17. Unravelling the alignment of the cervical spine: a new cervical classification system based on the MultiEthnic Alignment Normative Study

• Published in: The spine journal Vol. 24; no. 9; p. S9
• Main Authors: Miller, Roy, Reyes, Justin, Dionne, Alexandra, Coury, Josephine Rose, Malka, Matan, Huec, Jean-Charles Le, Bourret, Stéphane, Hasegawa, Kazuhiro, Wong, Hee-Kit, Liu, Gabriel Ka Po, Hey, Hwee Weng Dennis, Riahi, Hend, Kelly, Michael P., Lenke, Lawrence G., Sardar, Zeeshan
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.09.2024
• ISSN: 1529-9430
• DOI: 10.1016/j.spinee.2024.06.461

Determining normative cervical spine alignment is crucial for guiding corrective surgery in spinal deformities. Prior studies correlate T1S-CL>17 (cervical mismatch) as a threshold for defining deformity. Staub et al proposed a T1 Slope (T1S) formula predicting cervical lordosis (CL) with 40% accuracy and MAE of 7.8. A practical classification system for the cervical spine is still needed. This study assesses cervical mismatch rate and predictability of Staub's CL formula (CL = T1S - 16.5°) against linear regression, in an asymptomatic adult cohort as well as a new method of classifying cervical alignment. Retrospective cohort study/multicenter. A total of 468 asymptomatic adults (18-80 years) from 5 countries (USA, France, Japan, Singapore, Tunisia) formed the MultiEthnic Alignment Normative Study (MEANS). T1S and CL (C2C7°) were measured; cervical mismatch prevalence (T1S-CL > 17) was recorded. Positive values are kyphotic, and negative values are lordotic in this study. The validity of the Staub formula was assessed. Patients were classified under a new cervical classification system and radiographic parameters of the different alignment types were compared. T1S and CL (C2C7°) were measured; cervical mismatch prevalence (T1S-CL > 17) was recorded. Positive values are kyphotic, and negative values are lordotic in this study. The validity of the Staub formula was assessed. MEANS cohort data was used for linear regression to derive a new predictive formula for comparison. A classification system was then developed using the new predictive formula and cSVA. Modifiers were added for segmental subaxial sagittal cervical alignment (SCA). CL was evaluated based on its comparison to predicted. Thresholds were set based on the average+2SD in the MEANS cohort for cSVA, and T1S. The threshold for SCA was based on the segment with the highest mean+2SD. The resulting classification system (Sardar-Miller Classification) is shown in Table 1. Radiographic parameters of the different alignment types were also compared. Mean CL was -0.42 (12.67°), T1S was 23.0 (7.86), cSVA was 19.08 (9.75), and the highest mean SCA was 3.22(4.77) which was of segment C4/C5. T1S-CL was 22.58 (9.39) with an interquartile range of 9.5 - 35.7. Of all subjects, 71.4% exhibited cervical mismatch > 17°. Linear regression yielded CL= -1.085(T1S) +24.52 (R2=0.45, p<.0001) which was simplified to CL = 24.5 - T1S. This MEANS-derived formula predicted CL within 5° in 38.9% vs 35% with a MAE of 7.64 vs 8.99 when compared to Staub. 97% of the data was captured by Types 1A-C of the classification system, with the rest being captured by Types 2-3. Only 1% surpassed threshold for segmental kyphosis. Overall, age, CL, OC2-CL, cSVA, T1S, and TK were statistically different amongst the groups (p<.01), without significant differences in lumbar or pelvic parameters. In this normative cohort, there was a high prevalence of cervical mismatch > 17°, suggesting that the previous definition of cervical deformity needs to be reassessed. A new classification for cervical alignment was thus developed using this normative cohort. There were significant differences in all cervical alignment parameters between groups, indicating clear distinction of cervical alignment between types. This abstract does not discuss or include any applicable devices or drugs.