Multiscale Texture Analysis: From 18F-FDG PET Images to Histologic Images

Histology-Based Radiomics for [18F]FDG PET Identifies Tissue Heterogeneity in Pancreatic Cancer

Radiomics features can reveal hidden patterns in a tumor but usually lack an underlying biologic rationale. In this work, we aimed to investigate whether there is a correlation between radiomics features extracted from [18F]FDG PET images and histologic expression patterns of a glycolytic marker, monocarboxylate transporter-4 (MCT4), in pancreatic cancer. Methods: A cohort of pancreatic ductal adenocarcinoma patients (n = 29) for whom both tumor cross sections and [18F]FDG PET/CT scans were available was used to develop an [18F]FDG PET radiomics signature. By using immunohistochemistry for MCT4, we computed density maps of MCT4 expression and extracted pathomics features. Cluster analysis identified 2 subgroups with distinct MCT4 expression patterns. From corresponding [18F]FDG PET scans, radiomics features that associate with the predefined MCT4 subgroups were identified. Results: Complex heat map visualization showed that the MCT4-high/heterogeneous subgroup was correlating with a higher MCT4 expression level and local variation. This pattern linked to a specific [18F]FDG PET signature, characterized by a higher SUVmean and SUVmax and second-order radiomics features, correlating with local variation. This MCT4-based [18F]FDG PET signature of 7 radiomics features demonstrated prognostic value in an independent cohort of pancreatic cancer patients (n = 71) and identified patients with worse survival. Conclusion: Our cross-modal pipeline allows the development of PET scan signatures based on immunohistochemical analysis of markers of a particular biologic feature, here demonstrated on pancreatic cancer using intratumoral MCT4 expression levels to select [18F]FDG PET radiomics features. This study demonstrated the potential of radiomics scores to noninvasively capture intratumoral marker heterogeneity and identify a subset of pancreatic ductal adenocarcinoma patients with a poor prognosis.

Prediction of Histological Grade of Oral Squamous Cell Carcinoma Using Machine Learning Models Applied to 18F-FDG-PET Radiomics

The histological grade of oral squamous cell carcinoma affects the prognosis. In the present study, we performed a radiomics analysis to extract features from 18F-FDG PET image data, created machine learning models from the features, and verified the accuracy of the prediction of the histological grade of oral squamous cell carcinoma. The subjects were 191 patients in whom an 18F-FDG-PET examination was performed preoperatively and a histopathological grade was confirmed after surgery, and their tumor sizes were sufficient for a radiomics analysis. These patients were split in a 70%/30% ratio for use as training data and testing data, respectively. We extracted 2993 radiomics features from the PET images of each patient. Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), Naïve Bayes (NB), and K-Nearest Neighbor (KNN) machine learning models were created. The areas under the curve obtained from receiver operating characteristic curves for the prediction of the histological grade of oral squamous cell carcinoma were 0.72, 0.71, 0.84, 0.74, and 0.73 for LR, SVM, RF, NB, and KNN, respectively. We confirmed that a PET radiomics analysis is useful for the preoperative prediction of the histological grade of oral squamous cell carcinoma.

Characterization of Mediastinal Bulky Lymphomas with FDG-PET-Based Radiomics and Machine Learning Techniques

BACKGROUND

This study tested the diagnostic value of 18F-FDG PET/CT (FDG-PET) volumetric and texture parameters in the histological differentiation of mediastinal bulky disease due to classical Hodgkin lymphoma (cHL), primary mediastinal B-cell lymphoma (PMBCL) and grey zone lymphoma (GZL), using machine learning techniques.

METHODS

We reviewed 80 cHL, 29 PMBCL and 8 GZL adult patients with mediastinal bulky disease and histopathological diagnoses who underwent FDG-PET pre-treatment. Volumetric and radiomic parameters were measured using FDG-PET both for bulky lesions (BL) and for all lesions (AL) using LIFEx software (threshold SUV ≥ 2.5). Binary and multiclass classifications were performed with various machine learning techniques fed by a relevant subset of radiomic features.

RESULTS

The analysis showed significant differences between the lymphoma groups in terms of SUVmax, SUVmean, MTV, TLG and several textural features of both first- and second-order grey level. Among machine learning classifiers, the tree-based ensembles achieved the best performance both for binary and multiclass classifications in histological differentiation.

CONCLUSIONS

Our results support the value of metabolic heterogeneity as an imaging biomarker, and the use of radiomic features for early characterization of mediastinal bulky lymphoma.

Application of Machine Learning for Differentiating Bone Malignancy on Imaging: A Systematic Review

An accurate diagnosis of bone tumours on imaging is crucial for appropriate and successful treatment. The advent of Artificial intelligence (AI) and machine learning methods to characterize and assess bone tumours on various imaging modalities may assist in the diagnostic workflow. The purpose of this review article is to summarise the most recent evidence for AI techniques using imaging for differentiating benign from malignant lesions, the characterization of various malignant bone lesions, and their potential clinical application. A systematic search through electronic databases (PubMed, MEDLINE, Web of Science, and clinicaltrials.gov) was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A total of 34 articles were retrieved from the databases and the key findings were compiled and summarised. A total of 34 articles reported the use of AI techniques to distinguish between benign vs. malignant bone lesions, of which 12 (35.3%) focused on radiographs, 12 (35.3%) on MRI, 5 (14.7%) on CT and 5 (14.7%) on PET/CT. The overall reported accuracy, sensitivity, and specificity of AI in distinguishing between benign vs. malignant bone lesions ranges from 0.44–0.99, 0.63–1.00, and 0.73–0.96, respectively, with AUCs of 0.73–0.96. In conclusion, the use of AI to discriminate bone lesions on imaging has achieved a relatively good performance in various imaging modalities, with high sensitivity, specificity, and accuracy for distinguishing between benign vs. malignant lesions in several cohort studies. However, further research is necessary to test the clinical performance of these algorithms before they can be facilitated and integrated into routine clinical practice.

Primary tumor heterogeneity on pretreatment 18F-FDG PET/CT to predict outcome in patients with rectal cancer who underwent surgery after neoadjuvant therapy

PURPOSE

This retrospective study aimed to investigate the value of texture features of primary tumors in pretreatment ¹⁸F-FDG PET/CT in the prediction of response to treatment, progression, and overall survival in patients with rectal cancer who underwent surgery after neoadjuvant therapy(NAT).

METHODS

Patients with rectal cancer who had pretreatment ¹⁸F-FDG PET/CT, and underwent surgery after NAT were included in this study. Clinicopathologic features, date of last follow-up, progression, and death were recorded. Textural and conventional PET parameters(maximum standardized uptake value-SUVmax, metabolic tumor volume-MTV, total lesion glycolysis-TLG) were obtained from PET/CT images using LifeX program. Parameters were grouped using Youden index in ROC analysis. Factors predicting the pathological response to treatment, progression, and overall survival were determined using logistic regression and Cox regression analyses.

RESULTS

Forty-four patients (26(59%) male, 18(41%) female; 60.1±11.4 years) with rectal cancer were included in this study. The numbers of patients with responders and non-responders to NAT were 15(34.9%) and 28(65.1%), respectively. One patient' pathology report did not contain the response status to NAT. The median of follow-up duration was 29.9 months. 9(20.5%) showed disease progression, and 8(18.2%) died during the follow-up period. Difference entropy_GLCM and correlation_GLCM parameters were found as independent predictors for response to NAT. The positivity of surgical margin, intensity interquartile range_CONV and AUC-CSH_DISC texture parameters were independent predictors of progression, while normalized inverse difference_GLCM and LZLGE_GLZLM parameters were independent predictors of mortality.

CONCLUSION

The texture parameters obtained from pretreatment ¹⁸F-FDG PET/CT have presented a more robust predictive value than conventional parameters in patients with rectal cancer who underwent surgery after NAT.

Multimodal MRI radiomic models to predict genomic mutations in diffuse intrinsic pontine glioma with missing imaging modalities

Purpose

Predicting H3.1, TP53, and ACVR1 mutations in DIPG could aid in the selection of therapeutic options. The contribution of clinical data and multi-modal MRI were studied for these three predictive tasks. To keep the maximum number of subjects, which is essential for a rare disease, missing data were considered. A multi-modal model was proposed, collecting all available data for each patient, without performing any imputation.

Methods

A retrospective cohort of 80 patients with confirmed DIPG and at least one of the four MR modalities (T1w, T1c, T2w, and FLAIR), acquired with two different MR scanners was built. A pipeline including standardization of MR data and extraction of radiomic features within the tumor was applied. The values of radiomic features between the two MR scanners were realigned using the ComBat method. For each prediction task, the most robust features were selected based on a recursive feature elimination with cross-validation. Five different models, one based on clinical data and one per MR modality, were developed using logistic regression classifiers. The prediction of the multi-modal model was defined as the average of all possible prediction results among five for each patient. The performances of the models were compared using a leave-one-out approach.

Results

The percentage of missing modalities ranged from 6 to 11% across modalities and tasks. The performance of each individual model was dependent on each specific task, with an AUC of the ROC curve ranging from 0.63 to 0.80. The multi-modal model outperformed the clinical model for each prediction tasks, thus demonstrating the added value of MRI. Furthermore, regardless of performance criteria, the multi-modal model came in the first place or second place (very close to first). In the leave-one-out approach, the prediction of H3.1 (resp. ACVR1 and TP53) mutations achieved a balanced accuracy of 87.8% (resp. 82.1 and 78.3%).

Conclusion

Compared with a single modality approach, the multi-modal model combining multiple MRI modalities and clinical features was the most powerful to predict H3.1, ACVR1, and TP53 mutations and provided prediction, even in the case of missing modality. It could be proposed in the absence of a conclusive biopsy.

Spatially coherent modeling of 3D FDG-PET data for assessment of intratumoral heterogeneity and uptake gradients

Purpose: Radiomics have become invaluable for non-invasive cancer patient risk prediction, and the community now turns to exogenous assessment, e.g., from genomics, for interpretability of these agnostic analyses. Yet, some opportunities for clinically interpretable modeling of positron emission tomography (PET) imaging data remain unexplored, that could facilitate insightful characterization at voxel level. Approach: Here, we present a novel deformable tubular representation of the distribution of tracer uptake within a volume of interest, and derive interpretable prognostic summaries from it. This data-adaptive strategy yields a 3D-coherent and smooth model fit, and a profile curve describing tracer uptake as a function of voxel location within the volume. Local trends in uptake rates are assessed at each voxel via the calculation of gradients derived from this curve. Intratumoral heterogeneity can also be assessed directly from it. Results: We illustrate the added value of this approach over previous strategies, in terms of volume rendering and coherence of the structural representation of the data. We further demonstrate consistency of the implementation via simulations, and prognostic potential of heterogeneity and statistical summaries of the uptake gradients derived from the model on a clinical cohort of 158 sarcoma patients imaged withF 18 -fluorodeoxyglucose-PET, in multivariate prognostic models of patient survival. Conclusions: The proposed approach captures uptake characteristics consistently at any location, and yields a description of variations in uptake that holds prognostic value complementarily to structural heterogeneity. This creates opportunities for monitoring of local areas of greater interest within a tumor, e.g., to assess therapeutic response in avid locations.

Prognostic Value of Metabolic, Volumetric and Textural Parameters of Baseline [18F]FDG PET/CT in Early Triple-Negative Breast Cancer

Simple Summary

The aim of this study was to evaluate PET/CT parameters to determine different prognostic groups in TNBC, in order to select patients with a high risk of relapse, for whom therapeutic escalation can be considered. We have demonstrated that the MTV, TLG and entropy of the primary breast lesion could be of interest to predict the prognostic outcome of TNBC patients.

Abstract

(1) Background: triple-negative breast cancer (TNBC) remains a clinical and therapeutic challenge primarily affecting young women with poor prognosis. TNBC is currently treated as a single entity but presents a very diverse profile in terms of prognosis and response to treatment. Positron emission tomography/computed tomography (PET/CT) with ¹⁸F-fluorodeoxyglucose ([¹⁸F]FDG) is gaining importance for the staging of breast cancers. TNBCs often show high [¹⁸F]FDG uptake and some studies have suggested a prognostic value for metabolic and volumetric parameters, but no study to our knowledge has examined textural features in TNBC. The objective of this study was to evaluate the association between metabolic, volumetric and textural parameters measured at the initial [¹⁸F]FDG PET/CT and disease-free survival (DFS) and overall survival (OS) in patients with nonmetastatic TBNC. (2) Methods: all consecutive nonmetastatic TNBC patients who underwent a [¹⁸F]FDG PET/CT examination upon diagnosis between 2012 and 2018 were retrospectively included. The metabolic and volumetric parameters (SUV_max, SUV_mean, SUV_peak, MTV, and TLG) and the textural features (entropy, homogeneity, SRE, LRE, LGZE, and HGZE) of the primary tumor were collected. (3) Results: 111 patients were enrolled (median follow-up: 53.6 months). In the univariate analysis, high TLG, MTV and entropy values of the primary tumor were associated with lower DFS (p = 0.008, p = 0.006 and p = 0.025, respectively) and lower OS (p = 0.002, p = 0.001 and p = 0.046, respectively). The discriminating thresholds for two-year DFS were calculated as 7.5 for MTV, 55.8 for TLG and 2.6 for entropy. The discriminating thresholds for two-year OS were calculated as 9.3 for MTV, 57.4 for TLG and 2.67 for entropy. In the multivariate analysis, lymph node involvement in PET/CT was associated with lower DFS (p = 0.036), and the high MTV of the primary tumor was correlated with lower OS (p = 0.014). (4) Conclusions: textural features associated with metabolic and volumetric parameters of baseline [¹⁸F]FDG PET/CT have a prognostic value for identifying high-relapse-risk groups in early TNBC patients.

Association of 18F-FDG PET/CT textural features with immunohistochemical characteristics in invasive ductal breast cancer

OBJECTıVES: This study investigates whether textural features (TFs) extracted from ¹⁸F-FDG positron emission tomography/computed tomography (PET/CT) are associated with immunohistochemical characteristics (IHCs) of invasive ductal breast carcinoma (IDBC).

MATERIALS AND METHODS

The relationship of TFs with IHCs [estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor-2 (HER-2), Ki-67 proliferation index, and histological grades] from solely excised primary tumors were evaluated for a more accurate assessment. Therefore patients with early-stage IDBC who underwent pre-operative ¹⁸F-FDG PET/CT scan for staging were included in this retrospective study. The clinical staging was performed according to the 8th edition of the American Joint Committee on Cancer. Maximum standardized uptake value (SUVmax) and 37TFs of the primary tumor were extracted from ¹⁸F-FDG PET/CT. Spearman's rank correlation test was used to evaluate the correlation between TFs and SUVmax. Receiver operating characteristic curves were generated to define the diagnostic performance of each parameter. Among these parameters, those with the highest diagnostic performance were included in the multivariate logistic regression model to identify the independent predictors of histopathological characteristics.

RESULTS

A total of 124 patients were included. Histogram-uniformity, grey-level co-occurrence matrix (GLCM), GLCM-energy, and GLCM-homogeneity showed a strong negative correlation with SUVmax, while grey-level run-length matrix (GLRLM), GLRLM-SRHGE, grey-level zone length matrix (GLZLM), GLZLM-HGZE, GLRLM-HGRE, GLCM-entropy, GLCM-contrast, histogram-entropy, and GLCM-dissimilarity showed a strong positive correlation. Some of the TFs were independently associated with ER-negativity, PR-negativity, HER-2-positivity, and increased Ki-67 proliferation index (GLCM-contrast, GLZLM-GLNU, histogram-uniformity, and shape-sphericity respectively). While SUVmax had an independent association with high-grade and triple-negativity, GLZLM-SZLGE, a high-order TF that shows the distribution of the short homogeneous zones with low grey-levels, had an independent association with axillary lymph node metastasis.

CONCLUSIONS

ER-negative, PR-negative, HER-2-positive, triple-negative, high-grade, highly proliferative, and high-stage tumors were found to be more glycolytic and metabolically heterogeneous. These findings suggest that the use of TFs in addition to SUVmax may improve the prognostic value of ¹⁸F-FDG PET/CT in IDBC, as certain TFs were independently associated with many IHCs and predicted axillary lymph node involvement.

Improved 18-FDG PET/CT diagnosis of multiple myeloma diffuse disease by radiomics analysis

OBJECTIVES

In multiple myeloma, the diagnosis of diffuse bone marrow infiltration on 18-FDG PET/CT can be challenging. We aimed to develop a PET/CT radiomics-based model that could improve the diagnosis of multiple myeloma diffuse disease on 18-FDG PET/CT.

METHODS

We prospectively performed PET/CT and whole-body diffusion-weighted MRI in 30 newly diagnosed multiple myeloma. MRI was the reference standard for diffuse disease assessment. Twenty patients were randomly assigned to a training set and 10 to an independent test set. Visual analysis of PET/CT was performed by two nuclear medicine physicians. Spine volumes were automatically segmented, and a total of 174 Imaging Biomarker Standardisation Initiative-compliant radiomics features were extracted from PET and CT. Selection of best features was performed with random forest features importance and correlation analysis. Machine-learning algorithms were trained on the selected features with cross-validation and evaluated on the independent test set.

RESULTS

Out of the 30 patients, 18 had established diffuse disease on MRI. The sensitivity, specificity and accuracy of visual analysis were 67, 75 and 70%, respectively, with a moderate kappa coefficient of agreement of 0.6. Five radiomics features were selected. On the training set, random forest classifier reached a sensitivity, specificity and accuracy of 93, 86 and 91%, respectively, with an area under the curve of 0.90 (95% confidence interval, 0.89-0.91). On the independent test set, the model achieved an accuracy of 80%.

CONCLUSIONS

Radiomics analysis of 18-FDG PET/CT images with machine-learning overcame the limitations of visual analysis, providing a highly accurate and more reliable diagnosis of diffuse bone marrow infiltration in multiple myeloma patients.

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 different overview of staging PET/CT images in patients with esophageal cancer: the role of textural analysis with machine learning methods

OBJECTIVE

This study evaluates the ability of several machine learning (ML) algorithms, developed using volumetric and texture data extracted from baseline 18F-FDG PET/CT studies performed initial staging of patient with esophageal cancer (EC), to predict survival and histopathology.

METHODS

The initial staging 18F-FDG PET/CT images obtained on newly diagnosed EC patients between January 2008 and June 2019 were evaluated using LIFEx software. A region of interest (ROI) of the primary tumor was created and volumetric and textural features were obtained. A significant relationship between these features and pathological subtypes, 1-year, and 5-year survival was investigated. Due to the nonhomogeneity of the data, nonparametric test (The Mann-Whitney U test) was used for each feature, in pairwise comparisons of independent variables. A p value of < 0.05 was considered significant. Receiver operating curve (ROC) analysis was performed for features with p < 0.05. Correlation between the significant features was evaluated with Spearman correlation test; features with correlation coefficient < 0.8 were evaluated with several ML algorithms.

RESULTS

In predicting survival in a 1-year follow-up J48 was obtained as the most successful algorithm (AUC: 0.581, PRC: 0.565, MCC: 0.258, acc: 64.29%). 5-year survival results were more promising than 1-year survival results with (AUC: 0.820, PRC: 0.860, MCC: 271, acc: 81.36%) by logistic regression. It is revealed that the most successful algorithm was naive bayes (AUC: 0.680 PRC: 0.776, MCC: 0.298, acc: 82.66%) in the histopathological discrimination.

CONCLUSION

Texture analysis with ML algorithms could be predictive of overall survival and discriminating histopathological subtypes of EC.

The preoperative prognostic value of the radiomics nomogram based on CT combined with machine learning in patients with intrahepatic cholangiocarcinoma

Background

Intrahepatic cholangiocarcinoma is an aggressive liver carcinoma with increasing incidence and mortality. A good auxiliary prognostic prediction tool is desperately needed for the development of treatment strategies. The purpose of this study was to explore the prognostic value of the radiomics nomogram based on enhanced CT in intrahepatic cholangiocarcinoma.

Methods

In this retrospective study, 101 patients with pathological confirmation of intrahepatic cholangiocarcinoma were recruited. A radiomics nomogram was developed by radiomics score and independent clinical risk factors selecting from multivariate Cox regression. All patients were stratified as high risk and low risk by a nomogram. Model performance and clinical usefulness were assessed by calibration curve, ROC curve, and survival curve.

Results

A total of 101patients (mean age, 58.2 years old; range 36–79 years old) were included in the study. The 1-year, 3-year, and 5-year overall survival rates were 49.5%, 26.6%, and 14.4%, respectively, with a median survival time of 12.2 months in the whole set. The least absolute shrinkage and selection operator (LASSO) method selected 3 features. Multivariate Cox analysis found three independent prognostic factors. The radiomics nomogram showed a significant prognosis value with overall survival. There was a significant difference in the 1-year and 3-year survival rates of stratified high-risk and low-risk patients in the whole set (30.4% vs. 56.4% and 13.0% vs. 30.6%, respectively, p = 0.018).

Conclusions

This radiomics nomogram has potential application value in the preoperative prognostic prediction of intrahepatic cholangiocarcinoma and may facilitate in clinical decision-making.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12957-021-02162-0.

Relationship between tumor heterogeneity and volume in cervical cancer: Evidence from integrated fluorodeoxyglucose 18 PET/MR texture analysis

OBJECTIVE

The aim of this study was to evaluate the effect of cervical cancer volume on PET/magnetic resonance (MR) texture heterogeneity.

MATERIALS AND METHODS

We retrospectively analyzed the PET/MR images of 138 patients with pathologically diagnosed cervical squamous cell carcinoma, including 50 patients undergoing surgery and 88 patients receiving concurrent chemoradiotherapy. Fluorodeoxyglucose 18 (18FDG)-PET/MR examination were performed for each patient before treatment, and the PET and MR texture analysis were undertaken. The texture features of the tumor based on gray-level co-occurrence matrices were extracted, and the correlation between tumor texture features and volume parameters was analyzed using Spearman's rank correlation coefficient. Finally, the variation trend of tumor texture heterogeneity was analyzed as tumor volumes increased.

RESULTS

PET texture features were highly correlated with metabolic tumor volume (MTV), including entropy-log2, entropy-log10, energy, homogeneity, dissimilarity, contrast, correlation, and the correlation coefficients (rs) were 0.955, 0.955, -0.897, 0.883, -0.881, -0.876, and 0.847 (P < 0.001), respectively. In the range of smaller MTV, the texture heterogeneity of energy, entropy-log2, and entropy-log10 increases with an increase in tumor volume, whereas the texture heterogeneity of homogeneity, dissimilarity, contrast, and correlation decreases with an increase in tumor volume. Only homogeneity, contrast, correlation, and dissimilarity had high correlation with tumor volume on MRI. The correlation coefficients (rs) were 0.76, -0.737, 0.644, and -0.739 (P < 0.001), respectively. The texture heterogeneity of MRI features that are highly correlated with tumor volume decreases with increasing tumor volume.

CONCLUSION

In the small tumor volume range, the heterogeneity variation trend of PET texture features is inconsistent as the tumor volume increases, but the variation trend of MRI texture heterogeneity is consistent, and MRI texture heterogeneity decreases as tumor volume increases. These results suggest that MRI is a better imaging modality when compared with PET in determining tumor texture heterogeneity in the small tumor volume range.

Progression to Symptomatic Multiple Myeloma Predicted by Texture Analysis-Derived Parameters in Patients Without Focal Disease at 18F-FDG PET/CT

This retrospective study is focused on the possible clinical implications of texture analysis-derived PET parameters in patients with smoldering multiple myeloma. Several texture features are significantly associated with progression to symptomatic multiple myeloma and with a shorter time to progression. The results of this study may lead to early identification of patients who could benefit from specific therapies.

BACKGROUND

The aim of the study was to determine whether positron emission tomography parameters derived from texture analysis of axial and peripheral skeleton predict progression to symptomatic multiple myeloma (MM) in patients undergoing 18F-​fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) without evidence of focal sites of ¹⁸F-FDG uptake.

PATIENTS AND METHODS

Patients with smoldering MM who underwent ¹⁸F-FDG PET/CT from May 2014 to June 2018 were retrospectively reviewed. Volumes of interest (VOIs) were placed on T5-T7 and L2-L4, iliac crests, and femoral diaphyses. Dedicated software (LIFEx) allowed us to obtain PET-derived first-, second-, and higher order texture features. Possible associations between PET parameters and progression to symptomatic MM were determined. Kaplan-Meier curves allowed to assess time to progression (TTP) based on the PET parameters.

RESULTS

Forty-five patients were included: 26 patients (58%) did not meet the criteria for symptomatic MM, but 19 patients (42%) progressed to symptomatic MM. Several texture features extracted from VOIs placed on iliac crests and femoral diaphyses were significantly associated with progression to symptomatic MM and with a shorter TTP (P < .05); conversely, the above-mentioned parameters extracted from VOIs placed on T5-T7 and L2-L4 did not significantly differ among the patients with regard to their progression to symptomatic MM and length of TTP, except for the gray-level zone length matrix-short-zone low-gray-level emphasis and gray-level zone length matrix-low gray-level zone emphasis. Particularly, second- and higher order texture features showed a significant association with the above-mentioned outcomes.

CONCLUSION

Texture features derived from PET may be an expression of subtle disease distribution in the axial and peripheral bone marrow.

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.

Preoperative Texture Analysis Using 11C-Methionine Positron Emission Tomography Predicts Survival after Surgery for Glioma

BACKGROUND

Positron emission tomography with ¹¹C-methionine (MET) is well established in the diagnostic work-up of malignant brain tumors. Texture analysis is a novel technique for extracting information regarding relationships among surrounding voxels, in order to quantify their inhomogeneity. This study evaluated whether the texture analysis of MET uptake has prognostic value for patients with glioma.

METHODS

We retrospectively analyzed adults with glioma who had undergone preoperative metabolic imaging at a single center. Tumors were delineated using a threshold of 1.3-fold of the mean standardized uptake value for the contralateral cortex, and then processed to calculate the texture features in glioma.

RESULTS

The study included 42 patients (median age: 56 years). The World Health Organization classifications were grade II (7 patients), grade III (17 patients), and grade IV (18 patients). Sixteen (16.1%) all-cause deaths were recorded during the median follow-up of 18.8 months. The univariate analyses revealed that overall survival (OS) was associated with age (hazard ratio (HR) 1.04, 95% confidence interval (CI) 1.01-1.08, p = 0.0093), tumor grade (HR 3.64, 95% CI 1.63-9.63, p = 0.0010), genetic status (p < 0.0001), low gray-level run emphasis (LGRE, calculated from the gray-level run-length matrix) (HR 2.30 × 10¹¹, 95% CI 737.11-4.23 × 10¹⁹, p = 0.0096), and correlation (calculated from the gray-level co-occurrence matrix) (HR 5.17, 95% CI 1.07-20.93, p = 0.041). The multivariate analyses revealed OS was independently associated with LGRE and correlation. The survival curves were also significantly different (both log-rank p < 0.05).

CONCLUSION

Textural features obtained using preoperative MET positron emission tomography may compliment the semi-quantitative assessment for prognostication in glioma cases.

Texture analysis of delayed contrast-enhanced computed tomography to diagnose cardiac sarcoidosis

PURPOSE

To investigate the diagnostic value of texture analysis to differentiate cardiac sarcoidosis (CS) from other non-ischemic cardiomyopathies (non-CS).

MATERIALS AND METHODS

Twenty CS patients and 15 non-CS patients who had undergone myocardial CT delayed enhancement (CTDE) were included. A total of 36 texture features were calculated according to the CT attenuation of CTDE. We investigated the diagnostic value to differentiate CS from non-CS. We also assessed the intra- and inter-rater reproducibility for each feature and inter-observer agreement for visual assessment.

RESULTS

Seven extracted features had significantly higher run length non-uniformity (RLNU) values (5.4 × 10² ± 6.2 × 10² vs. 11.2 × 10² ± 4.9 × 10², p = 0.037) and significantly lower low gray-level zone emphasis (LGZE) values (7.1 × 10^-3 ± 8.6 × 10^-3 vs. 18.1 × 10^-3 ± 16.9 × 10^-3, p = 0.017) in CS than in non-CS. Intra- and inter-rater reproducibility of RLNU and LGZE were excellent (ICCs > 0.8), while inter-observer agreement of visual assessment was poor (kappa = 0.19). The accuracies of texture analysis were 69% with RLNU and 71% with LGZE, which were better than that of visual assessment.

CONCLUSION

Texture analysis of CTDE could differentiate CS from non-CS with high reproducibility.

Texture Analysis of 18F-FDG PET/CT for Differential Diagnosis Spinal Metastases

Purpose: To evaluate the value of texture analysis for the differential diagnosis of spinal metastases and to improve the diagnostic performance of 2-deoxy-2-[fluorine-18]fluoro-D-glucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) for spinal metastases.

Methods: This retrospective analysis of patients who underwent PET/CT between December 2015 and January 2020 at Shanghai Tenth People's Hospital due to high FDG uptake lesions in the spine included 45 cases of spinal metastases and 44 cases of benign high FDG uptake lesions in the spine. The patients were randomly divided into a training group of 65 and a test group of 24. Seventy-two PET texture features were extracted from each lesion, and the Mann-Whitney U-test was used to screen the training set for texture parameters that differed between the two groups in the presence or absence of spinal metastases. Then, the diagnostic performance of the texture parameters was screened out by receiver operating characteristic (ROC) curve analysis. Texture parameters with higher area under the curve (AUC) values than maximum standardized uptake values (SUVmax) were selected to construct classification models using logistic regression, support vector machines, and decision trees. The probability output of the model with high classification accuracy in the training set was used to compare the diagnostic performance of the classification model and SUVmax using the ROC curve. For all patients with spinal metastases, survival analysis was performed using the Kaplan-Meier method and Cox regression.

Results: There were 51 texture parameters that differed meaningfully between benign and malignant lesions, of which four had higher AUC than SUVmax. The texture parameters were input to build a classification model using logistic regression, support vector machine, and decision tree. The accuracy of classification was 87.5, 83.34, and 75%, respectively. The accuracy of the manual diagnosis was 84.27%. Single-factor survival analysis using the Kaplan-Meier method showed that intensity was correlated with patient survival.

Conclusion: Partial texture features showed higher diagnostic value for spinal metastases than SUVmax. The machine learning part of the model combined with the texture parameters was more accurate than manual diagnosis. Therefore, texture analysis may be useful to assist in the diagnosis of spinal metastases.

Single image super-resolution for whole slide image using convolutional neural networks and self-supervised color normalization

High-quality whole slide scanners used for animal and human pathology scanning are expensive and can produce massive datasets, which limits the access to and adoption of this technique. As a potential solution to these challenges, we present a deep learning-based approach making use of single image super-resolution (SISR) to reconstruct high-resolution histology images from low-resolution inputs. Such low-resolution images can easily be shared, require less storage, and can be acquired quickly using widely available low-cost slide scanners. The network consists of multi-scale fully convolutional networks capable of capturing hierarchical features. Conditional generative adversarial loss is incorporated to penalize blurriness in the output images. The network is trained using a progressive strategy where the scaling factor is sampled from a normal distribution with an increasing mean. The results are evaluated with quantitative metrics and are used in a clinical histopathology diagnosis procedure which shows that the SISR framework can be used to reconstruct high-resolution images with clinical level quality. We further propose a self-supervised color normalization method that can remove staining variation artifacts. Quantitative evaluations show that the SISR framework can generalize well on unseen data collected from other patient tissue cohorts by incorporating the color normalization method.

A Role for FDG PET Radiomics in Personalized Medicine?

Radiomics describes the extraction of multiple features from medical images, including molecular imaging modalities, that with bioinformatic approaches, provide additional clinically relevant information that may be invisible to the human eye. This information may complement standard radiological interpretation with data that may better characterize a disease or that may provide predictive or prognostic information. Progressing from predefined image features, often describing heterogeneity of voxel intensities within a volume of interest, there is increasing use of machine learning to classify disease characteristics and deep learning methods based on artificial neural networks that can learn features without a priori definition and without the need for preprocessing of images. There have been advances in standardization and harmonization of methods to a level that should support multicenter studies. However, in this relatively early phase of research in the field, there are limited aspects that have been adopted into routine practice. Most of the reports in the molecular imaging field describe radiomic approaches in cancer using 18F-fluorodeoxyglucose positron emission tomography (18F-FDG-PET). In this review, we will describe radiomics in molecular imaging and summarize the pertinent literature in lung cancer where reports are most prevalent and mature.

Value of volumetric and textural analysis in predicting the treatment response in patients with locally advanced rectal cancer

OBJECTIVE

The aim of this study was to assess the value of baseline 18F-FDG PET/CT in predicting the response to neoadjuvant chemo-radiotherapy (NCRT) in patients with locally advanced rectal cancer (LARC) via the volumetric and texture data obtained from 18F-FDG PET/CT images.

METHODS

 In total, 110 patients who had undergone NCRT after initial PET/CT and followed by surgical resection were included in this study. Patients were divided into two groups randomly as a train set (n: 88) and test set (n: 22). Pathological response using three-point tumor regression grade (TRG) and metastatic lymph nodes in PET/CT images were determined. TRG1 were accepted as responders and TRG2-3 as non-responders. Region of interest for the primary tumors was drawn and volumetric features (metabolic tumor volume (MTV) and total lesion glycolysis (TLG)) and texture features were calculated. In train set, the relationship between these features and TRG was investigated with Mann-Whitney U test. Receiver operating curve analysis was performed for features with p < 0.05. Correlation between features were evaluated with Spearman correlation test, features with correlation coefficient < 0.8 were evaluated with the logistic regression analysis for creating a model. The model obtained was tested with a test set that has not been used in modeling before.

RESULTS

 In train set 32 (36.4%) patients were responders. The rate of visually detected metastatic lymph node at baseline PET/CT was higher in non-responders than responders (71.4% and 46.9%, respectively, p = 0.022). There was a statistically significant difference between TLG, MTV, SHAPE_compacity, NGLDMcoarseness, GLRLM_GLNU, GLRLM_RLNU, GLZLM_LZHGE and GLZLM_GLNU between responders and non-responders. MTV and NGLDMcoarseness demonstrated the most significance (p = 0.011). A multivariate logistic regression analysis that included MTV, coarseness, GLZLM_LZHGE and lymph node metastasis was performed. Multivariate analysis demonstrated MTV and lymph node metastasis were the most meaningful parameters. The model's AUC was calculated as 0.714 (p = 0.001,0.606-0.822, 95% CI). In test set, AUC was determined 0.838 (p = 0.008,0.671-1.000, 95% CI) in discriminating non-responders.

CONCLUSIONS

Although there were points where textural features were found to be significant, multivariate analysis revealed no diagnostic superiority over MTV in predicting treatment response. In this study, it was thought higher MTV value and metastatic lymph nodes in PET/CT images could be a predictor of low treatment response in patients with LARC.

What can artificial intelligence teach us about the molecular mechanisms underlying disease?

While molecular imaging with positron emission tomography or single-photon emission computed tomography already reports on tumour molecular mechanisms on a macroscopic scale, there is increasing evidence that there are multiple additional features within medical images that can further improve tumour characterization, treatment prediction and prognostication. Early reports have already revealed the power of radiomics to personalize and improve patient management and outcomes. What remains unclear is how these additional metrics relate to underlying molecular mechanisms of disease. Furthermore, the ability to deal with increasingly large amounts of data from medical images and beyond in a rapid, reproducible and transparent manner is essential for future clinical practice. Here, artificial intelligence (AI) may have an impact. AI encompasses a broad range of 'intelligent' functions performed by computers, including language processing, knowledge representation, problem solving and planning. While rule-based algorithms, e.g. computer-aided diagnosis, have been in use for medical imaging since the 1990s, the resurgent interest in AI is related to improvements in computing power and advances in machine learning (ML). In this review we consider why molecular and cellular processes are of interest and which processes have already been exposed to AI and ML methods as reported in the literature. Non-small-cell lung cancer is used as an exemplar and the focus of this review as the most common tumour type in which AI and ML approaches have been tested and to illustrate some of the concepts.

Effects of alterations in positron emission tomography imaging parameters on radiomics features

Radiomics studies require large patient cohorts, which often include patients imaged using different imaging protocols. We aimed to determine the impact of variability in imaging protocol parameters and interscanner variability using a phantom that produced feature values similar to those of patients. Positron emission tomography (PET) scans of a Hoffman brain phantom were acquired on GE Discovery 710, Siemens mCT, and Philips Vereos scanners. A standard-protocol scan was acquired on each machine, and then each parameter that could be changed was altered individually. The phantom was contoured with 10 regions of interest (ROIs). Values for 45 features with 2 different preprocessing techniques were extracted for each image. To determine the impact of each parameter on the reliability of each radiomics feature, the intraclass correlation coefficient (ICC) was calculated with the ROIs as the subjects and the parameter values as the raters. For interscanner comparisons, we compared the standard deviation of each radiomics feature value from the standard-protocol images to the standard deviation of the same radiomics feature from PET scans of 224 patients with non-small cell lung cancer. When the pixel size was resampled prior to feature extraction, all features had good reliability (ICC > 0.75) for the field of view and matrix size. The time per bed position had excellent reliability (ICC > 0.9) on all features. When the filter cutoff was restricted to values below 6 mm, all features had good reliability. Similarly, when subsets and iterations were restricted to reasonable values used in clinics, almost all features had good reliability. The average ratio of the standard deviation of features on the phantom scans to that of the NSCLC patient scans was 0.73 using fixed-bin-width preprocessing and 0.92 using 64-level preprocessing. Most radiomics feature values had at least good reliability when imaging protocol parameters were within clinically used ranges. However, interscanner variability was about equal to interpatient variability; therefore, caution must be used when combining patients scanned on equipment from different vendors in radiomics data sets.

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

The sensitivity of radiomic features to several confounding factors, such as reconstruction settings, makes clinical use challenging. To investigate the impact of harmonized image reconstructions on feature consistency, a multicenter phantom study was performed using 3-dimensionally printed phantom inserts reflecting realistic tumor shapes and heterogeneity uptakes. Methods: Tumors extracted from real PET/CT scans of patients with non-small cell lung cancer served as model for three 3-dimensionally printed inserts. Different heterogeneity pattern were realized by printing separate compartments that could be filled with different activity solutions. The inserts were placed in the National Electrical Manufacturers Association image-quality phantom and scanned various times. First, a list-mode scan was acquired and 5 statistically equal replicates were reconstructed. Second, the phantom was scanned 4 times on the same scanner. Third, the phantom was scanned on 6 PET/CT systems. All images were reconstructed using EANM Research Ltd. (EARL)-compliant and locally clinically preferred reconstructions. EARL-compliant reconstructions were performed without (EARL1) or with (EARL2) point-spread function. Images were analyzed with and without resampling to 2-mm cubic voxels. Images were discretized with a fixed bin width (FBW) of 0.25 and a fixed bin number (FBN) of 64. The intraclass correlation coefficient (ICC) of each scan setup was calculated and compared across reconstruction settings. An ICC above 0.75 was regarded as high. Results: The percentage of features yielding a high ICC was largest for the statistically equal replicates (70%-91% for FBN; 90%-96% for FBW discretization). For scans acquired on the same system, the percentage decreased, but most features still resulted in a high ICC (FBN, 52%-63%; FBW, 75%-85%). The percentage of features yielding a high ICC decreased more in the multicenter setting. In this case, the percentage of features yielding a high ICC was larger for images reconstructed with EARL-compliant reconstructions: for example, 40% for EARL1 and 60% for EARL2 versus 21% for the clinically preferred setting for FBW discretization. When discretized with FBW and resampled to isotropic voxels, this benefit was more pronounced. Conclusion: EARL-compliant reconstructions harmonize a wide range of radiomic features. FBW discretization and a sampling to isotropic voxels enhances the benefits of EARL-compliant reconstructions.

Impact of intensity discretization on textural indices of [18F]FDG-PET tumour heterogeneity in lung cancer patients

Quantifying tumour heterogeneity from [¹⁸F]FDG-PET images promises benefits for treatment selection of cancer patients. Here, the calculation of texture parameters mandates an initial discretization step (binning) to reduce the number of intensity levels. Typically, three types of discrimination methods are used: lesion relative resampling (LRR) with fixed bin number, lesion absolute resampling (LAR) and absolute resampling (AR) with fixed bin widths. We investigated the effects of varying bin widths or bin number using 27 commonly cited local and regional texture indices (TIs) applied on lung tumour volumes. The data set were extracted from 58 lung cancer patients, with three different and robust tumour segmentation methods. In our cohort, the variations of the mean value as the function of the bin widths were similar for TIs calculated with LAR and AR quantification. The TI histograms calculated by LRR method showed distinct behaviour and its numerical values substantially effected by the selected bin number. The correlations of the AR and LAR based TIs demonstrated no principal differences between these methods. However, no correlation was found for the interrelationship between the TIs calculated by LRR and LAR (or AR) discretization method. Visual classification of the texture was also performed for each lesion. This classification analysis revealed that the parameters show statistically significant correlation with the visual score, if LAR or AR discretization method is considered, in contrast to LRR. Moreover, all the resulted tendencies were similar regardless the segmentation methods and the type of textural features involved in this work.

Volumetric and texture analysis of pretherapeutic 18F-FDG PET can predict overall survival in medullary thyroid cancer patients treated with Vandetanib

PURPOSE

The metabolically most active lesion in 2-deoxy-2-(¹⁸F)fluoro-D-glucose (¹⁸F-FDG) PET/CT can predict progression-free survival (PFS) in patients with medullary thyroid carcinoma (MTC) starting treatment with the tyrosine kinase inhibitor (TKI) vandetanib. However, this metric failed in overall survival (OS) prediction. In the present proof of concept study, we aimed to explore the prognostic value of intratumoral textural features (TF) as well as volumetric parameters (total lesion glycolysis, TLG) derived by pre-therapeutic ¹⁸F-FDG PET.

METHODS

Eighteen patients with progressive MTC underwent baseline ¹⁸F-FDG PET/CT prior to and 3 months after vandetanib initiation. By manual segmentation of the tumor burden at baseline and follow-up PET, intratumoral TF and TLG were computed. The ability of TLG, imaging-based TF, and clinical parameters (including age, tumor marker doubling times, prior therapies and RET (rearranged during transfection) mutational status) for prediction of both PFS and OS were evaluated.

RESULTS

The TF Complexity and the volumetric parameter TLG obtained at baseline prior to TKI initiation successfully differentiated between low- and high-risk patients. Complexity allocated 10/18 patients to the high-risk group with an OS of 3.3 y (vs. low-risk group, OS = 5.3 y, 8/18, AUC = 0.78, P = 0.03). Baseline TLG designated 11/18 patients to the high-risk group (OS = 3.5 y vs. low-risk group, OS = 5 y, 7/18, AUC = 0.83, P = 0.005). The Hazard Ratio for cancer-related death was 6.1 for Complexity (TLG, 9.5). Among investigated clinical parameters, the age at initiation of TKI treatment reached significance for PFS prediction (P = 0.02, OS, n.s.).

CONCLUSIONS

The TF Complexity and the volumetric parameter TLG are both independent parameters for OS prediction.

Repeatability of 18 F-FDG PET radiomic features: A phantom study to explore sensitivity to image reconstruction settings, noise, and delineation method

BACKGROUND

18 F-fluoro-2-deoxy-D-Glucose positron emission tomography (18 F-FDG PET) radiomics has the potential to guide the clinical decision making in cancer patients, but validation is required before radiomics can be implemented in the clinical setting. The aim of this study was to explore how feature space reduction and repeatability of 18 F-FDG PET radiomic features are affected by various sources of variation such as underlying data (e.g., object size and uptake), image reconstruction methods and settings, noise, discretization method, and delineation method.

METHODS

The NEMA image quality phantom was scanned with various sphere-to-background ratios (SBR), simulating different activity uptakes, including spheres with low uptake, that is, SBR smaller than 1. Furthermore, images of a phantom containing 3D printed inserts reflecting realistic heterogeneity uptake patterns were acquired. Data were reconstructed using various matrix sizes, reconstruction algorithms, and scan durations (noise). For every specific reconstruction and noise level, ten statistically equal replicates were generated. The phantom inserts were delineated using CT and PET-based segmentation methods. A total of 246 radiomic features was extracted from each image dataset. Images were discretized with a fixed number of 64 bins (FBN) and a fixed bin width (FBW) of 0.25 for the high and a FBW of 0.05 for the low uptake data. In terms of feature reduction, we determined the impact of these factors on the composition of feature clusters, which were defined on the basis of Spearman's correlation matrices. To assess feature repeatability, the intraclass correlation coefficient was calculated over the ten replicates.

RESULTS

In general, larger spheres with high uptake resulted in better repeatability compared to smaller low uptake spheres. In terms of repeatability, features extracted from heterogeneous phantom inserts were comparable to features extracted from bigger high uptake spheres. For example, for an EARL-compliant reconstruction, larger and smaller high uptake spheres yielded good repeatability for 32% and 30% of the features, while the heterogeneous inserts resulted in 34% repeatable features. For the low uptake spheres, this was the case for 22% and 20% of the features for bigger and smaller spheres, respectively. Images reconstructed with point-spread-function (PSF) resulted in the highest repeatability when compared with OSEM or time-of-flight, for example, 53%, 30%, and 32% of repeatable features, respectively (for unsmoothed data, discretized with FBN, 300 s scan duration). Reducing image noise (increasing scan duration and smoothing) and using CT-based segmentation for the low uptake spheres yielded improved repeatability. FBW discretization resulted in higher repeatability than FBN discretization, for example, 89% and 35% of the features, respectively (for the EARL-compliant reconstruction and larger high uptake spheres).

CONCLUSION

Feature space reduction and repeatability of 18 F-FDG PET radiomic features depended on all studied factors. The high sensitivity of PET radiomic features to image quality suggests that a high level of image acquisition and preprocessing standardization is required to be used as clinical imaging biomarker.

A feasibility study of spatiotemporally integrated radiotherapy using the LQ model

This paper investigates the feasibility of spatiotemporally modulated radiotherapy (STMRT)-integrated model with explicit constraints on the tumor dose heterogeneity. In particular, we demonstrate the effect of the tumor dose heterogeneity on the tumor biologically effective dose (BED) achievable and optimal fractionation. We propose an STMRT model that simultaneously optimizes the dose distributions and fractionation schedule for each individual case with the maximum and minimum constraints on the tumor BED to explicitly control the level of tumor dose heterogeneity. Sixteen thoracic phantom cases were planned using (1) STMRT and (2) standard fractionation (60 Gy in 30 fractions fixed) IMRT. Constraints on the organs-at-risk (OAR) BED were identical for both plans. BEDs were calculated using the [Formula: see text] ratio of 10 Gy for the tumor and 3 Gy for all OARs. The maximum tumor BED for STMRT plans was constrained to be less than 100%-150% of the maximum tumor BED resulted from the standard fractionation plans. The mean tumor BED from STMRT plans was up to 110.7%, 128.3%, 135.0% and 148.0% of that from the standard fractionation plans when the maximum tumor BED was constrained to be less than 100%, 120%, 130% and 150% of the maximum BED achieved using the standard plans. The optimal number of fractions varied widely for different phantom geometries for the same radiobiological parameter values. The increase in the tumor BED and the range of optimal fractionation was larger with a larger tumor dose heterogeneity allowed. The results have shown the feasibility of personalizing fractionation schedule using an STMRT integrated model to deliver a maximum feasible BED to the tumor for a fixed OAR BED. The potential increase in the tumor BED was positively correlated to the tumor dose heterogeneity allowed.

Investigating the histopathologic correlates of 18F-FDG PET heterogeneity in non-small-cell lung cancer

PURPOSE

Despite the growing use of fluorine-18-fluorodeoxyglucose (F-FDG) PET texture analysis to measure intratumoural heterogeneity in cancer research, the biologic basis of F-FDG PET-derived texture variables is poorly understood. We aimed to assess correlations between F-FDG PET-derived texture variables and whole-slide image (WSI)-derived metrics of tumour cellularity and spatial heterogeneity.

PATIENTS AND METHODS

Twenty-two patients with non-small-cell lung cancer prospectively underwent F-FDG PET imaging before tumour resection. We tested nine F-FDG PET parameters: metabolically active tumour volume, total lesion glycolysis, mean standardized uptake value (SUVmean), first-order entropy, energy, skewness, kurtosis, grey-level co-occurrence matrix entropy and lacunarity (SUV-lacunarity). From the haematoxylin and eosin-stained WSIs, we derived mean tumour-cell density (MCD) and lacunarity (path-lacunarity). Spearman's correlation analysis and agglomerative hierarchical clustering were performed to assess variable associations.

RESULTS

Tumour volumes ranged from 2.2 to 74 cm (median: 17.9 cm). MCD correlated positively with total lesion glycolysis (rs: 0.46, P: 0.007) and SUVmean (rs : 0.55; P: 0.008) and negatively with skewness and kurtosis (rs: -0.47 for both; P: 0.028 and 0.026, respectively). SUV-lacunarity and path-lacunarity were positively correlated (rs: 0.5; P: 0.018). On cluster analysis, larger tumours trended towards higher SUVmean and entropy with a predominance of tightly concentrated high SUV-voxels (negative skewness and low kurtosis on the histogram); on WSI analysis such larger tumours also displayed generally higher MCD and low SUV-lacunarity and path-lacunarity.

CONCLUSION

Our data suggest that histopathological MCD and lacunarity are associated with several commonly used F-FDG PET-derived indices including SUV-lacunarity, metabolically active tumour volume, SUVmean, entropy, skewness, and kurtosis, and thus may explain the biological basis of F-FDG PET-uptake heterogeneity in non-small-cell lung cancer.

Challenges and Promises of PET Radiomics

PURPOSE

Radiomics describes the extraction of multiple, otherwise invisible, features from medical images that, with bioinformatic approaches, can be used to provide additional information that can predict underlying tumor biology and behavior.

METHODS AND MATERIALS

Radiomic signatures can be used alone or with other patient-specific data to improve tumor phenotyping, treatment response prediction, and prognosis, noninvasively. The data describing 18F-fluorodeoxyglucose positron emission tomography radiomics, often using texture or heterogeneity parameters, are increasing rapidly.

RESULTS

In relation to radiation therapy practice, early data have reported the use of radiomic approaches to better define tumor volumes and predict radiation toxicity and treatment response.

CONCLUSIONS

Although at an early stage of development, with many technical challenges remaining and a need for standardization, promise nevertheless exists that PET radiomics will contribute to personalized medicine, especially with the availability of increased computing power and the development of machine-learning approaches for imaging.

Use of 18F-FDG PET/CT texture analysis to diagnose cardiac sarcoidosis

PURPOSE

¹⁸F-fluorodeoxyglocose positron emission tomography (FDG PET) plays a significant role in the diagnosis of cardiac sarcoidosis (CS). Texture analysis is a group of computational methods for evaluating the inhomogeneity among adjacent pixels or voxels. We investigated whether texture analysis applied to myocardial FDG uptake has diagnostic value in patients with CS.

METHODS

Thirty-seven CS patients (CS group), and 52 patients who underwent FDG PET/CT to detect malignant tumors with any FDG cardiac uptake (non-CS group) were studied. A total of 36 texture features from the histogram, gray-level co-occurrence matrix (GLCM), gray-level run length matrix (GLRLM), gray-level zone size matrix (GLZSM) and neighborhood gray-level difference matrix (NGLDM), were computed using polar map images. First, the inter-operator and inter-scan reproducibility of the texture features of the CS group were evaluated. Then, texture features of the patients with CS were compared to those without CS lesions.

RESULTS

Twenty-eight of the 36 texture features showed high inter-operator reproducibility with intraclass correlation coefficients (ICCs) over 0.80. In addition, 17 of the 36 showed high inter-scan reproducibility with ICCs over 0.80. The SUVmax showed no difference between the CS and non-CS group [7.36 ± 2.77 vs. 8.78 ± 4.65, p = 0.45, area under the curve (AUC) = 0.60]. By contrast, 16 of the 36 texture features could distinguish CS from non-CS grsoup with AUC > 0.80. Multivariate logistic regression analysis after hierarchical clustering concluded that long-run emphasis (LRE; P = 0.0004) and short-run low gray-level emphasis (SRLGE; P = 0.016) were significant independent factors that could distinguish between the CS and non-CS groups. Specifically, LRE was significantly higher in CS than in non-CS (30.1 ± 25.4 vs. 11.4 ± 4.6, P < 0.0001), with high diagnostic ability (AUC = 0.91), and had high inter-operator reproducibility (ICC = 0.98).

CONCLUSIONS

The texture analysis had high inter-operator and high inter-scan reproducibility. Some of texture features showed higher diagnostic value than SUVmax for CS diagnosis. Therefore, texture analysis may have a role in semi-automated systems for diagnosing CS.

Texture analysis of 3D dose distributions for predictive modelling of toxicity rates in radiotherapy

BACKGROUND AND PURPOSE

To explore the use of texture analysis (TA) features of patients' 3D dose distributions to improve prediction modelling of treatment complication rates in prostate cancer radiotherapy.

MATERIAL AND METHODS

Late toxicity scores, dose distributions, and non-treatment related (NTR) predictors for late toxicity, such as age and baseline symptoms, of 351 patients of the hypofractionation arm of the HYPRO randomized trial were used in this study. Apart from DVH parameters, also TA features of rectum and bladder 3D dose distributions were used for predictive modelling of gastrointestinal (GI) and genitourinary (GU) toxicities. Logistic Normal Tissue Complication Probability (NTCP) models were derived, using only NTR parameters, NTR + DVH, NTR + TA, and NTR + DVH + TA.

RESULTS

For rectal bleeding, the area under the curve (AUC) for using only NTR parameters was 0.58, which increased to 0.68, and 0.73, when adding DVH or TA parameters respectively. For faecal incontinence, the AUC went up from 0.63 (NTR only), to 0.68 (+DVH) and 0.73 (+TA). For nocturia, adding TA features resulted in an AUC increase from 0.64 to 0.66, while no improvement was seen when including DVH parameters in the modelling. For urinary incontinence, the AUC improved from 0.68 to 0.71 (+DVH) and 0.73 (+TA). For GI, model improvements resulting from adding TA parameters to NTR instead of DVH were statistically significant (p < 0.04).

CONCLUSION

Inclusion of 3D dosimetric texture analysis features in predictive modelling of GI and GU toxicity rates in prostate cancer radiotherapy improved prediction performance, which was statistically significant for GI.

LIFEx: A Freeware for Radiomic Feature Calculation in Multimodality Imaging to Accelerate Advances in the Characterization of Tumor Heterogeneity

Textural and shape analysis is gaining considerable interest in medical imaging, particularly to identify parameters characterizing tumor heterogeneity and to feed radiomic models. Here, we present a free, multiplatform, and easy-to-use freeware called LIFEx, which enables the calculation of conventional, histogram-based, textural, and shape features from PET, SPECT, MR, CT, and US images, or from any combination of imaging modalities. The application does not require any programming skills and was developed for medical imaging professionals. The goal is that independent and multicenter evidence of the usefulness and limitations of radiomic features for characterization of tumor heterogeneity and subsequent patient management can be gathered. Many options are offered for interactive textural index calculation and for increasing the reproducibility among centers. The software already benefits from a large user community (more than 800 registered users), and interactions within that community are part of the development strategy.Significance: This study presents a user-friendly, multi-platform freeware to extract radiomic features from PET, SPECT, MR, CT, and US images, or any combination of imaging modalities. Cancer Res; 78(16); 4786-9. ©2018 AACR.

Prediction of cervical cancer recurrence using textural features extracted from 18F-FDG PET images acquired with different scanners

Objectives

To identify an imaging signature predicting local recurrence for locally advanced cervical cancer (LACC) treated by chemoradiation and brachytherapy from baseline ¹⁸F-FDG PET images, and to evaluate the possibility of gathering images from two different PET scanners in a radiomic study.

Methods

118 patients were included retrospectively. Two groups (G1, G2) were defined according to the PET scanner used for image acquisition. Eleven radiomic features were extracted from delineated cervical tumors to evaluate: (i) the predictive value of features for local recurrence of LACC, (ii) their reproducibility as a function of the scanner within a hepatic reference volume, (iii) the impact of voxel size on feature values.

Results

Eight features were statistically significant predictors of local recurrence in G1 (p < 0.05). The multivariate signature trained in G2 was validated in G1 (AUC=0.76, p<0.001) and identified local recurrence more accurately than SUV_max (p=0.022). Four features were significantly different between G1 and G2 in the liver. Spatial resampling was not sufficient to explain the stratification effect.

Conclusion

This study showed that radiomic features could predict local recurrence of LACC better than SUV_max. Further investigation is needed before applying a model designed using data from one PET scanner to another.

Selected PET radiomic features remain the same

Purpose

We investigated whether PET radiomic features are affected by differences in the scanner, scan protocol, and lesion location using ¹⁸F-FDG PET/CT and PET/MR scans.

Results

SUV, TMR, skewness, kurtosis, entropy, and homogeneity strongly correlated between PET/CT and PET/MR images. SUVs were significantly higher on PET/MR_{0-2 min} and PET/MR_{0-10 min} than on PET/CT in gynecological cancer (p = 0.008 and 0.008, respectively), whereas no significant difference was observed between PET/CT, PET/MR_{0–2 min}, and PET/MR_{0–10 min} images in oral cavity/oropharyngeal cancer. TMRs on PET/CT, PET/MR_{0–2 min}, and PET/MR_{0–10 min} increased in this order in gynecological cancer and oral cavity/oropharyngeal cancer. In contrast to conventional and histogram indices, 4 textural features (entropy, homogeneity, SRE, and LRE) were not significantly different between PET/CT, PET/MR_{0–2 min}, and PET/MR_{0–10 min} images.

Conclusions

¹⁸F-FDG PET radiomic features strongly correlated between PET/CT and PET/MR images. Dixon-based attenuation correction on PET/MR images underestimated tumor tracer uptake more significantly in oral cavity/oropharyngeal cancer than in gynecological cancer. ¹⁸F-FDG PET textural features were affected less by differences in the scanner and scan protocol than conventional and histogram features, possibly due to the resampling process using a medium bin width.

Methods

Eight patients with gynecological cancer and 7 with oral cavity/oropharyngeal cancer underwent a whole-body ¹⁸F-FDG PET/CT scan and regional PET/MR scan in one day. PET/MR scans were performed for 10 minutes in the list mode, and PET/CT and 0–2 min and 0–10 min PET/MR images were reconstructed. The standardized uptake value (SUV), tumor-to-muscle SUV ratio (TMR), skewness, kurtosis, entropy, homogeneity, short-run emphasis (SRE), and long-run emphasis (LRE) were compared between PET/CT, PET/MR_{0-2 min}, and PET/MR_{0-10 min} images.

Clinical Utility of FDG PET/CT in Patients with Autoimmune Pancreatitis: a Case-Control Study

Autoimmune pancreatitis (AIP) shares overlapping clinical features with pancreatic cancer (PC). Importantly, treatment of the two conditions is different. We investigated the clinical usefulness of ¹⁸F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) in patients with suspected AIP before treatment. From September 2008 to July 2016, 53 patients with suspected AIP at National Taiwan University Hospital had PET/CT prior to therapy to exclude malignancy and evaluate the extent of inflammation. Their scans were compared with those from 61 PC patients. PET imaging features were analyzed using logistic regression. Significant differences in pancreatic tumor uptake morphology, maximum standardized uptake value, high-order primary tumor texture feature (i.e. high-gray level zone emphasis value), and numbers and location of extrapancreatic foci were found between AIP and PC. Using the prediction model, the area under curve of receiver-operator curve was 0.95 (P < 0.0001) with sensitivity, specificity, positive predictive, and negative predictive values of 90.6%, 84.0%, 87.9%, and 87.5% respectively, in differentiating AIP from PC. FDG PET/CT offers high sensitivity, albeit slightly lower specificity in differentiating AIP from PC. Nonetheless, additional systemic inflammatory foci detected by the whole body PET/CT help confirm diagnosis of AIP in these patients before initiating steroid therapy, especially when biopsy is inconclusive.

Heterogeneity and irregularity of pretreatment 18F-fluorodeoxyglucose positron emission tomography improved prognostic stratification of p16-negative high-risk squamous cell carcinoma of the oropharynx

OBJECTIVES

Human papillomavirus-negative oropharyngeal squamous cell carcinoma (OPSCC) has unfavorable survival outcomes. Two outcomes have been identified based on smoking history and tumor stage. We investigate the prognostic role of pre-treatment positron emission tomography (PET) in high-risk OPSCC.

MATERIALS AND METHODS

We enrolled 147 M0 OPSCC patients with p16-negative staining and a history of heavy smoking (>10 pack-years) or T4 disease. All patients completed primary chemoradiotherapy, and 42% maximum standard uptake values (SUV_max) were used as the threshold for primary tumor. Patients were classified into training and validation cohorts with a ratio of 1:1.5 according to the PET date. Heterogeneity and irregularity indices were obtained. PET parameters with significant impact on progression-free survival (PFS) in receiver operating characteristic curves and univariate Cox models were identified and included in recursive partitioning analysis (RPA) for constructing a prognostic model. The RPA-based prognostic model was further tested in the validation cohort using multivariate Cox models.

RESULTS

Fifty-eight and 89 patients were in the training and validation groups, respectively. Heterogeneity parameter, SUV-entropy (derived from histogram analysis), and irregularity index, and asphericity were significantly associated with PFS. The RPA model revealed that patients with both high SUV-entropy and high asphericity experienced the worst PFS. Results were confirmed in the validation group. The overall concordance index for PFS of the model was 0.75, which was higher than the clinical stages, performance status, SUV_max, and metabolic tumor volume of PET.

CONCLUSIONS

PET prognostic model provided useful prediction of PFS for patients with high-risk OPSCC.

Diagnostic and prognostic value of amyloid PET textural and shape features: comparison with classical semi-quantitative rating in 760 patients from the ADNI-2 database

We evaluated the performance of amyloid PET textural and shape features in discriminating normal and Alzheimer's disease (AD) subjects, and in predicting conversion to AD in subjects with mild cognitive impairment (MCI) or significant memory concern (SMC). Subjects from the Alzheimer's Disease Neuroimaging Initiative with available baseline ¹⁸F-florbetapir and T1-MRI scans were included. The cross-sectional cohort consisted of 181 controls and 148 AD subjects. The longitudinal cohort consisted of 431 SMC/MCI subjects, 85 of whom converted to AD during follow-up. PET images were normalized to MNI space and post-processed using in-house software. Relative retention indices (SUVr) were computed with respect to pontine, cerebellar, and composite reference regions. Several textural and shape features were extracted then combined using a support vector machine (SVM) to build a predictive model of AD conversion. Diagnostic and prognostic performance was evaluated using ROC analysis and survival analysis with the Cox proportional hazard model. The three SUVr and all the tested features effectively discriminated AD subjects in cross-sectional analysis (all p < 0.001). In longitudinal analysis, the variables with the highest prognostic value were composite SUVr (AUC 0.86; accuracy 81%), skewness (0.87; 83%), local minima (0.85; 79%), Geary's index (0.86; 81%), gradient norm maximal argument (0.83; 82%), and the SVM model (0.91; 86%). The adjusted hazard ratio for AD conversion was 5.5 for the SVM model, compared with 4.0, 2.6, and 3.8 for cerebellar, pontine and composite SUVr (all p < 0.001), indicating that appropriate amyloid textural and shape features predict conversion to AD with at least as good accuracy as classical SUVr.

A Postreconstruction Harmonization Method for Multicenter Radiomic Studies in PET

Several reports have shown that radiomic features are affected by acquisition and reconstruction parameters, thus hampering multicenter studies. We propose a method that, by removing the center effect while preserving patient-specific effects, standardizes features measured from PET images obtained using different imaging protocols. Methods: Pretreatment ¹⁸F-FDG PET images of patients with breast cancer were included. In one nuclear medicine department (department A), 63 patients were scanned on a time-of-flight PET/CT scanner, and 16 lesions were triple-negative (TN). In another nuclear medicine department (department B), 74 patients underwent PET/CT on a different brand of scanner and a different reconstruction protocol, and 15 lesions were TN. The images from department A were smoothed using a gaussian filter to mimic data from a third department (department A-S). The primary lesion was segmented to obtain a lesion volume of interest (VOI), and a spheric VOI was set in healthy liver tissue. Three SUVs and 6 textural features were computed in all VOIs. A harmonization method initially described for genomic data was used to estimate the department effect based on the observed feature values. Feature distributions in each department were compared before and after harmonization. Results: In healthy liver tissue, the distributions significantly differed for 4 of 9 features between departments A and B and for 6 of 9 between departments A and A-S (P < 0.05, Wilcoxon test). After harmonization, none of the 9 feature distributions significantly differed between 2 departments (P > 0.1). The same trend was observed in lesions, with a realignment of feature distributions between the departments after harmonization. Identification of TN lesions was largely enhanced after harmonization when the cutoffs were determined on data from one department and applied to data from the other department. Conclusion: The proposed harmonization method is efficient at removing the multicenter effect for textural features and SUVs. The method is easy to use, retains biologic variations not related to a center effect, and does not require any feature recalculation. Such harmonization allows for multicenter studies and for external validation of radiomic models or cutoffs and should facilitate the use of radiomic models in clinical practice.

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

OBJECTIVES

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

METHODS

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

RESULTS

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

CONCLUSIONS

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

KEY POINTS

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

Characterisation of malignant peripheral nerve sheath tumours in neurofibromatosis-1 using heterogeneity analysis of 18F-FDG PET

PURPOSE

Measurement of heterogeneity in 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) images is reported to improve tumour phenotyping and response assessment in a number of cancers. We aimed to determine whether measurements of 18F-FDG heterogeneity could improve differentiation of benign symptomatic neurofibromas from malignant peripheral nerve sheath tumours (MPNSTs).

METHODS

18F-FDG PET data from a cohort of 54 patients (24 female, 30 male, mean age 35.1 years) with neurofibromatosis-1 (NF1), and clinically suspected malignant transformation of neurofibromas into MPNSTs, were included. Scans were performed to a standard clinical protocol at 1.5 and 4 h post-injection. Six first-order [including three standardised uptake value (SUV) parameters], four second-order (derived from grey-level co-occurrence matrices) and four high-order (derived from neighbourhood grey-tone difference matrices) statistical features were calculated from tumour volumes of interest. Each patient had histological verification or at least 5 years clinical follow-up as the reference standard with regards to the characterisation of tumours as benign (n = 30) or malignant (n = 24).

RESULTS

There was a significant difference between benign and malignant tumours for all six first-order parameters (at 1.5 and 4 h; p < 0.0001), for second-order entropy (only at 4 h) and for all high-order features (at 1.5 h and 4 h, except contrast at 4 h; p < 0.0001-0.047). Similarly, the area under the receiver operating characteristic curves was high (0.669-0.997, p < 0.05) for the same features as well as 1.5-h second-order entropy. No first-, second- or high-order feature performed better than maximum SUV (SUVmax) at differentiating benign from malignant tumours.

CONCLUSIONS

18F-FDG uptake in MPNSTs is higher than benign symptomatic neurofibromas, as defined by SUV parameters, and more heterogeneous, as defined by first- and high-order heterogeneity parameters. However, heterogeneity analysis does not improve on SUVmax discriminative performance.