Schwab and colleagues tested whether the SRS‑Schwab adult spinal deformity classification, revised to add pelvic parameters, could be applied clearly and reliably.
Reliability study in which 9 readers each classified 21 premarked adult deformity cases on two occasions about 1 week apart.
Methods
Design
Inter- and intra‑rater reliability study
Initiated by the SRS Adult Deformity Committee
Patients
21 premarked radiographic cases
Curve types: 3 T, 6 L, 7 D, 5 N
Basis
Full‑length frontal and sagittal radiographs
Coronal curve type by Cobb angle
Three sagittal modifiers: PI‑LL, PT and SVA
Modifier cutoffs set from HRQOL analysis of a multicenter database
Reliability
9 readers
2 readings about 1 week apart
Fleiss' kappa for curve type and for each modifier
Key Findings
By the second reading, inter‑rater kappa was 0.87 for curve type, 0.86 for PI‑LL, 0.98 for PT and 0.96 for SVA, all in the almost perfect range.
PI‑LL was the weakest modifier at first, with inter‑rater kappa rising from 0.75 (substantial) to 0.86 on the second reading.
All 9 readers agreed in 95.2% of cases for PT and 90.5% for SVA by the second reading, vs 66.7% for curve type and 71.4% for PI‑LL.
Readers were consistent on repeat grading, with mean kappas of 0.94 (curve type), 0.88 (PI‑LL), 0.97 (PT) and 0.97 (SVA), though the lowest single‑reader PI‑LL kappa was 0.66.
For the complete grade (curve type plus all three modifiers), inter‑rater kappa rose from 0.70 to 0.79 between readings.
Readers worked from premarked cases, so the study measured agreement in classifying rather than in identifying radiographic landmarks.
Board pearl
The modifiers are anchored to disability: in the paper's threshold table, PI‑LL of 11°, PT of 22° and SVA of 46 mm predicted an ODI of 40.
SRS‑Schwab adult spinal deformity classificationSchwab, Spine (Phila Pa 1976) 2012
Coronal curve type
T
Thoracic major curve >30°, apex at T9 or higher
L
Lumbar or thoracolumbar major curve >30°, apex at T10 or lower
D
Double major curve, each >30°
N
No coronal curve >30°
PI‑LL modifier (pelvic incidence minus lumbar lordosis)
0
<10°
+
10° to 20°
++
>20°
Pelvic tilt (PT) modifier
0
<20°
+
20° to 30°
++
>30°
High PT means pelvic retroversion, a compensatory mechanism
SVA modifier (sagittal vertical axis)
0
<40 mm
+
40 to 95 mm
++
>95 mm
Authors' conclusion
Data from this study show that there is excellent inter- and intra-rater reliability and inter-rater agreement for curve type and each modifier. The high degree of reliability demonstrates that applying the classification system is easy and consistent.
Schwab and colleagues tested whether the SRS‑Schwab adult spinal deformity classification, revised to add pelvic parameters, could be applied clearly and reliably.
Reliability study in which 9 readers each classified 21 premarked adult deformity cases on two occasions about 1 week apart.
Methods
Design
Inter- and intra‑rater reliability study
Initiated by the SRS Adult Deformity Committee
Patients
21 premarked radiographic cases
Curve types: 3 T, 6 L, 7 D, 5 N
Basis
Full‑length frontal and sagittal radiographs
Coronal curve type by Cobb angle
Three sagittal modifiers: PI‑LL, PT and SVA
Modifier cutoffs set from HRQOL analysis of a multicenter database
Reliability
9 readers
2 readings about 1 week apart
Fleiss' kappa for curve type and for each modifier
Key Findings
By the second reading, inter‑rater kappa was 0.87 for curve type, 0.86 for PI‑LL, 0.98 for PT and 0.96 for SVA, all in the almost perfect range.
PI‑LL was the weakest modifier at first, with inter‑rater kappa rising from 0.75 (substantial) to 0.86 on the second reading.
All 9 readers agreed in 95.2% of cases for PT and 90.5% for SVA by the second reading, vs 66.7% for curve type and 71.4% for PI‑LL.
Readers were consistent on repeat grading, with mean kappas of 0.94 (curve type), 0.88 (PI‑LL), 0.97 (PT) and 0.97 (SVA), though the lowest single‑reader PI‑LL kappa was 0.66.
For the complete grade (curve type plus all three modifiers), inter‑rater kappa rose from 0.70 to 0.79 between readings.
Readers worked from premarked cases, so the study measured agreement in classifying rather than in identifying radiographic landmarks.
Board pearl
The modifiers are anchored to disability: in the paper's threshold table, PI‑LL of 11°, PT of 22° and SVA of 46 mm predicted an ODI of 40.
SRS‑Schwab adult spinal deformity classificationSchwab, Spine (Phila Pa 1976) 2012
Coronal curve type
T
Thoracic major curve >30°, apex at T9 or higher
L
Lumbar or thoracolumbar major curve >30°, apex at T10 or lower
D
Double major curve, each >30°
N
No coronal curve >30°
PI‑LL modifier (pelvic incidence minus lumbar lordosis)
0
<10°
+
10° to 20°
++
>20°
Pelvic tilt (PT) modifier
0
<20°
+
20° to 30°
++
>30°
High PT means pelvic retroversion, a compensatory mechanism
SVA modifier (sagittal vertical axis)
0
<40 mm
+
40 to 95 mm
++
>95 mm
Authors' conclusion
Data from this study show that there is excellent inter- and intra-rater reliability and inter-rater agreement for curve type and each modifier. The high degree of reliability demonstrates that applying the classification system is easy and consistent.