Section I Evaluation and Management


 

Joseph M. Lombardi and Lawrence G. Lenke


Summary


This chapter serves to educate the reader on the historical evolution, modern-day systems, and future applications of classification systems for the diagnosis, treatment, and prognosis of adolescent idiopathic scoliosis.




6 Classification of Adolescent Idiopathic Scoliosis for Surgical Intervention



6.1 Introduction


Adolescent idiopathic scoliosis (AIS) is a complex deformity of the spinal column that affects roughly 2 to 3% of the adolescent population between the ages of 10 and 18 years. 1 , 2 The etiology is multifactorial including genetics, environmental factors, and lifestyle choices, although the definitive cause is still not completely understood by the medical community. Although classically defined as any coronal curvature ≥10 degrees, the understanding of AIS has come to include that of a complex, multiplanar, three-dimensional (3D) deformity. Classification and treatment schema will vary depending on age, growth potential, the magnitude of curvature, and location of the curve. This chapter serves to educate the reader on the background and classification systems for AIS with an emphasis on how they guide surgical intervention.



6.2 Classification



6.2.1 Historical Perspective


JR Cobb was the first author to describe a classification system for scoliosis in the age of radiography. 3 His 1948 paper entitled “Outline for the study of scoliosis” set out to identify major or minor curves, etiology of curves, and treatment recommendations based upon etiology. The Scoliosis Research Society (SRS) was established roughly 20 years later as a platform for surgeons to further the etiological understanding and treatment options for this disease.



6.2.2 King Classification


The first attempt to describe a concise and reproducible classification system for AIS was performed by King et al in 1983. 4 King’s group retrospectively reviewed the images of 405 patients with AIS treated with Harrington rods and concluded that a selective thoracic fusion can be safely performed if the combined curves (lumbar and thoracic) measured less than 80 degrees. The subsequent King–Moe classification system was divided into five curve types (I–V), which focused on the coronal Cobb angles of the thoracic and lumbar curves from which fusion levels could be decided. This classification provided a number of vital paradigms to the understanding of scoliosis treatment: it was the first to describe the typical thoracic curve for idiopathic scoliosis and also offered recommendations for selective thoracic fusions. Additionally, it introduced important concepts that would be carried forward to future classification systems.



Description

King et al 4 were the first to introduce the concept of the stable vertebra, which was described as the vertebral body cephalad to the lumbar curve that is most closely bisected by a vertically directed central sacral vertical line (CSVL). Another important concept was that of curve flexibility and how it related to structural aspects of the scoliotic deformity. King et al defined flexibility by measuring the Cobb angle during bending films and dividing it by the Cobb angle in upright anteroposterior radiographs and then multiplying by 100. This was performed for both the thoracic and lumbar curves. The flexibility index was then defined as the percentage of correction (flexibility) of the thoracic curve minus the percentage of correction of the lumbar curve. This value was then used to help support the King–Moe classifications.


A type 1 curve was defined as an “S-shaped” curve in which both the thoracic and lumbar curves cross the CSVL with the lumbar curve being larger and possessing a negative flexibility index. It was recommended that the lowest instrumented vertebra (LIV) for type I curves be L4. Conversely, a type II curve is an S-shaped curve in which both thoracic and lumbar components cross the CSVL with a larger or equal thoracic curve and a positive flexibility index. These curves can be treated with a selective thoracic fusion ending at the stable vertebra. A type III curve is a thoracic dominant curve where the lumbar curve does not cross the CSVL. Type IV curves differ from type III curves in that the fifth lumbar vertebra is centered over the sacrum with the fourth vertebra tilted toward the thoracic dominant curve. In both instances, it was recommended that the LIV end with the stable vertebra. Finally, a type V curve encompasses double thoracic curves. These types require fusion of both thoracic curves with the LIV ending at the stable vertebra.


The King–Moe classification helped advance the nomenclature and surgical treatment guidelines for idiopathic scoliosis. These included the use of a stable vertebra in the selection of distal fusion levels, the utilization of selective thoracic fusions for King II (false double major) curve types, and the recognition of a double thoracic curve pattern (King V) to optimize shoulder balance when treating a major thoracic curve. Despite this, the system had some significant shortcomings, which were compounded by the advent of more powerful Cotrel–Dubousset (C-D) fixation techniques, which came about in the 1980s. These systems employed the use of segmental hooks and a 90-degree rod rotation maneuver, which attempted to provide axial derotation and 3D correction of the AIS deformity. The use of C-D fixation commonly resulted in decompensation of the caudal lumbar curve, especially with King type II deformities. Richards 5 reported on 24 patients with King type II curves who underwent a selective thoracic fusion with C-D fixation. They noted an absence of passive lumbar correction and persistent coronal imbalance in these patients, prompting them to recommend against selective thoracic fusion with a lumbar curve greater than 40 degrees. These findings were echoed by further studies, which found decompensation for King type II and III curves when using the stable vertebra as the LIV. 6 , 7


In addition to challenges with LIV selection, the King classification system was found to have questionable reproducibility and reliability among surgeons. Multiple studies published in 1998 found fair to poor intra- and interobserver reliability of the King classification system. 8 , 9 Lenke et al 9 demonstrated a kappa coefficient of 0.40 for interobserver reliability (poor) and 0.62 kappa coefficient for intraobserver reliability (fair).



6.2.3 Lenke Classification


In response to the shortcomings of the King classification system, Lawrence Lenke in conjunction with the Harms Study Group (HSG) proposed a new classification system in 2001. 10 The goals were to provide a descriptive and reproducible classification that would improve inter- and intraobserver reliability as well as advice on surgical intervention. This was accomplished by incorporating radiographic measurements of curves in the coronal plane, assessing curve flexibility, and introducing the concept of sagittal alignment in the identification and treatment of idiopathic scoliosis.



Description


Radiographic Evaluation

The triad Lenke classification requires three important components: identification of structural curves, the application of a coronal lumbar modifier, and the application of a thoracic sagittal modifier. The classification should always begin with the radiographic identification of curves. This requires four standard upright radiographs. These include a long cassette or EOS films in the coronal and sagittal planes (anteroposterior and lateral) as well as upright or supine right and left side-bending films. It should be noted that further imaging could be performed to assess the overall flexibility of the spine, which includes prone coronal, push-prone coronal, or traction films. Likewise, advanced imaging such as magnetic resonance imaging (MRI) should be obtained for any patients with atypical patterns; that is, sharp angular curves, left-sided thoracic curves, or any neurologic symptoms (e.g., abnormal abdominal reflex). 11 Cobb measurements should begin on the upright coronal films, identifying the proximal thoracic (PT), main thoracic (MT), and thoracolumbar–lumbar (TL/L) curves (Fig. 6‑1). The apex of the MT curve should fall between the T3 vertebral body and the T11–T12 disk space. Conversely, the TL/L apex is found between the T11–T12 disk and the L4 vertebral body. Prior to assigning a classification number, surgeons must familiarize themselves with nomenclature specific to the Lenke classification system. The major curve is the curve with the greatest magnitude. It should be noted that if the thoracic and lumbar curves are of equal magnitude, the thoracic curve will default as the major curve. The major curve is always considered structural and to be an involved segment in surgical correction. The remaining curves are termed minor curves. It is important to next determine whether these minor curves are structural versus nonstructural, which can be evaluated through side-bending coronal and long cassette lateral films. By consensus of the HSG surgeons, a curve was considered structural if its Cobb angle remains ≥25 degrees on side-bending films. It was decided not to specify bending film techniques but to recommend the worst residual Cobb angle on the bend film. Rotation was considered in the development of the classification and, although important, there was no good standardized measurement on which the HSG could find consensus. The PT curve may also be considered structural if kyphosis of ≥20 degrees is present from T2–T5 on lateral films. Likewise, both MT and TL/L curves are to be considered structural if kyphosis ≥20 degrees exists from T10–L2 on sagittal Cobb measurements. It is important to note that the lateral Cobb measurements are independent qualifiers for a structural curve regardless of whether the coronal side-bending measurements meet the criteria for structurality.

Fig. 6.1 Schematic diagram demonstrating typical regions of deformity as described in the Lenke classification.

Once the major and structural curves have been identified, the lumbar and sagittal modifiers should be applied to complete the classification. The coronal lumbar modifier is determined by drawing a CSVL, which extends proximal to the thoracolumbar junction. The CSVL is defined as a line perpendicular to the horizontal, which bisects the midpoint of the sacrum. An “A” modifier is designated when the CSVL lies between the pedicles of the apical vertebra in the lumbar curve. A “B” modifier is designated if the CSVL touches the apical pedicle and, finally, a “C” modifier is designated if the CSVL lies outside of the pedicles at the apical lumbar vertebra (Fig. 6‑2). The sagittal thoracic modifier is selected from the long cassette lateral radiographs with the T5–T12 sagittal Cobb measurement of central importance. This region is designated as “–” or hypokyphotic if it measures less than 10 degrees, normokyphotic if it is between 10 and 40 degrees, and “+” or hyperkyphotic if it is greater than 40 degrees (Fig. 6‑3).

Fig. 6.2 Lumbar modifiers as described by the Lenke classification.
Fig. 6.3 Schematic of the Lenke classification system.

Using these radiographic analyses, the Lenke triad classification can be determined. The system describes six elemental curve types depending on the location of the major curve and whether the accompanying minor curves are structural or nonstructural. The modifiers are then applied to those curves yielding a three-digit classification, for example, 1AN. Although this classification system is descriptive and systematic in its approach, it should be noted that there are many additional factors that play an important role in deciding specific fusion levels. 12 In general, the Lenke classification recommends fusion of all major curves as well as structural minor curves while excluding nonstructural curves. However, other radiographic and clinical considerations should be taken into account when deciding fusion levels including the ratio of thoracic to lumbar curves, skeletal maturity, preoperative shoulder balance, truncal balance, thoracic versus lumbar prominences, and, finally, desires of the patient. 13



Curve Types


Lenke Type 1

Type 1 curves are defined as having a major curve in the MT region with nonstructural minor curves in the PT and TL/L region (Fig. 6‑4). The mainstay for type 1 curves is therefore fusion of the MT curve only. The upper instrumented vertebra (UIV) is typically T3, T4, or T5, whereas the LIV will vary considerably from T11 to the sacrum depending on characteristics of the lumbar curve. Correction for the type 1 MT curve can be accomplished through a variety of maneuvers including a cantilever, in situ contouring, in situ translation, apical derotation, and stepwise compression distraction with the goal of maximally translating the apex while maintaining a neutral position of the LIV. The development of pedicle screw fixation has led to an increase in the utilization of apical derotation maneuvers resulting in improved cosmesis of thoracic and lumbar prominences associated with the rotational deformity and obviating the need for thoracoplasty procedures. 14 The use of soft-tissue releases or periapical Ponte or posterior column osteotomies (PCOs) may be necessary in instances of severe apical deformity to restore physiologic kyphosis. Likewise, osteotomies may be necessary in large or stiff curves (upright coronal Cobb ≥75 degrees or side-bending Cobb ≥50 degrees) to allow maximal correction while minimizing bone–screw interface stress during corrections. 15

Fig. 6.4 (a) Patient with a structural main thoracic (MT) curve and nonstructural, minor proximal thoracic (PT) and thoracolumbar–lumbar curve (TL/L) curves representing a Lenke 1AN. (b) After posterior spinal instrumentation and fusion (PSIF) T3–L1.

The selection of LIV in Lenke type 1 curves is strongly correlated to the lumbar modifier. In A and B curve patterns, the LIV selected is usually the most cephalad vertebra in the TL/L region that is least intersected by the CSVL on upright coronal radiographs and with neutral rotation. This is referred to as the last touched vertebra (TV). Commonly, the last TV will be one level caudal to the lower end vertebra of the MT curve and one to two levels cephalad to the true stable vertebra. The last touch rule will hold true as long as there is no thoracolumbar junctional kyphosis present. In instances where it is difficult to discern which vertebra is last touched, a supine radiograph can help provide further information. Generally, when selecting between two vertebrae as the LIV in the caudal thoracic spine, choosing the more distal segment is safer and does not require a significant sacrifice in motion.


Treatment of Lenke type 1C curves remains highly controversial. 16 By definition, these are curves in which the apex of the TL/L region will deviate entirely beyond the CSVL but corrects to ≤25 degrees on the coronal side-bending films. In such instances, a selective thoracic fusion can be considered. Lenke et al 17 recommended further analysis of radiographic and clinical ratios of the thoracic and lumbar spine to assist in deciding on selective thoracic fusion. They described three measurements that should be considered: ratio of thoracic Cobb angle to lumbar Cobb angle, the ratio of thoracic to lumbar apical vertebral rotation (AVR), and the ratio of thoracic to lumbar apical vertebral translation from the CSVL (AVT). When ratios exceed 1.2, a selective thoracic fusion can be successfully performed with the AVT ratio being the most important indicator. Additionally, it is important to evaluate the patient clinically for thoracic and lumbar prominence on upright and forward bending. In a suitable patient, a thoracic truncal shift should outweigh a lumbar shift. This can be evaluated by scoliometer measurements on forward bending or observation of curve correction on prone positioning. Behensky et al 18 demonstrated that 21 patients who met two or three of the above criteria (AVT, AVR, Cobb ratio) had good postoperative coronal balance and outcomes. The selective fusion criteria can be expanded beyond type 1C curves to include 1C to 4C curves. It is worth noting that a subset of these deformities will not meet selective fusion criteria. In such instances, fusion will extend to include the nonstructural lumbar curve in a 1C curve, which is a departure from the tenets of the Lenke classification system. These instances, while rare, have been termed “rule breakers” and demonstrate the importance of surgical decision-making when evaluating the compensatory nature of the nonstructural lumbar curve. 19



Lenke Type 2

Type 2 or “double thoracic” curves describe a major MT curve and a structural PT curve with a nonstructural TL/L region (Fig. 6‑5). Recommended treatment for type 2 curves involves a posterior spinal fusion (PSF) of the PT and MT curves. When selecting the UIV for type 2 curves, preoperative shoulder balance plays a critical role with the ultimate goal of establishing coronal balance after correction. 20 , 21 Typically, if the left shoulder is higher than the contralateral preoperatively (assuming a right-sided MT curve), it will be necessary to begin the fusion at T2 or, in rare circumstances T1, to restore shoulder balance and minimize further elevation. This can be achieved through stepwise compression and distraction of the convex and concave sides, respectively. If the shoulders are balanced preoperatively, it is recommended that the UIV start at the T2–T3 level. Conversely, if the left shoulder is depressed preoperatively, a UIV of T4 is recommended, taking care not to overcorrect the deformity.

Fig. 6.5 (a, b) A 14-year-old adolescent girl with structural proximal thoracic (PT) and main thoracic (MT) curves and nonstructural thoracolumbar–lumbar curve (TL/L) curves representing a Lenke type 2BN. (c) After posterior spinal instrumentation and fusion (PSIF) of T2–L1.


Lenke Type 3

Lenke type 3 or “double major” curves are defined by a major MT curve with a structural TL/L curve and nonstructural PT region (Fig. 6‑6). Therefore, recommended treatment involves fusion of the thoracic and TL/L regions for most of these curve patterns. The selection of the UIV is similar to that of a type 1 curve, typically beginning at T3–T5. This is dictated by the radiographic and clinical features of the nonstructural PT region as well as preoperative shoulder balance. Most Lenke type 3 curves are associated with a C lumbar modifier. Type 3 curves with A or B modifiers are usually characterized by very large MT curves, resulting in a compensatory structural TL/L curve but with less substantial apical translation. Likewise, a curve can be considered type 3 when TL/L junctional kyphosis exceeds 20 degrees despite a side-bending coronal Cobb measurement that would otherwise be considered nonstructural (<25 degrees). In such instances, a selective thoracic fusion may be considered if it meets the criteria as explained earlier. Selective thoracic fusion can additionally be considered in rare instances where type 3C deformities meet the criteria. 16

Fig. 6.6 (a, b) A 16-year-old adolescent boy with a major main thoracic (MT) curve, structural thoracolumbar–lumbar curve (TL/L) curve, and nonstructural proximal thoracic (PT) curve representing a Lenke type 3CN. (c) After posterior spinal instrumentation and fusion (PSIF) of T3–T12.

In instances where type 3 deformities do not meet the selective thoracic fusion criteria, the recommended LIV is typically at the L3 or L4 level. However, there remains much controversy when selecting between these levels. 22 , 23 Traditional teachings dictate the inclusion of convex or “open” disks into the fusion construct, that is, if the L3–L4 disk space is open on the convexity of the TL/L curve, then L4 should be included in the fusion. Conversely, if the L3–L4 disk is concave or “closed” on the convexity of the curve, then the fusion can be stopped at the L3 level provided that the deformity correction adequately centralizes and horizontalizes the L3 segment. This decision is most controversial when the L3–L4 disk is parallel on the upright coronal radiographs. In such instances, other factors including distance of L3 from the CSVL, the distance of L3 from the apex of the TL/L curve, the overall size (Cobb magnitude) of the curve, the amount of residual lumbosacral deformity as reflected by any fixed wedging of the L3–L4 disk or translation of L3 off the CSVL, centering of L3 on a push-prone and/or side-bending radiographs, and presence of an L5 and/or S1 fixed obliquity. All of these factors should be considered when choosing the LIV in Lenke type 3 curves.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Apr 30, 2022 | Posted by in ORTHOPEDIC | Comments Off on Section I Evaluation and Management

Full access? Get Clinical Tree

Get Clinical Tree app for offline access