Dataset Information

¹⁸F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.

ABSTRACT: Total metabolic tumor volume (TMTV) and tumor dissemination (Dmax) calculated from baseline ¹⁸F-FDG PET/CT images are prognostic biomarkers in diffuse large B-cell lymphoma (DLBCL) patients. Yet, their automated calculation remains challenging. The purpose of this study was to investigate whether TMTV and Dmax features could be replaced by surrogate features automatically calculated using an artificial intelligence (AI) algorithm from only 2 maximum-intensity projections (MIPs) of the whole-body ¹⁸F-FDG PET images. Methods: Two cohorts of DLBCL patients from the REMARC (NCT01122472) and LNH073B (NCT00498043) trials were retrospectively analyzed. Experts delineated lymphoma lesions from the baseline whole-body ¹⁸F-FDG PET/CT images, from which TMTV and Dmax were measured. Coronal and sagittal MIP images and associated 2-dimensional reference lesion masks were calculated. An AI algorithm was trained on the REMARC MIP data to segment lymphoma regions. The AI algorithm was then used to estimate surrogate TMTV (sTMTV) and surrogate Dmax (sDmax) on both datasets. The ability of the original and surrogate TMTV and Dmax to stratify patients was compared. Results: Three hundred eighty-two patients (mean age ± SD, 62.1 y ± 13.4 y; 207 men) were evaluated. sTMTV was highly correlated with TMTV for REMARC and LNH073B datasets (Spearman r = 0.878 and 0.752, respectively), and so were sDmax and Dmax (r = 0.709 and 0.714, respectively). The hazard ratios for progression free survival of volume and MIP-based features derived using AI were similar, for example, TMTV: 11.24 (95% CI: 2.10-46.20), sTMTV: 11.81 (95% CI: 3.29-31.77), and Dmax: 9.0 (95% CI: 2.53-23.63), sDmax: 12.49 (95% CI: 3.42-34.50). Conclusion: Surrogate TMTV and Dmax calculated from only 2 PET MIP images are prognostic biomarkers in DLBCL patients and can be automatically estimated using an AI algorithm.

SUBMITTER: Girum KB

PROVIDER: S-EPMC9730929 | biostudies-literature | 2022 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

<sup>18</sup>F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.

Girum Kibrom B KB Rebaud Louis L Cottereau Anne-Ségolène AS Meignan Michel M Clerc Jérôme J Vercellino Laetitia L Casasnovas Olivier O Morschhauser Franck F Thieblemont Catherine C Buvat Irène I

Journal of nuclear medicine : official publication, Society of Nuclear Medicine 20220616 12

Total metabolic tumor volume (TMTV) and tumor dissemination (Dmax) calculated from baseline <sup>18</sup>F-FDG PET/CT images are prognostic biomarkers in diffuse large B-cell lymphoma (DLBCL) patients. Yet, their automated calculation remains challenging. The purpose of this study was to investigate whether TMTV and Dmax features could be replaced by surrogate features automatically calculated using an artificial intelligence (AI) algorithm from only 2 maximum-intensity projections (MIPs) of the ...[more]

PMID: 35710733

Similar Datasets

Project description:Positron emission tomography (PET) of lung tumors suffers from breathing-motion induced blurring. Respiratory-correlated PET ameliorates motion blurring and enables visualization of lung tumor functional uptake throughout the breathing cycle but has achieved limited clinical use in radiotherapy planning. In this work, the authors propose a process for generating a gated PET maximum intensity projection (MIP), a breathing-phase projection of the 4D image set comprising gated PET images, as a technique to quantitatively and efficiently incorporate respiratory-correlated PET information into radiotherapy treatment planning.4D-CT and respiratory-gated PET using [(18)F]fluorodeoxyglucose (FDG) were acquired of three patients with a total of four small (4-18 cc), clearly defined lower-lobe lung tumors. Internal target volumes (ITVs) for the lung tumors were generated by threshold-based segmentation of PET-MIP images and ungated PET images (ITV(PET-MIP) and ITV(3D-PET), respectively), and by manual contouring of CT-MIP and end-exhale and end-inhale phases of 4D-CT (ITV(CT-MIP)) by a radiation oncologist. Because of the sensitivity of tumor segmentation to threshold value, several different thresholds were tested for ITV generation, including 40%, 30%, and 20% of maximum standardized uptake value (SUV(max)) for FDG as well as absolute SUV thresholds of 2.5 and 3.0. The normalized overlap and relative volumes of ITV(PET-MIP) and ITV(3D-PET) with respect to ITV(CT-MIP) were compared. The images were also visually compared. ITV(CT-MIP) was considered a gold standard for these tumors with CT-visible morphology.The mean and standard deviation normalized overlap and relative volumes between ITV(PET-MIP) and ITV(CT-MIP) were 0.68 ± 0.07 and 1.07 ± 0.42, respectively, averaged over all four tumors and all five threshold values. The mean and standard deviation normalized overlap and relative volumes of ITV(3D-PET) and ITV(CT-MIP) were 0.47 ± 0.12 and 0.69 ± 0.56, respectively.PET-MIP images better match CT-MIP images for this sample of four small CT-visible tumors as compared to ungated PET images, based on the metrics of volumetric overlap and relative volumes as well as visual interpretation. The PET-MIP is a way to incorporate 4D-PET imaging into the process of lung tumor contouring that is time-efficient for the radiation oncologist and involves minimal effort to implement in treatment planning software, because it requires only a single PET image beyond contouring on CT alone.

Project description:ObjectivesTo automatically detect MRI artifacts on dynamic contrast-enhanced (DCE) maximum intensity projections (MIPs) of the breast using deep learning.MethodsWomen who underwent clinically indicated breast MRI between October 2015 and December 2019 were included in this IRB-approved retrospective study. We employed two convolutional neural network architectures (ResNet and DenseNet) to detect the presence of artifacts on DCE MIPs of the left and right breasts. Networks were trained on images acquired up to and including the year 2018 using a 5-fold cross-validation (CV). Ensemble classifiers were built with the resulting CV models and applied to an independent holdout test dataset, which was formed by images acquired in 2019.ResultsOur study sample contained 2265 examinations from 1794 patients (median age at first acquisition: 50 years [IQR: 17 years]), corresponding to 1827 examinations of 1378 individuals in the training dataset and 438 examinations of 416 individuals in the holdout test dataset with a prevalence of image-level artifacts of 53% (1951/3654 images) and 43% (381/876 images), respectively. On the holdout test dataset, the ResNet and DenseNet ensembles demonstrated an area under the ROC curve of 0.92 and 0.94, respectively.ConclusionNeural networks are able to reliably detect artifacts that may impede the diagnostic assessment of MIPs derived from DCE subtraction series in breast MRI. Future studies need to further explore the potential of such neural networks to complement quality assurance and improve the application of DCE MIPs in a clinical setting, such as abbreviated protocols.Key points• Deep learning classifiers are able to reliably detect MRI artifacts in dynamic contrast-enhanced protocol-derived maximum intensity projections of the breast. • Automated quality assurance of maximum intensity projections of the breast may be of special relevance for abbreviated breast MRI, e.g., in high-throughput settings, such as cancer screening programs.

Project description:BackgroundIn the present study, we aimed to investigate the role of baseline (B), interim (I) and end-of-treatment (Eot) 18F-FDG PET/CT in assessing the prognosis of diffuse large B cell lymphoma (DLBCL), so as to identify patients who need intensive treatment at an early stage.MethodsA total of 127 DLBCL patients (62 men; 65 women; median age 62 years) were retrospectively analyzed in this study. Baseline (n = 127), interim (n = 127, after 3-4 cycles) and end-of-treatment (n = 53, after 6-8 cycles) PET/CT images were re-evaluated; semi-quantitative parameters such as maximum standardized uptake value of lesion-to-liver ratio (SUVmax(LLR)) and lesion-to-mediastinum ratio (SUVmax(LMR)), total metabolic tumor volume (TMTV) and total metabolic tumor volume (TLG) were recorded. ΔTLG1 was the change of interim relative to baseline TLG (I to B), ΔTLG2 (Eot to B). ΔSUVmax and ΔTMTV were the same algorithm. The visual Deauville 5-point scale (D-5PS) has been adopted as the major criterion for PET evaluation. Visual analysis (VA) and semi-quantitative parameters were assessed for the ability to predict progression-free survival (PFS) and overall survival (OS) by using Kaplan-Meier method, cox regression and logistic regression analysis. When visual and semi-quantitative analysis are combined, the result is only positive if both are positive.ResultsAt a median follow-up of 34 months, the median PFS and OS were 20 and 32 months. The survival curve analysis showed that advanced stage and IPI score with poor prognosis, ΔSUVmax(LLR)1 < 89.2%, ΔTMTV1 < 91.8% and ΔTLG1 < 98.8%, ΔSUVmax(LLR)2 < 86.4% were significantly related to the shortening of PFS in patient (p < 0.05). ΔSUVmax(LLR)1 < 83.2% and ΔTLG1 < 97.6% were significantly correlated with the shortening of OS in patients (p < 0.05). Visual analysis showed that incomplete metabolic remission at I-PET and Eot-PET increased the risk of progress and death. In terms of predicting recurrence by I-PET, the combination of visual and semi-quantitative parameters showed higher positive predictive value (PPV) and specificity than a single index.ConclusionThree to four cycles of R-CHOP treatment may be a time point for early prediction of early recurrence/refractory (R/R) patients and active preemptive treatment. Combined visual analysis with semi-quantitative parameters of 18F-FDG PET/CT at interim can improve prognostic accuracy and may allow for more precise screening of patients requiring early intensive therapy.

Project description:BackgroundThe pro-tumoral action of the cluster of differentiation 147 (CD147), which is associated with the chemotherapy resistance of lung adenocarcinoma, is partly due to accelerated tumor cell glycolysis. 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) metabolic parameters included maximal standardized uptake value (SUVmax), mean standardized uptake value (SUVmean), metabolic tumor volume (MTV), and total lesion glycolysis (TLG), which are non-invasive markers of the glucose metabolism of tumor cells in vivo. This study aimed to clarify the correlation between PET metabolic parameters and CD147 expression, and to evaluate the prognostic value of CD147 expression in resectable lung adenocarcinoma patients.MethodsA total of 89 lung adenocarcinoma chemotherapy-naive patients who underwent 18F-fluorodeoxyglucose positron emission tomography and computerized tomography scan before pulmonary surgery were retrospectively analyzed. The PET metabolic parameters were calculated by 18F-FDG PET imaging, and CD147 expression was analyzed by immunohistochemistry. SUVmax, SUVmean, MTV, and TLG compared for their performance in predicting the expression of CD147 were illustrated with statistical analysis. All patients were then followed-up for survival analysis.ResultsThe SUVmax was significantly correlated with the CD147 expression and was the primary predictor for the CD147 expression of lung adenocarcinoma. A cut-off value of the SUVmax, 9.77 allowed 85.1% sensitivity and 64.3% specificity for predicting the CD147 positive lung adenocarcinoma. CD147 expression was correlated with tumor differentiation and metastasis. Univariate survival analysis showed that CD147 expression was significantly associated with a shorter overall survival (OS) time. Multivariate analysis revealed that CD147 was an independent prognostic factor of lung adenocarcinoma patients.ConclusionThe SUVmax of a primary tumor measured with 18F-FDG PET may be a simple and non-invasive marker for predicting CD147 expression in lung adenocarcinoma. CD147 is an independent prognostic factor related to OS of postoperative lung adenocarcinoma patients.

Project description:We introduce multiple-organ objective segmentation (MOOSE) software that generates subject-specific, multiorgan segmentation using data-centric artificial intelligence principles to facilitate high-throughput systemic investigations of the human body via whole-body PET imaging. Methods: Image data from 2 PET/CT systems were used in training MOOSE. For noncerebral structures, 50 whole-body CT images were used, 30 of which were acquired from healthy controls (14 men and 16 women), and 20 datasets were acquired from oncology patients (14 men and 6 women). Noncerebral tissues consisted of 13 abdominal organs, 20 bone segments, subcutaneous fat, visceral fat, psoas muscle, and skeletal muscle. An expert panel manually segmented all noncerebral structures except for subcutaneous fat, visceral fat, and skeletal muscle, which were semiautomatically segmented using thresholding. A majority-voting algorithm was used to generate a reference-standard segmentation. From the 50 CT datasets, 40 were used for training and 10 for testing. For cerebral structures, 34 18F-FDG PET/MRI brain image volumes were used from 10 healthy controls (5 men and 5 women imaged twice) and 14 nonlesional epilepsy patients (7 men and 7 women). Only 18F-FDG PET images were considered for training: 24 and 10 of 34 volumes were used for training and testing, respectively. The Dice score coefficient (DSC) was used as the primary metric, and the average symmetric surface distance as a secondary metric, to evaluate the automated segmentation performance. Results: An excellent overlap between the reference labels and MOOSE-derived organ segmentations was observed: 92% of noncerebral tissues showed DSCs of more than 0.90, whereas a few organs exhibited lower DSCs (e.g., adrenal glands [0.72], pancreas [0.85], and bladder [0.86]). The median DSCs of brain subregions derived from PET images were lower. Only 29% of the brain segments had a median DSC of more than 0.90, whereas segmentation of 60% of regions yielded a median DSC of 0.80-0.89. The results of the average symmetric surface distance analysis demonstrated that the average distance between the reference standard and the automatically segmented tissue surfaces (organs, bones, and brain regions) lies within the size of image voxels (2 mm). Conclusion: The proposed segmentation pipeline allows automatic segmentation of 120 unique tissues from whole-body 18F-FDG PET/CT images with high accuracy.

Dataset Information

¹⁸F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.

Publications

<sup>18</sup>F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets

Dataset Information

18F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.

Publications

<sup>18</sup>F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets

¹⁸F-FDG PET Maximum-Intensity Projections and Artificial Intelligence: A Win-Win Combination to Easily Measure Prognostic Biomarkers in DLBCL Patients.