Explainable AI to improve acceptance of convolutional neural networks for automatic classification of dopamine transporter SPECT in the diagnosis of clinically uncertain parkinsonian syndromes

Nazari, Mahmood; Kluge, Andreas; Apostolova, Ivayla; Klutmann, Susanne; Kimiaei, Sharok; Schroeder, Michael; Buchert, Ralph

doi:10.1007/s00259-021-05569-9

Cited by 27 publications

(11 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Figure S8d-p, DL activated attention to the periphery of alveolar bone, which is difficult for humans in making a rapid assessment. It is well known that the attention/saliency of DL models sometimes does not reveal an obvious image pattern for humans to interpret even if they make a correct prediction (Ding et al, 2019;Chereda et al, 2021;Nazari et al, 2021).…”

Section: Discussionmentioning

confidence: 99%

“…In Figure S8d–p, DL activated attention to the periphery of alveolar bone, which is difficult for humans in making a rapid assessment. It is well known that the attention/saliency of DL models sometimes does not reveal an obvious image pattern for humans to interpret even if they make a correct prediction (Ding et al, 2019; Chereda et al, 2021; Nazari et al, 2021). Encouragingly, these discrepancies reveal the potential to integrate both machine‐learned features and human expert experience, which may further optimize the performance of future prediction models.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting the risk of dental implant loss using deep learning

Huang

Liu

Yan

et al. 2022

J Clinic Periodontology

View full text Add to dashboard Cite

Aim To investigate the feasibility of predicting dental implant loss risk with deep learning (DL) based on preoperative cone‐beam computed tomography. Materials and Methods Six hundred and three patients who underwent implant surgery (279 high‐risk patients who did and 324 low‐risk patients who did not experience implant loss within 5 years) between January 2012 and January 2020 were enrolled. Three models, a logistic regression clinical model (CM) based on clinical features, a DL model based on radiography features, and an integrated model (IM) developed by combining CM with DL, were developed to predict the 5‐year implant loss risk. The area under the receiver operating characteristic curve (AUC) was used to evaluate the model performance. Time to implant loss was considered for both groups, and Kaplan–Meier curves were created and compared by the log‐rank test. Results The IM exhibited the best performance in predicting implant loss risk (AUC = 0.90, 95% confidence interval [CI] 0.84–0.95), followed by the DL model (AUC = 0.87, 95% CI 0.80–0.92) and the CM (AUC = 0.72, 95% CI 0.63–0.79). Conclusions Our study offers preliminary evidence that both the DL model and the IM performed well in predicting implant fate within 5 years and thus may greatly facilitate implant practitioners in assessing preoperative risks.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Predicting the risk of dental implant loss using deep learning

Huang

Liu

Yan

et al. 2022

J Clinic Periodontology

View full text Add to dashboard Cite

show abstract

“… Nazari et al (2022) tested layer-wise relevance propagation (LRP) based CNN for classification of normal and reduced patients. The study achieved a sensitivity and specificity of 92.8% and 98.7% respectively.…”

Section: Resultsmentioning

confidence: 99%

Soft Attention Based DenseNet Model for Parkinson’s Disease Classification Using SPECT Images

Thakur

Kuresan

Dhanalakshmi

et al. 2022

Front. Aging Neurosci.

View full text Add to dashboard Cite

ObjectiveDeep learning algorithms have long been involved in the diagnosis of severe neurological disorders that interfere with patients’ everyday tasks, such as Parkinson’s disease (PD). The most effective imaging modality for detecting the condition is DaTscan, a variety of single-photon emission computerized tomography (SPECT) imaging method. The goal is to create a convolutional neural network that can specifically identify the region of interest following feature extraction.MethodsThe study comprised a total of 1,390 DaTscan imaging groups with PD and normal classes. The architecture of DenseNet-121 is leveraged with a soft-attention block added before the final classification layer. For visually analyzing the region of interest (ROI) from the images after classification, Soft Attention Maps and feature map representation are used.OutcomesThe model obtains an overall accuracy of 99.2% and AUC-ROC score 99%. A sensitivity of 99.2%, specificity of 99.4% and f1-score of 99.1% is achieved that surpasses all prior research findings. Soft-attention map and feature map representation aid in highlighting the ROI, with a specific attention on the putamen and caudate regions.ConclusionWith the deep learning framework adopted, DaTscan images reveal the putamen and caudate areas of the brain, which aid in the distinguishing of normal and PD cohorts with high accuracy and sensitivity.

show abstract

“…About two thirds of the DAT-SPECT had been included in previous studies on deep learning-based classi cation of DAT-SPECT [17] and data-driven identi cation of diagnostically useful extrastriatal signal in DAT-SPECT [18].…”

Section: Patientsmentioning

confidence: 99%

Stereotactical Normalization with Multiple Templates Representative of Normal and Parkinson-Typical Reduction of Striatal Uptake Improves the Diagnostic Power of Automatic Semi-Quantitative Analysis in Dopamine Transporter SPECT

Apostolova

Schiebler

Lange

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Background The specific binding ratio (SBR) of 123I-FP-CIT in the putamen is widely used to support the interpretation of dopamine transporter (DAT) SPECT. Automatic methods for computation of the putamen SBR often include stereotactical normalization of the individual DAT-SPECT image to an anatomical standard space. This study compared using a single 123I-FP-CIT template image as target for stereotactical normalization versus multiple templates representative of normal and different levels of Parkinson-typical reduction of striatal 123I-FP-CIT uptake. Methods 123I-FP-CIT SPECT images of 1738 patients (66.2 ± 11.6y) were stereotactically normalized (affine) to the anatomical space of the Montreal Neurological Institute (MNI) with SPM12 either using a single custom-made FP-CIT template representative of normal striatal uptake or using eight different templates representative of normal and different levels of Parkinson-typical reduction of striatal FP-CIT uptake with and without attenuation and scatter correction. In the latter case, SPM finds the linear combination of these templates that best matches the patient’s image. The putamen SBR was obtained using hottest voxels analysis in large unilateral regions of interest predefined in MNI space. The histogram of the putamen SBR in the whole sample was fitted by the sum of two Gaussian functions. The power to differentiate between reduced and normal SBR was estimated by the effect size of the distance between the two Gaussian functions computed as the differences between their mean values scaled to their pooled standard deviation. Results The effect size of the distance between the two Gaussian functions was 3.84 with the single template versus 3.96 with multiple templates for stereotactical normalization. Conclusions Multiple templates representative of normal and different levels of Parkinson-typical reduction for stereotactical normalization of DAT-SPECT provide improved separation between normal and reduced putamen SBR, suggesting slightly improved power for the detection of nigrostriatal degeneration.

show abstract

Explainable AI to improve acceptance of convolutional neural networks for automatic classification of dopamine transporter SPECT in the diagnosis of clinically uncertain parkinsonian syndromes

Cited by 27 publications

References 53 publications

Predicting the risk of dental implant loss using deep learning

Predicting the risk of dental implant loss using deep learning

Soft Attention Based DenseNet Model for Parkinson’s Disease Classification Using SPECT Images

Stereotactical Normalization with Multiple Templates Representative of Normal and Parkinson-Typical Reduction of Striatal Uptake Improves the Diagnostic Power of Automatic Semi-Quantitative Analysis in Dopamine Transporter SPECT

Contact Info

Product

Resources

About