Systematic volumetric analysis predicts response to CSF drainage and outcome to shunt surgery in idiopathic normal pressure hydrocephalus

Abstract

Objectives: Idiopathic normal pressure hydrocephalus (INPH) is a neurodegenerative disorder characterized by excess cerebrospinal fluid (CSF) in the ventricles, which can be diagnosed by invasive CSF drainage test and treated by shunt placement. Here, we aim to investigate the diagnostic and prognostic power of systematic volumetric analysis based on brain structural MRI for INPH.

Methods: We performed a retrospective study with a cohort of 104 probable INPH patients who underwent CSF drainage tests and another cohort of 41 INPH patients who had shunt placement. High-resolution T1-weighted images of the patients were segmented using an automated pipeline into 283 structures that are grouped into different granularity levels for volumetric analysis. Volumes at multi-granularity levels were used in a recursive feature elimination model to classify CSF drainage responders and non-responders. We then used pre-surgical brain volumes to predict Tinetti and MMSE scores after shunting, based on the least absolute shrinkage and selection operator.

Results: The classification accuracy of differentiating the CSF drainage responders and non-responders increased as the granularity increased. The highest diagnostic accuracy was achieved at the finest segmentation with a sensitivity/specificity/precision/accuracy of 0.89/0.91/0.84/0.90 and an area under the curve of 0.94. The predicted post-surgical neurological scores showed high correlations with the ground truth, with r = 0.80 for Tinetti and r = 0.88 for MMSE. The anatomical features that played important roles in the diagnostic and prognostic tasks were also illustrated.

Conclusions: We demonstrated that volumetric analysis with fine segmentation could reliably differentiate CSF drainage responders from other INPH-like patients, and it could accurately predict the neurological outcomes after shunting.

Key points: • We performed a fully automated segmentation of brain MRI at multiple granularity levels for systematic volumetric analysis of idiopathic normal pressure hydrocephalus (INPH) patients. • We were able to differentiate patients that responded to CSF drainage test with an accuracy of 0.90 and area under the curve of 0.94 in a cohort of 104 probable INPH patients, as well as to predict the post-shunt gait and cognitive scores with a coefficient of 0.80 for Tinetti and 0.88 for MMSE. • Feature analysis showed the inferior lateral ventricle, bilateral hippocampus, and orbital cortex are positive indicators of CSF drainage responders, whereas the posterior deep white matter and parietal subcortical white matter were negative predictors.

Keywords: Algorithm; Hydrocephalus; Normal pressure; Segmentation; Volume.

Fig. 1

Multi-atlas-based segmentation of the brain of an INPH patient. Structural labels are shown at different granularity levels from level 1 (7 labels) to level 5 (283 labels), in transverse and coronal views

Fig. 2

Classification of the CSF drainage responders and non-responders. a Classification accuracy at different levels of segmentation with varying number of features. The solid arrows point to the optimal number of features on each curve, and the hollow purple arrow points to a local peak at level 5. b Receiver operation characteristic (ROC) of the classifiers with the optimal sets of features at granularity levels 1–5. ROC of Evan’s index is also compared. The area under the curve (AUC) values are denoted correspondingly

Fig. 3

Weight maps of the brain regions that contributed to the classification of the responders and non-responders to CSF drainage at level 4 (a) and level 5 (b). The color map indicates the weights of selected regions in the RFE model. Abbreviations: post-DPWM, posterior deep white matter; parietal WM, parietal white matter; SFG, superior frontal gyrus; IFG, inferior frontal gyrus; SPG, superior parietal gyrus; Fx/ST, fornix/stria terminalis; HP, hippocampus; inferior LV, inferior lateral ventricle; OG, orbital gyrus

Fig. 4

Prediction of Tinetti and MMSE scores after shunting in INPH patients. a Correlation between the pre-shunt and post-shunt Tinetti scores in 47 INPH patients. b Correlation between the model predicted and clinically measured post-shunt Tinetti scores. c Weights of ROIs selected by the LASSO model to predict the post-surgical Tinetti scores. d Correlation between the pre-shunt and post-shunt MMSE scores in 47 INPH patients. e Correlation between the predicted and measured post-shunt MMSE scores. f Weights of ROIs selected by LASSO to predict the post-surgical MMSE scores. Abbreviations: MOG, middle occipital gyrus; CGC, Cingulum (cingulate gyrus part); DPWM, deep WM; PLIC, posterior limb of internal capsule; LV, lateral ventricle; ITG, inferior temporal gyrus; Fx/ST, fornix/stria terminalis