Comparative sensitivity and specificity of change in third ventricular diameter or global ventricular change for...

Clinical features of shunt malfunction are often nonspecific and neuroimaging is a well-established aid in diagnosis. In practice, qualitative assessment of overall ventricular change guides management; however, change in third ventricular diameter (TVD) alone has been proposed as a low-complexity quantitative index of overall ventricular change. This study aimed to evaluate the diagnostic utility of change in TVD compared with global assessment of ventricular caliber change by a neurosurgeon in detecting shunt malfunction.

A retrospective review of all pediatric ventricular shunt revisions performed at a single center (November 2014-September 2024) was conducted. TVDs were measured when the patient was last known well and on preoperative imaging. Quantitative change was compared to the overall impression of change in ventricular caliber by neurosurgeons for detecting shunt malfunction. Shunt malfunction was defined as the need to replace one or more shunt components at the time of surgery (diagnostic gold standard). Diagnostic performance was assessed using comparison of areas under the receiver operating characteristic (ROC) curve.

A total of 422 shunt revisions were performed during the study period, of which 315 (75%) were found to have shunt malfunction. ROC analysis utilizing a cutoff of ≥ 2-mm TVD increase detected shunt malfunction with 63% sensitivity and 58% specificity (area under the curve [AUC] 0.62, 95% CI 0.55-0.69; p < 0.001). Pragmatic refinement by restricting the cohort to the subset of 136 patients known to have a change in ventricular caliber at the time of previous shunt failure and selecting a ≥ 1-mm cutoff increased the discriminative power of change in TVD to 92% sensitivity and 21% specificity. In this restricted cohort, the overall assessment of ventricular change by neurosurgeons yielded 81% sensitivity and 62% specificity (AUC 0.72, 95% CI 0.59-0.84) and the AUC difference between the two techniques was not statistically significant (AUC difference 0.08, 95% CI -0.03 to 0.20; p = 0.16).

In a pragmatically selected cohort, an increase ≥ 1 mm from baseline TVD alone provides sensitivity comparable to that of overall assessment of ventricular change by a neurosurgeon identifying shunt malfunction in a cohort undergoing shunt revision. This simple linear measurement could be integrated as a screening test in clinical, imaging, and/or AI algorithms to facilitate rapid recognition of pediatric shunt failure.