SVM-Based Normal Pressure Hydrocephalus Detection

Rau, Alexander; Kim, Suam; Yang, Shan; Reisert, Marco; Kellner, Elias; Duman, Ikram Eda; Stieltjes, Bram; Hohenhaus, Marc; Beck, Jürgen; Urbach, Horst; Egger, Karl

doi:10.1007/s00062-020-00993-0

Cited by 20 publications

(21 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In addition, we evaluated this subjective scoring using an SVM-based approach 25 using a prediction score of 0.42 [Range 0–1] as cut-off value, whereas > 0.42 corresponds to a high probability of iNPH and < 0.42 to a low probability of iNPH (in our study = control group).…”

Section: Methodsmentioning

confidence: 99%

Increased interstitial fluid in periventricular and deep white matter hyperintensities in patients with suspected idiopathic normal pressure hydrocephalus

Rau

Reisert

Kellner

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

Periventricular white matter changes are common in patients with idiopathic normal pressure hydrocephalus (iNPH) and considered to represent focally elevated interstitial fluid. We compared diffusion measures in periventricular hyperintensities in patients with imaging features of iNPH to patients without. The hypothesis is that periventricular hyperintensities in patients with presumed iNPH show higher water content than in patients without imaging features of iNPH. 21 patients with iNPH Radscale 7–12 (“high probability of iNPH”) and 10 patients with iNPH Radscale 2–4 (“low probability of iNPH”) were examined with a neurodegeneration imaging protocol including a diffusion microstructure imaging sequence. Periventricular hyperintensities and deep white matter hyperintensities were segmented and diffusion measures were compared. In patients with imaging features of iNPH, the free water content in periventricular hyperintensities was significantly higher compared to the control group (p = 0.005). This effect was also detectable in deep white matter hyperintensities (p = 0.024). Total brain volumes and total gray or white matter volumes did not differ between the groups. Periventricular cap free water fraction was highly discriminative regarding patients with presumed iNPH and controls with an ROC AUC of 0.933. Quantitative diffusion microstructure imaging shows elevated water content in periventricular hyperintensities in patients with imaging features of iNPH, which could be the imaging correlate for pathologic fluid accumulation and may be used as an imaging biomarker in the future.

show abstract

Section: Methodsmentioning

confidence: 99%

Increased interstitial fluid in periventricular and deep white matter hyperintensities in patients with suspected idiopathic normal pressure hydrocephalus

Rau

Reisert

Kellner

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…Using the SVM, we predicted the response with an AUC of 0.8, which is acceptable for a challenging task and better than many prior studies. Interestingly, in one recently published study, the SVM had a better accuracy of iNPH diagnosis in comparison to radiologists [ 29 ]. In our study, the patients' age and sex were not predictive (neither in statistical analysis nor in machine learning models), which matches results from previous studies [ 6 ].…”

Section: Discussionmentioning

confidence: 99%

The Role of Machine Learning and Radiomics for Treatment Response Prediction in Idiopathic Normal Pressure Hydrocephalus

Sotoudeh

Sadaatpour

Rezaei

et al. 2021

Cureus

View full text Add to dashboard Cite

IntroductionVentricular shunting remains the standard of care for patients with idiopathic normal pressure hydrocephalus (iNPH); however, not all patients benefit from the shunting. Prediction of response in advance can result in improved patient selection for ventricular shunting. This study aims to develop a machine learning predictive model for treatment response after shunt placement using the clinical and radiomics features. MethodsIn this retrospective pilot study, the medical records of iNPH patients who underwent ventricular shunting were evaluated. In each patient, the "idiopathic normal pressure hydrocephalus grading scale" (iNPHGS) and a "Modified Rankin Scale" were calculated before and after surgery. The subsequent treatment response was calculated as the difference between the iNPHGS scores before and after surgery. iNPHGS score reduction of two or more than two were considered as treatment response. The presurgical MRI scans were evaluated by radiologists, the ventricular systems were segmented on the T2-weighted images, and the radiomics features were extracted from the segmented ventricular system. Using Orange data mining open-source platform, different machine learning models were then developed based on the presurgical clinical features and the selected radiomics features to predict treatment response after shunt placement. ResultsAfter the implementation of the inclusion criteria, 78 patients were included in this study. One hundred twenty radiomics features were extracted, and the 12 best predictive radiomics features were selected. Using only clinical data (iNPHGS and Modified Rankin Scale), the random forest model achieved the best performance in treatment prediction with an area under the curve (AUC) of 0.71. Adding the Radiomics analysis to the clinical data improved the prediction performance, with the support vector machine (SVM) achieving the highest rank in treatment prediction with an AUC of 0.8. Adding age and sex to the analysis did not improve the prediction. ConclusionUsing machine learning models for treatment response prediction in patients with iNPH is feasible with acceptable accuracy. Adding the Radiomics analysis to the clinical features can further improve the predictive performance. SVM is likely the best model for this task.

show abstract

“…An important design choice was to break up the detection problem into segmentation, quantification of anatomical features, and classification based on these features. This is not a new approach (4,7,8,17,32), but it does have the benefit of straightforward interpretation of the results. More recently, the trend in the AI literature is to have a single deep network provide a final output for the likelihood of hydrocephalus without intermediate steps (16,39,40).…”

Section: Discussionmentioning

confidence: 99%

“…Previous machine learning efforts to diagnose hydrocephalus using MRI exams have compared NPH with healthy volunteers or NPH within specific patient populations ( e.g ., Alzheimer’s Disease, AD). In these specific populations and using small datasets (< 100 patients), accuracies of over 90% have been reported (4,7,8,16,17). However, these methods have not been tested in a broader clinical population with heterogeneous conditions typically observed in general neuroradiology practice.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning Achieves Neuroradiologist-Level Performance in Detecting Hydrocephalus Requiring Treatment

Huang

Moreno

Malani

et al. 2021

Preprint

View full text Add to dashboard Cite

PurposeWe aim to develop automated detection of hydrocephalus requiring treatment in a heterogeneous patient population referred for MRI brain scans, and compare performance to that of neuroradiologists.Materials and MethodsWe leveraged 496 clinical MRI brain scans (259 hydrocephalus) collected retrospectively at a single clinical site from patients aged 2–90 years (mean 54) referred for any reason. Sixteen MRI scans (ten hydrocephalus) were segmented semi-automatically in 3D to delineate ventricles, extraventricular CSF, and brain tissues. A 3D CNN was trained on these segmentations and subsequently used to automatically segment the remaining 480 scans. To detect hydrocephalus, volumetric features such as volumes of ventricles and temporal horns were computed from the segmentation and were used to train a linear classifier. Machine performance was evaluated in a diagnosis dataset where hydrocephalus was confirmed as requiring surgical intervention, and compared to four neuroradiologists on a random subset of 240 scans. The pipeline was tested on a separate screening dataset of 205 scans collected from a routine clinical population aged 1–95 years (mean 56) to predict the majority reading from four neuroradiologists using images alone.ResultsWhen compared to the neuroradiologists at a matched sensitivity, the machine did not show a significant difference in specificity (proportions test, p > 0.05). The machine demonstrated comparable performance in independent diagnosis and screening datasets. Overall ROC performance compared favorably with the state-of-the-art (AUC 0.82–0.93).ConclusionHydrocephalus can be detected automatically from MRI in a heterogeneous patient population with performance equivalent to that of neuroradiologists.Summary statementA two-stage automated pipeline was developed to segment head MRI and extract volumetric features to accurately and efficiently detect hydrocephalus that required shunting and achieved performance comparable to that of trained neuroradiologists.Key PointsWe developed a state-of-the-art 3D deep convolutional network to perform fully automated segmentation of the ventricles, extraventricular cerebrospinal fluid, and brain tissues in anisotropic MRI brain scans in a heterogeneous patient population.Volumetric features extracted from anatomical segmentations can be used to classify hydrocephalus (which may require neurosurgical intervention) vs. non-hydrocephalus.When tested in an independent dataset, the network achieved performance comparable to that of expert neuroradiologists.

show abstract

SVM-Based Normal Pressure Hydrocephalus Detection

Cited by 20 publications

References 25 publications

Increased interstitial fluid in periventricular and deep white matter hyperintensities in patients with suspected idiopathic normal pressure hydrocephalus

Increased interstitial fluid in periventricular and deep white matter hyperintensities in patients with suspected idiopathic normal pressure hydrocephalus

The Role of Machine Learning and Radiomics for Treatment Response Prediction in Idiopathic Normal Pressure Hydrocephalus

Deep Learning Achieves Neuroradiologist-Level Performance in Detecting Hydrocephalus Requiring Treatment

Contact Info

Product

Resources

About