Search Results

You are looking at 1 - 6 of 6 items for

  • Author or Editor: Nicolai Maldaner x
Clear All Modify Search
Free access

Victor E. Staartjes, Carlo Serra, Giovanni Muscas, Nicolai Maldaner, Kevin Akeret, Christiaan H. B. van Niftrik, Jorn Fierstra, David Holzmann and Luca Regli

OBJECTIVE

Gross-total resection (GTR) is often the primary surgical goal in transsphenoidal surgery for pituitary adenoma. Existing classifications are effective at predicting GTR but are often hampered by limited discriminatory ability in moderate cases and by poor interrater agreement. Deep learning, a subset of machine learning, has recently established itself as highly effective in forecasting medical outcomes. In this pilot study, the authors aimed to evaluate the utility of using deep learning to predict GTR after transsphenoidal surgery for pituitary adenoma.

METHODS

Data from a prospective registry were used. The authors trained a deep neural network to predict GTR from 16 preoperatively available radiological and procedural variables. Class imbalance adjustment, cross-validation, and random dropout were applied to prevent overfitting and ensure robustness of the predictive model. The authors subsequently compared the deep learning model to a conventional logistic regression model and to the Knosp classification as a gold standard.

RESULTS

Overall, 140 patients who underwent endoscopic transsphenoidal surgery were included. GTR was achieved in 95 patients (68%), with a mean extent of resection of 96.8% ± 10.6%. Intraoperative high-field MRI was used in 116 (83%) procedures. The deep learning model achieved excellent area under the curve (AUC; 0.96), accuracy (91%), sensitivity (94%), and specificity (89%). This represents an improvement in comparison with the Knosp classification (AUC: 0.87, accuracy: 81%, sensitivity: 92%, specificity: 70%) and a statistically significant improvement in comparison with logistic regression (AUC: 0.86, accuracy: 82%, sensitivity: 81%, specificity: 83%) (all p < 0.001).

CONCLUSIONS

In this pilot study, the authors demonstrated the utility of applying deep learning to preoperatively predict the likelihood of GTR with excellent performance. Further training and validation in a prospective multicentric cohort will enable the development of an easy-to-use interface for use in clinical practice.

Free access

Michel Roethlisberger, Lara Gut, Daniel Walter Zumofen, Urs Fisch, Oliver Boss, Nicolai Maldaner, Davide Marco Croci, Ethan Taub, Natascia Corti, Jan-Karl Burkhardt, Raphael Guzman, Oliver Bozinov and Luigi Mariani

OBJECTIVE

Women taking combined hormonal contraceptives (CHCs) are generally considered to be at low risk for cerebral venous thrombosis (CVT). When it does occur, however, intensive care and neurosurgical management may, in rare cases, be needed for the control of elevated intracranial pressure (ICP). The use of nonsurgical strategies such as barbiturate coma and induced hypothermia has never been reported in this context. The objective of this study is to determine predictive factors for invasive or surgical ICP treatment and the potential complications of nonsurgical strategies in this population.

METHODS

The authors conducted a 2-center, retrospective chart review of 168 cases of CVT in women between 2000 and 2012. Eligible patients were classified as having had a mild or a severe clinical course, the latter category including all patients who underwent invasive or surgical ICP treatment and all who had an unfavorable outcome (modified Rankin Scale score ≥ 3 or Glasgow Outcome Scale score ≤ 3). The Mann-Whitney U-test was used for continuous parameters and Fisher’s exact test for categorical parameters, and odds ratios were calculated with statistical significance set at p ≤ 0.05.

RESULTS

Of the 168 patients, 57 (age range 16–49 years) were determined to be eligible for the study. Six patients (10.5%) required invasive or surgical ICP treatment. Three patients (5.3%) developed refractory ICP > 30 mm Hg despite early surgical decompression; 2 of them (3.5%) were treated with barbiturate coma and induced hypothermia, with documented infectious, thromboembolic, and hemorrhagic complications. Coma on admission, thrombosis of the deep venous system with consecutive hydrocephalus, intraventricular hemorrhage, and hemorrhagic venous infarction were associated with a higher frequency of surgical intervention. Coma, quadriparesis on admission, and hydrocephalus were more commonly seen among women with unfavorable outcomes. Thrombosis of the transverse sinus was less common in patients with an unfavorable outcome, with similar distribution in patients needing invasive or surgical ICP treatment.

CONCLUSIONS

The need for invasive or surgical ICP treatment in women taking CHCs who have CVT is partly predictable on the basis of the clinical and radiological findings on admission. The use of nonsurgical treatments for refractory ICP, such as barbiturate coma and induced hypothermia, is associated with systemic infectious and hematological complications and may worsen morbidity in this patient population. The significance of these factors should be studied in larger multicenter cohorts.

Free access

Martin N. Stienen, Nicolai Maldaner, Holger Joswig, Marco V. Corniola, David Bellut, Peter Prömmel, Luca Regli, Astrid Weyerbrock, Karl Schaller and Oliver P. Gautschi

OBJECTIVE

Patient-reported outcome measures (PROMs) are standard of care for the assessment of functional impairment. Subjective outcome measures are increasingly complemented by objective ones, such as the “Timed Up and Go” (TUG) test. Currently, only a few studies report pre- and postoperative TUG test assessments in patients with lumbar spinal stenosis (LSS).

METHODS

A prospective two-center database was reviewed to identify patients with LSS who underwent lumbar decompression with or without fusion. The subjective functional status was estimated using PROMs for pain (visual analog scale [VAS]), disability (Roland-Morris Disability Index [RMDI] and Oswestry Disability Index [ODI]), and health-related quality of life (HRQoL; 12-Item Short-Form Physical Component Summary [SF-12 PCS] and the EQ-5D) preoperatively, as well as on postoperative day 3 (D3) and week 6 (W6). Objective functional impairment (OFI) was measured using age- and sex-standardized TUG test results.

RESULTS

Sixty-four patients (n = 32 [50%] male, mean age 66.8 ± 11.7 years) were included. Preoperatively, they reported a mean VAS back pain score of 4.1 ± 2.7, VAS leg pain score of 5.4 ± 2.7, RMDI of 10.4 ± 5.3, ODI of 41.9 ± 16.2, SF-12 PCS score of 32.7 ± 8.3, and an EQ-5D index of 0.517 ± 0.226. The preoperative rates of severe, moderate, and mild OFI were 4.7% (n = 3), 12.5% (n = 8), and 7.8% (n = 5), respectively, and the mean OFI T-score was 116.3 ± 23.7. At W6, 60 (93.8%) of 64 patients had a TUG test result within the normal population range (no OFI); 3 patients (4.7%) had mild and 1 patient (1.6%) severe OFI. The mean W6 OFI T-score was significantly decreased (103.1 ± 13.6; p < 0.001). Correspondingly, the PROMs showed a decrease in subjective VAS back pain (1.6 ± 1.7, p < 0.001) and leg pain (1.0 ± 1.8, p < 0.001) scores, disability (RMDI 5.3 ± 4.7, p < 0.001; ODI 21.3 ± 16.1, p < 0.001), and increase in HRQoL (SF-12 PCS 40.1 ± 8.3, p < 0.001; EQ-5D 0.737 ± 0.192, p < 0.001) at W6. The W6 responder status (clinically meaningful improvement) ranged between 81.3% (VAS leg pain) and 29.7% (EQ-5D index) of patients.

CONCLUSIONS

The TUG test is a quick and easily applicable tool that reliably measures OFI in patients with LSS. Objective tests incorporating longer walking time should be considered if OFI is suspected but fails to be proven by the TUG test, taking into account that neurogenic claudication may not clinically manifest during the brief TUG examination. Objective tests do not replace the subjective PROM-based assessment, but add valuable information to a comprehensive patient evaluation.

Restricted access

Victor E. Staartjes, Costanza M. Zattra, Kevin Akeret, Nicolai Maldaner, Giovanni Muscas, Christiaan Hendrik Bas van Niftrik, Jorn Fierstra, Luca Regli and Carlo Serra

OBJECTIVE

Although rates of postoperative morbidity and mortality have become relatively low in patients undergoing transnasal transsphenoidal surgery (TSS) for pituitary adenoma, cerebrospinal fluid (CSF) fistulas remain a major driver of postoperative morbidity. Persistent CSF fistulas harbor the potential for headache and meningitis. The aim of this study was to investigate whether neural network–based models can reliably identify patients at high risk for intraoperative CSF leakage.

METHODS

From a prospective registry, patients who underwent endoscopic TSS for pituitary adenoma were identified. Risk factors for intraoperative CSF leaks were identified using conventional statistical methods. Subsequently, the authors built a prediction model for intraoperative CSF leaks based on deep learning.

RESULTS

Intraoperative CSF leaks occurred in 45 (29%) of 154 patients. No risk factors for CSF leaks were identified using conventional statistical methods. The deep neural network–based prediction model classified 88% of patients in the test set correctly, with an area under the curve of 0.84. Sensitivity (83%) and specificity (89%) were high. The positive predictive value was 71%, negative predictive value was 94%, and F1 score was 0.77. High suprasellar Hardy grade, prior surgery, and older age contributed most to the predictions.

CONCLUSIONS

The authors trained and internally validated a robust deep neural network–based prediction model that identifies patients at high risk for intraoperative CSF. Machine learning algorithms may predict outcomes and adverse events that were previously nearly unpredictable, thus enabling safer and improved patient care and better patient counseling.

Restricted access

Pavlina Lenga, Christian Hohaus, Bujung Hong, Adisa Kursumovic, Nicolai Maldaner, Jan-Karl Burkhardt, Philippe Bijlenga, Daniel A. Rüfenacht, Nils O. Schmidt, Peter Vajkoczy and Julius Dengler

OBJECTIVE

Giant posterior circulation aneurysms (GPCirAs) usually cause substantial mass effect on the brainstem, which may lead to neurological deficits. So far, there has been no systematic investigation of factors associated with such deficits in GPCirA. The authors aim to examine the risk factors for cranial nerve deficit (CND), motor deficit, and disability in patients with GPCirA.

METHODS

Using MR images obtained in 30 patients with unruptured GPCirA, the authors examined GPCirA volume, presence of hydrocephalus or partial thrombosis (PT) of the aneurysm, and the degree of brainstem displacement measured by the distance between the McRae line and the tip of the GPCirA (∆MT). They evaluated associations between these factors and neurological deficits.

RESULTS

Thirty GPCirAs in 30 patients were included. The prevalence of CNDs was 50%. Patients with CNDs significantly differed from those without CNDs in terms of age (mean 51.0 years [SD 15.0 years] vs 69.0 years [SD 21.0 years], p = 0.01) and in ∆MT (median 50.7 mm [IQR 39.2–53.9 mm] vs 39.0 mm [IQR 32.3–45.9 mm], p = 0.02). The prevalence of motor deficits was 33.3%. Patients with motor deficits showed a larger ∆MT (median 50.5 mm [IQR 40.8–54.6 mm]) compared with those without (∆MT: median 39.1 mm [IQR 32.8–50.5 mm], p = 0.04). GPCirA volume was larger in patients with poor modified Rankin Scale (mRS) scores (median 14.9 cm3 [IQR 8.6–18.7 cm3]) than in those with mRS scores of 0–2 (median 6.8 cm3 [IQR 4.4–11.7 cm3], p = 0.03). After adjusting for patient age and the occurrence of hydrocephalus or PT, the authors found that higher degrees of disability were significantly associated with aneurysm volume (OR 1.13, 95% CI 1.0–1.3; p = 0.04), but not with ∆MT. The occurrence of CND or motor deficit was not associated with any of the examined variables. There was no correlation between GPCirA volume and ∆MT (rs = 0.01, p = 0.96). The prevalence of neurological deficits did not differ between GPCirA at the basilar apex, the basilar trunk, the vertebrobasilar junction, or the vertebral artery.

CONCLUSIONS

In this study, the neurological condition of the patients was associated only with GPCirA volume and not with the degree of brainstem displacement, the occurrence of PT or hydrocephalus, or the exact location of the GPCirA. These findings highlight the clinical relevance of GPCirA volume and suggest that factors such as brainstem displacement or PT should play less of a role when finding arguments for or against treatment of GPCirA.

Clinical trial registration no.: NCT02066493 (clinicaltrials.gov)

Full access

Julius Dengler, Nicolai Maldaner, Philippe Bijlenga, Jan-Karl Burkhardt, Alexander Graewe, Susanne Guhl, Bujung Hong, Christian Hohaus, Adisa Kursumovic, Dorothee Mielke, Karl-Michael Schebesch, Maria Wostrack, Daniel Rufenacht, Peter Vajkoczy, Nils Ole Schmidt and Giant Intracranial Aneurysm Study Group

OBJECT

The underlying mechanisms causing intracranial perianeurysmal edema (PAE) are still poorly understood. Since PAE is most frequently observed in giant intracranial aneurysms (GIAs), the authors designed a study to examine the occurrence of PAE in relation to the location, size, and partial thrombosis (PT) of GIAs along with the clinical impact of PAE.

METHODS

Magnetic resonance imaging data for patients with a diagnosis of unruptured GIA from the international multicenter Giant Intracranial Aneurysm Registry were retrospectively analyzed with regard to location and size of the GIA, PAE volume, and the presence of PT. The occurrence of PAE was correlated to clinical findings.

RESULTS

Imaging data for 69 GIAs were eligible for inclusion in this study. Perianeurysmal edema was observed in 33.3% of all cases, with the highest frequency in GIAs of the middle cerebral artery (MCA; 68.8%) and the lowest frequency in GIAs of the cavernous internal carotid artery (ICA; 0.0%). Independent predictors of PAE formation were GIA volume (OR 1.13, p = 0.02) and the occurrence of PT (OR 9.84, p = 0.04). Giant intracranial aneurysm location did not predict PAE occurrence. Giant aneurysms with PAE were larger than GIAs without PAE (p < 0.01), and GIA volume correlated with PAE volume (rs = 0.51, p = 0.01). Perianeurysmal edema had no influence on the modified Rankin Scale score (p = 0.30 or the occurrence of aphasia (p = 0.61) or hemiparesis (p = 0.82).

CONCLUSIONS

Perianeurysmal edema was associated with GIA size and the presence of PT. As no PAE was observed in cavernous ICA aneurysms, even though they exerted mass effect on the brain and also displayed PT, the dura mater may serve as a barrier protecting the brain from PAE formation.