Decentralized Distributed Multi-institutional PET Image Segmentation Using a Federated Deep Learning Framework

The effect of harmonization on the variability of PET radiomic features extracted using various segmentation methods

PURPOSE

This study aimed to examine the robustness of positron emission tomography (PET) radiomic features extracted via different segmentation methods before and after ComBat harmonization in patients with non-small cell lung cancer (NSCLC).

METHODS

We included 120 patients (positive recurrence = 46 and negative recurrence = 74) referred for PET scanning as a routine part of their care. All patients had a biopsy-proven NSCLC. Nine segmentation methods were applied to each image, including manual delineation, K-means (KM), watershed, fuzzy-C-mean, region-growing, local active contour (LAC), and iterative thresholding (IT) with 40, 45, and 50% thresholds. Diverse image discretizations, both without a filter and with different wavelet decompositions, were applied to PET images. Overall, 6741 radiomic features were extracted from each image (749 radiomic features from each segmented area). Non-parametric empirical Bayes (NPEB) ComBat harmonization was used to harmonize the features. Linear Support Vector Classifier (LinearSVC) with L1 regularization For feature selection and Support Vector Machine classifier (SVM) with fivefold nested cross-validation was performed using StratifiedKFold with 'n_splits' set to 5 to predict recurrence in NSCLC patients and assess the impact of ComBat harmonization on the outcome.

RESULTS

From 749 extracted radiomic features, 206 (27%) and 389 (51%) features showed excellent reliability (ICC ≥ 0.90) against segmentation method variation before and after NPEB ComBat harmonization, respectively. Among all, 39 features demonstrated poor reliability, which declined to 10 after ComBat harmonization. The 64 fixed bin widths (without any filter) and wavelets (LLL)-based radiomic features set achieved the best performance in terms of robustness against diverse segmentation techniques before and after ComBat harmonization. The first-order and GLRLM and also first-order and NGTDM feature families showed the largest number of robust features before and after ComBat harmonization, respectively. In terms of predicting recurrence in NSCLC, our findings indicate that using ComBat harmonization can significantly enhance machine learning outcomes, particularly improving the accuracy of watershed segmentation, which initially had fewer reliable features than manual contouring. Following the application of ComBat harmonization, the majority of cases saw substantial increase in sensitivity and specificity.

CONCLUSION

Radiomic features are vulnerable to different segmentation methods. ComBat harmonization might be considered a solution to overcome the poor reliability of radiomic features.

Artificial intelligence-based analysis of whole-body bone scintigraphy: The quest for the optimal deep learning algorithm and comparison with human observer performance

PURPOSE

Whole-body bone scintigraphy (WBS) is one of the most widely used modalities in diagnosing malignant bone diseases during the early stages. However, the procedure is time-consuming and requires vigour and experience. Moreover, interpretation of WBS scans in the early stages of the disorders might be challenging because the patterns often reflect normal appearance that is prone to subjective interpretation. To simplify the gruelling, subjective, and prone-to-error task of interpreting WBS scans, we developed deep learning (DL) models to automate two major analyses, namely (i) classification of scans into normal and abnormal and (ii) discrimination between malignant and non-neoplastic bone diseases, and compared their performance with human observers.

MATERIALS AND METHODS

After applying our exclusion criteria on 7188 patients from three different centers, 3772 and 2248 patients were enrolled for the first and second analyses, respectively. Data were split into two parts, including training and testing, while a fraction of training data were considered for validation. Ten different CNN models were applied to single- and dual-view input (posterior and anterior views) modes to find the optimal model for each analysis. In addition, three different methods, including squeeze-and-excitation (SE), spatial pyramid pooling (SPP), and attention-augmented (AA), were used to aggregate the features for dual-view input models. Model performance was reported through area under the receiver operating characteristic (ROC) curve (AUC), accuracy, sensitivity, and specificity and was compared with the DeLong test applied to ROC curves. The test dataset was evaluated by three nuclear medicine physicians (NMPs) with different levels of experience to compare the performance of AI and human observers.

RESULTS

DenseNet121_AA (DensNet121, with dual-view input aggregated by AA) and InceptionResNetV2_SPP achieved the highest performance (AUC = 0.72) for the first and second analyses, respectively. Moreover, on average, in the first analysis, Inception V3 and InceptionResNetV2 CNN models and dual-view input with AA aggregating method had superior performance. In addition, in the second analysis, DenseNet121 and InceptionResNetV2 as CNN methods and dual-view input with AA aggregating method achieved the best results. Conversely, the performance of AI models was significantly higher than human observers for the first analysis, whereas their performance was comparable in the second analysis, although the AI model assessed the scans in a drastically lower time.

CONCLUSION

Using the models designed in this study, a positive step can be taken toward improving and optimizing WBS interpretation. By training DL models with larger and more diverse cohorts, AI could potentially be used to assist physicians in the assessment of WBS images.

Differential privacy preserved federated learning for prognostic modeling in COVID-19 patients using large multi-institutional chest CT dataset

BACKGROUND

Notwithstanding the encouraging results of previous studies reporting on the efficiency of deep learning (DL) in COVID-19 prognostication, clinical adoption of the developed methodology still needs to be improved. To overcome this limitation, we set out to predict the prognosis of a large multi-institutional cohort of patients with COVID-19 using a DL-based model.

PURPOSE

This study aimed to evaluate the performance of deep privacy-preserving federated learning (DPFL) in predicting COVID-19 outcomes using chest CT images.

METHODS

After applying inclusion and exclusion criteria, 3055 patients from 19 centers, including 1599 alive and 1456 deceased, were enrolled in this study. Data from all centers were split (randomly with stratification respective to each center and class) into a training/validation set (70%/10%) and a hold-out test set (20%). For the DL model, feature extraction was performed on 2D slices, and averaging was performed at the final layer to construct a 3D model for each scan. The DensNet model was used for feature extraction. The model was developed using centralized and FL approaches. For FL, we employed DPFL approaches. Membership inference attack was also evaluated in the FL strategy. For model evaluation, different metrics were reported in the hold-out test sets. In addition, models trained in two scenarios, centralized and FL, were compared using the DeLong test for statistical differences.

RESULTS

The centralized model achieved an accuracy of 0.76, while the DPFL model had an accuracy of 0.75. Both the centralized and DPFL models achieved a specificity of 0.77. The centralized model achieved a sensitivity of 0.74, while the DPFL model had a sensitivity of 0.73. A mean AUC of 0.82 and 0.81 with 95% confidence intervals of (95% CI: 0.79-0.85) and (95% CI: 0.77-0.84) were achieved by the centralized model and the DPFL model, respectively. The DeLong test did not prove statistically significant differences between the two models (p-value = 0.98). The AUC values for the inference attacks fluctuate between 0.49 and 0.51, with an average of 0.50 ± 0.003 and 95% CI for the mean AUC of 0.500 to 0.501.

CONCLUSION

The performance of the proposed model was comparable to centralized models while operating on large and heterogeneous multi-institutional datasets. In addition, the model was resistant to inference attacks, ensuring the privacy of shared data during the training process.

PRIMIS: Privacy-preserving medical image sharing via deep sparsifying transform learning with obfuscation

OBJECTIVE

The primary objective of our study is to address the challenge of confidentially sharing medical images across different centers. This is often a critical necessity in both clinical and research environments, yet restrictions typically exist due to privacy concerns. Our aim is to design a privacy-preserving data-sharing mechanism that allows medical images to be stored as encoded and obfuscated representations in the public domain without revealing any useful or recoverable content from the images. In tandem, we aim to provide authorized users with compact private keys that could be used to reconstruct the corresponding images.

METHOD

Our approach involves utilizing a neural auto-encoder. The convolutional filter outputs are passed through sparsifying transformations to produce multiple compact codes. Each code is responsible for reconstructing different attributes of the image. The key privacy-preserving element in this process is obfuscation through the use of specific pseudo-random noise. When applied to the codes, it becomes computationally infeasible for an attacker to guess the correct representation for all the codes, thereby preserving the privacy of the images.

RESULTS

The proposed framework was implemented and evaluated using chest X-ray images for different medical image analysis tasks, including classification, segmentation, and texture analysis. Additionally, we thoroughly assessed the robustness of our method against various attacks using both supervised and unsupervised algorithms.

CONCLUSION

This study provides a novel, optimized, and privacy-assured data-sharing mechanism for medical images, enabling multi-party sharing in a secure manner. While we have demonstrated its effectiveness with chest X-ray images, the mechanism can be utilized in other medical images modalities as well.

Information fusion for fully automated segmentation of head and neck tumors from PET and CT images

BACKGROUND

PET/CT images combining anatomic and metabolic data provide complementary information that can improve clinical task performance. PET image segmentation algorithms exploiting the multi-modal information available are still lacking.

PURPOSE

Our study aimed to assess the performance of PET and CT image fusion for gross tumor volume (GTV) segmentations of head and neck cancers (HNCs) utilizing conventional, deep learning (DL), and output-level voting-based fusions.

METHODS

The current study is based on a total of 328 histologically confirmed HNCs from six different centers. The images were automatically cropped to a 200 × 200 head and neck region box, and CT and PET images were normalized for further processing. Eighteen conventional image-level fusions were implemented. In addition, a modified U2-Net architecture as DL fusion model baseline was used. Three different input, layer, and decision-level information fusions were used. Simultaneous truth and performance level estimation (STAPLE) and majority voting to merge different segmentation outputs (from PET and image-level and network-level fusions), that is, output-level information fusion (voting-based fusions) were employed. Different networks were trained in a 2D manner with a batch size of 64. Twenty percent of the dataset with stratification concerning the centers (20% in each center) were used for final result reporting. Different standard segmentation metrics and conventional PET metrics, such as SUV, were calculated.

RESULTS

In single modalities, PET had a reasonable performance with a Dice score of 0.77 ± 0.09, while CT did not perform acceptably and reached a Dice score of only 0.38 ± 0.22. Conventional fusion algorithms obtained a Dice score range of [0.76-0.81] with guided-filter-based context enhancement (GFCE) at the low-end, and anisotropic diffusion and Karhunen-Loeve transform fusion (ADF), multi-resolution singular value decomposition (MSVD), and multi-level image decomposition based on latent low-rank representation (MDLatLRR) at the high-end. All DL fusion models achieved Dice scores of 0.80. Output-level voting-based models outperformed all other models, achieving superior results with a Dice score of 0.84 for Majority_ImgFus, Majority_All, and Majority_Fast. A mean error of almost zero was achieved for all fusions using SUVpeak , SUVmean and SUVmedian .

CONCLUSION

PET/CT information fusion adds significant value to segmentation tasks, considerably outperforming PET-only and CT-only methods. In addition, both conventional image-level and DL fusions achieve competitive results. Meanwhile, output-level voting-based fusion using majority voting of several algorithms results in statistically significant improvements in the segmentation of HNC.

Operational greenhouse-gas emissions of deep learning in digital pathology: a modelling study

BACKGROUND

Deep learning is a promising way to improve health care. Image-processing medical disciplines, such as pathology, are expected to be transformed by deep learning. The first clinically applicable deep-learning diagnostic support tools are already available in cancer pathology, and their number is increasing. However, data on the environmental sustainability of these tools are scarce. We aimed to conduct an environmental-sustainability analysis of a theoretical implementation of deep learning in patient-care pathology.

METHODS

For this modelling study, we first assembled and calculated relevant data and parameters of a digital-pathology workflow. Data were breast and prostate specimens from the university clinic at the Institute of Pathology of the Rheinisch-Westfälische Technische Hochschule Aachen (Aachen, Germany), for which commercially available deep learning was already available. Only specimens collected between Jan 1 and Dec 31, 2019 were used, to omit potential biases due to the COVID-19 pandemic. Our final selection was based on 2 representative weeks outside holidays, covering different types of specimens. To calculate carbon dioxide (CO2) or CO2 equivalent (CO2 eq) emissions of deep learning in pathology, we gathered relevant data for exact numbers and sizes of whole-slide images (WSIs), which were generated by scanning histopathology samples of prostate and breast specimens. We also evaluated different data input scenarios (including all slide tiles, only tiles containing tissue, or only tiles containing regions of interest). To convert estimated energy consumption from kWh to CO2 eq, we used the internet protocol address of the computational server and the Electricity Maps database to obtain information on the sources of the local electricity grid (ie, renewable vs non-renewable), and estimated the number of trees and proportion of the local and world's forests needed to sequester the CO2 eq emissions. We calculated the computational requirements and CO2 eq emissions of 30 deep-learning models that varied in task and size. The first scenario represented the use of one commercially available deep-learning model for one task in one case (1-task), the second scenario considered two deep-learning models for two tasks per case (2-task), the third scenario represented a future, potentially automated workflow that could handle 7 tasks per case (7-task), and the fourth scenario represented the use of a single potential, large, computer-vision model that could conduct multiple tasks (multitask). We also compared the performance (ie, accuracy) and CO2 eq emissions of different deep-learning models for the classification of renal cell carcinoma on WSIs, also from Rheinisch-Westfälische Technische Hochschule Aachen. We also tested other approaches to reducing CO2 eq emissions, including model pruning and an alternative method for histopathology analysis (pathomics).

FINDINGS

The pathology database contained 35 552 specimens (237 179 slides), 6420 of which were prostate specimens (10 115 slides) and 11 801 of which were breast specimens (19 763 slides). We selected and subsequently digitised 140 slides from eight breast-cancer cases and 223 slides from five prostate-cancer cases. Applying large deep-learning models on all WSI tiles of prostate and breast pathology cases would result in yearly CO2 eq emissions of 7·65 metric tons (t; 95% CI 7·62-7·68) with the use of a single deep-learning model per case; yearly CO2 eq emissions were up to 100·56 t (100·21-100·99) with the use of seven deep-learning models per case. CO2 eq emissions for different deep-learning model scenarios, data inputs, and deep-learning model sizes for all slides varied from 3·61 t (3·59-3·63) to 2795·30 t (1177·51-6482·13. For the estimated number of overall pathology cases worldwide, the yearly CO2 eq emissions varied, reaching up to 16 megatons (Mt) of CO2 eq, requiring up to 86 590 km2 (0·22%) of world forest to sequester the CO2 eq emissions. Use of the 7-task scenario and small deep-learning models on slides containing tissue only could substantially reduce CO2 eq emissions worldwide by up to 141 times (0·1 Mt, 95% CI 0·1-0·1). Considering the local environment in Aachen, Germany, the maximum CO2 eq emission from the use of deep learning in digital pathology only would require 32·8% (95% CI 13·8-76·6) of the local forest to sequester the CO2 eq emissions. A single pathomics run on a tissue could provide information that was comparable to or even better than the output of multitask deep-learning models, but with 147 times reduced CO2 eq emissions.

INTERPRETATION

Our findings suggest that widespread use of deep learning in pathology might have considerable global-warming potential. The medical community, policy decision makers, and the public should be aware of this potential and encourage the use of CO2 eq emissions reduction strategies where possible.

FUNDING

German Research Foundation, European Research Council, German Federal Ministry of Education and Research, Health, Economic Affairs and Climate Action, and the Innovation Fund of the Federal Joint Committee.

Differential privacy preserved federated transfer learning for multi-institutional 68Ga-PET image artefact detection and disentanglement

PURPOSE

Image artefacts continue to pose challenges in clinical molecular imaging, resulting in misdiagnoses, additional radiation doses to patients and financial costs. Mismatch and halo artefacts occur frequently in gallium-68 (68Ga)-labelled compounds whole-body PET/CT imaging. Correcting for these artefacts is not straightforward and requires algorithmic developments, given that conventional techniques have failed to address them adequately. In the current study, we employed differential privacy-preserving federated transfer learning (FTL) to manage clinical data sharing and tackle privacy issues for building centre-specific models that detect and correct artefacts present in PET images.

METHODS

Altogether, 1413 patients with 68Ga prostate-specific membrane antigen (PSMA)/DOTA-TATE (TOC) PET/CT scans from 3 countries, including 8 different centres, were enrolled in this study. CT-based attenuation and scatter correction (CT-ASC) was used in all centres for quantitative PET reconstruction. Prior to model training, an experienced nuclear medicine physician reviewed all images to ensure the use of high-quality, artefact-free PET images (421 patients' images). A deep neural network (modified U2Net) was trained on 80% of the artefact-free PET images to utilize centre-based (CeBa), centralized (CeZe) and the proposed differential privacy FTL frameworks. Quantitative analysis was performed in 20% of the clean data (with no artefacts) in each centre. A panel of two nuclear medicine physicians conducted qualitative assessment of image quality, diagnostic confidence and image artefacts in 128 patients with artefacts (256 images for CT-ASC and FTL-ASC).

RESULTS

The three approaches investigated in this study for 68Ga-PET imaging (CeBa, CeZe and FTL) resulted in a mean absolute error (MAE) of 0.42 ± 0.21 (CI 95%: 0.38 to 0.47), 0.32 ± 0.23 (CI 95%: 0.27 to 0.37) and 0.28 ± 0.15 (CI 95%: 0.25 to 0.31), respectively. Statistical analysis using the Wilcoxon test revealed significant differences between the three approaches, with FTL outperforming CeBa and CeZe (p-value < 0.05) in the clean test set. The qualitative assessment demonstrated that FTL-ASC significantly improved image quality and diagnostic confidence and decreased image artefacts, compared to CT-ASC in 68Ga-PET imaging. In addition, mismatch and halo artefacts were successfully detected and disentangled in the chest, abdomen and pelvic regions in 68Ga-PET imaging.

CONCLUSION

The proposed approach benefits from using large datasets from multiple centres while preserving patient privacy. Qualitative assessment by nuclear medicine physicians showed that the proposed model correctly addressed two main challenging artefacts in 68Ga-PET imaging. This technique could be integrated in the clinic for 68Ga-PET imaging artefact detection and disentanglement using multicentric heterogeneous datasets.

PET/MR Imaging in Head and Neck Cancer

Head and neck squamous cell carcinoma (HNSCC) can either be examined with hybrid PET/MR imaging systems or sequentially, using PET/CT and MR imaging. Regardless of the acquisition technique, the superiority of MR imaging compared to CT lies in its potential to interrogate tumor and surrounding tissues with different sequences, including perfusion and diffusion. For this reason, PET/MR imaging is preferable for the detection and assessment of locoregional residual/recurrent HNSCC after therapy. In addition, MR imaging interpretation is facilitated when combined with PET. Nevertheless, distant metastases and distant second primary tumors are detected equally well with PET/MR imaging and PET/CT.

Whole-body tumor segmentation from PET/CT images using a two-stage cascaded neural network with camouflaged object detection mechanisms

BACKGROUND

Whole-body Metabolic Tumor Volume (MTVwb) is an independent prognostic factor for overall survival in lung cancer patients. Automatic segmentation methods have been proposed for MTV calculation. Nevertheless, most of existing methods for patients with lung cancer only segment tumors in the thoracic region.

PURPOSE

In this paper, we present a Two-Stage cascaded neural network integrated with Camouflaged Object Detection mEchanisms (TS-Code-Net) for automatic segmenting tumors from whole-body PET/CT images.

METHODS

Firstly, tumors are detected from the Maximum Intensity Projection (MIP) images of PET/CT scans, and tumors' approximate localizations along z-axis are identified. Secondly, the segmentations are performed on PET/CT slices that contain tumors identified by the first step. Camouflaged object detection mechanisms are utilized to distinguish the tumors from their surrounding regions that have similar Standard Uptake Values (SUV) and texture appearance. Finally, the TS-Code-Net is trained by minimizing the total loss that incorporates the segmentation accuracy loss and the class imbalance loss.

RESULTS

The performance of the TS-Code-Net is tested on a whole-body PET/CT image data-set including 480 Non-Small Cell Lung Cancer (NSCLC) patients with five-fold cross-validation using image segmentation metrics. Our method achieves 0.70, 0.76, and 0.70, for Dice, Sensitivity and Precision, respectively, which demonstrates the superiority of the TS-Code-Net over several existing methods related to metastatic lung cancer segmentation from whole-body PET/CT images.

CONCLUSIONS

The proposed TS-Code-Net is effective for whole-body tumor segmentation of PET/CT images. Codes for TS-Code-Net are available at: https://github.com/zyj19/TS-Code-Net.

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.

Machine learning-based diagnosis and risk classification of coronary artery disease using myocardial perfusion imaging SPECT: A radiomics study

This study aimed to investigate the diagnostic performance of machine learning-based radiomics analysis to diagnose coronary artery disease status and risk from rest/stress Myocardial Perfusion Imaging (MPI) single-photon emission computed tomography (SPECT). A total of 395 patients suspicious of coronary artery disease who underwent 2-day stress-rest protocol MPI SPECT were enrolled in this study. The left ventricle myocardium, excluding the cardiac cavity, was manually delineated on rest and stress images to define a volume of interest. Added to clinical features (age, sex, family history, diabetes status, smoking, and ejection fraction), a total of 118 radiomics features, were extracted from rest and stress MPI SPECT images to establish different feature sets, including Rest-, Stress-, Delta-, and Combined-radiomics (all together) feature sets. The data were randomly divided into 80% and 20% subsets for training and testing, respectively. The performance of classifiers built from combinations of three feature selections, and nine machine learning algorithms was evaluated for two different diagnostic tasks, including 1) normal/abnormal (no CAD vs. CAD) classification, and 2) low-risk/high-risk CAD classification. Different metrics, including the area under the ROC curve (AUC), accuracy (ACC), sensitivity (SEN), and specificity (SPE), were reported for models' evaluation. Overall, models built on the Stress feature set (compared to other feature sets), and models to diagnose the second task (compared to task 1 models) revealed better performance. The Stress-mRMR-KNN (feature set-feature selection-classifier) reached the highest performance for task 1 with AUC, ACC, SEN, and SPE equal to 0.61, 0.63, 0.64, and 0.6, respectively. The Stress-Boruta-GB model achieved the highest performance for task 2 with AUC, ACC, SEN, and SPE of 0.79, 0.76, 0.75, and 0.76, respectively. Diabetes status from the clinical feature family, and dependence count non-uniformity normalized, from the NGLDM family, which is representative of non-uniformity in the region of interest were the most frequently selected features from stress feature set for CAD risk classification. This study revealed promising results for CAD risk classification using machine learning models built on MPI SPECT radiomics. The proposed models are helpful to alleviate the labor-intensive MPI SPECT interpretation process regarding CAD status and can potentially expedite the diagnostic process.

Myocardial Perfusion SPECT Imaging Radiomic Features and Machine Learning Algorithms for Cardiac Contractile Pattern Recognition

A U-shaped contraction pattern was shown to be associated with a better Cardiac resynchronization therapy (CRT) response. The main goal of this study is to automatically recognize left ventricular contractile patterns using machine learning algorithms trained on conventional quantitative features (ConQuaFea) and radiomic features extracted from Gated single-photon emission computed tomography myocardial perfusion imaging (GSPECT MPI). Among 98 patients with standard resting GSPECT MPI included in this study, 29 received CRT therapy and 69 did not (also had CRT inclusion criteria but did not receive treatment yet at the time of data collection, or refused treatment). A total of 69 non-CRT patients were employed for training, and the 29 were employed for testing. The models were built utilizing features from three distinct feature sets (ConQuaFea, radiomics, and ConQuaFea + radiomics (combined)), which were chosen using Recursive feature elimination (RFE) feature selection (FS), and then trained using seven different machine learning (ML) classifiers. In addition, CRT outcome prediction was assessed by different treatment inclusion criteria as the study's final phase. The MLP classifier had the highest performance among ConQuaFea models (AUC, SEN, SPE = 0.80, 0.85, 0.76). RF achieved the best performance in terms of AUC, SEN, and SPE with values of 0.65, 0.62, and 0.68, respectively, among radiomic models. GB and RF approaches achieved the best AUC, SEN, and SPE values of 0.78, 0.92, and 0.63 and 0.74, 0.93, and 0.56, respectively, among the combined models. A promising outcome was obtained when using radiomic and ConQuaFea from GSPECT MPI to detect left ventricular contractile patterns by machine learning.

Deep Learning-based Non-rigid Image Registration for High-dose Rate Brachytherapy in Inter-fraction Cervical Cancer

In this study, an inter-fraction organ deformation simulation framework for the locally advanced cervical cancer (LACC), which considers the anatomical flexibility, rigidity, and motion within an image deformation, was proposed. Data included 57 CT scans (7202 2D slices) of patients with LACC randomly divided into the train (n = 42) and test (n = 15) datasets. In addition to CT images and the corresponding RT structure (bladder, cervix, and rectum), the bone was segmented, and the coaches were eliminated. The correlated stochastic field was simulated using the same size as the target image (used for deformation) to produce the general random deformation. The deformation field was optimized to have a maximum amplitude in the rectum region, a moderate amplitude in the bladder region, and an amplitude as minimum as possible within bony structures. The DIRNet is a convolutional neural network that consists of convolutional regressors, spatial transformation, as well as resampling blocks. It was implemented by different parameters. Mean Dice indices of 0.89 ± 0.02, 0.96 ± 0.01, and 0.93 ± 0.02 were obtained for the cervix, bladder, and rectum (defined as at-risk organs), respectively. Furthermore, a mean average symmetric surface distance of 1.61 ± 0.46 mm for the cervix, 1.17 ± 0.15 mm for the bladder, and 1.06 ± 0.42 mm for the rectum were achieved. In addition, a mean Jaccard of 0.86 ± 0.04 for the cervix, 0.93 ± 0.01 for the bladder, and 0.88 ± 0.04 for the rectum were observed on the test dataset (15 subjects). Deep learning-based non-rigid image registration is, therefore, proposed for the high-dose-rate brachytherapy in inter-fraction cervical cancer since it outperformed conventional algorithms.

Clinical application of AI-based PET images in oncological patients

Based on the advantages of revealing the functional status and molecular expression of tumor cells, positron emission tomography (PET) imaging has been performed in numerous types of malignant diseases for diagnosis and monitoring. However, insufficient image quality, the lack of a convincing evaluation tool and intra- and interobserver variation in human work are well-known limitations of nuclear medicine imaging and restrict its clinical application. Artificial intelligence (AI) has gained increasing interest in the field of medical imaging due to its powerful information collection and interpretation ability. The combination of AI and PET imaging potentially provides great assistance to physicians managing patients. Radiomics, an important branch of AI applied in medical imaging, can extract hundreds of abstract mathematical features of images for further analysis. In this review, an overview of the applications of AI in PET imaging is provided, focusing on image enhancement, tumor detection, response and prognosis prediction and correlation analyses with pathology or specific gene mutations in several types of tumors. Our aim is to describe recent clinical applications of AI-based PET imaging in malignant diseases and to focus on the description of possible future developments.

Decentralized collaborative multi-institutional PET attenuation and scatter correction using federated deep learning

PURPOSE

Attenuation correction and scatter compensation (AC/SC) are two main steps toward quantitative PET imaging, which remain challenging in PET-only and PET/MRI systems. These can be effectively tackled via deep learning (DL) methods. However, trustworthy, and generalizable DL models commonly require well-curated, heterogeneous, and large datasets from multiple clinical centers. At the same time, owing to legal/ethical issues and privacy concerns, forming a large collective, centralized dataset poses significant challenges. In this work, we aimed to develop a DL-based model in a multicenter setting without direct sharing of data using federated learning (FL) for AC/SC of PET images.

METHODS

Non-attenuation/scatter corrected and CT-based attenuation/scatter corrected (CT-ASC) ¹⁸F-FDG PET images of 300 patients were enrolled in this study. The dataset consisted of 6 different centers, each with 50 patients, with scanner, image acquisition, and reconstruction protocols varying across the centers. CT-based ASC PET images served as the standard reference. All images were reviewed to include high-quality and artifact-free PET images. Both corrected and uncorrected PET images were converted to standardized uptake values (SUVs). We used a modified nested U-Net utilizing residual U-block in a U-shape architecture. We evaluated two FL models, namely sequential (FL-SQ) and parallel (FL-PL) and compared their performance with the baseline centralized (CZ) learning model wherein the data were pooled to one server, as well as center-based (CB) models where for each center the model was built and evaluated separately. Data from each center were divided to contribute to training (30 patients), validation (10 patients), and test sets (10 patients). Final evaluations and reports were performed on 60 patients (10 patients from each center).

RESULTS

In terms of percent SUV absolute relative error (ARE%), both FL-SQ (CI:12.21-14.81%) and FL-PL (CI:11.82-13.84%) models demonstrated excellent agreement with the centralized framework (CI:10.32-12.00%), while FL-based algorithms improved model performance by over 11% compared to CB training strategy (CI: 22.34-26.10%). Furthermore, the Mann-Whitney test between different strategies revealed no significant differences between CZ and FL-based algorithms (p-value > 0.05) in center-categorized mode. At the same time, a significant difference was observed between the different training approaches on the overall dataset (p-value < 0.05). In addition, voxel-wise comparison, with respect to reference CT-ASC, exhibited similar performance for images predicted by CZ (R² = 0.94), FL-SQ (R² = 0.93), and FL-PL (R² = 0.92), while CB model achieved a far lower coefficient of determination (R² = 0.74). Despite the strong correlations between CZ and FL-based methods compared to reference CT-ASC, a slight underestimation of predicted voxel values was observed.

CONCLUSION

Deep learning-based models provide promising results toward quantitative PET image reconstruction. Specifically, we developed two FL models and compared their performance with center-based and centralized models. The proposed FL-based models achieved higher performance compared to center-based models, comparable with centralized models. Our work provided strong empirical evidence that the FL framework can fully benefit from the generalizability and robustness of DL models used for AC/SC in PET, while obviating the need for the direct sharing of datasets between clinical imaging centers.

Federated Learning in Ocular Imaging: Current Progress and Future Direction

Advances in artificial intelligence deep learning (DL) have made tremendous impacts on the field of ocular imaging over the last few years. Specifically, DL has been utilised to detect and classify various ocular diseases on retinal photographs, optical coherence tomography (OCT) images, and OCT-angiography images. In order to achieve good robustness and generalisability of model performance, DL training strategies traditionally require extensive and diverse training datasets from various sites to be transferred and pooled into a "centralised location". However, such a data transferring process could raise practical concerns related to data security and patient privacy. Federated learning (FL) is a distributed collaborative learning paradigm which enables the coordination of multiple collaborators without the need for sharing confidential data. This distributed training approach has great potential to ensure data privacy among different institutions and reduce the potential risk of data leakage from data pooling or centralisation. This review article aims to introduce the concept of FL, provide current evidence of FL in ocular imaging, and discuss potential challenges as well as future applications.

[18F]FDG-PET/CT radiomics and artificial intelligence in lung cancer: Technical aspects and potential clinical applications

Lung cancer is the second most common cancer and the leading cause of cancer-related death worldwide. Molecular imaging using [¹⁸F]fluorodeoxyglucose Positron Emission Tomography and/or Computed Tomography ([¹⁸F]FDG-PET/CT) plays an essential role in the diagnosis, evaluation of response to treatment, and prediction of outcomes. The images are evaluated using qualitative and conventional quantitative indices. However, there is far more information embedded in the images, which can be extracted by sophisticated algorithms. Recently, the concept of uncovering and analyzing the invisible data extracted from medical images, called radiomics, is gaining more attention. Currently, [¹⁸F]FDG-PET/CT radiomics is growingly evaluated in lung cancer to discover if it enhances the diagnostic performance or implication of [¹⁸F]FDG-PET/CT in the management of lung cancer. In this review, we provide a short overview of the technical aspects, as they are discussed in different articles of this special issue. We mainly focus on the diagnostic performance of the [¹⁸F]FDG-PET/CT-based radiomics and the role of artificial intelligence in non-small cell lung cancer, impacting the early detection, staging, prediction of tumor subtypes, biomarkers, and patient's outcomes.