ATLAAS: an automatic decision tree-based learning algorithm for advanced image segmentation in positron emission tomography

Implications of the Harmonization of [18F]FDG-PET/CT Imaging for Response Assessment of Treatment in Radiotherapy Planning

The purpose of this work is to present useful recommendations for the use of [¹⁸F]FDG-PET/CT imaging in radiotherapy planning and monitoring under different versions of EARL accreditation for harmonization of PET devices. A proof-of-concept experiment designed on an anthropomorphic phantom was carried out to establish the most suitable interpolation methods of the PET images in the different steps of the planning procedure. Based on PET/CT images obtained by using these optimal interpolations for the old EARL accreditation (EARL1) and for the new one (EARL2), the treatment plannings of representative actual clinical cases were calculated, and the clinical implications of the resulting differences were analyzed. As expected, EARL2 provided smaller volumes with higher resolution than EARL1. The increase in the size of the reconstructed volumes with EARL1 accreditation caused high doses in the organs at risk and in the regions adjacent to the target volumes. EARL2 accreditation allowed an improvement in the accuracy of the PET imaging precision, allowing more personalized radiotherapy. This work provides recommendations for those centers that intend to benefit from the new accreditation, EARL2, and can help build confidence of those that must continue working under the EARL1 accreditation.

Online Learning Behavior Feature Mining Method Based on Decision Tree

This research mainly discusses the design of online learning behavior feature mining method based on decision tree. Data collection is the real-time collection of online learning behavior data from distance learning websites. OWC (Office Web Component) technology is used to draw real-time charts on the page. Online learning students are selected as the research object, and the student's system log data and questionnaire data are selected. When combining the pre-pruning method and the post-pruning method to make decisions after the tree is pruned, the same source data is used to adjust, test and evaluate the decision tree model. The evaluation process to generate a complete decision tree is completed by the c4.5tree algorithm in C4.5, which can be named with a suffix of .names The type definition file is used to record the type of each attribute item or the range of possible values. In the study, the prediction accuracy rate of predicting learning effect based on "online learning behavior" reached more than 66%.

Joint EANM/SNMMI/ESTRO practice recommendations for the use of 2-[18F]FDG PET/CT external beam radiation treatment planning in lung cancer V1.0

Purpose

2-[¹⁸F]FDG PET/CT is of utmost importance for radiation treatment (RT) planning and response monitoring in lung cancer patients, in both non-small and small cell lung cancer (NSCLC and SCLC). This topic has been addressed in guidelines composed by experts within the field of radiation oncology. However, up to present, there is no procedural guideline on this subject, with involvement of the nuclear medicine societies.

Methods

A literature review was performed, followed by a discussion between a multidisciplinary team of experts in the different fields involved in the RT planning of lung cancer, in order to guide clinical management. The project was led by experts of the two nuclear medicine societies (EANM and SNMMI) and radiation oncology (ESTRO).

Results and conclusion

This guideline results from a joint and dynamic collaboration between the relevant disciplines for this topic. It provides a worldwide, state of the art, and multidisciplinary guide to 2-[¹⁸F]FDG PET/CT RT planning in NSCLC and SCLC. These practical recommendations describe applicable updates for existing clinical practices, highlight potential flaws, and provide solutions to overcome these as well. Finally, the recent developments considered for future application are also reviewed.

Artificial intelligence in radiation oncology: A review of its current status and potential application for the radiotherapy workforce

OBJECTIVE

Radiation oncology is a continually evolving speciality. With the development of new imaging modalities and advanced imaging processing techniques, there is an increasing amount of data available to practitioners. In this narrative review, Artificial Intelligence (AI) is used as a reference to machine learning, and its potential, along with current problems in the field of radiation oncology, are considered from a technical position.

KEY FINDINGS

AI has the potential to harness the availability of data for improving patient outcomes, reducing toxicity, and easing clinical burdens. However, problems including the requirement of complexity of data, undefined core outcomes and limited generalisability are apparent.

CONCLUSION

This original review highlights considerations for the radiotherapy workforce, particularly therapeutic radiographers, as there will be an increasing requirement for their familiarity with AI due to their unique position as the interface between imaging technology and patients.

IMPLICATIONS FOR PRACTICE

Collaboration between AI experts and the radiotherapy workforce are required to overcome current issues before clinical adoption. The development of educational resources and standardised reporting of AI studies may help facilitate this.

A review of the application of machine learning in molecular imaging

Molecular imaging (MI) is a science that uses imaging methods to reflect the changes of molecular level in living state and conduct qualitative and quantitative studies on its biological behaviors in imaging. Optical molecular imaging (OMI) and nuclear medical imaging are two key research fields of MI. OMI technology refers to the optical information generated by the imaging target (such as tumors) due to drug intervention and other reasons. By collecting the optical information, researchers can track the motion trajectory of the imaging target at the molecular level. Owing to its high specificity and sensitivity, OMI has been widely used in preclinical research and clinical surgery. Nuclear medical imaging mainly detects ionizing radiation emitted by radioactive substances. It can provide molecular information for early diagnosis, effective treatment and basic research of diseases, which has become one of the frontiers and hot topics in the field of medicine in the world today. Both OMI and nuclear medical imaging technology require a lot of data processing and analysis. In recent years, artificial intelligence technology, especially neural network-based machine learning (ML) technology, has been widely used in MI because of its powerful data processing capability. It provides a feasible strategy to deal with large and complex data for the requirement of MI. In this review, we will focus on the applications of ML methods in OMI and nuclear medical imaging.

Artificial intelligence for molecular neuroimaging

In recent years, artificial intelligence (AI) or the study of how computers and machines can gain intelligence, has been increasingly applied to problems in medical imaging, and in particular to molecular imaging of the central nervous system. Many AI innovations in medical imaging include improving image quality, segmentation, and automating classification of disease. These advances have led to an increased availability of supportive AI tools to assist physicians in interpreting images and making decisions affecting patient care. This review focuses on the role of AI in molecular neuroimaging, primarily applied to positron emission tomography (PET) and single photon emission computed tomography (SPECT). We emphasize technical innovations such as AI in computed tomography (CT) generation for the purposes of attenuation correction and disease localization, as well as applications in neuro-oncology and neurodegenerative diseases. Limitations and future prospects for AI in molecular brain imaging are also discussed. Just as new equipment such as SPECT and PET revolutionized the field of medical imaging a few decades ago, AI and its related technologies are now poised to bring on further disruptive changes. An understanding of these new technologies and how they work will help physicians adapt their practices and succeed with these new tools.

Interobserver agreement in automated metabolic tumor volume measurements of Deauville score 4 and 5 lesions at interim 18F-FDG PET in DLBCL

Introduction: Metabolic tumor volume (MTV) on interim-PET (I-PET) is a potential prognostic biomarker for diffuse large B-cell lymphoma (DLBCL). Implementation of MTV on I-PET requires consensus which semi-automated segmentation method delineates lesions most successfully with least user interaction. Methods used for baseline PET are not necessarily optimal for I-PET due to lower lesional standardized uptake values (SUV) at I-PET. Therefore, we aimed to evaluate which method provides the best delineation quality of Deauville-score (DS) 4-5 DLBCL lesions on I-PET at best interobserver agreement on delineation quality and, secondly, to assess the effect of lesional SUV_max on delineation quality and performance agreements. Methods: DS4-5 lesions from 45 I-PET scans were delineated using six semi-automated methods i) SUV 2.5, ii) SUV 4.0, iii) adaptive threshold [A50%peak], iv) 41% of maximum SUV [41%max], v) majority vote including voxels detected by ≥2 methods [MV2] and vi) detected by ≥3 methods [MV3]. Delineation quality per MTV was rated by three independent observers as acceptable or non-acceptable. For each method, observer scores on delineation quality, specific agreements and MTV were assessed for all lesions, and per category of lesional SUV_max (<5, 5-10, >10). Results: In 60 DS4-5 lesions on I-PET, MV3 performed best, with acceptable delineation in 90% of lesions, with a positive agreement (PA) of 93%. Delineation quality scores and agreements per method strongly depended on lesional SUV: the best delineation quality scores were obtained using MV3 in lesions with SUV_max<10 and SUV4.0 in more FDG-avid lesions. Consequently, overall delineation quality and PA improved by applying the most preferred method per SUV category instead of using MV3 as single best method. MV3- and SUV4.0-derived MTVs of lesions with SUV_max>10, were comparable after excluding visually failed MV3 contouring. For lesions with SUV_max<10, MTVs using different methods correlated poorly. Conclusion: On I-PET, MV3 performed best and provided the highest interobserver agreement regarding acceptable delineations of DS4-5 DLBCL lesions. However, delineation method preference strongly depended on lesional SUV. Therefore, we suggest to explore an approach that identifies the optimal delineation method per lesion as function of tumor FDG uptake characteristics, i.e. SUV_max.

Prediction of lymph node metastases using pre-treatment PET radiomics of the primary tumour in esophageal adenocarcinoma: an external validation study

OBJECTIVES

To improve clinical lymph node staging (cN-stage) in oesophageal adenocarcinoma by developing and externally validating three prediction models; one with clinical variables only, one with positron emission tomography (PET) radiomics only, and a combined clinical and radiomics model.

METHODS

Consecutive patients with fluorodeoxyglucose (FDG) avid tumours treated with neoadjuvant therapy between 2010 and 2016 in two international centres (n = 130 and n = 60, respectively) were included. Four clinical variables (age, gender, clinical T-stage and tumour regression grade) and PET radiomics from the primary tumour were used for model development. Diagnostic accuracy, area under curve (AUC), discrimination and calibration were calculated for each model. The prognostic significance was also assessed.

RESULTS

The incidence of lymph node metastases was 58% in both cohorts. The areas under the curve of the clinical, radiomics and combined models were 0.79, 0.69 and 0.82 in the developmental cohort, and 0.65, 0.63 and 0.69 in the external validation cohort, with good calibration demonstrated. The area under the curve of current cN-stage in development and validation cohorts was 0.60 and 0.66, respectively. For overall survival, the combined clinical and radiomics model achieved the best discrimination performance in the external validation cohort (X² = 6.08, df = 1, p = 0.01).

CONCLUSION

Accurate diagnosis of lymph node metastases is crucial for prognosis and guiding treatment decisions. Despite finding improved predictive performance in the development cohort, the models using PET radiomics derived from the primary tumour were not fully replicated in an external validation cohort.

ADVANCES IN KNOWLEDGE

This international study attempted to externally validate a new prediction model for lymph node metastases using PET radiomics. A model combining clinical variables and PET radiomics improved discrimination of lymph node metastases, but these results were not externally replicated.

Radiomics in PET/CT: Current Status and Future AI-Based Evolutions

This short review aims at providing the readers with an update on the current status, as well as future perspectives in the quickly evolving field of radiomics applied to the field of PET/CT imaging. Numerous pitfalls have been identified in study design, data acquisition, segmentation, features calculation and modeling by the radiomics community, and these are often the same issues across all image modalities and clinical applications, however some of these are specific to PET/CT (and SPECT/CT) imaging and therefore the present paper focuses on those. In most cases, recommendations and potential methodological solutions do exist and should therefore be followed to improve the overall quality and reproducibility of published studies. In terms of future evolutions, the techniques from the larger field of artificial intelligence (AI), including those relying on deep neural networks (also known as deep learning) have already shown impressive potential to provide solutions, especially in terms of automation, but also to maybe fully replace the tools the radiomics community has been using until now in order to build the usual radiomics workflow. Some important challenges remain to be addressed before the full impact of AI may be realized but overall the field has made striking advances over the last few years and it is expected advances will continue at a rapid pace.

Improving PET Imaging Acquisition and Analysis With Machine Learning: A Narrative Review With Focus on Alzheimer's Disease and Oncology

Machine learning (ML) algorithms have found increasing utility in the medical imaging field and numerous applications in the analysis of digital biomarkers within positron emission tomography (PET) imaging have emerged. Interest in the use of artificial intelligence in PET imaging for the study of neurodegenerative diseases and oncology stems from the potential for such techniques to streamline decision support for physicians providing early and accurate diagnosis and allowing personalized treatment regimens. In this review, the use of ML to improve PET image acquisition and reconstruction is presented, along with an overview of its applications in the analysis of PET images for the study of Alzheimer's disease and oncology.

Assessing radiomic feature robustness to interpolation in 18F-FDG PET imaging

Radiomic studies link quantitative imaging features to patient outcomes in an effort to personalise treatment in oncology. To be clinically useful, a radiomic feature must be robust to image processing steps, which has made robustness testing a necessity for many technical aspects of feature extraction. We assessed the stability of radiomic features to interpolation processing and categorised features based on stable, systematic, or unstable responses. Here, ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) PET images for 441 oesophageal cancer patients (split: testing = 353, validation = 88) were resampled to 6 isotropic voxel sizes (1.5 mm, 1.8 mm, 2.0 mm, 2.2 mm, 2.5 mm, 2.7 mm) and 141 features were extracted from each volume of interest (VOI). Features were categorised into four groups with two statistical tests. Feature reliability was analysed using an intraclass correlation coefficient (ICC) and patient ranking consistency was assessed using a Spearman’s rank correlation coefficient (ρ). We categorised 93 features robust and 6 limited robustness (stable responses), 34 potentially correctable (systematic responses), and 8 not robust (unstable responses). We developed a correction technique for features with potential systematic variation that used surface fits to link voxel size and percentage change in feature value. Twenty-nine potentially correctable features were re-categorised to robust for the validation dataset, after applying corrections defined by surface fits generated on the testing dataset. Furthermore, we found the choice of interpolation algorithm alone (spline vs trilinear) resulted in large variation in values for a number of features but the response categorisations remained constant. This study attempted to quantify the diverse response of radiomics features commonly found in ¹⁸F-FDG PET clinical modelling to isotropic voxel size interpolation.

Cyclo-oxygenase-2 expression is associated with mean standardised uptake value on 18F-Fluorodeoxyglucose positron emission tomography in oesophageal adenocarcinoma

OBJECTIVES

This pilot study investigated the association of four PET image features and cyclo-oxygenase-2 (COX-2) expression in patients with oesophageal adenocarcinoma. The prognostic significance of these biomarkers was also assessed.

METHODS

50 consecutive patients [median age = 68 (range 47 - 84), males = 45) with oesophageal adenocarcinoma had PET/CT staging between January 2011 and July 2015. The maximum and mean standardised uptake values (SUV_max and SUV_mean), metabolic tumour volume (MTV) and tumour lesion glycolysis (TLG) were calculated from the primary tumour. Their association with COX-2 status was assessed using Mann-Whitney U tests. Kaplan-Meier and Cox regression analysis tested their prognostic significance. A p-value < 0.05 was considered statistically significant.

RESULTS

32 tumours (64.0%) were COX-2 positive. There was a significant association between SUV_mean and COX-2 status (p = 0.019). TLG (hazard ratio (HR) 1.001, 95 % confidence intervals (CI) 1.000 - 1.002, p = 0.018) was significantly associated with overall survival on multivariable analysis.

CONCLUSIONS

This study investigated the association between PET image features and COX-2 expression in oesophageal adenocarcinoma. The preliminary results signal that a combination of TLG (calculated as product of MTV and SUV_mean) and COX-2 status may be a strong and clinically important prognostic biomarker. Our research group are planning a prospective, multi-centre study to validate these findings.

ADVANCES IN KNOWLEDGE

Mean standardised uptake value (SUV_mean) on PET imaging is associated with COX-2 expression in oesophageal adenocarcinoma.

Machine-learned target volume delineation of 18F-FDG PET images after one cycle of induction chemotherapy

Biological tumour volume (GTV_PET) delineation on ¹⁸F-FDG PET acquired during induction chemotherapy (ICT) is challenging due to the reduced metabolic uptake and volume of the GTV_PET. Automatic segmentation algorithms applied to ¹⁸F-FDG PET (PET-AS) imaging have been used for GTV_PET delineation on ¹⁸F-FDG PET imaging acquired before ICT. However, their role has not been investigated in ¹⁸F-FDG PET imaging acquired after ICT. In this study we investigate PET-AS techniques, including ATLAAS a machine learned method, for accurate delineation of the GTV_PET after ICT. Twenty patients were enrolled onto a prospective phase I study (FiGaRO). PET/CT imaging was acquired at baseline and 3 weeks following 1 cycle of induction chemotherapy. The GTV_PET was manually delineated by a nuclear medicine physician and clinical oncologist. The resulting GTV_PET was used as the reference contour. The ATLAAS original statistical model was expanded to include images of reduced metabolic activity and the ATLAAS algorithm was re-trained on the new reference dataset. Estimated GTV_PET contours were derived using sixteen PET-AS methods and compared to the GTV_PET using the Dice Similarity Coefficient (DSC). The mean DSC for ATLAAS, 60% Peak Thresholding (PT60), Adaptive Thresholding (AT) and Watershed Thresholding (WT) was 0.72, 0.61, 0.63 and 0.60 respectively. The GTV_PET generated by ATLAAS compared favourably with manually delineated volumes and in comparison, to other PET-AS methods, was more accurate for GTV_PET delineation after ICT. ATLAAS would be a feasible method to reduce inter-observer variability in multi-centre trials.

Time to Prepare for Risk Adaptation in Lymphoma by Standardizing Measurement of Metabolic Tumor Burden

Increased tumor burden is associated with inferior outcomes in many lymphoma subtypes. Surrogates of tumor burden that are easy to measure, such as the maximum tumor dimension of the bulkiest lesion on CT, have been used as prognostic indices for many years. Recently, total metabolic tumor volume (MTV) and tumor lesion glycolysis have emerged as promising and robust biomarkers of outcome in various lymphomas. The median MTV and the optimal cutoffs to separate patients into risk groups in a study population are, however, highly dependent on the population characteristics and the delineation method used to outline tumor on the PET image. This issue has precluded the use of MTV for risk stratification in trials and clinical practice. Standardization of the methodology is timely to allow the potential for risk adaptation to be explored in addition to response adaptation using PET. Meetings between representatives from research groups active in the field were held under the auspices of the PET International Lymphoma and Myeloma Workshop. A summary of those discussions, which included a review of the literature and a practical assessment of methods used for outlining, including various software options, is presented. Finally, a proposal is made to perform a technical validation of MTV measurement enabling benchmark reference ranges to be derived for published delineation approaches used for outlining with various software. This process would require collation of representative imaging data sets of the most common lymphoma subtypes; agreement on pragmatic criteria for the selection of lesions; generation of a range of MTVs, with consensus to be reached on final contours in a training set; and development of automated software solutions with a set of minimum functionalities to reduce measurement variability. Methods developed in the above training exercise could then be applied to another data set, with a final set of contours and values generated. This final data set would provide a benchmark against which end-users could test their ability to measure MTVs that are consistent with expected values. The data set and automated software solutions could be shared with manufacturers with the aim of including these in standard workflows to allow standardization of MTV measurement across the world.

Joint Tumor Segmentation in PET-CT Images Using Co-Clustering and Fusion Based on Belief Functions

Precise delineation of target tumor is a key factor to ensure the effectiveness of radiation therapy. While hybrid positron emission tomography-computed tomography (PET-CT) has become a standard imaging tool in the practice of radiation oncology, many existing automatic/semi-automatic methods still perform tumor segmentation on mono-modal images. In this paper, a co-clustering algorithm is proposed to concurrently segment 3D tumors in PET-CT images, considering that the two complementary imaging modalities can combine functional and anatomical information to improve segmentation performance. The theory of belief functions is adopted in the proposed method to model, fuse, and reason with uncertain and imprecise knowledge from noisy and blurry PET-CT images. To ensure reliable segmentation for each modality, the distance metric for the quantification of clustering distortions and spatial smoothness is iteratively adapted during the clustering procedure. On the other hand, to encourage consistent segmentation between different modalities, a specific context term is proposed in the clustering objective function. Moreover, during the iterative optimization process, clustering results for the two distinct modalities are further adjusted via a belief-functions-based information fusion strategy. The proposed method has been evaluated on a data set consisting of 21 paired PET-CT images for non-small cell lung cancer patients. The quantitative and qualitative evaluations show that our proposed method performs well compared with the state-of-the-art methods.

External validation of a prognostic model incorporating quantitative PET image features in oesophageal cancer

AIM

Enhanced prognostic models are required to improve risk stratification of patients with oesophageal cancer so treatment decisions can be optimised. The primary aim was to externally validate a published prognostic model incorporating PET image features. Transferability of the model was compared using only clinical variables.

METHODS

This was a Transparent Reporting of a multivariate prediction model for Individual Prognosis Or Diagnosis (TRIPOD) type 3 study. The model was validated against patients treated with neoadjuvant chemoradiotherapy according to the Neoadjuvant chemoradiotherapy plus surgery versus surgery alone for oesophageal or junctional cancer (CROSS) trial regimen using pre- and post-harmonised image features. The Kaplan-Meier method with log-rank significance tests assessed risk strata discrimination. A Cox proportional hazards model assessed model calibration. Primary outcome was overall survival (OS).

RESULTS

Between 2010 and 2015, 449 patients were included in the development (n = 302), internal validation (n = 101) and external validation (n = 46) cohorts. No statistically significant difference in OS between patient quartiles was demonstrated in prognostic models incorporating PET image features (X² = 1.42, df = 3, p = 0.70) or exclusively clinical variables (age, disease stage and treatment; X² = 1.19, df = 3, p = 0.75). The calibration slope β of both models was not significantly different from unity (p = 0.29 and 0.29, respectively). Risk groups defined using only clinical variables suggested differences in OS, although these were not statistically significant (X² = 0.71, df = 2, p = 0.70).

CONCLUSION

The prognostic model did not enable significant discrimination between the validation risk groups, but a second model with exclusively clinical variables suggested some transferable prognostic ability. PET harmonisation did not significantly change the results of model validation.

Radiomics in Nuclear Medicine Applied to Radiation Therapy: Methods, Pitfalls, and Challenges

Radiomics is a recent area of research in precision medicine and is based on the extraction of a large variety of features from medical images. In the field of radiation oncology, comprehensive image analysis is crucial to personalization of treatments. A better characterization of local heterogeneity and the shape of the tumor, depicting individual cancer aggressiveness, could guide dose planning and suggest volumes in which a higher dose is needed for better tumor control. In addition, noninvasive imaging features that could predict treatment outcome from baseline scans could help the radiation oncologist to determine the best treatment strategies and to stratify patients as at low risk or high risk of recurrence. Nuclear medicine molecular imaging reflects information regarding biological processes in the tumor thanks to a wide range of radiotracers. Many studies involving ¹⁸F-fluorodeoxyglucose positron emission tomography suggest an added value of radiomics compared with the use of conventional PET metrics such as standardized uptake value for both tumor diagnosis and prediction of recurrence or treatment outcome. However, these promising results should not hide technical difficulties that still currently prevent the approach from being widely studied or clinically used. These difficulties mostly pertain to the variability of the imaging features as a function of the acquisition device and protocol, the robustness of the models with respect to that variability, and the interpretation of the radiomic models. Addressing the impact of the variability in acquisition and reconstruction protocols is needed, as is harmonizing the radiomic feature calculation methods, to ensure the reproducibility of studies in a multicenter context and their implementation in a clinical workflow. In this review, we explain the potential impact of positron emission tomography radiomics for radiation therapy and underline the various aspects that need to be carefully addressed to make the most of this promising approach.

Evaluation of prognostic models developed using standardised image features from different PET automated segmentation methods

Background

Prognosis in oesophageal cancer (OC) is poor. The 5-year overall survival (OS) rate is approximately 15%. Personalised medicine is hoped to increase the 5- and 10-year OS rates. Quantitative analysis of PET is gaining substantial interest in prognostic research but requires the accurate definition of the metabolic tumour volume. This study compares prognostic models developed in the same patient cohort using individual PET segmentation algorithms and assesses the impact on patient risk stratification.

Consecutive patients (n = 427) with biopsy-proven OC were included in final analysis. All patients were staged with PET/CT between September 2010 and July 2016. Nine automatic PET segmentation methods were studied. All tumour contours were subjectively analysed for accuracy, and segmentation methods with < 90% accuracy were excluded. Standardised image features were calculated, and a series of prognostic models were developed using identical clinical data. The proportion of patients changing risk classification group were calculated.

Results

Out of nine PET segmentation methods studied, clustering means (KM2), general clustering means (GCM3), adaptive thresholding (AT) and watershed thresholding (WT) methods were included for analysis. Known clinical prognostic factors (age, treatment and staging) were significant in all of the developed prognostic models. AT and KM2 segmentation methods developed identical prognostic models. Patient risk stratification was dependent on the segmentation method used to develop the prognostic model with up to 73 patients (17.1%) changing risk stratification group.

Conclusion

Prognostic models incorporating quantitative image features are dependent on the method used to delineate the primary tumour. This has a subsequent effect on risk stratification, with patients changing groups depending on the image segmentation method used.

Electronic supplementary material

The online version of this article (10.1186/s13550-018-0379-3) contains supplementary material, which is available to authorized users.

Development and validation of a prognostic model incorporating texture analysis derived from standardised segmentation of PET in patients with oesophageal cancer

OBJECTIVES

This retrospective cohort study developed a prognostic model incorporating PET texture analysis in patients with oesophageal cancer (OC). Internal validation of the model was performed.

METHODS

Consecutive OC patients (n = 403) were chronologically separated into development (n = 302, September 2010-September 2014, median age = 67.0, males = 227, adenocarcinomas = 237) and validation cohorts (n = 101, September 2014-July 2015, median age = 69.0, males = 78, adenocarcinomas = 79). Texture metrics were obtained using a machine-learning algorithm for automatic PET segmentation. A Cox regression model including age, radiological stage, treatment and 16 texture metrics was developed. Patients were stratified into quartiles according to a prognostic score derived from the model. A p-value < 0.05 was considered statistically significant. Primary outcome was overall survival (OS).

RESULTS

Six variables were significantly and independently associated with OS: age [HR =1.02 (95% CI 1.01-1.04), p < 0.001], radiological stage [1.49 (1.20-1.84), p < 0.001], treatment [0.34 (0.24-0.47), p < 0.001], log(TLG) [5.74 (1.44-22.83), p = 0.013], log(Histogram Energy) [0.27 (0.10-0.74), p = 0.011] and Histogram Kurtosis [1.22 (1.04-1.44), p = 0.017]. The prognostic score demonstrated significant differences in OS between quartiles in both the development (X2 143.14, df 3, p < 0.001) and validation cohorts (X2 20.621, df 3, p < 0.001).

CONCLUSIONS

This prognostic model can risk stratify patients and demonstrates the additional benefit of PET texture analysis in OC staging.

KEY POINTS

• PET texture analysis adds prognostic value to oesophageal cancer staging. • Texture metrics are independently and significantly associated with overall survival. • A prognostic model including texture analysis can help risk stratify patients.

The first MICCAI challenge on PET tumor segmentation

INTRODUCTION

Automatic functional volume segmentation in PET images is a challenge that has been addressed using a large array of methods. A major limitation for the field has been the lack of a benchmark dataset that would allow direct comparison of the results in the various publications. In the present work, we describe a comparison of recent methods on a large dataset following recommendations by the American Association of Physicists in Medicine (AAPM) task group (TG) 211, which was carried out within a MICCAI (Medical Image Computing and Computer Assisted Intervention) challenge.

MATERIALS AND METHODS

Organization and funding was provided by France Life Imaging (FLI). A dataset of 176 images combining simulated, phantom and clinical images was assembled. A website allowed the participants to register and download training data (n = 19). Challengers then submitted encapsulated pipelines on an online platform that autonomously ran the algorithms on the testing data (n = 157) and evaluated the results. The methods were ranked according to the arithmetic mean of sensitivity and positive predictive value.

RESULTS

Sixteen teams registered but only four provided manuscripts and pipeline(s) for a total of 10 methods. In addition, results using two thresholds and the Fuzzy Locally Adaptive Bayesian (FLAB) were generated. All competing methods except one performed with median accuracy above 0.8. The method with the highest score was the convolutional neural network-based segmentation, which significantly outperformed 9 out of 12 of the other methods, but not the improved K-Means, Gaussian Model Mixture and Fuzzy C-Means methods.

CONCLUSION

The most rigorous comparative study of PET segmentation algorithms to date was carried out using a dataset that is the largest used in such studies so far. The hierarchy amongst the methods in terms of accuracy did not depend strongly on the subset of datasets or the metrics (or combination of metrics). All the methods submitted by the challengers except one demonstrated good performance with median accuracy scores above 0.8.

Toward a standard for the evaluation of PET-Auto-Segmentation methods following the recommendations of AAPM task group No. 211: Requirements and implementation

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 by inclusion 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 was 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 with more evaluation datasets.

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

PURPOSE

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.

APPROACH

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.

FINDINGS

A large number of PET-AS algorithms have been developed within the last 20 years. 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.

CONCLUSIONS

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 in situations 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.

SPECT image segmentation for estimation of tumour volume and activity concentration in 177Lu-DOTATATE radionuclide therapy

Background

Dosimetry in radionuclide therapy has the potential to allow for a treatment tailored to the individual patient. One therapeutic radiopharmaceutical where patient-specific dosimetry is feasible is ¹⁷⁷Lu-DOTATATE, used for the treatment of neuroendocrine tumours. The emission of gamma photons by ¹⁷⁷Lu allows for imaging with SPECT (single photon emission computed tomography). One important step for dosimetry using this imaging technique is the SPECT image segmentation, which needs to be robust and accurate for the estimated quantities to be reliable. This work investigates different methods for automatic tumour delineation in ¹⁷⁷Lu-DOTATATE SPECT images.

Three segmentation methods are considered: a fixed 42% threshold (FT), the Otsu method (OM) and a method based on Fourier surfaces (FS). Effects of including resolution compensation in the iterative SPECT image reconstruction are also studied. Evaluation is performed based on Monte Carlo-simulated SPECT images from 24 h and 336 h post injection (p.i.), for determination of the volume, activity concentration and dice similarity coefficient. In addition, patient data are used to investigate the correspondence of tumour volumes when delineated in SPECT or morphological CT or MR images. Patient data are also used to examine the sensitivity to the operator-dependent initialization.

Results

For simulated images from 24 h p.i. reconstructed without resolution compensation, a volume and activity-concentration root-mean-square error below 15% is typically obtained for tumours above approximately 10 cm³ when using OM or FS, while FT performs considerably worse. When including resolution compensation, the tumour volume becomes underestimated and the activity concentration overestimated. The FS method appears to be robust to noise, as seen for the 336 h images. The differences between the tumour volumes estimated from the SPECT images and the volumes estimated from morphological images are generally larger than the discrepancies seen for the simulated data sets.

Conclusions

Segmentation results are encouraging for future dosimetry of tumours with volumes above approximately 10 cm³. Using resolution compensation in the reconstruction may have a negative effect on volume estimation.

Head and neck target delineation using a novel PET automatic segmentation algorithm

PURPOSE

To evaluate the feasibility and impact of using a novel advanced PET auto-segmentation method in Head and Neck (H&N) radiotherapy treatment (RT) planning.

METHODS

ATLAAS, Automatic decision Tree-based Learning Algorithm for Advanced Segmentation, previously developed and validated on pre-clinical data, was applied to ¹⁸F-FDG-PET/CT scans of 20 H&N patients undergoing Intensity Modulated Radiation Therapy. Primary Gross Tumour Volumes (GTVs) manually delineated on CT/MRI scans (GTVp_CT/MRI), together with ATLAAS-generated contours (GTVp_ATLAAS) were used to derive the RT planning GTV (GTVp_final). ATLAAS outlines were compared to CT/MRI and final GTVs qualitatively and quantitatively using a conformity metric.

RESULTS

The ATLAAS contours were found to be reliable and useful. The volume of GTVp_ATLAAS was smaller than GTVp_CT/MRI in 70% of the cases, with an average conformity index of 0.70. The information provided by ATLAAS was used to grow the GTVp_CT/MRI in 10 cases (up to 10.6mL) and to shrink the GTVp_CT/MRI in 7 cases (up to 12.3mL). ATLAAS provided complementary information to CT/MRI and GTVp_ATLAAS contributed to up to 33% of the final GTV volume across the patient cohort.

CONCLUSIONS

ATLAAS can deliver operator independent PET segmentation to augment clinical outlining using CT and MRI and could have utility in future clinical studies.