Vishal Sarwahi1, Terry Amaral2, Stephen Wendolowski2, Rachel Gecelter2, Etan Sugarman3, Yungtai Lo4, Dan Wang4, Beverly Thornhill5. 1. Billie and George Ross Center for Advanced Pediatric Orthopaedics and Minimally Invasive Spinal Surgery, Cohen Children's Medical Center, 7 Vermont Drive, Lake Success, NY 11042, USA. Electronic address: vsarwahi@yahoo.com. 2. Billie and George Ross Center for Advanced Pediatric Orthopaedics and Minimally Invasive Spinal Surgery, Cohen Children's Medical Center, 7 Vermont Drive, Lake Success, NY 11042, USA. 3. Department of Orthopaedic Surgery, Montefiore Medical Center, 111 East 210th Street Bronx, NY, 10467, USA. 4. Department of Epidemiology & Population Health, Albert Einstein College of Medicine, 1300 Morris Park Avenue Bronx, NY 10461, USA. 5. Department of Radiology, Montefiore Medical Center, 111 East 210th Street Bronx, NY, 10467, USA.
Abstract
STUDY DESIGN: Retrospective review of magnetic resonance imaging (MRI) and computed tomographic (CT) scan imaging modalities. OBJECTIVE: To determine MRI's capability of identifying pedicle morphology. SUMMARY OF BACKGROUND DATA: Understanding pedicle morphology is important for accurate placement of pedicle screws. The gold standard modality to assess pedicle morphology is CT scan. However, CT scans carry the risk of radiation exposure. We have studied MRI as a potential alternative to CT scan. METHODS: Nine hundred seventy pedicles in 33 spinal deformity patients were reviewed. Pedicle morphology was classified as follows: Type A (normal pedicle): >4-mm cancellous channel; Type B: 2-4-mm channel; Type C: any size cortical channel; and Type D: <2-mm cortical or cancellous channel. Pedicles in the same patients were classified on both low-dose CT scan and MRI. Concordance and discordance rates of MRI relative to CT scan in classification of pedicles into types A, B, C, and D were calculated for the entire length of the thoracolumbar spine and subgrouped into spinal sections. All images were evaluated by a single fellowship-trained musculoskeletal radiologist. RESULTS: CT scan had 809 Type A, 126 Type B, 29 Type C, and 6 Type D pedicles. Group II (MRI) had 735 Type A, 203 Type B, 30 Type C, and 2 Type D pedicles. Analysis of the entire spinal column showed a concordance rate of 86.7% in classification of the pedicles into the 4 types. In the upper thoracic region, the concordance rate was 77.1%, main thoracic 85.5%, thoracolumbar 96%, and lumbar 98.1%. MRI has a poor overall accuracy for detecting Type C pedicles, only a 44.8% concordance with CT scan. MRI overcalls Type B pedicles, often calling Type A pedicles Type B. CONCLUSIONS: MRI is an inferior alternative to CT scan as it has poor accuracy to properly detect pedicle abnormalities. The more severe the pedicle abnormality, the less diagnostic value the MRI has. LEVEL OF EVIDENCE: Level III, diagnostic.
STUDY DESIGN: Retrospective review of magnetic resonance imaging (MRI) and computed tomographic (CT) scan imaging modalities. OBJECTIVE: To determine MRI's capability of identifying pedicle morphology. SUMMARY OF BACKGROUND DATA: Understanding pedicle morphology is important for accurate placement of pedicle screws. The gold standard modality to assess pedicle morphology is CT scan. However, CT scans carry the risk of radiation exposure. We have studied MRI as a potential alternative to CT scan. METHODS: Nine hundred seventy pedicles in 33 spinal deformitypatients were reviewed. Pedicle morphology was classified as follows: Type A (normal pedicle): >4-mm cancellous channel; Type B: 2-4-mm channel; Type C: any size cortical channel; and Type D: <2-mm cortical or cancellous channel. Pedicles in the same patients were classified on both low-dose CT scan and MRI. Concordance and discordance rates of MRI relative to CT scan in classification of pedicles into types A, B, C, and D were calculated for the entire length of the thoracolumbar spine and subgrouped into spinal sections. All images were evaluated by a single fellowship-trained musculoskeletal radiologist. RESULTS: CT scan had 809 Type A, 126 Type B, 29 Type C, and 6 Type D pedicles. Group II (MRI) had 735 Type A, 203 Type B, 30 Type C, and 2 Type D pedicles. Analysis of the entire spinal column showed a concordance rate of 86.7% in classification of the pedicles into the 4 types. In the upper thoracic region, the concordance rate was 77.1%, main thoracic 85.5%, thoracolumbar 96%, and lumbar 98.1%. MRI has a poor overall accuracy for detecting Type C pedicles, only a 44.8% concordance with CT scan. MRI overcalls Type B pedicles, often calling Type A pedicles Type B. CONCLUSIONS: MRI is an inferior alternative to CT scan as it has poor accuracy to properly detect pedicle abnormalities. The more severe the pedicle abnormality, the less diagnostic value the MRI has. LEVEL OF EVIDENCE: Level III, diagnostic.