation approaches for PET: Report of AAPM Task Group No. 211. In Press. Med Phys. 2017. |
| Choice A: | i,ii,v |
| Choice B: | i,iii,vi |
| Choice C: | vi,vii |
| Choice D: | iii,iv,vii |
| Question 10: Which would be the most appropriate test to evaluate the robustness of a PET-AS method? |
| Reference: | Hatt M, Lee J, Schmidtlein CR, El Naqa I, Caldwell C, De Bernardi E, et al. Classification and evaluation strategies of auto‐segmentation approaches for PET: Report of AAPM Task Group No. 211. In Press. Med Phys. 2017. |
| Choice A: | Introduce differences in the same image by varying the image reconstruction and noise levels. |
| Choice B: | Introduce subtle changes in patient anatomy and physiology by varying the time of acquisition. |
| Choice C: | Introduce variations in GTV shape/size by looking at scans from different patients. |
