Maria Berndt1, Nicole Lange2, Yu-Mi Ryang2, Bernhard Meyer2, Claus Zimmer3, Alexander Hapfelmeier4, Nina Wantia5, Jens Gempt2, Nina Lummel3. 1. Department of Neuroradiology, Technical University of Munich, Munich, Germany. Electronic address: maria.berndt@tum.de. 2. Department of Neurosurgery, Technical University of Munich, Munich, Germany. 3. Department of Neuroradiology, Technical University of Munich, Munich, Germany. 4. Institute of Medical Informatics, Statistics, and Epidemiology, Technical University of Munich, Munich, Germany. 5. Institute of Medical Microbiology, Immunology, and Hygiene, Technical University of Munich, Munich, Germany.
Abstract
BACKGROUND: Magnetic resonance imaging (MRI), including diffusion-weighted imaging (DWI), is an excellent tool for diagnosing intracranial infection, with limitations in previous neurosurgical intervention. This study aimed to evaluate the accuracy of DWI in the diagnosis of postoperative intracranial infection. METHODS: We retrospectively evaluated all consecutive patients with intracranial infection undergoing a neurosurgical intervention who had preoperative MRI, including DWI. Patients were divided into 2 groups: spontaneous intracranial infection (SI) and postoperative intracranial infection (PI). A control group (CG) of patients who had undergone brain surgery without any signs of subsequent infection was also included. Qualitatively, MRI data were evaluated for the presence of intracranial infection. Sensitivity, specificity, and positive and negative predictive values for PI as opposed to no infection were determined. Quantitatively, areas with diffusion restriction within the surgery/abscess cavity were identified for the 3 groups using semiautomated segmentation. Group differences regarding apparent diffusion coefficient (ADC) ratios were evaluated. Receiver operating characteristic curve analysis was used to identify a point in time beyond which ADC ratios might show reasonable discriminatory power between the PI and CG groups. RESULTS: A total of 78 patients were included (38 in the SI group, 20 in the PI group, 20 in the CG group). Sensitivity, specificity, and positive and negative predictive values in the diagnosis of PI were 80%, 95%, 4%, and 100%, respectively. Median ADC was significantly higher in the PI group compared with the SI group (0.98 vs. 0.69; P < 0.001) but lower compared with the CG group (1.24; P = 0.16). The analysis suggested a possible differentiation of PI and CG after a time interval of approximately 100 days/3 months. CONCLUSIONS: DWI is of limited value in evaluating postoperative brain infection. Our data show a trend toward DWI regaining its validity at approximately 3 months after surgery.
BACKGROUND: Magnetic resonance imaging (MRI), including diffusion-weighted imaging (DWI), is an excellent tool for diagnosing intracranial infection, with limitations in previous neurosurgical intervention. This study aimed to evaluate the accuracy of DWI in the diagnosis of postoperative intracranial infection. METHODS: We retrospectively evaluated all consecutive patients with intracranial infection undergoing a neurosurgical intervention who had preoperative MRI, including DWI. Patients were divided into 2 groups: spontaneous intracranial infection (SI) and postoperative intracranial infection (PI). A control group (CG) of patients who had undergone brain surgery without any signs of subsequent infection was also included. Qualitatively, MRI data were evaluated for the presence of intracranial infection. Sensitivity, specificity, and positive and negative predictive values for PI as opposed to no infection were determined. Quantitatively, areas with diffusion restriction within the surgery/abscess cavity were identified for the 3 groups using semiautomated segmentation. Group differences regarding apparent diffusion coefficient (ADC) ratios were evaluated. Receiver operating characteristic curve analysis was used to identify a point in time beyond which ADC ratios might show reasonable discriminatory power between the PI and CG groups. RESULTS: A total of 78 patients were included (38 in the SI group, 20 in the PI group, 20 in the CG group). Sensitivity, specificity, and positive and negative predictive values in the diagnosis of PI were 80%, 95%, 4%, and 100%, respectively. Median ADC was significantly higher in the PI group compared with the SI group (0.98 vs. 0.69; P < 0.001) but lower compared with the CG group (1.24; P = 0.16). The analysis suggested a possible differentiation of PI and CG after a time interval of approximately 100 days/3 months. CONCLUSIONS: DWI is of limited value in evaluating postoperative brain infection. Our data show a trend toward DWI regaining its validity at approximately 3 months after surgery.