Imaging of pediatric thyroid tumors: A COG Diagnostic Imaging Committee/SPR Oncology Committee White Paper.

Pediatric thyroid cancer is rare in children; however, incidence is increasing. Papillary thyroid cancer and follicular thyroid cancer are the most common subtypes, comprising about 90% and 10% of cases, respectively. This paper provides consensus imaging recommendations for evaluation of pediatric patients with thyroid cancer at diagnosis and during follow-up.

Added Value of FDG PET/MRI in Gynecologic Oncology: A Pictorial Review.

Fluorine 18-fluorodeoxyglucose (FDG) PET and MRI independently play a valuable role in the management of patients with gynecologic malignancies, particularly endometrial and cervical cancer. The PET/MRI hybrid imaging technique combines the metabolic information obtained from PET with the excellent soft-tissue resolution and anatomic details provided by MRI in a single examination. MRI is the modality of choice for assessment of local tumor extent in the pelvis, whereas PET is used to assess for local-regional spread and distant metastases. The authors discuss the added value of FDG PET/MRI in imaging gynecologic malignancies of the pelvis, with a focus on the role of FDG PET/MRI in diagnosis, staging, assessing treatment response, and characterizing complications. PET/MRI allows better localization and demarcation of the extent of disease, characterization of lesions and involvement of adjacent organs and lymph nodes, and improved differentiation of benign from malignant tissues, as well as detection of the presence of distant metastasis. It also has the advantages of decreased radiation dose and a higher signal-to-noise ratio of a prolonged PET examination of the pelvis contemporaneous with MRI. The authors provide a brief technical overview of PET/MRI, highlight how simultaneously performed PET/MRI can improve stand-alone MRI and PET/CT in gynecologic malignancies, provide an image-rich review to illustrate practical and clinically relevant applications of this imaging technique, and review common pitfalls encountered in clinical practice. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material.

18 F-FDG positron emission tomography-computed tomography has a low positive predictive value for detecting occult recurrence in asymptomatic patients with high-risk Stages IIB, IIC, and IIIA melanoma.

BACKGROUND AND OBJECTIVES: High recurrence rates of Stages II and IIIA melanoma make close follow-up essential, especially with new adjuvant therapies for metastatic disease. However, there are currently no consensus guidelines for routine imaging for Stages IIB, IIC, and IIIA melanoma. The study's aim is to determine the utility of 18 F-fluorodeoxyglucose (18 F-FDG) positron emission tomography/computed tomography (PET/CT) for detecting asymptomatic recurrence of melanoma after primary surgical resection. METHODS: This retrospective cohort study included 158 patients with the American Joint Committee on Cancer 8th edition Stages IIB, IIC, or IIIA cutaneous melanoma who underwent an 18 F-FDG PET/CT from 2010 to 2020. We retrospectively analyzed clinical data after a median follow-up time of 39 months. RESULTS: We calculated a positive predictive value (PPV) of 32% (95% confidence interval: 11%-53%) for 154 routine PET/CTs, including six true positives and 13 false positives (FPs). PPV was 33% for Stage IIB, 50% for Stage IIC, and 14% for Stage IIIA. FPs were mostly benign or inflammatory foci (75%), and some other malignancies were found (21%). CONCLUSIONS: This cohort of patients imaged for high-risk melanoma demonstrated a high FP rate and low PPV. These findings suggest that routine surveillance with 18 F-FDG PET/CT may not be indicated for monitoring recurrence in this population.

Failure modes and effects analysis of pediatric I-131 MIBG therapy: Program design and potential pitfalls.

BACKGROUND: There is growing interest among pediatric institutions for implementing iodine-131 (I-131) meta-iodobenzylguanidine (MIBG) therapy for treating children with high-risk neuroblastoma. Due to regulations on the medical use of radioactive material (RAM), and the complexity and safety risks associated with the procedure, a multidisciplinary team involving radiation therapy/safety experts is required. Here, we describe methods for implementing pediatric I-131 MIBG therapy and evaluate our program's robustness via failure modes and effects analysis (FMEA). METHODS: We formed a multidisciplinary team, involving pediatric oncology, radiation oncology, and radiation safety staff. To evaluate the robustness of the therapy workflow and quantitatively assess potential safety risks, an FMEA was performed. Failure modes were scored (1-10) for their risk of occurrence (O), severity (S), and being undetected (D). Risk priority number (RPN) was calculated from a product of these scores and used to identify high-risk failure modes. RESULTS: A total of 176 failure modes were identified and scored. The majority (94%) of failure modes scored low (RPN <100). The highest risk failure modes were related to training and to drug-infusion procedures, with the highest S scores being (a) caregivers did not understand radiation safety training (O = 5.5, S = 7, D = 5.5, RPN = 212); (b) infusion training of staff was inadequate (O = 5, S = 8, D = 5, RPN = 200); and (c) air in intravenous lines/not monitoring for air in lines (O = 4.5, S = 8, D = 5, RPN = 180). CONCLUSION: Through use of FMEA methodology, we successfully identified multiple potential points of failure that have allowed us to proactively mitigate risks when implementing a pediatric MIBG program.

Comparison of radiation dose and image quality between contrast-enhanced single- and dual-energy abdominopelvic computed tomography in children as a function of patient size.

BACKGROUND: Widespread adoption of dual-energy computed tomography (DECT) requires evidence it does not cause higher radiation dose than conventional single-energy CT (SECT). While a few publications involving pediatric patients exist, most have focused on small cohorts. Hence, there is still a need for studies that ascertain what radiation doses are expected in larger populations that include representative ranges of patient sizes and ages. OBJECTIVE: To compare radiation dose and image quality of DECT and SECT abdominopelvic examinations in children as a function of patient size. MATERIALS AND METHODS: This retrospective study included 860 children (age range: 12.3±5.3 years) who underwent contrast-enhanced abdominopelvic exams on second-generation dual-source CT in a five-year period. Two groups, SECT and DECT, consisting of 430 children each, were matched by 5 effective diameters. Volume CT dose index (CTDIvol) and size-specific dose estimate (SSDE) were analyzed as a function of effective diameter. Objective image quality was compared between the groups. RESULTS: DECT SSDEs were lower across all effective patient diameters compared with SECT (mean: 8.5±1.8 mGv vs. 9.3±2.0 mGv, respectively, P≤0.001). DECT CTDIvol was lower compared to SECT (mean: 5.6±2.4 mGv vs. 6.1±2.7 mGv, respectively, P≤0.001) except in the smallest diameter group (<15 cm) where it was comparable to SECT (P=0.065). Objective image quality versus effective diameter between the two CT groups was comparable (P>0.05). CONCLUSION: In children, regardless of effective diameter, contrast-enhanced abdominopelvic DECT can be performed with a similar or lower dose and similar image quality compared with SECT examinations.

18F-FDG PET of the hands with a dedicated high-resolution PEM system (arthro-PET): correlation with PET/CT, radiography and clinical parameters.

PURPOSE: The aim of this study was to prospectively determine the feasibility and compare the novel use of a positron emission mammography (PEM) scanner with standard PET/CT for evaluating hand osteoarthritis (OA) with (18)F-FDG. METHODS: Institutional review board approval and written informed consent were obtained for this HIPAA-compliant prospective study in which 14 adults referred for oncological (18)F-FDG PET/CT underwent dedicated hand PET/CT followed by arthro-PET using the PEM device. Hand radiographs were obtained and scored for the presence and severity of OA. Summed qualitative and quantitative joint glycolytic scores for each modality were compared with the findings on plain radiography and clinical features. RESULTS: Eight patients with clinical and/or radiographic evidence of OA comprised the OA group (mean age 73 ± 7.7 years). Six patients served as the control group (53.7 ± 9.3 years). Arthro-PET quantitative and qualitative joint glycolytic scores were highly correlated with PET/CT findings in the OA patients (r = 0.86. p = 0.007; r = 0.94, p = 0.001). Qualitative arthro-PET and PET/CT joint scores were significantly higher in the OA patients than in controls (38.7 ± 6.6 vs. 32.2 ± 0.4, p = 0.02; 37.5 ± 5.4 vs. 32.2 ± 0.4, p = 0.03, respectively). Quantitative arthro-PET and PET/CT maximum SUV-lean joint scores were higher in the OA patients, although they did not reach statistical significance (20.8 ± 4.2 vs. 18 ± 1.8, p = 0.13; 22.8 ± 5.38 vs. 20.1 ± 1.54, p = 0.21). By definition, OA patients had higher radiographic joint scores than controls (30.9 ± 31.3 vs. 0, p = 0.03). CONCLUSION: Hand imaging using a small field of view PEM system (arthro-PET) with FDG is feasible, performing comparably to PET/CT in assessing metabolic joint activity. Arthro-PET and PET/CT showed higher joint FDG uptake in OA. Further exploration of arthro-PET in arthritis management is warranted.

Differentiation of HIV-associated lymphoma from HIV-associated reactive adenopathy using quantitative FDG PET and symmetry.

PURPOSE: To determine the diagnostic accuracy of a semiautomated (18)F-FDG PET/CT measurement of total lesion glycolysis (TLG), maximum and peak standardized uptake value at lean body mass (SUL-Max and SUL-Peak), qualitative estimates of left/right nodal symmetry and FDG uptake for differentiating lymphoma from reactive adenopathy in HIV-infected patients. METHODS: We retrospectively analyzed 41 whole-body (18)F-FDG PET/CT studies performed in HIV-infected patients for clinical reasons. The study received institutional review board approval. Of the 41 patients, 19 had biopsy-proven untreated lymphoma, and 22 with reactive adenopathy without malignancy on follow-up were used as controls. Nodal and extranodal visual qualitative metabolic scores, SUL-Max, SUL-Peak, CT nodal size, and PERCIST 1.0 threshold-based TLG and metabolic tumor volume (MTV) were determined. The qualitative intensity of nodal involvement and symmetry of uptake were compared using receiver operator curve (ROC) analysis. HIV plasma viral RNA measurements were also obtained. RESULTS: All of the quantitative PET metrics performed well in differentiating lymphoma from reactive adenopathy and performed better than qualitative visual intensity scores. The areas under the ROC curves (AUC) were significantly higher for TLG = 0.96, single SUL-Peak = 0.96, single SUL-Max = 0.97, and MTV = 0.96, compared to 0.67 for CT nodal size (p < 0.001). These PET metrics performed best in separating the two populations in aviremic patients, with AUCs of 1 (AUC 0.91 for CT nodal size). TLG, MTV, SUL-Peak and SUL-Max were more reliable markers among viremic individuals, with AUCs between 0.84 and 0.93, compared to other metrics. PET metrics were significantly correlated with plasma viral load in HIV-reactive adenopathy controls. Asymmetrical FDG uptake had an accuracy of 90.4 % for differentiating lymphoma from reactive adenopathy in HIV-infected patients. CONCLUSION: Quantitative PET metabolic metrics as well as the qualitative assessment of symmetry of nodal uptake appear to be valuable tools for differentiating lymphoma from reactive adenopathy in HIV-infected patients using FDG PET. These parameters appear more robust in aviremic patients.

Need for Objective Task-Based Evaluation of Image Segmentation Algorithms for Quantitative PET: A Study with ACRIN 6668/RTOG 0235 Multicenter Clinical Trial Data.

Reliable performance of PET segmentation algorithms on clinically relevant tasks is required for their clinical translation. However, these algorithms are typically evaluated using figures of merit (FoMs) that are not explicitly designed to correlate with clinical task performance. Such FoMs include the Dice similarity coefficient (DSC), the Jaccard similarity coefficient (JSC), and the Hausdorff distance (HD). The objective of this study was to investigate whether evaluating PET segmentation algorithms using these task-agnostic FoMs yields interpretations consistent with evaluation on clinically relevant quantitative tasks. Methods: We conducted a retrospective study to assess the concordance in the evaluation of segmentation algorithms using the DSC, JSC, and HD and on the tasks of estimating the metabolic tumor volume (MTV) and total lesion glycolysis (TLG) of primary tumors from PET images of patients with non-small cell lung cancer. The PET images were collected from the American College of Radiology Imaging Network 6668/Radiation Therapy Oncology Group 0235 multicenter clinical trial data. The study was conducted in 2 contexts: (1) evaluating conventional segmentation algorithms, namely those based on thresholding (SUVmax40% and SUVmax50%), boundary detection (Snakes), and stochastic modeling (Markov random field-Gaussian mixture model); (2) evaluating the impact of network depth and loss function on the performance of a state-of-the-art U-net-based segmentation algorithm. Results: Evaluation of conventional segmentation algorithms based on the DSC, JSC, and HD showed that SUVmax40% significantly outperformed SUVmax50%. However, SUVmax40% yielded lower accuracy on the tasks of estimating MTV and TLG, with a 51% and 54% increase, respectively, in the ensemble normalized bias. Similarly, the Markov random field-Gaussian mixture model significantly outperformed Snakes on the basis of the task-agnostic FoMs but yielded a 24% increased bias in estimated MTV. For the U-net-based algorithm, our evaluation showed that although the network depth did not significantly alter the DSC, JSC, and HD values, a deeper network yielded substantially higher accuracy in the estimated MTV and TLG, with a decreased bias of 91% and 87%, respectively. Additionally, whereas there was no significant difference in the DSC, JSC, and HD values for different loss functions, up to a 73% and 58% difference in the bias of the estimated MTV and TLG, respectively, existed. Conclusion: Evaluation of PET segmentation algorithms using task-agnostic FoMs could yield findings discordant with evaluation on clinically relevant quantitative tasks. This study emphasizes the need for objective task-based evaluation of image segmentation algorithms for quantitative PET.

The role of PET/CT in the management of cervical cancer.

OBJECTIVE. Cervical cancer is the second most common malignancy in women worldwide and the third most common cause of cancer mortality in the United States. The aim of this article is to describe cervical cancer and outline the value of (18)F-FDG PET/CT in the management of cervical malignancy. CONCLUSION. The value of PET/CT has been found in staging and treatment strategy for cervical cancer. FDG PET/CT facilitates decision-making and radiation treatment planning and provides important information about treatment response, disease recurrence, and long-term survival.

Need for objective task-based evaluation of AI-based segmentation methods for quantitative PET.

Artificial intelligence (AI)-based methods are showing substantial promise in segmenting oncologic positron emission tomography (PET) images. For clinical translation of these methods, assessing their performance on clinically relevant tasks is important. However, these methods are typically evaluated using metrics that may not correlate with the task performance. One such widely used metric is the Dice score, a figure of merit that measures the spatial overlap between the estimated segmentation and a reference standard (e.g., manual segmentation). In this work, we investigated whether evaluating AI-based segmentation methods using Dice scores yields a similar interpretation as evaluation on the clinical tasks of quantifying metabolic tumor volume (MTV) and total lesion glycolysis (TLG) of primary tumor from PET images of patients with non-small cell lung cancer. The investigation was conducted via a retrospective analysis with the ECOG-ACRIN 6668/RTOG 0235 multi-center clinical trial data. Specifically, we evaluated different structures of a commonly used AI-based segmentation method using both Dice scores and the accuracy in quantifying MTV/TLG. Our results show that evaluation using Dice scores can lead to findings that are inconsistent with evaluation using the task-based figure of merit. Thus, our study motivates the need for objective task-based evaluation of AI-based segmentation methods for quantitative PET.

Need for objective task-based evaluation of AI-based segmentation methods for quantitative PET.

Artificial intelligence (AI)-based methods are showing substantial promise in segmenting oncologic positron emission tomography (PET) images. For clinical translation of these methods, assessing their performance on clinically relevant tasks is important. However, these methods are typically evaluated using metrics that may not correlate with the task performance. One such widely used metric is the Dice score, a figure of merit that measures the spatial overlap between the estimated segmentation and a reference standard (e.g., manual segmentation). In this work, we investigated whether evaluating AI-based segmentation methods using Dice scores yields a similar interpretation as evaluation on the clinical tasks of quantifying metabolic tumor volume (MTV) and total lesion glycolysis (TLG) of primary tumor from PET images of patients with non-small cell lung cancer. The investigation was conducted via a retrospective analysis with the ECOG-ACRIN 6668/RTOG 0235 multi-center clinical trial data. Specifically, we evaluated different structures of a commonly used AI-based segmentation method using both Dice scores and the accuracy in quantifying MTV/TLG. Our results show that evaluation using Dice scores can lead to findings that are inconsistent with evaluation using the task-based figure of merit. Thus, our study motivates the need for objective task-based evaluation of AI-based segmentation methods for quantitative PET.

Observer-study-based approaches to quantitatively evaluate the realism of synthetic medical images.

Objective.Synthetic images generated by simulation studies have a well-recognized role in developing and evaluating imaging systems and methods. However, for clinically relevant development and evaluation, the synthetic images must be clinically realistic and, ideally, have the same distribution as that of clinical images. Thus, mechanisms that can quantitatively evaluate this clinical realism and, ideally, the similarity in distributions of the real and synthetic images, are much needed.Approach.We investigated two observer-study-based approaches to quantitatively evaluate the clinical realism of synthetic images. In the first approach, we presented a theoretical formalism for the use of an ideal-observer study to quantitatively evaluate the similarity in distributions between the real and synthetic images. This theoretical formalism provides a direct relationship between the area under the receiver operating characteristic curve, AUC, for an ideal observer and the distributions of real and synthetic images. The second approach is based on the use of expert-human-observer studies to quantitatively evaluate the realism of synthetic images. In this approach, we developed a web-based software to conduct two-alternative forced-choice (2-AFC) experiments with expert human observers. The usability of this software was evaluated by conducting a system usability scale (SUS) survey with seven expert human readers and five observer-study designers. Further, we demonstrated the application of this software to evaluate a stochastic and physics-based image-synthesis technique for oncologic positron emission tomography (PET). In this evaluation, the 2-AFC study with our software was performed by six expert human readers, who were highly experienced in reading PET scans, with years of expertise ranging from 7 to 40 years (median: 12 years, average: 20.4 years).Main results.In the ideal-observer-study-based approach, we theoretically demonstrated that the AUC for an ideal observer can be expressed, to an excellent approximation, by the Bhattacharyya distance between the distributions of the real and synthetic images. This relationship shows that a decrease in the ideal-observer AUC indicates a decrease in the distance between the two image distributions. Moreover, a lower bound of ideal-observer AUC = 0.5 implies that the distributions of synthetic and real images exactly match. For the expert-human-observer-study-based approach, our software for performing the 2-AFC experiments is available athttps://apps.mir.wustl.edu/twoafc. Results from the SUS survey demonstrate that the web application is very user friendly and accessible. As a secondary finding, evaluation of a stochastic and physics-based PET image-synthesis technique using our software showed that expert human readers had limited ability to distinguish the real images from the synthetic images.Significance.This work addresses the important need for mechanisms to quantitatively evaluate the clinical realism of synthetic images. The mathematical treatment in this paper shows that quantifying the similarity in the distribution of real and synthetic images is theoretically possible by using an ideal-observer-study-based approach. Our developed software provides a platform for designing and performing 2-AFC experiments with human observers in a highly accessible, efficient, and secure manner. Additionally, our results on the evaluation of the stochastic and physics-based image-synthesis technique motivate the application of this technique to develop and evaluate a wide array of PET imaging methods.

Importance of Central Imaging Review in a Pediatric Hodgkin Lymphoma Trial Using Positron Emission Tomography Response Adapted Radiation Therapy.

PURPOSE: We investigated the effects of central review of the interim fluorodeoxyglucose-positron emission tomography/computed tomography (FDG-PET/CT) scan response (iPET) assessment on treatment allocation in the risk-based, response-adapted, Children's Oncology Group study AHOD1331 (ClinicalTrials.gov identifier: NCT02166463) for pediatric patients with high-risk Hodgkin lymphoma. METHODS AND MATERIALS: Per protocol, after 2 cycles of systemic therapy, patients underwent iPET, with visual response assessment by 5-point Deauville score (DS) at their treating institution and a real-time central review, with the latter considered the reference standard. An area of disease with a DS of 1 to 3 was considered a rapid-responding lesion, whereas a DS of 4 to 5 was considered a slow-responding lesion (SRL). Patients with 1 or more SRLs were considered iPET positive, whereas patients with only rapid-responding lesions were considered iPET negative. We conducted a predefined exploratory evaluation of concordance in iPET response assessment between institutional and central reviews of 573 patients. The concordance rate was evaluated using the Cohen κ statistic (κ > 0.80 was considered very good agreement and κ > 0.60-0.80, good agreement). RESULTS: The concordance rate (514 of 573 [89.7%]) had a κ of 0.685 (95% CI, 0.610-0.759), consistent with good agreement. In terms of the direction of discordance, among the 126 patients who were considered iPET positive by institutional review, 38 (30.2%) were categorized as iPET negative by central review, preventing overtreatment with radiation therapy. Conversely, among the 447 patients who were considered iPET negative by institutional review, 21 patients (4.7%) were categorized as iPET positive by the central review and would have been undertreated without radiation therapy. CONCLUSIONS: Central review is integral to PET response-adapted clinical trials for children with Hodgkin lymphoma. Continued support of central imaging review and education about DS are needed.

Tazemetostat for tumors harboring SMARCB1/SMARCA4 or EZH2 alterations: results from NCI-COG pediatric MATCH APEC1621C.

BACKGROUND: National Cancer Institute-Children's Oncology Group Pediatric Molecular Analysis for Therapy Choice assigns patients aged 1-21 years with refractory solid tumors, brain tumors, lymphomas, and histiocytic disorders to phase II trials of molecularly targeted therapies based on detection of predefined genetic alterations. Patients whose tumors harbored EZH2 mutations or loss of SMARCB1 or SMARCA4 by immunohistochemistry were treated with EZH2 inhibitor tazemetostat. METHODS: Patients received tazemetostat for 28-day cycles until disease progression or intolerable toxicity (max 26 cycles). The primary endpoint was objective response rate; secondary endpoints included progression-free survival and tolerability of tazemetostat. RESULTS: Twenty patients (median age = 5 years) enrolled, all evaluable for response and toxicities. The most frequent diagnoses were atypical teratoid rhabdoid tumor (n = 8) and malignant rhabdoid tumor (n = 4). Actionable alterations consisted of SMARCB1 loss (n = 16), EZH2 mutation (n = 3), and SMARCA4 loss (n = 1). One objective response was observed in a patient with non-Langerhans cell histiocytosis with SMARCA4 loss (26 cycles, 1200 mg/m2/dose twice daily). Four patients with SMARCB1 loss had a best response of stable disease: epithelioid sarcoma (n = 2), atypical teratoid rhabdoid tumor (n = 1), and renal medullary carcinoma (n = 1). Six-month progression-free survival was 35% (95% confidence interval [CI] = 15.7% to 55.2%) and 6-month overall survival was 45% (95% CI = 23.1% to 64.7%). Treatment-related adverse events were consistent with prior tazemetostat reports. CONCLUSIONS: Although tazemetostat did not meet its primary efficacy endpoint in this population of refractory pediatric tumors (objective response rate = 5%, 90% CI = 1% to 20%), 25% of patients with multiple histologic diagnoses experienced prolonged stable disease of 6 months and over (range = 9-26 cycles), suggesting a potential effect of tazemetostat on disease stabilization.

No-gold-standard evaluation of quantitative imaging methods in the presence of correlated noise.

Objective evaluation of quantitative imaging (QI) methods with patient data is highly desirable, but is hindered by the lack or unreliability of an available gold standard. To address this issue, techniques that can evaluate QI methods without access to a gold standard are being actively developed. These techniques assume that the true and measured values are linearly related by a slope, bias, and Gaussian-distributed noise term, where the noise between measurements made by different methods is independent of each other. However, this noise arises in the process of measuring the same quantitative value, and thus can be correlated. To address this limitation, we propose a no-gold-standard evaluation (NGSE) technique that models this correlated noise by a multi-variate Gaussian distribution parameterized by a covariance matrix. We derive a maximum-likelihood-based approach to estimate the parameters that describe the relationship between the true and measured values, without any knowledge of the true values. We then use the estimated slopes and diagonal elements of the covariance matrix to compute the noise-to-slope ratio (NSR) to rank the QI methods on the basis of precision. The proposed NGSE technique was evaluated with multiple numerical experiments. Our results showed that the technique reliably estimated the NSR values and yielded accurate rankings of the considered methods for 83% of 160 trials. In particular, the technique correctly identified the most precise method for ∼ 97% of the trials. Overall, this study demonstrates the efficacy of the NGSE technique to accurately rank different QI methods when correlated noise is present, and without access to any knowledge of the ground truth. The results motivate further validation of this technique with realistic simulation studies and patient data.

A Bayesian approach to tissue-fraction estimation for oncological PET segmentation.

Tumor segmentation in oncological PET is challenging, a major reason being the partial-volume effects (PVEs) that arise due to low system resolution and finite voxel size. The latter results in tissue-fraction effects (TFEs), i.e. voxels contain a mixture of tissue classes. Conventional segmentation methods are typically designed to assign each image voxel as belonging to a certain tissue class. Thus, these methods are inherently limited in modeling TFEs. To address the challenge of accounting for PVEs, and in particular, TFEs, we propose a Bayesian approach to tissue-fraction estimation for oncological PET segmentation. Specifically, this Bayesian approach estimates the posterior mean of the fractional volume that the tumor occupies within each image voxel. The proposed method, implemented using a deep-learning-based technique, was first evaluated using clinically realistic 2D simulation studies with known ground truth, in the context of segmenting the primary tumor in PET images of patients with lung cancer. The evaluation studies demonstrated that the method accurately estimated the tumor-fraction areas and significantly outperformed widely used conventional PET segmentation methods, including a U-net-based method, on the task of segmenting the tumor. In addition, the proposed method was relatively insensitive to PVEs and yielded reliable tumor segmentation for different clinical-scanner configurations. The method was then evaluated using clinical images of patients with stage IIB/III non-small cell lung cancer from ACRIN 6668/RTOG 0235 multi-center clinical trial. Here, the results showed that the proposed method significantly outperformed all other considered methods and yielded accurate tumor segmentation on patient images with Dice similarity coefficient (DSC) of 0.82 (95% CI: 0.78, 0.86). In particular, the method accurately segmented relatively small tumors, yielding a high DSC of 0.77 for the smallest segmented cross-section of 1.30 cm2. Overall, this study demonstrates the efficacy of the proposed method to accurately segment tumors in PET images.

Quantitative PET/CT in clinical practice: assessing the agreement of PET tumor indices using different clinical reading platforms.

OBJECTIVE: The aim of this study was to determine whether various fluorine-18-fluorodeoxyglucose PET/CT-derived parameters used in oncology vary significantly depending on the interpretation software systems used in clinical practice for multiple human solid tumors. PATIENTS AND METHODS: A total of 120 fluorine-18-fluorodeoxyglucose PET/CT studies carried out in patients with pancreatic, lung, colorectal, and head and neck cancers were evaluated retrospectively on two different vendor software platforms including Mirada and MIMVista. Regions of interest were placed on the liver to determine the liver mean standardized uptake value at lean body mass (SUL) and on each tumor to determine the SULmax, SULpeak. Total lesion glycolysis (TLG) and metabolic tumor volume (MTV) were determined using fixed thresholds of 50% of SULmax and SULpeak. Inter-reader, intersystem intraclass correlations, systematic bias, and variability reflected by the 95% limits of agreement, and precision were determined. RESULTS: There was excellent inter-reader reliability between the readers and the two software systems, with intraclass correlations more than 0.9 for all PET metrics, with P values less than 0.0001. The bias and SD on Bland-Altman analysis between the two software platforms for tumor SULmax, SULpeak, Max50MTV, and Peak50MTV, respectively, for Reader 1 were -1.52±2.24, 0.80±3.67, -0.80±13.01, and -4.49±20.6. For Reader 2, the biases were -1.62±1.95, 0.18±3.60, -0.27±4.64, and -3.13±8.30. The precision between the two systems was better for SULmax and SULpeak, with less variance observed, than for volume-based metrics such as Max50MTV and Peak50MTV or TLG. CONCLUSION: Excellent correlation has been found between two tested software reading platforms for all PET-derived metrics in a dual-reader analysis. Overall, the SULmax and SULpeak values had less bias and better precision compared with the MTV and TLG.

Effect of patient arm motion in whole-body PET/CT.

UNLABELLED: Arm motion during whole-body PET/CT acquisition is not uncommon and can give rise to striking cold artifacts on PET images. We investigated the mechanisms that underlie these artifacts and proposed a potential solution. METHODS: A phantom experiment based on 5 clinical cases of suspected arm motion was designed. The experiment involved a central 20-cm-diameter (68)Ge/(68)Ga cylinder simulating the neck and 2 peripheral 10-cm-diameter (18)F cylinders simulating arms. After motion-free CT and PET on a whole-body PET/CT system, the position of the arms was altered so as to introduce different amounts of misalignment. Twenty sequential PET scans were acquired in this position, alternating between 2-dimensional (2D) and 3-dimensional (3D) acquisition, as the (18)F decayed. Decay of (18)F in the arms, while the activity in the (68)Ge/(68)Ga cylinder remained approximately constant, allowed the relative impact of scatter and attenuation-correction errors to be determined. RESULTS: Image artifacts were largely confined to the local region of motion in 2D but extended throughout the affected slices in 3D, where they manifested as a striking underestimation of radiotracer concentration that became more significant with increasing misalignment. For 3D, scatter-correction error depended on activity in the arms, but for typical activity concentrations scatter-correction error was more significant than attenuation-correction error. 3D image reconstruction without scatter correction substantially eliminated these artifacts in both phantom and patient images. CONCLUSION: Reconstruction artifacts due to patient arm motion can be substantial and should be recognized because they can affect both qualitative and quantitative assessment of PET.