Multilayer Diagnostic Battery for Normal Pressure Hydrocephalus

 

The proposed study addresses the open issues of the normal pressure hydrocephalus (NPH) management. NPH is a clinical syndrome characteristic with the classical Hakim’s symptoms triad consisting of cognitive deficit, gait impairment and urinary incontinence combined with ventriculomegaly, but without signs of obstruction of cerebrospinal fluid (CSF) flow or raised intracranial pressure (ICP). The main problem for clinicians is the poor differential diagnostics among neurodegenerative diseases, yet fundamental for adequate selection of active treatment candidates – the most used therapeutic approach for NPH is the surgical insertion of a ventriculoperitoneal shunt, a system draining excess CSF from the lateral ventricles of the brain to the peritoneal cavity. At the point when NPH is properly diagnosed, between 70–90% of shunted patients improve in their clinical status in contrast to other, often untreatable forms of neurodegenerative disorders.

 

Our work includes devising and implementing a reliable and robust multidisciplinary battery in order to advance the up-to-date diagnostic performance, as a result aiming to improve the treatment efficacy and overall patient prognosis. In contrast to the tightly focused and highly specific diagnostic and prognostic models available so far, our comprehensive battery is based on a carefully chosen set of predictor markers derived from various medical research domains including 1) mathematical investigation of the intracranial pressure waveform morphology and non-linear quasi-chaotic behaviour, 2) endocrinological evaluation of the derived cerebrospinal fluid analyte levels, 3) modern neuroimaging examination, and 4) neurological/neuropsychological consideration of the patients' clinical symptoms and their mental state. The acquired are statistically processed and classified using state-of-the-art intensive computational techniques including artificial intelligence neural networks and deep-learning convolution networks.

 

The construction of the classification algorithms is followed by an exhaustive validation and implementation as an available and easy-to-use software code designed for real-time application in clinical practice. Simultaneously, we are identifing and interpreting the NPH-specific classification patterns, thus shedding light on the CSF dynamics and concomitant ICP fluctuations at the very basic level – overall contributing to a better understanding of the NPH.