Experimental Multicenter and Multivendor Evaluation of the Performance of PET Radiomic Features Using 3-Dimensionally Printed Phantom Inserts

Value of PET radiomic features for diagnosis and reccurence prediction of newly diagnosed oral squamous cell carcinoma

Oral Squamous Cell Carcinoma (OSCC) represents more than 90% of oral cancers. The usefulness of radiomic features extracted from PET images of OSCC patients to predict tumor characteristics such as primary tumor stage (T-stage), or tumor grade has not been investigated yet. In this prospective study, 112 patients with newly diagnosed, treatment-naïve OSCC were included. Tumor segmentation was performed using three strategies, the majority vote of these segmentations was used to calculate 445 radiomic features. Features instable over segmentation methods and features highly correlated with volume, SUVmax, and SUVmean were eliminated. A Random Forest classifier was trained to predict T-stage, tumor grade, lymph node involvement, and tumor recurrence. Stratified 10-fold cross-validation was performed. Evaluation metrics such as accuracy and area under the curve (AUC) were reported. SHAP dependence plots were generated to understand classifier decisions. The classifier reached a mean cross-validation AUC of 0.83 for predicting T-stage, an AUC of 0.56 for the grading of the primary tumor, a mean AUC of 0.64 for lymph node involvement, and a mean AUC of 0.63 for recurrence. In patients with newly-diagnosed OSCC, radiomics might have some potential to predict T-stage. These results need to be validated in a larger patient cohort.

Robustness of textural analysis features in quantitative 99 mTc and 177Lu SPECT-CT phantom acquisitions

BACKGROUND

Textural Analysis features in molecular imaging require to be robust under repeat measurement and to be independent of volume for optimum use in clinical studies. Recent EANM and SNMMI guidelines for radiomics provide advice on the potential use of phantoms to identify robust features (Hatt in EJNMMI, 2022). This study applies the suggested phantoms to use in SPECT quantification for two radionuclides, 99 mTc and 177Lu.

METHODS

Acquisitions were made with a uniform phantom to test volume dependency and with a customised 'Revolver' phantom, based on the PET phantom described in Hatt (EJNMMI, 2022) but with local adaptations for SPECT. Each phantom was filled separately with 99 mTc and 177Lu. Sixty-seven Textural Analysis features were extracted and tested for robustness and volume dependency.

RESULTS

Features showing high volume dependency or high Coefficient of Variation (indicating poor repeatability) were removed from the list of features that may be suitable for use in clinical studies. After feature reduction, there were 39 features for 99 mTc and 33 features for 177Lu remaining.

CONCLUSION

The use of a uniform phantom to test volume dependency and a Revolver phantom to identify repeatable Textural Analysis features is possible for quantitative SPECT using 99 mTc or 177Lu. Selection of such features is likely to be centre-dependent due to differences in camera performance as well as acquisition and reconstruction protocols.

Deep learning image enhancement algorithms in PET/CT imaging: a phantom and sarcoma patient radiomic evaluation

PURPOSE

PET/CT imaging data contains a wealth of quantitative information that can provide valuable contributions to characterising tumours. A growing body of work focuses on the use of deep-learning (DL) techniques for denoising PET data. These models are clinically evaluated prior to use, however, quantitative image assessment provides potential for further evaluation. This work uses radiomic features to compare two manufacturer deep-learning (DL) image enhancement algorithms, one of which has been commercialised, against 'gold-standard' image reconstruction techniques in phantom data and a sarcoma patient data set (N=20).

METHODS

All studies in the retrospective sarcoma clinical [18 F]FDG dataset were acquired on either a GE Discovery 690 or 710 PET/CT scanner with volumes segmented by an experienced nuclear medicine radiologist. The modular heterogeneous imaging phantom used in this work was filled with [18 F]FDG, and five repeat acquisitions of the phantom were acquired on a GE Discovery 710 PET/CT scanner. The DL-enhanced images were compared to 'gold-standard' images the algorithms were trained to emulate and input images. The difference between image sets was tested for significance in 93 international biomarker standardisation initiative (IBSI) standardised radiomic features.

RESULTS

Comparing DL-enhanced images to the 'gold-standard', 4.0% and 9.7% radiomic features measured significantly different (pcritical < 0.0005) in the phantom and patient data respectively (averaged over the two DL algorithms). Larger differences were observed comparing DL-enhanced images to algorithm input images with 29.8% and 43.0% of radiomic features measuring significantly different in the phantom and patient data respectively (averaged over the two DL algorithms).

CONCLUSION

DL-enhanced images were found to be similar to images generated using the 'gold-standard' target image reconstruction method with more than 80% of radiomic features not significantly different in all comparisons across unseen phantom and sarcoma patient data. This result offers insight into the performance of the DL algorithms, and demonstrate potential applications for DL algorithms in harmonisation for radiomics and for radiomic features in quantitative evaluation of DL algorithms.

Long-axial field-of-view PET/CT improves radiomics feature reliability

PURPOSE

To assess the influence of long-axial field-of-view (LAFOV) PET/CT systems on radiomics feature reliability, to assess the suitability for short-duration or low-activity acquisitions for textural feature analysis and to investigate the influence of acceptance angle.

METHODS

34 patients were analysed: twelve patients underwent oncological 2-[18F]-FDG PET/CT, fourteen [18F]PSMA-1007 and eight [68Ga]Ga-DOTATOC. Data were obtained using a 106 cm LAFOV system for 10 min. Sinograms were generated from list-mode data corresponding to scan durations of 2, 5, 10, 20, 30, 60, 120, 240, 360 and 600s using both standard (minimum ring difference MRD 85 crystals) and maximum acceptance angles (MRD 322). Target lesions were segmented and radiomics features were calculated. To assess feature correlation, Pearson's product-moment correlation coefficient (PPMCC) was calculated with respect to the full duration acquisition for MRD 85 and 322 respectively. The number of features with excellent (r > 0.9), moderate (r > 0.7 and < 0.9) and poor (r ≤ 0.7) correlation was compared as a measure of feature stability. Intra-class heterogeneity was assessed by means of the quartile coefficient of dispersion.

RESULTS

As expected, PPMCC improved with acquisition time for all features. By 240s almost all features showed at least moderate agreement with the full count (C100%) data, and by 360s almost all showed excellent agreement. Compared to standard-axial field of view (SAFOV) equivalent scans, fewer features showed moderate or poor agreement, and this was most pronounced for [68Ga]Ga-DOTATOC. Data obtained at C100% at MRD 322 were better able to capture between-patient heterogeneities.

CONCLUSION

The improved feature reliability at longer acquisition times and higher MRD demonstrate the advantages of high sensitivity LAFOV systems for reproducible and low-noise data. High fidelity between MRD 85 and MRD 322 was seen at all scan durations > 2s. When contrasted with data comparable to a simulated SAFOV acquisition, full-count and full-MRD data were better able to capture underlying feature heterogeneities.

Outcome prediction based on [18F]FDG PET/CT in patients with pleural mesothelioma treated with ipilimumab and nivolumab +/- UV1 telomerase vaccine

PURPOSE

The introduction of immunotherapy in pleural mesothelioma (PM) has highlighted the need for effective outcome predictors. This study explores the role of [18F]FDG PET/CT in predicting outcomes in PM treated with immunotherapy.

METHODS

Patients from the NIPU trial, receiving ipilimumab and nivolumab +/- telomerase vaccine in second-line, were included. [18F]FDG PET/CT was obtained at baseline (n = 100) and at week-5 (n = 76). Metabolic tumour volume (MTV) and peak standardised uptake value (SUVpeak) were evaluated in relation to survival outcomes. Wilcoxon rank-sum test was used to assess differences in MTV, total lesion glycolysis (TLG), maximum standardised uptake value (SUVmax) and SUVpeak between patients exhibiting an objective response, defined as either partial response or complete response according to the modified Response Criteria in Solid Tumours (mRECIST) and immune RECIST (iRECIST), and non-responders, defined as either stable disease or progressive disease as their best overall response.

RESULTS

Univariate Cox regression revealed significant associations of MTV with OS (HR 1.36, CI: 1.14, 1.62, p < 0.001) and PFS (HR 1.18, CI: 1.03, 1.34, p = 0.02), while multivariate analysis showed a significant association with OS only (HR 1.35, CI: 1.09, 1.68, p = 0.007). While SUVpeak was not significantly associated with OS or PFS in univariate analyses, it was significantly associated with OS in multivariate analysis (HR 0.43, CI: 0.23, 0.80, p = 0.008). Objective responders had significant reductions in TLG, SUVmax and SUVpeak at week-5.

CONCLUSION

MTV provides prognostic value in PM treated with immunotherapy. High SUVpeak was not associated with inferior outcomes, which could be attributed to the distinct mechanisms of immunotherapy. Early reductions in PET metrics correlated with treatment response.

STUDY REGISTRATION

The NIPU trial (NCT04300244) is registered at clinicaltrials.gov. https://classic.

CLINICALTRIALS

gov/ct2/show/NCT04300244?cond=Pleural+Mesothelioma&cntry=NO&draw=2&rank=4.

Enhancing Radiomics Reproducibility: Deep Learning-Based Harmonization of Abdominal Computed Tomography (CT) Images

We assessed the feasibility of using deep learning-based image harmonization to improve the reproducibility of radiomics features in abdominal CT scans. In CT imaging, harmonization adjusts images from different institutions to ensure consistency despite variations in scanners and acquisition protocols. This process is essential because such differences can lead to variability in radiomics features, affecting reproducibility and accuracy. Harmonizing images minimizes these inconsistencies, supporting more reliable and clinically applicable results across diverse settings. A pre-trained harmonization algorithm was applied to 63 dual-energy abdominal CT images, which were reconstructed into four different types, and 10 regions of interest (ROIs) were analyzed. From the original 455 radiomics features per ROI, 387 were used after excluding redundant features. Reproducibility was measured using the intraclass correlation coefficient (ICC), with a threshold of ICC ≥ 0.85 indicating acceptable reproducibility. The region-based analysis revealed significant improvements in reproducibility post-harmonization, especially in vessel features, which increased from 14% to 69%. Other regions, including the spleen, kidney, muscle, and liver parenchyma, also saw notable improvements, although air reproducibility slightly decreased from 95% to 94%, impacting only a few features. In patient-based analysis, reproducible features increased from 18% to 65%, with an average of 179 additional reproducible features per patient after harmonization. These results demonstrate that deep learning-based harmonization can significantly enhance the reproducibility of radiomics features in abdominal CT, offering promising potential for advancing radiomics development and its clinical applications.

Multicentric study on the reproducibility and robustness of PET-based radiomics features with a realistic activity painting phantom

Previously, we developed an "activity painting" tool for PET image simulation; however, it could simulate heterogeneous patterns only in the air. We aimed to improve this phantom technique to simulate arbitrary lesions in a radioactive background to perform relevant multi-center radiomic analysis. We conducted measurements moving a 22Na point source in a 20-liter background volume filled with 5 kBq/mL activity with an adequately controlled robotic system to prevent the surge of the water. Three different lesion patterns were "activity-painted" in five PET/CT cameras, resulting in 8 different reconstructions. We calculated 46 radiomic indeces (RI) for each lesion and imaging setting, applying absolute and relative discretization. Reproducibility and reliability were determined by the inter-setting coefficient of variation (CV) and the intraclass correlation coefficient (ICC). Hypothesis tests were used to compare RI between lesions. By simulating precisely the same lesions, we confirmed that the reconstructed voxel size and the spatial resolution of different PET cameras were critical for higher order RI. Considering conventional RIs, the SUVpeak and SUVmean proved the most reliable (CV<10%). CVs above 25% are more common for higher order RIs, but we also found that low CVs do not necessarily imply robust parameters but often rather insensitive RIs. Based on the hypothesis test, most RIs could clearly distinguish between the various lesions using absolute resampling. ICC analysis also revealed that most RIs were more reproducible with absolute discretization. The activity painting method in a real radioactive environment proved suitable for precisely detecting the radiomic differences derived from the different camera settings and texture characteristics. We also found that inter-setting CV is not an appropriate metric for analyzing RI parameters' reliability and robustness. Although multicentric cohorts are increasingly common in radiomics analysis, realistic texture phantoms can provide indispensable information on the sensitivity of an RI and how an individual RI parameter measures the texture.

PET radiomics in lung cancer: advances and translational challenges

Radiomics is an emerging field of medical imaging that aims at improving the accuracy of diagnosis, prognosis, treatment planning and monitoring non-invasively through the automated or semi-automated quantitative analysis of high-dimensional image features. Specifically in the field of nuclear medicine, radiomics utilizes imaging methods such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) to evaluate biomarkers related to metabolism, blood flow, cellular activity and some biological pathways. Lung cancer ranks among the leading causes of cancer-related deaths globally, and radiomics analysis has shown great potential in guiding individualized therapy, assessing treatment response, and predicting clinical outcomes. In this review, we summarize the current state-of-the-art radiomics progress in lung cancer, highlighting the potential benefits and existing limitations of this approach. The radiomics workflow was introduced first including image acquisition, segmentation, feature extraction, and model building. Then the published literatures were described about radiomics-based prediction models for lung cancer diagnosis, differentiation, prognosis and efficacy evaluation. Finally, we discuss current challenges and provide insights into future directions and potential opportunities for integrating radiomics into routine clinical practice.

Distributed Harmonization: Federated Clustered Batch Effect Adjustment and Generalization

Independent and identically distributed (i.i.d.) data is essential to many data analysis and modeling techniques. In the medical domain, collecting data from multiple sites or institutions is a common strategy that guarantees sufficient clinical diversity, determined by the decentralized nature of medical data. However, data from various sites are easily biased by the local environment or facilities, thereby violating the i.i.d. rule. A common strategy is to harmonize the site bias while retaining important biological information. The COMBAT is among the most popular harmonization approaches and has recently been extended to handle distributed sites. However, when faced with situations involving newly joined sites in training or evaluating data from unknown/unseen sites, COMBAT lacks compatibility and requires retraining with data from all the sites. The retraining leads to significant computational and logistic overhead that is usually prohibitive. In this work, we develop a novel Cluster ComBat harmonization algorithm, which leverages cluster patterns of the data in different sites and greatly advances the usability of COMBAT harmonization. We use extensive simulation and real medical imaging data from ADNI to demonstrate the superiority of the proposed approach. Our codes are provided in https://github.com/illidanlab/distributed-cluster-harmonization.

Review on radiomic analysis in 18F-fluorodeoxyglucose positron emission tomography for prediction of melanoma outcomes

Over the past decade, several strategies have revolutionized the clinical management of patients with cutaneous melanoma (CM), including immunotherapy and targeted tyrosine kinase inhibitor (TKI)-based therapies. Indeed, immune checkpoint inhibitors (ICIs), alone or in combination, represent the standard of care for patients with advanced disease without an actionable mutation. Notably BRAF combined with MEK inhibitors represent the therapeutic standard for disease disclosing BRAF mutation. At the same time, FDG PET/CT has become part of the routine staging and evaluation of patients with cutaneous melanoma. There is growing interest in using FDG PET/CT measurements to predict response to ICI therapy and/or target therapy. While semiquantitative values such as standardized uptake value (SUV) are limited for predicting outcome, new measures including tumor metabolic volume, total lesion glycolysis and radiomics seem promising as potential imaging biomarkers for nuclear medicine. The aim of this review, prepared by an interdisciplinary group of experts, is to take stock of the current literature on radiomics approaches that could improve outcomes in CM.

Is PET Radiomics Useful to Predict Pathologic Tumor Response and Prognosis in Locally Advanced Cervical Cancer?

This study investigated whether radiomic features extracted from pretreatment [18F]FDG PET could improve the prediction of both histopathologic tumor response and survival in patients with locally advanced cervical cancer (LACC) treated with neoadjuvant chemoradiotherapy followed by surgery compared with conventional PET parameters and histopathologic features. Methods: The medical records of all consecutive patients with LACC referred between July 2010 and July 2016 were reviewed. [18F]FDG PET/CT was performed before neoadjuvant chemoradiotherapy. Radiomic features were extracted from the primary tumor volumes delineated semiautomatically on the PET images and reduced by factor analysis. A receiver-operating-characteristic analysis was performed, and conventional and radiomic features were dichotomized with Liu's method according to pathologic response (pR) and cancer-specific death. According to the study protocol, only areas under the curve of more than 0.70 were selected for further analysis, including logistic regression analysis for response prediction and Cox regression analysis for survival prediction. Results: A total of 195 patients fulfilled the inclusion criteria. At pathologic evaluation after surgery, 131 patients (67.2%) had no or microscopic (≤3 mm) residual tumor (pR0 or pR1, respectively); 64 patients (32.8%) had macroscopic residual tumor (>3 mm, pR2). With a median follow-up of 76.0 mo (95% CI, 70.7-78.7 mo), 31.3% of patients had recurrence or progression and 20.0% died of the disease. Among conventional PET parameters, SUVmean significantly differed between pathologic responders and nonresponders. Among radiomic features, 1 shape and 3 textural features significantly differed between pathologic responders and nonresponders. Three radiomic features significantly differed between presence and absence of recurrence or progression and between presence and absence of cancer-specific death. Areas under the curve were less than 0.70 for all parameters; thus, univariate and multivariate regression analyses were not performed. Conclusion: In a large series of patients with LACC treated with neoadjuvant chemoradiotherapy followed by surgery, PET radiomic features could not predict histopathologic tumor response and survival. It is crucial to further explore the biologic mechanism underlying imaging-derived parameters and plan a large, prospective, multicenter study with standardized protocols for all phases of the process of radiomic analysis to validate radiomics before its use in clinical routine.

3D printed non-uniform anthropomorphic phantoms for quantitative SPECT

BACKGROUND

A 3D printing grid-based method was developed to construct anthropomorphic phantoms with non-uniform activity distributions, to be used for evaluation of quantitative SPECT images. The aims were to characterize the grid-based method and to evaluate its capability to provide realistically shaped phantoms with non-uniform activity distributions.

METHODS

Characterization of the grid structures was performed by printing grid-filled spheres. Evaluation was performed by micro-CT imaging to investigate the printing accuracy and by studying the modulation contrast ([Formula: see text]) in SPECT images for 177Lu and 99mTc as a function of the grid fillable-volume fraction (FVF) determined from weighing. The grid-based technique was applied for the construction of two kidney phantoms and two thyroid phantoms, designed using templates from the XCAT digital phantoms. The kidneys were constructed with a hollow outer container shaped as cortex, an inner grid-based structure representing medulla and a solid section representing pelvis. The thyroids consisted of two lobes printed as grid-based structures, with void hot spots within the lobes. The phantoms were filled with solutions of 177Lu (kidneys) or 99mTc (thyroids) and imaged with SPECT. For verification, Monte Carlo simulations of SPECT imaging were performed for activity distributions corresponding to those of the printed phantoms. Measured and simulated SPECT images were compared qualitatively and quantitatively.

RESULTS

Micro-CT images showed that printing inaccuracies were mainly uniform across the grid. The relationships between the FVF from weighing and [Formula: see text] were found to be linear (r = 0.9995 and r = 0.9993 for 177Lu and 99mTc, respectively). The FVF-deviations from the design were up to 15% for thyroids and 4% for kidneys, mainly related to possibilities of cleaning after printing. Measured and simulated SPECT images of kidneys and thyroids exhibited similar activity distributions and quantitative comparisons agreed well, thus verifying the grid-based method.

CONCLUSIONS

We find the grid-based technique useful for the provision of 3D printed, realistically shaped, phantoms with non-uniform activity distributions, which can be used for evaluation of different quantitative methods in SPECT imaging.

Impact of respiratory motion on 18 F-FDG PET radiomics stability: Clinical evaluation with a digital PET scanner

PURPOSE

18 F-FDG PET quantitative features are susceptible to respiratory motion. However, studies using clinical patient data to explore the impact of respiratory motion on 18 F-FDG PET radiomic features are limited. In this study, we investigated the impact of respiratory motion on radiomics stability with clinical 18 F-FDG PET images using a data-driven gating (DDG) algorithm on the digital PET scanner.

MATERIALS AND METHODS

A total of 101 patients who underwent oncological 18 F-FDG PET scans were retrospectively included. A DDG algorithm combined with a motion compensation technique was used to extract the PET images with respiratory motion correction. 18 F-FDG-avid lesions from the thorax to the upper abdomen were analyzed on the non-DDG and DDG PET images. The lesions were segmented with a 40% threshold of the maximum standardized uptake. A total of 725 radiomic features were computed from the segmented lesions, including first-order, shape, texture, and wavelet features. The intraclass correlation coefficient (ICC) and coefficient of variation (COV) were calculated to evaluate feature stability. An ICC above 0.9 and a COV below 5% were considered high stability.

RESULTS

In total, 168 lesions with and without respiratory motion correction were analyzed. Our results indicated that most 18 F-FDG PET radiomic features are sensitive to respiratory motion. Overall, only 27 out of 725 (3.72%) radiomic features were identified as highly stable, including one from the first-order features (entropy), one from the shape features (sphericity), four from the gray-level co-occurrence matrix features (normalized and unnormalized inverse difference moment, joint entropy, and sum entropy), one from the gray-level run-length matrix features (run entropy), and 20 from the wavelet filter-based features.

CONCLUSION

Respiratory motion has a significant impact on 18 F-FDG PET radiomics stability. The highly stable features identified in our study may serve as potential candidates for further applications, such as machine learning modeling.

Nuclear medicine radiomics in digestive system tumors: Concept, applications, challenges, and future perspectives

Radiomics aims to develop novel biomarkers and provide relevant deeper subvisual information about pathology, immunophenotype, and tumor microenvironment. It uses automated or semiautomated quantitative analysis of high-dimensional images to improve characterization, diagnosis, and prognosis. Recent years have seen a rapid increase in radiomics applications in nuclear medicine, leading to some promising research results in digestive system oncology, which have been driven by big data analysis and the development of artificial intelligence. Although radiomics advances one step further toward the non-invasive precision medical analysis, it is still a step away from clinical application and faces many challenges. This review article summarizes the available literature on digestive system tumors regarding radiomics in nuclear medicine. First, we describe the workflow and steps involved in radiomics analysis. Subsequently, we discuss the progress in clinical application regarding the utilization of radiomics for distinguishing between various diseases and evaluating their prognosis, and demonstrate how radiomics advances this field. Finally, we offer our viewpoint on how the field can progress by addressing the challenges facing clinical implementation.

Enhancing Interoperability and Harmonisation of Nuclear Medicine Image Data and Associated Clinical Data

Nuclear imaging techniques such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) in combination with computed tomography (CT) are established imaging modalities in clinical practice, particularly for oncological problems. Due to a multitude of manufacturers, different measurement protocols, local demographic or clinical workflow variations as well as various available reconstruction and analysis software, very heterogeneous datasets are generated. This review article examines the current state of interoperability and harmonisation of image data and related clinical data in the field of nuclear medicine. Various approaches and standards to improve data compatibility and integration are discussed. These include, for example, structured clinical history, standardisation of image acquisition and reconstruction as well as standardised preparation of image data for evaluation. Approaches to improve data acquisition, storage and analysis will be presented. Furthermore, approaches are presented to prepare the datasets in such a way that they become usable for projects applying artificial intelligence (AI) (machine learning, deep learning, etc.). This review article concludes with an outlook on future developments and trends related to AI in nuclear medicine, including a brief research of commercial solutions.

Stacking Ensemble Learning-Based [18F]FDG PET Radiomics for Outcome Prediction in Diffuse Large B-Cell Lymphoma

This study aimed to develop an analytic approach based on [18F]FDG PET radiomics using stacking ensemble learning to improve the outcome prediction in diffuse large B-cell lymphoma (DLBCL). Methods: In total, 240 DLBCL patients from 2 medical centers were divided into the training set (n = 141), internal testing set (n = 61), and external testing set (n = 38). Radiomics features were extracted from pretreatment [18F]FDG PET scans at the patient level using 4 semiautomatic segmentation methods (SUV threshold of 2.5, SUV threshold of 4.0 [SUV4.0], 41% of SUVmax, and SUV threshold of mean liver uptake [PERCIST]). All extracted features were harmonized with the ComBat method. The intraclass correlation coefficient was used to evaluate the reliability of radiomics features extracted by different segmentation methods. Features from the most reliable segmentation method were selected by Pearson correlation coefficient analysis and the LASSO (least absolute shrinkage and selection operator) algorithm. A stacking ensemble learning approach was applied to build radiomics-only and combined clinical-radiomics models for prediction of 2-y progression-free survival and overall survival based on 4 machine learning classifiers (support vector machine, random forests, gradient boosting decision tree, and adaptive boosting). Confusion matrix, receiver-operating-characteristic curve analysis, and survival analysis were used to evaluate the model performance. Results: Among 4 semiautomatic segmentation methods, SUV4.0 segmentation yielded the highest interobserver reliability, with 830 (66.7%) selected radiomics features. The combined model constructed by the stacking method achieved the best discrimination performance. For progression-free survival prediction in the external testing set, the areas under the receiver-operating-characteristic curve and accuracy of the stacking-based combined model were 0.771 and 0.789, respectively. For overall survival prediction, the stacking-based combined model achieved an area under the curve of 0.725 and an accuracy of 0.763 in the external testing set. The combined model also demonstrated a more distinct risk stratification than the International Prognostic Index in all sets (log-rank test, all P < 0.05). Conclusion: The combined model that incorporates [18F]FDG PET radiomics and clinical characteristics based on stacking ensemble learning could enable improved risk stratification in DLBCL.

Baseline PET radiomics outperforms the IPI risk score for prediction of outcome in diffuse large B-cell lymphoma

The objective of this study is to externally validate the clinical positron emission tomography (PET) model developed in the HOVON-84 trial and to compare the model performance of our clinical PET model using the international prognostic index (IPI). In total, 1195 patients with diffuse large B-cell lymphoma (DLBCL) were included in the study. Data of 887 patients from 6 studies were used as external validation data sets. The primary outcomes were 2-year progression-free survival (PFS) and 2-year time to progression (TTP). The metabolic tumor volume (MTV), maximum distance between the largest lesion and another lesion (Dmaxbulk), and peak standardized uptake value (SUVpeak) were extracted. The predictive values of the IPI and clinical PET model (MTV, Dmaxbulk, SUVpeak, performance status, and age) were tested. Model performance was assessed using the area under the curve (AUC), and diagnostic performance, using the positive predictive value (PPV). The IPI yielded an AUC of 0.62. The clinical PET model yielded a significantly higher AUC of 0.71 (P < .001). Patients with high-risk IPI had a 2-year PFS of 61.4% vs 51.9% for those with high-risk clinical PET, with an increase in PPV from 35.5% to 49.1%, respectively. A total of 66.4% of patients with high-risk IPI were free from progression or relapse vs 55.5% of patients with high-risk clinical PET scores, with an increased PPV from 33.7% to 44.6%, respectively. The clinical PET model remained predictive of outcome in 6 independent first-line DLBCL studies, and had higher model performance than the currently used IPI in all studies.

Predictive value of 18F-FDG PET/CT-based radiomics model for neoadjuvant chemotherapy efficacy in breast cancer: a multi-scanner/center study with external validation

PURPOSE

To develop and validate the predictive value of an 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) model for breast cancer neoadjuvant chemotherapy (NAC) efficacy based on the tumor-to-liver ratio (TLR) radiomic features and multiple data pre-processing methods.

METHODS

One hundred and ninety-three breast cancer patients from multiple centers were retrospectively included in this study. According to the endpoint of NAC, we divided the patients into pathological complete remission (pCR) and non-pCR groups. All patients underwent 18F-FDG PET/CT imaging before NAC treatment, and CT and PET images volume of interest (VOI) segmentation by manual segmentation and semi-automated absolute threshold segmentation, respectively. Then, feature extraction of VOI was performed with the pyradiomics package. A total of 630 models were created based on the source of radiomic features, the elimination of the batch effect approach, and the discretization method. The differences in data pre-processing approaches were compared and analyzed to identify the best-performing model, which was further tested by the permutation test.

RESULTS

A variety of data pre-processing methods contributed in varying degrees to the improvement of model effects. Among them, TLR radiomic features and Combat and Limma methods that eliminate batch effects could enhance the model prediction overall, and data discretization could be used as a potential method that can further optimize the model. A total of seven excellent models were selected and then based on the AUC of each model in the four test sets and their standard deviations, we selected the optimal model. The optimal model predicted AUC between 0.7 and 0.77 for the four test groups, with p-values less than 0.05 for the permutation test.

CONCLUSION

It is necessary to enhance the predictive effect of the model by eliminating confounding factors through data pre-processing. The model developed in this way is effective in predicting the efficacy of NAC for breast cancer.

Prognostic Value of [18F]FDG PET Radiomics to Detect Peritoneal and Distant Metastases in Locally Advanced Gastric Cancer-A Side Study of the Prospective Multicentre PLASTIC Study

AIM

To improve identification of peritoneal and distant metastases in locally advanced gastric cancer using [¹⁸F]FDG-PET radiomics.

METHODS

[¹⁸F]FDG-PET scans of 206 patients acquired in 16 different Dutch hospitals in the prospective multicentre PLASTIC-study were analysed. Tumours were delineated and 105 radiomic features were extracted. Three classification models were developed to identify peritoneal and distant metastases (incidence: 21%): a model with clinical variables, a model with radiomic features, and a clinicoradiomic model, combining clinical variables and radiomic features. A least absolute shrinkage and selection operator (LASSO) regression classifier was trained and evaluated in a 100-times repeated random split, stratified for the presence of peritoneal and distant metastases. To exclude features with high mutual correlations, redundancy filtering of the Pearson correlation matrix was performed (r = 0.9). Model performances were expressed by the area under the receiver operating characteristic curve (AUC). In addition, subgroup analyses based on Lauren classification were performed.

RESULTS

None of the models could identify metastases with low AUCs of 0.59, 0.51, and 0.56, for the clinical, radiomic, and clinicoradiomic model, respectively. Subgroup analysis of intestinal and mixed-type tumours resulted in low AUCs of 0.67 and 0.60 for the clinical and radiomic models, and a moderate AUC of 0.71 in the clinicoradiomic model. Subgroup analysis of diffuse-type tumours did not improve the classification performance.

CONCLUSION

Overall, [¹⁸F]FDG-PET-based radiomics did not contribute to the preoperative identification of peritoneal and distant metastases in patients with locally advanced gastric carcinoma. In intestinal and mixed-type tumours, the classification performance of the clinical model slightly improved with the addition of radiomic features, but this slight improvement does not outweigh the laborious radiomic analysis.

Development and External Validation of a PET Radiomic Model for Prognostication of Head and Neck Cancer

AIM

To build and externally validate an [¹⁸F]FDG PET radiomic model to predict overall survival in patients with head and neck squamous cell carcinoma (HNSCC).

METHODS

Two multicentre datasets of patients with operable HNSCC treated with preoperative afatinib who underwent a baseline and evaluation [¹⁸F]FDG PET/CT scan were included (EORTC: n = 20, Unicancer: n = 34). Tumours were delineated, and radiomic features were extracted. Each cohort served once as a training and once as an external validation set for the prediction of overall survival. Supervised feature selection was performed using variable hunting with variable importance, selecting the top two features. A Cox proportional hazards regression model using selected radiomic features and clinical characteristics was fitted on the training dataset and validated in the external validation set. Model performances are expressed by the concordance index (C-index).

RESULTS

In both models, the radiomic model surpassed the clinical model with validation C-indices of 0.69 and 0.79 vs. 0.60 and 0.67, respectively. The model that combined the radiomic features and clinical variables performed best, with validation C-indices of 0.71 and 0.82.

CONCLUSION

Although assessed in two small but independent cohorts, an [¹⁸F]FDG-PET radiomic signature based on the evaluation scan seems promising for the prediction of overall survival for HNSSC treated with preoperative afatinib. The robustness and clinical applicability of this radiomic signature should be assessed in a larger cohort.

Radiomics prognostic analysis of PET/CT images in a multicenter head and neck cancer cohort: investigating ComBat strategies, sub-volume characterization, and automatic segmentation

PURPOSE

This study aimed to investigate the impact of several ComBat harmonization strategies, intra-tumoral sub-volume characterization, and automatic segmentations for progression-free survival (PFS) prediction through radiomics modeling for patients with head and neck cancer (HNC) in PET/CT images.

METHODS

The HECKTOR MICCAI 2021 challenge set containing PET/CT images and clinical data of 325 oropharynx HNC patients was exploited. A total of 346 IBSI-compliant radiomic features were extracted for each patient's primary tumor volume defined by the reference manual contours. Modeling relied on least absolute shrinkage Cox regression (Lasso-Cox) for feature selection (FS) and Cox proportional-hazards (CoxPH) models were built to predict PFS. Within this methodological framework, 8 different strategies for ComBat harmonization were compared, including before or after FS, in feature groups separately or all features directly, and with center or clustering-determined labels. Features extracted from tumor sub-volume clustering were also investigated for their prognostic additional value. Finally, 3 automatic segmentations (2 threshold-based and a 3D U-Net) were also compared. All results were evaluated with the concordance index (C-index).

RESULTS

Radiomics features without harmonization, combined with clinical factors, led to models with C-index values of 0.69 in the testing set. The best version of ComBat harmonization, i.e., after FS, for feature groups separately and relying on clustering-determined labels, achieved a C-index of 0.71. The use of features extracted from tumor sub-volumes further improved the C-index to 0.72. Models that relied on the automatic segmentations yielded close but slightly lower prognostic performance (0.67-0.70) compared to reference contours.

CONCLUSION

A standard radiomics pipeline allowed for prediction of PFS in a multicenter HNC cohort. Applying a specific strategy of ComBat harmonization improved the performance. The extraction of intra-tumoral sub-volume features and automatic segmentation could contribute to the improvement and automation of prognosis modeling, respectively.

Harmonization Strategies in Multicenter MRI-Based Radiomics

Radiomics analysis is a powerful tool aiming to provide diagnostic and prognostic patient information directly from images that are decoded into handcrafted features, comprising descriptors of shape, size and textural patterns. Although radiomics is gaining momentum since it holds great promise for accelerating digital diagnostics, it is susceptible to bias and variation due to numerous inter-patient factors (e.g., patient age and gender) as well as inter-scanner ones (different protocol acquisition depending on the scanner center). A variety of image and feature based harmonization methods has been developed to compensate for these effects; however, to the best of our knowledge, none of these techniques has been established as the most effective in the analysis pipeline so far. To this end, this review provides an overview of the challenges in optimizing radiomics analysis, and a concise summary of the most relevant harmonization techniques, aiming to provide a thorough guide to the radiomics harmonization process.

Baseline radiomics features and MYC rearrangement status predict progression in aggressive B-cell lymphoma

We investigated whether the outcome prediction of patients with aggressive B-cell lymphoma can be improved by combining clinical, molecular genotype, and radiomics features. MYC, BCL2, and BCL6 rearrangements were assessed using fluorescence in situ hybridization. Seventeen radiomics features were extracted from the baseline positron emission tomography-computed tomography of 323 patients, which included maximum standardized uptake value (SUVmax), SUVpeak, SUVmean, metabolic tumor volume (MTV), total lesion glycolysis, and 12 dissemination features pertaining to distance, differences in uptake and volume between lesions, respectively. Logistic regression with backward feature selection was used to predict progression after 2 years. The predictive value of (1) International Prognostic Index (IPI); (2) IPI plus MYC; (3) IPI, MYC, and MTV; (4) radiomics; and (5) MYC plus radiomics models were tested using the cross-validated area under the curve (CV-AUC) and positive predictive values (PPVs). IPI yielded a CV-AUC of 0.65 ± 0.07 with a PPV of 29.6%. The IPI plus MYC model yielded a CV-AUC of 0.68 ± 0.08. IPI, MYC, and MTV yielded a CV-AUC of 0.74 ± 0.08. The highest model performance of the radiomics model was observed for MTV combined with the maximum distance between the largest lesion and another lesion, the maximum difference in SUVpeak between 2 lesions, and the sum of distances between all lesions, yielding an improved CV-AUC of 0.77 ± 0.07. The same radiomics features were retained when adding MYC (CV-AUC, 0.77 ± 0.07). PPV was highest for the MYC plus radiomics model (50.0%) and increased by 20% compared with the IPI (29.6%). Adding radiomics features improved model performance and PPV and can, therefore, aid in identifying poor prognosis patients.

Mitigation of noise-induced bias of PET radiomic features

Introduction

One major challenge in PET radiomics is its sensitivity to noise. Low signal-to-noise ratio (SNR) affects not only the precision but also the accuracy of quantitative metrics extracted from the images resulting in noise-induced bias. This phantom study aims to identify the radiomic features that are robust to noise in terms of precision and accuracy and to explore some methods that might help to correct noise-induced bias.

Methods

A phantom containing three ¹⁸F-FDG filled 3D printed inserts, reflecting heterogeneous tracer uptake and realistic tumor shapes, was used in the study. The three different phantom inserts were filled and scanned with three different tumor-to-background ratios, simulating a total of nine different tumors. From the 40-minute list-mode data, ten frames each for 5 s, 10 s, 30 s, and 120 s frame duration were reconstructed to generate images with different noise levels. Under these noise conditions, the precision and accuracy of the radiomic features were analyzed using intraclass correlation coefficient (ICC) and similarity distance metric (SDM) respectively. Based on the ICC and SDM values, the radiomic features were categorized into four groups: poor, moderate, good, and excellent precision and accuracy. A “difference image” created by subtracting two statistically equivalent replicate images was used to develop a model to correct the noise-induced bias. Several regression methods (e.g., linear, exponential, sigmoid, and power-law) were tested. The best fitting model was chosen based on Akaike information criteria.

Results

Several radiomic features derived from low SNR images have high repeatability, with 68% of radiomic features having ICC ≥ 0.9 for images with a frame duration of 5 s. However, most features show a systematic bias that correlates with the increase in noise level. Out of 143 features with noise-induced bias, the SDM values were improved based on a regression model (53 features to excellent and 67 to good) indicating that the noise-induced bias of these features can be, at least partially, corrected.

Conclusion

To have a predictive value, radiomic features should reflect tumor characteristics and be minimally affected by noise. The present study has shown that it is possible to correct for noise-induced bias, at least in a subset of the features, using a regression model based on the local image noise estimates.

[18F]FDG-PET/CT radiomics for the identification of genetic clusters in pheochromocytomas and paragangliomas

Objectives

Based on germline and somatic mutation profiles, pheochromocytomas and paragangliomas (PPGLs) can be classified into different clusters. We investigated the use of [¹⁸F]FDG-PET/CT radiomics, SUV_max and biochemical profile for the identification of the genetic clusters of PPGLs.

Methods

In this single-centre cohort, 40 PPGLs (13 cluster 1, 18 cluster 2, 9 sporadic) were delineated using a 41% adaptive threshold of SUV_peak ([¹⁸F]FDG-PET) and manually (low-dose CT; ldCT). Using PyRadiomics, 211 radiomic features were extracted. Stratified 5-fold cross-validation for the identification of the genetic cluster was performed using multinomial logistic regression with dimensionality reduction incorporated per fold. Classification performances of biochemistry, SUV_max and PET(/CT) radiomic models were compared and presented as mean (multiclass) test AUCs over the five folds. Results were validated using a sham experiment, randomly shuffling the outcome labels.

Results

The model with biochemistry only could identify the genetic cluster (multiclass AUC 0.60). The three-factor PET model had the best classification performance (multiclass AUC 0.88). A simplified model with only SUV_max performed almost similarly. Addition of ldCT features and biochemistry decreased the classification performances. All sham AUCs were approximately 0.50.

Conclusion

PET radiomics achieves a better identification of PPGLs compared to biochemistry, SUV_max, ldCT radiomics and combined approaches, especially for the differentiation of sporadic PPGLs. Nevertheless, a model with SUV_max alone might be preferred clinically, weighing model performances against laborious radiomic analysis. The limited added value of radiomics to the overall classification performance for PPGL should be validated in a larger external cohort.

Key Points

• Radiomics derived from [¹⁸F]FDG-PET/CT has the potential to improve the identification of the genetic clusters of pheochromocytomas and paragangliomas.

• A simplified model with SUV_maxonly might be preferred clinically, weighing model performances against the laborious radiomic analysis.

• Cluster 1 and 2 PPGLs generally present distinctive characteristics that can be captured using [¹⁸F]FDG-PET imaging. Sporadic PPGLs appear more heterogeneous, frequently resembling cluster 2 PPGLs and occasionally resembling cluster 1 PPGLs.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00330-022-09034-5.

Convolutional neural networks for automatic image quality control and EARL compliance of PET images

Background

Machine learning studies require a large number of images often obtained on different PET scanners. When merging these images, the use of harmonized images following EARL-standards is essential. However, when including retrospective images, EARL accreditation might not have been in place. The aim of this study was to develop a convolutional neural network (CNN) that can identify retrospectively if an image is EARL compliant and if it is meeting older or newer EARL-standards.

Materials and methods

96 PET images acquired on three PET/CT systems were included in the study. All images were reconstructed with the locally clinically preferred, EARL1, and EARL2 compliant reconstruction protocols. After image pre-processing, one CNN was trained to separate clinical and EARL compliant reconstructions. A second CNN was optimized to identify EARL1 and EARL2 compliant images. The accuracy of both CNNs was assessed using fivefold cross-validation. The CNNs were validated on 24 images acquired on a PET scanner not included in the training data. To assess the impact of image noise on the CNN decision, the 24 images were reconstructed with different scan durations.

Results

In the cross-validation, the first CNN classified all images correctly. When identifying EARL1 and EARL2 compliant images, the second CNN identified 100% EARL1 compliant and 85% EARL2 compliant images correctly. The accuracy in the independent dataset was comparable to the cross-validation accuracy. The scan duration had almost no impact on the results.

Conclusion

The two CNNs trained in this study can be used to retrospectively include images in a multi-center setting by, e.g., adding additional smoothing. This method is especially important for machine learning studies where the harmonization of images from different PET systems is essential.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40658-022-00468-w.

Impact of image quality on radiomics applications

Radiomics features extracted from medical images have been widely reported to be useful in the patient specific outcome modeling for variety of assessment and prediction purposes. Successful application of radiomics features as imaging biomarkers, however, is dependent on the robustness of the approach to the variation in each step of the modeling workflow. Variation in the input image quality is one of the main sources that impacts the reproducibility of radiomics analysis when a model is applied to broader range of medical imaging data. The quality of medical image is generally affected by both the scanner related factors such as image acquisition/reconstruction settings and the patient related factors such as patient motion. This article aimed to review the published literatures in this field that reported the impact of various imaging factors on the radiomics features through the change in image quality. The literatures were categorized by different imaging modalities and also tabulated based on the imaging parameters and the class of radiomics features included in the study. Strategies for image quality standardization were discussed based on the relevant literatures and recommendations for reducing the impact of image quality variation on the radiomics in multi-institutional clinical trial were summarized at the end of this article.

Comparing lesion and feature selections to predict progression in newly diagnosed DLBCL patients with FDG PET/CT radiomics features

PURPOSE

Biomarkers that can accurately predict outcome in DLBCL patients are urgently needed. Radiomics features extracted from baseline [¹⁸F]-FDG PET/CT scans have shown promising results. This study aims to investigate which lesion- and feature-selection approaches/methods resulted in the best prediction of progression after 2 years.

METHODS

A total of 296 patients were included. 485 radiomics features (n = 5 conventional PET, n = 22 morphology, n = 50 intensity, n = 408 texture) were extracted for all individual lesions and at patient level, where all lesions were aggregated into one VOI. 18 features quantifying dissemination were extracted at patient level. Several lesion selection approaches were tested (largest or hottest lesion, patient level [all with/without dissemination], maximum or median of all lesions) and compared to the predictive value of our previously published model. Several data reduction methods were applied (principal component analysis, recursive feature elimination (RFE), factor analysis, and univariate selection). The predictive value of all models was tested using a fivefold cross-validation approach with 50 repeats with and without oversampling, yielding the mean cross-validated AUC (CV-AUC). Additionally, the relative importance of individual radiomics features was determined.

RESULTS

Models with conventional PET and dissemination features showed the highest predictive value (CV-AUC: 0.72-0.75). Dissemination features had the highest relative importance in these models. No lesion selection approach showed significantly higher predictive value compared to our previous model. Oversampling combined with RFE resulted in highest CV-AUCs.

CONCLUSION

Regardless of the applied lesion selection or feature selection approach and feature reduction methods, patient level conventional PET features and dissemination features have the highest predictive value. Trial registration number and date: EudraCT: 2006-005174-42, 01-08-2008.

Quantitative classification and radiomics of [18F]FDG-PET/CT in indeterminate thyroid nodules

PURPOSE

To evaluate whether quantitative [18F]FDG-PET/CT assessment, including radiomic analysis of [18F]FDG-positive thyroid nodules, improved the preoperative differentiation of indeterminate thyroid nodules of non-Hürthle cell and Hürthle cell cytology.

METHODS

Prospectively included patients with a Bethesda III or IV thyroid nodule underwent [18F]FDG-PET/CT imaging. Receiver operating characteristic (ROC) curve analysis was performed for standardised uptake values (SUV) and SUV-ratios, including assessment of SUV cut-offs at which a malignant/borderline neoplasm was reliably ruled out (≥ 95% sensitivity). [18F]FDG-positive scans were included in radiomic analysis. After segmentation at 50% of SUVpeak, 107 radiomic features were extracted from [18F]FDG-PET and low-dose CT images. Elastic net regression classifiers were trained in a 20-times repeated random split. Dimensionality reduction was incorporated into the splits. Predictive performance of radiomics was presented as mean area under the ROC curve (AUC) across the test sets.

RESULTS

Of 123 included patients, 84 (68%) index nodules were visually [18F]FDG-positive. The malignant/borderline rate was 27% (33/123). SUV-metrices showed AUCs ranging from 0.705 (95% CI, 0.601-0.810) to 0.729 (0.633-0.824), 0.708 (0.580-0.835) to 0.757 (0.650-0.864), and 0.533 (0.320-0.747) to 0.700 (0.502-0.898) in all (n = 123), non-Hürthle (n = 94), and Hürthle cell (n = 29) nodules, respectively. At SUVmax, SUVpeak, SUVmax-ratio, and SUVpeak-ratio cut-offs of 2.1 g/mL, 1.6 g/mL, 1.2, and 0.9, respectively, sensitivity of [18F]FDG-PET/CT was 95.8% (95% CI, 78.9-99.9%) in non-Hürthle cell nodules. In Hürthle cell nodules, cut-offs of 5.2 g/mL, 4.7 g/mL, 3.4, and 2.8, respectively, resulted in 100% sensitivity (95% CI, 66.4-100%). Radiomic analysis of 84 (68%) [18F]FDG-positive nodules showed a mean test set AUC of 0.445 (95% CI, 0.290-0.600) for the PET model.

CONCLUSION

Quantitative [18F]FDG-PET/CT assessment ruled out malignancy in indeterminate thyroid nodules. Distinctive, higher SUV cut-offs should be applied in Hürthle cell nodules to optimize rule-out ability. Radiomic analysis did not contribute to the additional differentiation of [18F]FDG-positive nodules.

TRIAL REGISTRATION NUMBER

This trial is registered with ClinicalTrials.gov: NCT02208544 (5 August 2014), https://clinicaltrials.gov/ct2/show/NCT02208544 .

Prediction of Non-Response to Neoadjuvant Chemoradiotherapy in Esophageal Cancer Patients with 18F-FDG PET Radiomics Based Machine Learning Classification

Background: Approximately 26% of esophageal cancer (EC) patients do not respond to neoadjuvant chemoradiotherapy (nCRT), emphasizing the need for pre-treatment selection. The aim of this study was to predict non-response using a radiomic model on baseline ¹⁸F-FDG PET. Methods: Retrospectively, 143 ¹⁸F-FDG PET radiomic features were extracted from 199 EC patients (T1N1-3M0/T2–4aN0-3M0) treated between 2009 and 2019. Non-response (n = 57; 29%) was defined as Mandard Tumor Regression Grade 4–5 (n = 44; 22%) or interval progression (n = 13; 7%). Randomly, 139 patients (70%) were allocated to explore all combinations of 24 feature selection strategies and 6 classification methods towards the cross-validated average precision (AP). The predictive value of the best-performing model, i.e AP and area under the ROC curve analysis (AUC), was evaluated on an independent test subset of 60 patients (30%). Results: The best performing model had an AP (mean ± SD) of 0.47 ± 0.06 on the training subset, achieved by a support vector machine classifier trained on five principal components of relevant clinical and radiomic features. The model was externally validated with an AP of 0.66 and an AUC of 0.67. Conclusion: In the present study, the best-performing model on pre-treatment ¹⁸F-FDG PET radiomics and clinical features had a small clinical benefit to identify non-responders to nCRT in EC.

Design and evaluation of a modular multimodality imaging phantom to simulate heterogeneous uptake and enhancement patterns for radiomic quantification in hybrid imaging; a feasibility study

Background

Accuracy and precision assessment in radiomic features is important for the determination of their potential to characterize cancer lesions. In this regard, simulation of different imaging conditions using specialized phantoms is increasingly being investigated. In this study, the design and evaluation of a modular multimodality imaging phantom to simulate heterogeneous uptake and enhancement patterns for radiomics quantification in hybrid imaging is presented.

Methods

A modular multimodality imaging phantom was constructed that could simulate different patterns of heterogeneous uptake and enhancement patterns in positron emission tomography (PET), single‐photon emission computed tomography (SPECT), computed tomography (CT), and magnetic resonance (MR) imaging. The phantom was designed to be used as an insert in the standard NEMA‐NU2 IEC body phantom casing. The entire phantom insert is composed of three segments, each containing three separately fillable compartments. The fillable compartments between segments had different sizes in order to simulate heterogeneous patterns at different spatial scales. The compartments were separately filled with different ratios of ^99mTc‐pertechnetate, ¹⁸F‐fluorodeoxyglucose ([¹⁸F]FDG), iodine‐ and gadolinium‐based contrast agents for SPECT, PET, CT, and T₁‐weighted MR imaging respectively. Image acquisition was performed using standard oncological protocols on all modalities and repeated five times for repeatability assessment. A total of 93 radiomic features were calculated. Variability was assessed by determining the coefficient of quartile variation (CQV) of the features. Comparison of feature repeatability at different modalities and spatial scales was performed using Kruskal‐Wallis‐, Mann‐Whitney U‐, one‐way ANOVA‐ and independent t‐tests.

Results

Heterogeneous uptake and enhancement could be simulated on all four imaging modalities. Radiomic features in SPECT were significantly less stable than in all other modalities. Features in PET were significantly less stable than in MR and CT. A total of 20 features, particularly in the gray‐level co‐occurrence matrix (GLCM) and gray‐level run‐length matrix (GLRLM) class, were found to be relatively stable in all four modalities for all three spatial scales of heterogeneous patterns (with CQV < 10%).

Conclusion

The phantom was suitable for simulating heterogeneous uptake and enhancement patterns in [¹⁸F]FDG‐PET, ^99mTc‐SPECT, CT, and T₁‐weighted MR images. The results of this work indicate that the phantom might be useful for the further development and optimization of imaging protocols for radiomic quantification in hybrid imaging modalities.

Radiomic analysis for predicting prognosis of colorectal cancer from preoperative 18F-FDG PET/CT

Background

To develop and validate a survival model with clinico-biological features and ¹⁸F- FDG PET/CT radiomic features via machine learning, and for predicting the prognosis from the primary tumor of colorectal cancer.

Methods

A total of 196 pathologically confirmed patients with colorectal cancer (stage I to stage IV) were included. Preoperative clinical factors, serum tumor markers, and PET/CT radiomic features were included for the recurrence-free survival analysis. For the modeling and validation, patients were randomly divided into the training (n = 137) and validation (n = 59) set, while the 78 stage III patients [training (n = 55), and validation (n = 23)] was divided for the further experiment. After selecting features by the log-rank test and variable-hunting methods, random survival forest (RSF) models were built on the training set to analyze the prognostic value of selected features. The performance of models was measured by C-index and was tested on the validation set with bootstrapping. Feature importance and the Pearson correlation were also analyzed.

Results

Radiomics signature (containing four PET/CT features and four clinical factors) achieved the best result for prognostic prediction of 196 patients (C-index 0.780, 95% CI 0.634–0.877). Moreover, four features (including two clinical features and two radiomics features) were selected for prognostic prediction of the 78 stage III patients (C-index was 0.820, 95% CI 0.676–0.900). K–M curves of both models significantly stratified low-risk and high-risk groups (P < 0.0001). Pearson correlation analysis demonstrated that selected radiomics features were correlated with tumor metabolic factors, such as SUVmean, SUVmax.

Conclusion

This study presents integrated clinico-biological-radiological models that can accurately predict the prognosis in colorectal cancer using the preoperative ¹⁸F-FDG PET/CT radiomics in colorectal cancer. It is of potential value in assisting the management and decision making for precision treatment in colorectal cancer.

Trial registration The retrospectively registered study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center (No. 1909207-14-1910) and the data were analyzed anonymously.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-022-03262-5.

Comparison between Two Different Scanners for the Evaluation of the Role of 18F-FDG PET/CT Semiquantitative Parameters and Radiomics Features in the Prediction of Final Diagnosis of Thyroid Incidentalomas

The aim of this study was to compare two different tomographs for the evaluation of the role of semiquantitative PET/CT parameters and radiomics features (RF) in the prediction of thyroid incidentalomas (TIs) at ¹⁸F-FDG imaging. A total of 221 patients with the presence of TIs were retrospectively included. After volumetric segmentation of each TI, semiquantitative parameters and RF were extracted. All of the features were tested for significant differences between the two PET scanners. The performances of all of the features in predicting the nature of TIs were analyzed by testing three classes of final logistic regression predictive models, one for each tomograph and one with both scanners together. Some RF resulted significantly different between the two scanners. PET/CT semiquantitative parameters were not able to predict the final diagnosis of TIs while GLCM-related RF (in particular GLCM entropy_log2 e GLCM entropy_log10) together with some GLRLM-related and GLZLM-related features presented the best predictive performances. In particular, GLCM entropy_log2, GLCM entropy_log10, GLZLM SZHGE, GLRLM HGRE and GLRLM HGZE resulted the RF with best performances. Our study enabled the selection of some RF able to predict the final nature of TIs discovered at ¹⁸F-FDG PET/CT imaging. Classic semiquantitative and volumetric PET/CT parameters did not reveal these abilities. Furthermore, a good overlap in the extraction of RF between the two scanners was underlined.

Effects of tracer uptake time in non-small cell lung cancer 18F-FDG PET radiomics

Positron emission tomography (PET) radiomics applied to oncology allows the measurement of intra-tumoral heterogeneity. This quantification can be affected by image protocols hence there is an increased interest in understanding how radiomic expression on PET images is affected by different imaging conditions. To address that, this study explores how radiomic features are affected by changes in ¹⁸F-FDG uptake time, image reconstruction, lesion delineation, and radiomics binning settings. Methods: Ten non-small cell lung cancer (NSCLC) patients underwent ¹⁸F-FDG PET scans on two consecutive days. On each day, scans were obtained at 60min and 90min post-injection and reconstructed following EARL version 1 (EARL1) and with point-spread-function resolution modelling (PSF-EARL2). Lesions were delineated using thresholds at SUV=4.0, 40% of SUV_max, and with a contrast-based isocontour. PET image intensity was discretized with both fixed bin width (FBW) and fixed bin number (FBN) before the calculation of the radiomic features. Repeatability of features was measured with intraclass correlation (ICC), and the change in feature value over time was calculated as a function of its repeatability. Features were then classified on use-case scenarios based on their repeatability and susceptibility to tracer uptake time. Results: With PSF-EARL2 reconstruction, 40% of SUV_max lesion delineation, and FBW intensity discretization, most features (94%) were repeatable at both uptake times (ICC>0.9), 39% being classified for dual-time-point use-case for being sensitive to changes in uptake time, 39% were classified for cross-sectional studies with unclear dependency on time, 20% classified for cross-sectional use while being robust to tracer uptake time changes, and 6% were discarded for poor repeatability. EARL1 images had one less repeatable feature than PSF-EARL2 (Neighborhood Gray-Level Different Matrix Coarseness), the contrast-based delineation had the poorest repeatability of the delineation methods with 45% features being discarded, and FBN resulted in lower repeatability than FBW (45% and 6% features were discarded, respectively). Conclusion: Repeatability was maximized with PSF-EARL2 reconstruction, lesion delineation at 40% of SUV_max, and FBW intensity discretization. Based on their susceptibility to tracer uptake time, radiomic features were classified into specific NSCLC PET radiomics use-cases.

Machine learning in the differentiation of follicular lymphoma from diffuse large B-cell lymphoma with radiomic [18F]FDG PET/CT features

BACKGROUND

One of the challenges in the management of patients with follicular lymphoma (FL) is the identification of individuals with histological transformation, most commonly into diffuse large B-cell lymphoma (DLBCL). [¹⁸F]FDG-PET/CT is used for staging of patients with lymphoma, but visual interpretation cannot reliably discern FL from DLBCL. This study evaluated whether radiomic features extracted from clinical baseline [¹⁸F]FDG PET/CT and analyzed by machine learning algorithms may help discriminate FL from DLBCL.

MATERIALS AND METHODS

Patients were selected based on confirmed histopathological diagnosis of primary FL (n=44) or DLBCL (n=76) and available [¹⁸F]FDG PET/CT with EARL reconstruction parameters within 6 months of diagnosis. Radiomic features were extracted from the volume of interest on co-registered [¹⁸F]FDG PET and CT images. Analysis of selected radiomic features was performed with machine learning classifiers based on logistic regression and tree-based ensemble classifiers (AdaBoosting, Gradient Boosting, and XG Boosting). The performance of radiomic features was compared with a SUV_max-based logistic regression model.

RESULTS

From the segmented lesions, 121 FL and 227 DLBCL lesions were included for radiomic feature extraction. In total, 79 radiomic features were extracted from the SUVmap, 51 from CT, and 6 shape features. Machine learning classifier Gradient Boosting achieved the best discrimination performance using 136 radiomic features (AUC of 0.86 and accuracy of 80%). SUV_max-based logistic regression model achieved an AUC of 0.79 and an accuracy of 70%. Gradient Boosting classifier had a significantly greater AUC and accuracy compared to the SUV_max-based logistic regression (p≤0.01).

CONCLUSION

Machine learning analysis of radiomic features may be of diagnostic value for discriminating FL from DLBCL tumor lesions, beyond that of the SUV_max alone.

Role of 18F-FDG PET/CT Radiomics Features in the Differential Diagnosis of Solitary Pulmonary Nodules: Diagnostic Accuracy and Comparison between Two Different PET/CT Scanners

The aim of this retrospective study was to investigate the ability of 18 fluorine-fluorodeoxyglucose positron emission tomography/CT (¹⁸F-FDG-PET/CT) metrics and radiomics features (RFs) in predicting the final diagnosis of solitary pulmonary nodules (SPN). We retrospectively recruited 202 patients who underwent a ¹⁸F-FDG-PET/CT before any treatment in two PET scanners. After volumetric segmentation of each lung nodule, 8 PET metrics and 42 RFs were extracted. All the features were tested for significant differences between the two PET scanners. The performances of all features in predicting the nature of SPN were analyzed by testing three classes of final logistic regression predictive models: two were built/trained through exploiting the separate data from the two scanners, and the other joined the data together. One hundred and twenty-seven patients had a final diagnosis of malignancy, while 64 were of a benign nature. Comparing the two PET scanners, we found that all metabolic features and most of RFs were significantly different, despite the cross correlation being quite similar. For scanner 1, a combination between grey level co-occurrence matrix (GLCM), histogram, and grey-level zone length matrix (GLZLM) related features presented the best performances to predict the diagnosis; for scanner 2, it was GLCM and histogram-related features and metabolic tumour volume (MTV); and for scanner 1 + 2, it was histogram features, standardized uptake value (SUV) metrics, and MTV. RFs had a significant role in predicting the diagnosis of SPN, but their accuracies were directly related to the scanner.

Radiomics feature reliability assessed by intraclass correlation coefficient: a systematic review

Radiomics research is rapidly growing in recent years, but more concerns on radiomics reliability are also raised. This review attempts to update and overview the current status of radiomics reliability research in the ever expanding medical literature from the perspective of a single reliability metric of intraclass correlation coefficient (ICC). To conduct this systematic review, Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. After literature search and selection, a total of 481 radiomics studies using CT, PET, or MRI, covering a wide range of subject and disease types, were included for review. In these highly heterogeneous studies, feature reliability to image segmentation was much more investigated than reliability to other factors, such as image acquisition, reconstruction, post-processing, and feature quantification. The reported ICCs also suggested high radiomics feature reliability to image segmentation. Image acquisition was found to introduce much more feature variability than image segmentation, in particular for MRI, based on the reported ICC values. Image post-processing and feature quantification yielded different levels of radiomics reliability and might be used to mitigate image acquisition-induced variability. Some common flaws and pitfalls in ICC use were identified, and suggestions on better ICC use were given. Due to the extremely high study heterogeneities and possible risks of bias, the degree of radiomics feature reliability that has been achieved could not yet be safely synthesized or derived in this review. More future researches on radiomics reliability are warranted.

Radiomics in PET Imaging:: A Practical Guide for Newcomers

Radiomics has undergone considerable development in recent years. In PET imaging, very promising results concerning the ability of handcrafted features to predict the biological characteristics of lesions and to assess patient prognosis or response to treatment have been reported in the literature. This article presents a checklist for designing a reliable radiomic study, gives an overview of the steps of the pipeline, and outlines approaches for data harmonization. Tips are provided for critical reading of the content of articles. The advantages and limitations of handcrafted radiomics compared with deep-learning approaches for the characterization of PET images are also discussed.

A guide to ComBat harmonization of imaging biomarkers in multicenter studies

The impact of PET image acquisition and reconstruction parameters on SUV measurements or radiomic feature values is widely documented. This "scanner" effect is detrimental to the design and validation of predictive or prognostic models and limits the use of large multicenter cohorts. To reduce the impact of this scanner effect, the ComBat method has been proposed and is now used in various contexts. The purpose of this article is to explain and illustrate the use of ComBat based on practical examples. We also give examples in which the ComBat assumptions are not met; thus, ComBat should not be used.

Making Radiomics More Reproducible across Scanner and Imaging Protocol Variations: A Review of Harmonization Methods

Radiomics converts medical images into mineable data via a high-throughput extraction of quantitative features used for clinical decision support. However, these radiomic features are susceptible to variation across scanners, acquisition protocols, and reconstruction settings. Various investigations have assessed the reproducibility and validation of radiomic features across these discrepancies. In this narrative review, we combine systematic keyword searches with prior domain knowledge to discuss various harmonization solutions to make the radiomic features more reproducible across various scanners and protocol settings. Different harmonization solutions are discussed and divided into two main categories: image domain and feature domain. The image domain category comprises methods such as the standardization of image acquisition, post-processing of raw sensor-level image data, data augmentation techniques, and style transfer. The feature domain category consists of methods such as the identification of reproducible features and normalization techniques such as statistical normalization, intensity harmonization, ComBat and its derivatives, and normalization using deep learning. We also reflect upon the importance of deep learning solutions for addressing variability across multi-centric radiomic studies especially using generative adversarial networks (GANs), neural style transfer (NST) techniques, or a combination of both. We cover a broader range of methods especially GANs and NST methods in more detail than previous reviews.

18F-FDG PET baseline radiomics features improve the prediction of treatment outcome in diffuse large B-cell lymphoma

Purpose

Accurate prognostic markers are urgently needed to identify diffuse large B-Cell lymphoma (DLBCL) patients at high risk of progression or relapse. Our purpose was to investigate the potential added value of baseline radiomics features to the international prognostic index (IPI) in predicting outcome after first-line treatment.

Methods

Three hundred seventeen newly diagnosed DLBCL patients were included. Lesions were delineated using a semi-automated segmentation method (standardized uptake value ≥ 4.0), and 490 radiomics features were extracted. We used logistic regression with backward feature selection to predict 2-year time to progression (TTP). The area under the curve (AUC) of the receiver operator characteristic curve was calculated to assess model performance. High-risk groups were defined based on prevalence of events; diagnostic performance was assessed using positive and negative predictive values.

Results

The IPI model yielded an AUC of 0.68. The optimal radiomics model comprised the natural logarithms of metabolic tumor volume (MTV) and of SUV_peak and the maximal distance between the largest lesion and any other lesion (Dmax_bulk, AUC 0.76). Combining radiomics and clinical features showed that a combination of tumor- (MTV, SUV_peak and Dmax_bulk) and patient-related parameters (WHO performance status and age > 60 years) performed best (AUC 0.79). Adding radiomics features to clinical predictors increased PPV with 15%, with more accurate selection of high-risk patients compared to the IPI model (progression at 2-year TTP, 44% vs 28%, respectively).

Conclusion

Prediction models using baseline radiomics combined with currently used clinical predictors identify patients at risk of relapse at baseline and significantly improved model performance.

Trial registration number and date

EudraCT: 2006–005,174-42, 01–08-2008.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00259-021-05480-3.

Incorporating radiomics into clinical trials: expert consensus endorsed by the European Society of Radiology on considerations for data-driven compared to biologically driven quantitative biomarkers

Existing quantitative imaging biomarkers (QIBs) are associated with known biological tissue characteristics and follow a well-understood path of technical, biological and clinical validation before incorporation into clinical trials. In radiomics, novel data-driven processes extract numerous visually imperceptible statistical features from the imaging data with no a priori assumptions on their correlation with biological processes. The selection of relevant features (radiomic signature) and incorporation into clinical trials therefore requires additional considerations to ensure meaningful imaging endpoints. Also, the number of radiomic features tested means that power calculations would result in sample sizes impossible to achieve within clinical trials. This article examines how the process of standardising and validating data-driven imaging biomarkers differs from those based on biological associations. Radiomic signatures are best developed initially on datasets that represent diversity of acquisition protocols as well as diversity of disease and of normal findings, rather than within clinical trials with standardised and optimised protocols as this would risk the selection of radiomic features being linked to the imaging process rather than the pathology. Normalisation through discretisation and feature harmonisation are essential pre-processing steps. Biological correlation may be performed after the technical and clinical validity of a radiomic signature is established, but is not mandatory. Feature selection may be part of discovery within a radiomics-specific trial or represent exploratory endpoints within an established trial; a previously validated radiomic signature may even be used as a primary/secondary endpoint, particularly if associations are demonstrated with specific biological processes and pathways being targeted within clinical trials. KEY POINTS: • Data-driven processes like radiomics risk false discoveries due to high-dimensionality of the dataset compared to sample size, making adequate diversity of the data, cross-validation and external validation essential to mitigate the risks of spurious associations and overfitting. • Use of radiomic signatures within clinical trials requires multistep standardisation of image acquisition, image analysis and data mining processes. • Biological correlation may be established after clinical validation but is not mandatory.

Quantitative Prediction of Microsatellite Instability in Colorectal Cancer With Preoperative PET/CT-Based Radiomics

OBJECTIVES

Microsatellite instability (MSI) status is an important hallmark for prognosis prediction and treatment recommendation of colorectal cancer (CRC). To address issues due to the invasiveness of clinical preoperative evaluation of microsatellite status, we investigated the value of preoperative 18F-FDG PET/CT radiomics with machine learning for predicting the microsatellite status of colorectal cancer patients.

METHODS

A total of 173 patients that underwent 18F-FDG PET/CT scans before operations were retrospectively analyzed in this study. The microsatellite status for each patient was identified as microsatellite instability-high (MSI-H) or microsatellite stable (MSS), according to the test for mismatch repair gene proteins with immunohistochemical staining methods. There were 2,492 radiomic features in total extracted from 18F-FDG PET/CT imaging. Then, radiomic features were selected through multivariate random forest selection and univariate relevancy tests after handling the imbalanced dataset through the random under-sampling method. Based on the selected features, we constructed a BalancedBagging model based on Adaboost classifiers to identify the MSI status in patients with CRC. The model performance was evaluated by the area under the curve (AUC), sensitivity, specificity, and accuracy on the validation dataset.

RESULTS

The ensemble model was constructed based on two radiomic features and achieved an 82.8% AUC for predicting the MSI status of colorectal cancer patients. The sensitivity, specificity, and accuracy were 83.3, 76.3, and 76.8%, respectively. The significant correlation of the selected two radiomic features with multiple effective clinical features was identified (p < 0.05).

CONCLUSION

18F-FDG PET/CT radiomics analysis with the machine learning model provided a quantitative, efficient, and non-invasive mechanism for identifying the microsatellite status of colorectal cancer patients, which optimized the treatment decision support.

Quantitative radiomics features in diffuse large B-cell lymphoma: does segmentation method matter?

INTRODUCTION

Radiomics features may predict outcome in diffuse large B-cell lymphoma (DLBCL). Currently, multiple segmentation methods are used to calculate metabolic tumor volume (MTV). We assessed the influence of segmentation method on the discriminative power of radiomics features in DLBCL for patient level and for the largest lesion. Methods: 50 baseline ¹⁸F-fluorodeoxyglucose positron emission tomography computed tomography (PET/CT) scans of DLBCL patients who progressed or relapsed within 2 years after diagnosis were matched on uptake time and reconstruction method with 50 baseline PET/CT scans of DLBCL patients without progression. Scans were analysed using 6 semi-automatic segmentation methods (standardized uptake value (SUV)4.0, SUV2.5, 41% of the maximum SUV, 50% of the SUVpeak, majority vote (MV)2 and MV3, respectively). Based on these segmentations, 490 radiomics features were extracted at patient level and 486 features for the largest lesion. To quantify the agreement between features extracted from different segmentation methods, the intra-class correlation (ICC) agreement was calculated for each method compared to SUV4.0. The feature space was reduced by deleting features that had high Pearson correlations (≥0.7) with the previously established predictors MTV and/or SUVpeak. Model performance was assessed using stratified repeated cross-validation with 5 folds and 2000 repeats yielding the mean receiver-operating characteristics curve integral (CV-AUC) for all segmentation methods using logistic regression with backward feature selection. Results: The percentage of features yielding an ICC ≥0.75 compared to the SUV4.0 segmentation was lowest for A50P both at patient level and for the largest lesion, with 77.3% and 66.7% of the features yielding an ICC ≥0.75, respectively. Features were not highly correlated with MTV, with at least 435 features at patient level and 409 features for the largest lesion for all segmentation methods with a correlation coefficient <0.7. Features were highly correlated with SUVpeak (at least 190 and 134 were uncorrelated, respectively). CV-AUCs ranged between 0.69±0.11 and 0.84±0.09 for patient level, and between 0.69±0.11 and 0.73±0.10 for lesion level. Conclusion: Even though there are differences in the actual radiomics feature values derived and selected features between segmentation methods, there is no substantial difference in the discriminative power of radiomics features between segmentation methods.

A transfer learning approach to facilitate ComBat-based harmonization of multicentre radiomic features in new datasets

PURPOSE

To facilitate the demonstration of the prognostic value of radiomics, multicenter radiomics studies are needed. Pooling radiomic features of such data in a statistical analysis is however challenging, as they are sensitive to the variability in scanner models, acquisition protocols and reconstruction settings, which is often unavoidable in a multicentre retrospective analysis. A statistical harmonization strategy called ComBat was utilized in radiomics studies to deal with the "center-effect". The goal of the present work was to integrate a transfer learning (TL) technique within ComBat-and recently developed alternate versions of ComBat with improved flexibility (M-ComBat) and robustness (B-ComBat)-to allow the use of a previously determined harmonization transform to the radiomic feature values of new patients from an already known center.

MATERIAL AND METHODS

The proposed TL approach were incorporated in the four versions of ComBat (standard, B, M, and B-M ComBat). The proposed approach was evaluated using a dataset of 189 locally advanced cervical cancer patients from 3 centers, with magnetic resonance imaging (MRI) and positron emission tomography (PET) images, with the clinical endpoint of predicting local failure. The impact performance of the TL approach was evaluated by comparing the harmonization achieved using only parts of the data to the reference (harmonization achieved using all the available data). It was performed through three different machine learning pipelines.

RESULTS

The proposed TL technique was successful in harmonizing features of new patients from a known center in all versions of ComBat, leading to predictive models reaching similar performance as the ones developed using the features harmonized with all the data available.

CONCLUSION

The proposed TL approach enables applying a previously determined ComBat transform to new, previously unseen data.

Evaluation and optimization of radiomics features stability to respiratory motion in 18 F-FDG 3D PET imaging

PURPOSE

To evaluate the impact of respiratory motion on radiomics features in ¹⁸ F-fluoro-2-deoxy-D-glucose three dimensional positron emission tomography (¹⁸ F-FDG 3D PET) imaging and optimize feature stability by combining preprocessing configurations and aggregation strategies.

METHODS

An in-house developed respiratory motion phantom was imaged in 3D PET scanner under nine respiratory patterns including one reference pattern. In total, 487 radiomics features were extracted for each respiratory pattern. Feature stability to respiratory motion was first evaluated by metrics of coefficient of variation (COV) and relative difference (RD) in a fixed preprocessing configuration. Further, one-way ANOVA and trend analysis were performed to evaluate the impact of preprocessing configuration (voxel size, discretization scheme) and aggregation strategy on feature stability. Finally, an optimization framework was proposed by selected feature-specific configurations with minimum COV value, and the diagnostic performance was validated in stable versus unstable features and fixed versus optimal features by a set of 46 patients with lung disease.

RESULTS

PET radiomics features were sensitive to respiratory motion, only 79/487 (16%) features were identified to be very stable in the fixed configuration. Preprocessing configuration and aggregation strategy had an impact on feature stability. For different voxel size, bin number, bin size and aggregation strategy, 188/487 (39%), 70/487 (15%), 148/487 (30%), and 38/95 (29%) features appeared significant changes in feature stability. The optimized configuration had the potential to improve feature stability compared to fixed configuration, with the COV decreased from 22% ±24% to 16% ±20%. Regarding the diagnostic performance, the stable and optimal configuration features outperformed the unstable and fixed configuration features, respectively (AUC 0.88, 0.87 vs. 0.83, 0.85).

CONCLUSIONS

Respiratory motion shows considerable impact on feature stability in 3D PET imaging, while optimizing preprocessing configuration may improve feature stability and diagnostic performance.

Impact of segmentation and discretization on radiomic features in 68Ga-DOTA-TOC PET/CT images of neuroendocrine tumor

OBJECTIVE

To identify the impact of segmentation methods and intensity discretization on radiomic features (RFs) extraction from 68Ga-DOTA-TOC PET images in patients with neuroendocrine tumors.

METHODS

Forty-nine patients were retrospectively analyzed. Tumor contouring was performed manually by four different operators and with a semi-automatic edge-based segmentation (SAEB) algorithm. Three SUVmax fixed thresholds (20, 30, 40%) were applied. Fifty-one RFs were extracted applying two different intensity rescale factors for gray-level discretization: one absolute (AR60 = SUV from 0 to 60) and one relative (RR = min-max of the VOI SUV). Dice similarity coefficient (DSC) was calculated to quantify segmentation agreement between different segmentation methods. The impact of segmentation and discretization on RFs was assessed by intra-class correlation coefficients (ICC) and the coefficient of variance (COVL). The RFs' correlation with volume and SUVmax was analyzed by calculating Pearson's correlation coefficients.

RESULTS

DSC mean value was 0.75 ± 0.11 (0.45-0.92) between SAEB and operators and 0.78 ± 0.09 (0.36-0.97), among the four manual segmentations. The study showed high robustness (ICC > 0.9): (a) in 64.7% of RFs for segmentation methods using AR60, improved by applying SUVmax threshold of 40% (86.5%); (b) in 50.9% of RFs for different SUVmax thresholds using AR60; and (c) in 37% of RFs for discretization settings using different segmentation methods. Several RFs were not correlated with volume and SUVmax.

CONCLUSIONS

RFs robustness to manual segmentation resulted higher in NET 68Ga-DOTA-TOC images compared to 18F-FDG PET/CT images. Forty percent SUVmax thresholds yield superior RFs stability among operators, however leading to a possible loss of biological information. SAEB segmentation appears to be an optimal alternative to manual segmentation, but further validations are needed. Finally, discretization settings highly impacted on RFs robustness and should always be stated.