Fully automatic segmentation of diffuse large B cell lymphoma lesions on 3D FDG-PET/CT for total metabolic tumour volume prediction using a convolutional neural network.

Blanc-Durand, Paul; Jégou, Simon; Kanoun, Salim; Berriolo‐Riedinger, Alina; Bodet-Milin, Caroline; Kraeber-Bodéré, Françoise; Carlier, Thomas; Gouill, Steven Le; Casasnovas, Olivier; Meignan, Michel; Itti, Emmanuel

doi:10.1007/s00259-020-05080-7

Cited by 82 publications

(82 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our results are consistent with a recent study (34), in which the performances of a CNN model, based on nnU-Net, were investigated to automatically segment TMTV in patients with DLBCL. A first cohort of 639 patients with pretherapeutic FDG PET/CT was used to train the model.…”

Section: Discussionsupporting

confidence: 92%

Evaluation of an Automatic Classification Algorithm Using Convolutional Neural Networks in Oncological Positron Emission Tomography

Pinochet¹,

Eude²,

Becker

et al. 2021

Front. Med.

View full text Add to dashboard Cite

Introduction: Our aim was to evaluate the performance in clinical research and in clinical routine of a research prototype, called positron emission tomography (PET) Assisted Reporting System (PARS) (Siemens Healthineers) and based on a convolutional neural network (CNN), which is designed to detect suspected cancer sites in fluorine-18 fluorodeoxyglucose (18F-FDG) PET/computed tomography (CT).Method: We retrospectively studied two cohorts of patients. The first cohort consisted of research-based patients who underwent PET scans as part of the initial workup for diffuse large B-cell lymphoma (DLBCL). The second cohort consisted of patients who underwent PET scans as part of the evaluation of miscellaneous cancers in clinical routine. In both cohorts, we assessed the correlation between manually and automatically segmented total metabolic tumor volumes (TMTVs), and the overlap between both segmentations (Dice score). For the research cohort, we also compared the prognostic value for progression-free survival (PFS) and overall survival (OS) of manually and automatically obtained TMTVs.Results: For the first cohort (research cohort), data from 119 patients were retrospectively analyzed. The median Dice score between automatic and manual segmentations was 0.65. The intraclass correlation coefficient between automatically and manually obtained TMTVs was 0.68. Both TMTV results were predictive of PFS (hazard ratio: 2.1 and 3.3 for automatically based and manually based TMTVs, respectively) and OS (hazard ratio: 2.4 and 3.1 for automatically based and manually based TMTVs, respectively). For the second cohort (routine cohort), data from 430 patients were retrospectively analyzed. The median Dice score between automatic and manual segmentations was 0.48. The intraclass correlation coefficient between automatically and manually obtained TMTVs was 0.61.Conclusion: The TMTVs determined for the research cohort remain predictive of total and PFS for DLBCL. However, the segmentations and TMTVs determined automatically by the algorithm need to be verified and, sometimes, corrected to be similar to the manual segmentation.

show abstract

Section: Discussionsupporting

confidence: 92%

Evaluation of an Automatic Classification Algorithm Using Convolutional Neural Networks in Oncological Positron Emission Tomography

Pinochet¹,

Eude²,

Becker

et al. 2021

Front. Med.

View full text Add to dashboard Cite

show abstract

“…Indeed, SDmax_Vox slightly majored the maximal distance. This is important as some tumor segmentation methods involving deep learning only highlight tumor voxels without assigning voxels to a specific lesion [12,20]. In addition, this could make theoretically possible a rough evaluation of the distance on a Maximum Intensity Projection image displayed in 2D.…”

Section: Discussionmentioning

confidence: 99%

New Approaches in Characterization of Lesions Dissemination in DLBCL Patients on Baseline PET/CT

Meignan

Nioche

Clerc

et al. 2021

Cancers

Self Cite

View full text Add to dashboard Cite

Dissemination, expressed recently by the largest Euclidian distance between lymphoma sites (SDmax), appeared a promising risk factor in DLBCL patients. We investigated alternative distance metrics to characterize the robustness of the dissemination information. In 290 patients from the REMARC trial (NCT01122472), the Euclidean (Euc), Manhattan (Man), and Tchebychev (Tch) distances between the furthest lesions, firstly based on the centroid of each lesion and then directly from the two most distant tumor voxels and the Travelling Salesman Problem distance (TSP) were calculated. For PFS, the areas under the ROC curves were between 0.63 and 0.64, and between 0.62 and 0.65 for OS. Patients with high SDmax whatever the method of calculation or high SD_TSP had a significantly poorer outcome than patients with low SDmax or SD_TSP (p < 0.001 for both PFS and OS), with significance maintained in Ann Arbor advanced-stage patients. In multivariate analysis with total metabolic tumor volume and ECOG, each distance feature had an independent prognostic value for PFS. For OS, only SDmax_Tch, SDmax_Euc _Vox, and SDmax_Man _Vox reached significance. The spread of DLBCL lesions measured by the largest distance between lymphoma sites is a strong independent prognostic factor and could be measured directly from tumor voxels, allowing its development in the area of the deep learning segmentation methods.

show abstract

“…Of note, the integration of molecular imaging and AI algorithms will, probably, constitute an essential pillar in LN management in the near future, not only with diagnostic and staging purposes, but also as a prognostic marker. Regarding this, several AI algorithms have been tested for this purpose, such as deep learning (DL) for the reconstruction of positron emission tomography (PET) image in Hodgkin lymphoma (HL) [46], convolutional neural networks (CNNs) for the prediction of diffuse large B-cell lymphoma (DLBCL) total metabolic tumor based on PET/computed tomography (CT) [47] and support vector machine (SVM) to discriminate hypermetabolic lymphomatous lesions and noncancerous processes [48].…”

Section: Discussionmentioning

confidence: 99%

“…As mentioned, when analyzing research areas and source titles, functional imaging analysis and AI constitutes a growing area of interest. Regarding this, it could be defined a cognitive evolution from general concepts such as medical image analysis (first and second periods) to more specific tools such as magnetic resonance and CT. Herein, deep CNNs have been employed to discriminate patterns of tumor infiltration in PET/CT in 327 patients with NHL [66], and prediction of response to conventional chemotherapy by integrating AI and molecular techniques also constitutes a growing area of interest in recent years [47,67].…”

Section: Discussionmentioning

confidence: 99%

Translational Applications of Artificial Intelligence and Machine Learning for Diagnostic Pathology in Lymphoid Neoplasms: A Comprehensive and Evolutive Analysis

et al. 2021

View full text Add to dashboard Cite

Genomic analysis and digitalization of medical records have led to a big data scenario within hematopathology. Artificial intelligence and machine learning tools are increasingly used to integrate clinical, histopathological, and genomic data in lymphoid neoplasms. In this study, we identified global trends, cognitive, and social framework of this field from 1990 to 2020. Metadata were obtained from the Clarivate Analytics Web of Science database in January 2021. A total of 525 documents were assessed by document type, research areas, source titles, organizations, and countries. SciMAT and VOSviewer package were used to perform scientific mapping analysis. Geographical distribution showed the USA and People’s Republic of China as the most productive countries, reporting up to 190 (36.19%) of all documents. A third-degree polynomic equation predicts that future global production in this area will be three-fold the current number, near 2031. Thematically, current research is focused on the integration of digital image analysis and genomic sequencing in Non-Hodgkin lymphomas, prediction of chemotherapy response and validation of new prognostic models. These findings can serve pathology departments to depict future clinical and research avenues, but also, public institutions and administrations to promote synergies and optimize funding allocation.

show abstract

Fully automatic segmentation of diffuse large B cell lymphoma lesions on 3D FDG-PET/CT for total metabolic tumour volume prediction using a convolutional neural network.

Cited by 82 publications

References 28 publications

Evaluation of an Automatic Classification Algorithm Using Convolutional Neural Networks in Oncological Positron Emission Tomography

Evaluation of an Automatic Classification Algorithm Using Convolutional Neural Networks in Oncological Positron Emission Tomography

New Approaches in Characterization of Lesions Dissemination in DLBCL Patients on Baseline PET/CT

Translational Applications of Artificial Intelligence and Machine Learning for Diagnostic Pathology in Lymphoid Neoplasms: A Comprehensive and Evolutive Analysis

Contact Info

Product

Resources

About