Diagnostic accuracy of an artificial intelligence algorithm versus radiologists for fracture detection on cervical spine CT.

OBJECTIVES: To compare diagnostic accuracy of a deep learning artificial intelligence (AI) for cervical spine (C-spine) fracture detection on CT to attending radiologists and assess which undetected fractures were injuries in need of stabilising therapy (IST). METHODS: This single-centre, retrospective diagnostic accuracy study included consecutive patients (age ≥18 years; 2007-2014) screened for C-spine fractures with CT. To validate ground truth, one radiologist and three neurosurgeons independently examined scans positive for fracture. Negative scans were followed up until 2022 through patient files and two radiologists reviewed negative scans that were flagged positive by AI. The neurosurgeons determined which fractures were ISTs. Diagnostic accuracy of AI and attending radiologists (index tests) were compared using McNemar. RESULTS: Of the 2368 scans (median age, 48, interquartile range 30-65; 1441 men) analysed, 221 (9.3%) scans contained C-spine fractures with 133 IST. AI detected 158/221 scans with fractures (sensitivity 71.5%, 95% CI 65.5-77.4%) and 2118/2147 scans without fractures (specificity 98.6%, 95% CI 98.2-99.1). In comparison, attending radiologists detected 195/221 scans with fractures (sensitivity 88.2%, 95% CI 84.0-92.5%, p < 0.001) and 2130/2147 scans without fracture (specificity 99.2%, 95% CI 98.8-99.6, p = 0.07). Of the fractures undetected by AI 30/63 were ISTs versus 4/26 for radiologists. AI detected 22/26 fractures undetected by the radiologists, including 3/4 undetected ISTs. CONCLUSION: Compared to attending radiologists, the artificial intelligence has a lower sensitivity and a higher miss rate of fractures in need of stabilising therapy; however, it detected most fractures undetected by the radiologists, including fractures in need of stabilising therapy. Clinical relevance statement The artificial intelligence algorithm missed more cervical spine fractures on CT than attending radiologists, but detected 84.6% of fractures undetected by radiologists, including fractures in need of stabilising therapy. KEY POINTS: The impact of artificial intelligence for cervical spine fracture detection on CT on fracture management is unknown. The algorithm detected less fractures than attending radiologists, but detected most fractures undetected by the radiologists including almost all in need of stabilising therapy. The artificial intelligence algorithm shows potential as a concurrent reader.

Bone visualization of the cervical spine with deep learning-based synthetic CT compared to conventional CT: A single-center noninferiority study on image quality.

PURPOSE: To investigate whether the image quality of a specific deep learning-based synthetic CT (sCT) of the cervical spine is noninferior to conventional CT. METHOD: Paired MRI and CT data were collected from 25 consecutive participants (≥ 50 years) with cervical radiculopathy. The MRI exam included a T1-weighted multiple gradient echo sequence for sCT reconstruction. For qualitative image assessment, four structures at two vertebral levels were evaluated on sCT and compared with CT by three assessors using a four-point scale (range 1-4). The noninferiority margin was set at 0.5 point on this scale. Additionally, acceptable image quality was defined as a score of 3-4 in ≥ 80% of the scans. Quantitative assessment included geometrical analysis and voxelwise comparisons. RESULTS: Qualitative image assessment showed that sCT was noninferior to CT for overall bone image quality, artifacts, imaging of intervertebral joints and neural foramina at levels C3-C4 and C6-C7, and cortical delineation at C6-C7. Noninferiority was weak to absent for cortical delineation at level C3-C4 and trabecular bone at both levels. Acceptable image quality was achieved for all structures in sCT and CT, except for trabecular bone in sCT and level C6-C7 in CT. Geometrical analysis of the sCT showed good to excellent agreement with CT. Voxelwise comparisons showed a mean absolute error of 80.05 (±6.12) HU, dice similarity coefficient (cortical bone) of 0.84 (±0.04) and structural similarity index of 0.86 (±0.02). CONCLUSIONS: This deep learning-based sCT was noninferior to conventional CT for the general visualization of bony structures of the cervical spine, artifacts, and most detailed structure assessments.

Comparison of the Minimally Invasive Tubular Transmuscular Approach with the Conventional Microsurgical Approach for Microsurgical Treatment of Lumbar Disk Herniation: A Prospective Randomized Study.

BACKGROUND: The aim of this study is to compare the outcome of the minimally invasive transmuscular approach using a tubular retractor system (Metrx) with the conventional microsurgical standard approach (CM) for microsurgical treatment of lumbar disk herniation. METHODS: This is a prospective randomized controlled study with a 1:1 distribution of patients in CM and Metrx study groups. Two hundred and twenty-seven (117 CM and 110 Metrx) patients were included. The primary outcome parameters are postoperative pain intensity reduction, length of hospitalization, postoperative quality of life, and daily life performance based on the standardized questionnaires: Visual Analog Scale (VAS), 36-Item Short Form Survey (SF-36), Oswestry Disability Index (ODI), and Prolo scores. The secondary outcome parameters are intraoperative variables: surgery duration, blood loss, and fluoroscopy dose. RESULTS: There were no significant statistical differences in the primary outcome measures between the two groups with respect to postoperative pain relief (median VAS pre-op to 3 months post-op for sciatica: 9-2 [CM] vs. 8-2 [Metrx]; for lumbago: 7-2.5 [CM] vs. 6-3 [Metrx]), the length of hospitalization (median of 5 days), or the frequency of occupational reintegration after 3 months (59.1 vs. 60.7%). CONCLUSION: The microsurgical therapy of lumbar disk herniation via a Metrx approach is a safe and effective treatment option and is equivalent to the CM approach.

Vertebral artery injuries following cervical spine trauma: a prospective observational study.

PURPOSE: The purpose of this study was to report on the incidence, diagnosis and clinical manifestation of VAI following cervical spine injuries observed in a prospective observational study with a standardized clinical and radiographical protocol. METHODS: During a 16-year period, 69 (mean age: 43 ± 20.7 years; 25 female, 44 male) of 599 patients had cervical spine injury suspicious for VAI due to facet luxation and/or fractures extending into the transverse foramen. Diagnosis and management of these patients followed a previously published protocol (Kral in Zentralbl Neurochir 63:153-158, 2002). Digital subtraction angiography (DSA) was performed in all 69 patients. Injury grading of VAI was done according to Biffl et al. (Ann Surg 231:672-681, 2000). All patients with VAI were treated with anticoagulation (heparin followed by ASS) for 6 months. RESULTS: In cases suspicious for VAI, the incidence of VAI detected by DSA was 27.5% (n = 19 of 69 patients). VAI Grade I occurred in 15.8%, Grade II in 26.3%, Grade IV in 52.6% and Grade V in 5.2%. Of 19 patients, 4 (21%) had clinical signs of vertebrobasilar ischemia. Two patients died in hospital after 4 and 21 days respectively. Of 69 patients, 33 (47.8%) with suspected VAI had unstable spine injuries and were treated surgically. CONCLUSION: In patients with cervical spine fractures or dislocations crossing the course of the vertebral artery, VAI are relatively frequent and may be associated with significant morbidity and mortality. VAI were identified by DSA in 27.5%. Despite anticoagulation therapy, 5.8% became clinically symptomatic and 2.9% died due to cerebrovascular ischemia.

Accuracy and complications of transpedicular C2 screw placement without the use of spinal navigation.

The objective of the study was to describe the technique, accuracy of placement and complications of transpedicular C2 screw fixation without spinal navigation. Patients treated by C2 pedicle screw fixations were identified from the surgical log book of the department. Clinical data were extracted retrospectively from the patients' charts. Pedicle screw placement accuracy was assessed on postoperative CT scans according to Gertzbein and Robbins (GRGr). A total of 27 patients were included in the study. The mean age of the patients was 56 +/- 22.0 years; 51.9% of them were female. As much as 17 patients suffered from trauma, 5 of degenerative disease, 3 of inflammations and 2 of metastatic disease. A total of 47 C2 transpedicular screw fixations were performed. The canulated screws were inserted under visual control following the preparation of the superior surface of the isthmus and of the medial surface of the pedicles of the C2. Intraoperative fluoroscopy was additionally used. The postoperative CT findings showed in 55.3% GRGr 1, in 27.7% GRGr 2, in 10.6% GRGr 3, and in 6.3% GRGr 4 pedicle screw insertion accuracy. Screw insertions GRGr 5 were not observed. Screw malpositioning (i.e., GRGr 3 and 4) was significantly associated with thin (<5 mm) pedicle diameters and with surgery for C2 fractures. In the three patients with screw insertions GRGr 4, postoperative angiographies were performed to exclude vertebral artery affections. In one of these three cases, the screw caused a clinically asymptomatic vertebral artery compression. Hardware failures did not occur. In one patient, postoperative pneumonia resulted in the death of the patient. Careful patient selection and surgical technique is necessary to avoid vertebral artery injury in C2 pedicle screw fixation without spinal navigation. A slight opening of the vertebral artery canal (Gertzbein and Robbins grade < or =3) does not seem to put the artery at risk. However, the high rate of misplaced screws when inserted without spinal navigation, despite the fact that no neurovascular injury occurred, supports the use of spinal navigation in C2 pedicle screw insertions.

Surgical treatment of parietal lobe epilepsy.

OBJECT: Parietal lobe epilepsy (PLE) accounts for a small percentage of extratemporal epilepsies, and only a few and mostly smaller series have been reported. Preoperative findings, surgical strategies, pathological bases, and postoperative outcomes for PLE remain to be elucidated. METHODS: Patients with PLE were identified by screening a prospective epilepsy surgery database established in 1989 at the University of Bonn. Charts, preoperative imaging studies, surgical reports, and neuropathological findings were reviewed. Seizure outcome was classified according to Engel class (I-IV). RESULTS: Forty patients (23 females and 17 males) with PLE were identified and had a mean age of 25.0 years and a mean preoperative epilepsy duration of 13.7 years. Nine patients had a significant medical history (for example, trauma, meningitis/encephalitis, or perinatal hypoxia). Preoperative MR imaging abnormalities were identified in 38 (95%) of 40 patients; 26 patients (65%) underwent invasive electroencephalography evaluation. After lesionectomy of the dominant (in 20 patients) or nondominant (in 20 patients) parietal lobe and additional multiple subpial transections (in 11 patients), 2 patients suffered from surgical and 12 from neurological complications, including temporary partial Gerstmann syndrome. There were no deaths. Histopathological analysis revealed 16 low-grade tumors, 11 cortical dysplasias, 9 gliotic scars, 2 cavernous vascular malformations, and 1 granulomatous inflammation. In 1 case, no histopathological diagnosis could be made. After a mean follow-up of 45 months, 27 patients (67.5%) became seizure free or had rare seizures (57.5% Engel Class I; 10% Engel Class II; 27.5% Engel Class III; and 5% Engel Class IV). CONCLUSIONS: Parietal lobe epilepsy is an infrequent cause of extratemporal epilepsy. Satisfactory results (Engel Classes I and II) were obtained in 67.5% of patients in our series. A temporary partial hemisensory or Gerstmann syndrome occurs in a significant number of patients.

Correlation of MRI and histopathology in epileptogenic parietal and occipital lobe lesions.

INTRODUCTION: To analyze the diagnostic accuracy of MRI in patients undergoing parietal and occipital lobe epilepsy surgery. METHODS: In a retrospective study, we analyzed MRI scans and neuropathology reports of 42 patients who had undergone resective epilepsy surgery in the parietal and occipital lobe between 1998 and 2003. We evaluated, whether lesions were precisely characterized by MRI and whether lesion characterization allowed to estimate postsurgical seizure outcome. RESULTS: Within the categories epilepsy associated tumors, focal cortical dysplasias, vascular malformations, scarring, and others, MRI was concordant with histopathology in 36 of 42 (86%) lesions. Among the discordant lesions, one lesion was re-classified following MRI-histopathology synopsis, another two lesions represented new tumor entities (angiocentric neuroepithelial tumor, isomorphic astrocytoma) which have been described recently. Seizure freedom (Engel class I) one year following surgery was achieved in 25 patients (60%). Seizure outcome was different for lesion categories (Engel class I: epilepsy associated tumors, 62%; focal cortical dysplasias, 71%; vascular malformations, 75%; scarring, 40%), and was unchanged if no lesion was found on preoperative MRI. CONCLUSION: If MRI and histopathology are discordant, not only the MRI findings may be debatable. MRI lesion detection is important, since chance of seizure freedom is low if no lesion is detected.