Best imaging signs identified by radiomics could outperform the model: application to differentiating lung carcinoid tumors from atypical hamartomas

Insights into Imaging

Original Article

Laure Fournier¹  Armelle Arnoux²  Laure Gibault³  Arthur Varoquaux⁴  Lilia Chermati⁵  Kathia Chaumoitre⁵  Paul Habert⁶  Jean-Yves Gaubert⁷  Antoine Decoux⁸  Loïc Duron⁹  Françoise Le Pimpec-Barthes^1,10  Pascal Thomas^1,11  Loïc Juquel^1,12  Stéphane Garcia^1,12 

[1] AP-HP, Hopital Européen Georges Pompidou, PARCC UMRS 970, INSERM, Université Paris Cité, Paris, France;AP-HP, Hopital Européen Georges Pompidou, Unité de Recherche Clinique, Centre d’Investigation Clinique 1418 Épidémiologie Clinique, INSERM, Université Paris Cité, Paris, France;Department of Pathology, Hôpital Européen Georges Pompidou, Assistance, Publique Hôpitaux de Paris, Paris, France;Department of Radiology, La Conception Hospital, Assistance Publique-Hôpitaux de Marseille, Aix-Marseille University, 13005, Marseille, France;Imaging Department, Hopital Nord, APHM, Aix Marseille University, Marseille, France;Imaging Department, Hopital Nord, APHM, Aix Marseille University, Marseille, France;LIIE, Aix Marseille Univ, Marseille, France;PARCC UMRS 970, INSERM, Université Paris Cité, Paris, France;LIIE, Aix Marseille Univ, Marseille, France;Department of Radiology, AP-HM, Hôpital La Timone, 13005, Marseille, France;PARCC UMRS 970, INSERM, Université Paris Cité, Paris, France;PARCC UMRS 970, INSERM, Université Paris Cité, Paris, France;Department of Neuroradiology, Alphonse de Rothschild Foundation Hospital, 75019, Paris, France;Service de Chirurgie Thoracique Hopital Européen Georges Pompidou, Université Paris Cité, Paris, France;Service de Chirurgie Thoracique et Transplantation Pulmonaire, Hôpital Nord, Chemin des Bourrely, Aix Marseille Université, 13015, Marseille, France;Service d’anatomie et Cytologie Pathologiques, Hôpital Nord, Chemin Des Bourrely, 13015, Marseille, France;U1068-CRCM, Aix Marseille Université, 13015, Marseille, France;

关键词: Carcinoid tumors;    Hamartomas;    Pulmonary neoplasms;    X-ray;    Computed tomography;    Radiomics;   

DOI  :  10.1186/s13244-023-01484-9

 received in 2023-05-12, accepted in 2023-07-17,  发布年份 2023

来源: Springer

【 摘 要 】

ObjectivesLung carcinoids and atypical hamartomas may be difficult to differentiate but require different treatment. The aim was to differentiate these tumors using contrast-enhanced CT semantic and radiomics criteria.MethodsBetween November 2009 and June 2020, consecutives patient operated for hamartomas or carcinoids with contrast-enhanced chest-CT were retrospectively reviewed. Semantic criteria were recorded and radiomics features were extracted from 3D segmentations using Pyradiomics. Reproducible and non-redundant radiomics features were used to training a random forest algorithm with cross-validation. A validation-set from another institution was used to evaluate of the radiomics signature, the 3D ‘median’ attenuation feature (3D-median) alone and the mean value from 2D-ROIs.ResultsSeventy-three patients (median 58 years [43‒70]) were analyzed (16 hamartomas; 57 carcinoids). The radiomics signature predicted hamartomas vs carcinoids on the external dataset (22 hamartomas; 32 carcinoids) with an AUC = 0.76. The 3D-median was the most important in the model. Density thresholds < 10 HU to predict hamartoma and > 60 HU to predict carcinoids were chosen for their high specificity > 0.90. On the external dataset, sensitivity and specificity of the 3D-median and 2D-ROIs were, respectively, 0.23, 1.00 and 0.13, 1.00 < 10 HU; 0.63, 0.95 and 0.69, 0.91 > 60 HU. The 3D-median was more reproducible than 2D-ROIs (ICC = 0.97 95% CI [0.95‒0.99]; bias: 3 ± 7 HU limits of agreement (LoA) [− 10‒16] vs. ICC = 0.90 95% CI [0.85‒0.94]; bias: − 0.7 ± 21 HU LoA [− 4‒40], respectively).ConclusionsA radiomics signature can distinguish hamartomas from carcinoids with an AUC = 0.76. Median density < 10 HU and > 60 HU on 3D or 2D-ROIs may be useful in clinical practice to diagnose these tumors with confidence, but 3D is more reproducible.Critical relevance statementRadiomic features help to identify the most discriminating imaging signs using random forest. ‘Median’ attenuation value (Hounsfield units), extracted from 3D-segmentations on contrast-enhanced chest-CTs, could distinguish carcinoids from atypical hamartomas (AUC = 0.85), was reproducible (ICC = 0.97), and generalized to an external dataset.Key points• 3D-‘Median’ was the best feature to differentiate carcinoids from atypical hamartomas (AUC = 0.85).• 3D-‘Median’ feature is reproducible (ICC = 0.97) and was generalized to an external dataset.• Radiomics signature from 3D-segmentations differentiated carcinoids from atypical hamartomas with an AUC = 0.76.• 2D-ROI value reached similar performance to 3D-‘median’ but was less reproducible (ICC = 0.90).Graphical Abstract

【 授权许可】

CC BY   
© European Society of Radiology (ESR) 2023

【 图 表 】

12888_2023_5172_Article_IEq1.gif

12888_2023_5142_Article_IEq23.gif

Fig. 1

13690_2023_1170_Article_IEq194.gif

12888_2023_5142_Article_IEq4.gif

Fig. 4

Fig. 3