Automatic Diagnosis of Chronic Thromboembolic Pulmonary Hypertension Based on Volumetric Data from SPECT Ventilation and Perfusion Images

Seiffert, Alexander P.; Gómez‐Grande, Adolfo; Pilkington, Patrick; Cara, Paula; Bueno, Héctor; Estenoz, J.; Gómez, Enrique J.; Sánchez-González, Patricia

doi:10.3390/app10155360

Cited by 3 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is only few data in the litterature about lung V/Q SPECT delineation in the context of PE. Derlin et al [21] and Seiffert et al [24] tested several maximum intensity thresholds and compared them to manual delineation of functional lungs. The best threshold was much higher as compared with our study (40% vs 15%) for Derlin, but closer for Seiffert [24] (18% for left lung and 21% for right lung), illustrating the lake of robustness of the method, especially if there are hot spots.…”

Section: Discussionmentioning

confidence: 99%

“…Derlin et al [21] and Seiffert et al [24] tested several maximum intensity thresholds and compared them to manual delineation of functional lungs. The best threshold was much higher as compared with our study (40% vs 15%) for Derlin, but closer for Seiffert [24] (18% for left lung and 21% for right lung), illustrating the lake of robustness of the method, especially if there are hot spots. Wang et al [25][26][27] proposed three segmentation methods (one threshold based method and two adaptive contouring methods), tested on Monte-Carlo simulations of homogeneous distributions of radioactivity within the lungs.…”

Section: Discussionmentioning

confidence: 99%

“…The first delineation method consisted in applying to all pixels a fixed intensity threshold defined as a percentage of the maximal pixel value. This has been the most commonly used method for lung segmentation mainly because of its ease of use [ 21 – 24 ]. The overall results showed relatively good spatial agreement with a mean DICE PE = 0.78.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Automatic delineation and quantification of pulmonary vascular obstruction index in patients with pulmonary embolism using Perfusion SPECT-CT: a simulation study

et al. 2021

View full text Add to dashboard Cite

Background In patients with pulmonary embolism (PE), there is a growing interest in quantifying the pulmonary vascular obtruction index (PVOI), which may be an independent risk factor for PE recurrence. Perfusion SPECT/CT is a very attractive tool to provide an accurate quantification of the PVOI. However, there is currently no reliable method to automatically delineate and quantify it. The aim of this phantom study was to assess and compare 3 segmentation methods for PVOI quantification with perfusion SPECT/CT imaging. Methods Three hundred ninety-six SPECT/CT scans, with various PE scenarios (n = 44), anterior to posterior perfusion gradients (n = 3), and lung volumes (n = 3) were simulated using Simind software. Three segmentation methods were assesssed: (1) using an intensity threshold expressed as a percentage of the maximal voxel value (MaxTh), (2) using a Z-score threshold (ZTh) after building a Z-score parametric lung map, and (3) using a relative difference threshold (RelDiffTh) after building a relative difference parametric map. Ninety randomly selected simulations were used to define the optimal threshold, and 306 simulations were used for the complete analysis. Spacial correlation between PE volumes from the phantom data and the delineated PE volumes was assessed by computing DICEPE indices. Bland-Altman statistics were used to calculate agreement for PVOI between the phantom data and the segmentation methods. Results Mean DICEPE index was higher with the RelDiffTh method (0.85 ± 0.08), as compared with the MaxTh method (0.78 ± 0.16) and the ZTh method (0.67 ± 0.15). Using the RelDiffTh method, mean DICEPE index remained high (> 0.81) regardless of the perfusion gradient and the lung volumes. Using the RelDiffTh method, mean relative difference in PVOI was − 12%, and the limits of agreement were − 40% to 16%. Values were 3% (− 75% to 81%) for MaxTh method and 0% (− 120% to 120%) for ZTh method. Graphycal analysis of the Bland-Altman graph for the RelDiffTh method showed very close estimation of the PVOI for small and medium PE, and a trend toward an underestimation of large PE. Conclusion In this phantom study, a delineation method based on a relative difference parametric map provided a good estimation of the PVOI, regardless of the extent of PE, the intensity of the anterior to posterior gradient, and the whole lung volumes.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Automatic delineation and quantification of pulmonary vascular obstruction index in patients with pulmonary embolism using Perfusion SPECT-CT: a simulation study

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Both perfusion lung volumes and perfusion index decreased with an increase in mPAP [31]. Following on from this work, a quantitative metric based on the difference in ventilation and perfusion volumes (V-P) has been used to identify discordant defects on SPECT images in CTEPH before and after treatment [32].…”

Section: Nuclear Medicinementioning

confidence: 99%

From Early Morphometrics to Machine Learning—What Future for Cardiovascular Imaging of the Pulmonary Circulation?

Gopalan

Gibbs

2020

Diagnostics

View full text Add to dashboard Cite

Imaging plays a cardinal role in the diagnosis and management of diseases of the pulmonary circulation. Behind the picture itself, every digital image contains a wealth of quantitative data, which are hardly analysed in current routine clinical practice and this is now being transformed by radiomics. Mathematical analyses of these data using novel techniques, such as vascular morphometry (including vascular tortuosity and vascular volumes), blood flow imaging (including quantitative lung perfusion and computational flow dynamics), and artificial intelligence, are opening a window on the complex pathophysiology and structure–function relationships of pulmonary vascular diseases. They have the potential to make dramatic alterations to how clinicians investigate the pulmonary circulation, with the consequences of more rapid diagnosis and a reduction in the need for invasive procedures in the future. Applied to multimodality imaging, they can provide new information to improve disease characterization and increase diagnostic accuracy. These new technologies may be used as sophisticated biomarkers for risk prediction modelling of prognosis and for optimising the long-term management of pulmonary circulatory diseases. These innovative techniques will require evaluation in clinical trials and may in themselves serve as successful surrogate end points in trials in the years to come.

show abstract

Non-contrasted computed tomography (NCCT) based chronic thromboembolic pulmonary hypertension (CTEPH) automatic diagnosis using cascaded network with multiple instance learning

Zhao,

Song,

Zhu

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective The diagnosis of chronic thromboembolic pulmonary hypertension (CTEPH) is challenging due to nonspecific early symptoms, complex diagnostic processes, and small lesion sizes. This study aims to develop an automatic diagnosis method for CTEPH using non-contrasted computed tomography (NCCT) scans, enabling automated diagnosis without precise lesion annotation.Approach A novel Cascade Network with Multiple Instance Learning (CNMIL) framework was developed to improve the diagnosis of CTEPH. This method uses a cascade network architecture combining two Resnet-18 CNN networks to progressively distinguish between normal and CTEPH cases. Multiple Instance Learning (MIL) is employed to treat each 3D CT case as a "bag" of image slices, using attention scoring to identify the most important slices. An attention module helps the model focus on diagnostically relevant regions within each slice. The dataset comprised NCCT scans from 300 subjects, including 117 males and 183 females, with an average age of 52.5 ± 20.9 years, consisting of 132 normal cases and 168 cases of lung diseases, including 88 instances of CTEPH. The CNMIL framework was evaluated using sensitivity, specificity, and the area under the curve (AUC) metrics, and compared with common 3D supervised classification networks and existing CTEPH automatic diagnosis networks.Main Results The CNMIL framework demonstrated high diagnostic performance, achieving an AUC of 0.807, accuracy of 0.833, sensitivity of 0.795, and specificity of 0.849 in distinguishing CTEPH cases. Ablation studies revealed that integrating MIL and the cascade network significantly enhanced performance, with the model achieving an AUC of 0.993 and perfect sensitivity (1.000) in normal classification. Comparisons with other 3D network architectures confirmed that the integrated model outperformed others, achieving the highest AUC of 0.8419.Significance The CNMIL network requires no additional scans or annotations, relying solely on NCCT. This approach can improve timely and accurate CTEPH detection, resulting in better patient outcomes.

show abstract

Automatic Diagnosis of Chronic Thromboembolic Pulmonary Hypertension Based on Volumetric Data from SPECT Ventilation and Perfusion Images

Cited by 3 publications

References 21 publications

Automatic delineation and quantification of pulmonary vascular obstruction index in patients with pulmonary embolism using Perfusion SPECT-CT: a simulation study

Automatic delineation and quantification of pulmonary vascular obstruction index in patients with pulmonary embolism using Perfusion SPECT-CT: a simulation study

From Early Morphometrics to Machine Learning—What Future for Cardiovascular Imaging of the Pulmonary Circulation?

Non-contrasted computed tomography (NCCT) based chronic thromboembolic pulmonary hypertension (CTEPH) automatic diagnosis using cascaded network with multiple instance learning

Contact Info

Product

Resources

About