The Spine Journal 14 (2014) 315–325

2013 Outstanding Paper Runner-up

Assessment of skeletal maturity in scoliosis patients to determineclinical management: a new classification scheme using distal radius

and ulna radiographs

Keith D.K. Luk, MCh(Orth), FRCSE, FRCSG, FRACS, FHKAM(Orth)*,Lim Beng Saw, MBBS, Samuel Grozman, MD,

Kenneth M.C. Cheung, MBBS, MD, FRCS, FHKCOS, FHKAM(Orth),Dino Samartzis, DSc

Department of Orthopaedics & Traumatology, The University of Hong Kong, Professorial Block, 5th Floor, 102 Pokfulam Rd, Pokfulam,

Hong Kong, SAR, China

Received 2 February 2013; accepted 21 October 2013

Abstract BACKGROUND: Assessment of skeletal mat

FDA device/drug

Author disclosure

Paid directly to instit

(H, Paid directly to in

Nothing to disclose. K

(None); Consultancy:

Grants/grants pending:

Travel: AOSpine, SRS

directly to institution)

Journal of Spinal Diso

1529-9430/$ - see fro

http://dx.doi.org/10.10

urity in patients with adolescent idiopathic scoliosis(AIS) is important to guide clinical management. Understanding growth peak and cessation is cru-cial to determine clinical observational intervals, timing to initiate or end bracing therapy, and whento instrument and fuse. The commonly used clinical or radiologic methods to assess skeletal ma-turity are still deficient in predicting the growth peak and cessation among adolescents, and boneage is too complicated to apply.PURPOSE: To address these concerns, we describe a new distal radius and ulna (DRU) classifi-cation scheme to assess skeletal maturity.STUDY DESIGN: A prospective study.PATIENT SAMPLE: One hundred fifty young, female AIS patients with hand x-rays and no pre-vious history of spine surgery from a single institute were assessed.OUTCOME MEASURES: Radius and ulna plain radiographs, and various anthropomorphic pa-rameters were assessed.METHODS: We identified various stages of radius and ulna epiphysis maturity, which weregraded as R1–R11 for the radius and U1–U9 for the ulna. The bone age, development of sexualcharacteristics, standing height, sitting height, arm span, radius length, and tibia length were studiedprospectively at each stage of these epiphysis changes.RESULTS: Standing height, sitting height, and arm span growth were at their peak during stagesR7 (mean, 11.4 years old) and U5 (mean, 11.0 years old). The long bone growths also demonstrateda common peak at R7 and U5. Cessation of height and arm span growth was noted after stages R10(mean, 15.6 years old) and U9 (mean, 17.3 years old).CONCLUSIONS: The new DRU classification is a practical and easy-to-use scheme that can pro-vide skeletal maturation status. This classification scheme provides close relationship with adoles-cent growth spurt and cessation of growth. This classification may have a tremendous utility in

status: Not applicable.

s: KDKL: Endowments: Endowed professorship (E,

ution/employer); Grants: University of Hong Kong

stitution/employer). LBS: Nothing to disclose. SG:

MCC: Board of Directors: Scoliosis Research Society

Ellipse technologies (B, Paid directly to institution);

RGC (F, Paid directly to institution/employer); Trips/

(B); Endowments: Endowed Professorship (H, Paid

. DS: Board of Directors: The Spine Journal (None),

rders and Techniques (None), Journal of Orthopaedic

Surgery (None), Spine (None); Grants/grants pending: RGC, AOSpine (I,

Paid directly to institution); Trips/Travel: AOSpine (B, travel as council

board member of AOSpine East Asia).

The disclosure key can be found on the Table of Contents and at www.

TheSpineJournalOnline.com.

KDKL and LBS contributed equally to this work.

* Corresponding author. Department of Orthopaedics & Traumatology,

Queen Mary Hospital, Professorial Block, 5th Floor, 102 Pokfulam Rd, Pok-

fulam, Hong Kong SAR, China. Tel.: (852) 2255-4254; fax: (852) 2817-4392.

E-mail address: hrmoldk@hku.hk (K.D.K. Luk)

nt matter � 2014 Elsevier Inc. All rights reserved.

16/j.spinee.2013.10.045

316 K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

improving clinical-decision making in the conservative and operative management of scoliosis pa-tients. � 2014 Elsevier Inc. All rights reserved.

Keywords: Skeletal; Maturity; Classification; Distal; Radius; Spine; Scoliosis; Adolescent

Introduction

Assessment of skeletal growth in children and adoles-cents is an important tool in managing growth related prob-lems, such as scoliosis, hormonal disorders and in theapplication of expandable prosthesis in young children withsevere deformities [1–4]. Many studies on adolescent idio-pathic scoliosis (AIS) further highlight the importance ofgrowth and rapid progression of the curve during the peakof growth spurt [5–9]. The understanding of the patient’sgrowth potential and end-stage, or near end-stage, growthcarries an important implication in the prognosis and pro-vides guidance to the treating physician in deciding thetreatment modality (eg, observation interval, timing to ini-tiate bracing therapy, cessation of bracing, timing of instru-mentation, and fusion).

Clinical or radiologic methods are commonly used to as-sess the patient’s growth potential (eg, menarche age, trir-adiate cartilage fusion, Risser’s sign) [10]. However, mostof these measures have their own deficits and are not ableto directly indicate the peak growth spurt or when thegrowth has ended. Although the measurements of standingheight or arm span directly measure growth, they requirefew serial follow-ups to determine growth trends. Thoughdigital skeletal age assessment using the Tanner and White-house (TW3) [11–13] or the Greulich and Pyle [14] meth-ods have improved the assessment accuracy of the patient’sskeletal maturity, these techniques are time-consuming anddifficult to use in the outpatient clinical setting.

Sanders et al. [15] reported the close relationship of peakheight velocity (PHV) to various stages of finger epiphysismaturation according to the TW3 method. The authorsnoted that the PHV occurred at early adolescents and thatthe start of menarche and the Risser sign appear after thepeak of adolescent growth spurt. However, the authorsclaimed that distal radius and ulna (DRU) epiphysis werefound to have the least correlation with growth. Thesestages of DRU in the TW3 method were originally designedto be used in combination with the epiphysis of the fingerphalanges and its use alone in predicting growth spurtmay not be accurate because of the wide interval betweeneach stage.

Because the DRU physes progression spans the wholeperiod of skeletal growth and is the last one to close, wewant to refine the TW3 classification and see whether itcan be used in isolation. The following descriptive studyaddressed a new and practical classification scheme to as-sess skeletal maturity in early adolescents by utilizing amodified TW3 method for the radius and ulna. Based onthis new classification, known as the DRU classification

scheme, we further assessed its utility in relation with com-mon clinical growth parameters, such as standing height,sitting height, arm span, menarche age, bone age, and longbone growth of the radius and tibia.

Methods

Based on prospective anthropometry records of scoliosispatients at the Duchess of Kent Children’s Hospital, Pokfu-lam, Hong Kong, between 1975 and 2000, we studied theDRU epiphysis radiologic morphology and the relationshipof our new classification with growth trend among femaleChinese AIS patients. The exclusion criteria were patientswho were male, had undergone surgical treatment for sco-liosis, had less than 4 years of follow-up, were older than 17years of age, incomplete data (missing hand x-ray, dataform), and scoliosis with secondary causes.

All cases included in this study had spine and left handwrist x-rays taken during each visit. Biometric measure-ment, such as standing height, sitting height, arm span, ra-dial length, and tibial length, was recorded. The radiallength was taken with the elbow in 90� flexion and restedon a table, and the length between the radial head and radi-us styloid process was measured. While in the sitting posi-tion, tibial length was assessed as the as length between themedial tibial joint line and medial maleolus. In addition, allpatients also had records on their menarche age, the de-velopment of the breast according to the four-stage scaledescribed by Reed and Stuart [16]; the rating of develop-ment of the pubic hair was made following the five-stagescale by Reynolds and Wines [17]. A bone age assessmentwas performed according to the method described by Greu-lich and Pyle [14]. Scoliosis curve types and alignments(based on the Cobb’s angle) were determined based onposteroanterior standing radiographs. Radiographic meas-urements were in millimeters and were obtained using a DI-COM based Radworks 5.1 (Applicare Medical Imaging BV,Zeist, The Netherlands) computer software program. Twoindividuals (LBS, SG) conducted the radiographic dataassessment and on two separate occasions. Our new classi-fication for DRU radiologic morphology (DRU classifica-tion) is illustrated in Table 1 and based on clinicalobservation. Because the aim of the study was to assessskeletal changes at early adolescents, our study focusedon radius stages R5 to R11 and ulna stages U2 to U9.

Because records were taken as part of the clinicalfollow-up for AIS patients, the interval of follow-up varied;the younger patients had follow-up between 3 and 6months, whereas an older child may have had a longer

Table 1

Distal radius and ulna (DRU) classification of radiologic morphology and its similarity to original TW3 stages

DRU TW3 (RUS) Characterization of distal radius stages Illustration Radiograph

R1 B Epiphysis appears as single or multiple spots.

R2 C Distinct and oval shaped epiphysis.

R3 D Maximal diameter is more than half the width of metaphyseal.

R4 E Double line at the distal border of epiphysis, represent palmar

and dorsal surface.

R5 F Proximal border of epiphysis visible as an irregular thickened

white line, more pronounced on the medial side; proximal

boarder assuming concave shape; width of epiphysia not as

wide as metaphysis.

R6 G Either medial part of epiphysis overhangs the metaphysis,

lateral border wider than metaphysis, or both; medial border

develops articulation with the ulna.

(Continued)

317K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

Table 1

(Continued )

DRU TW3 (RUS) Characterization of distal radius stages Illustration Radiograph

R7 H Epiphysis capping on the medial side while the lateral side is

rounded

R8 H Squaring/cap seen on lateral proximal corner; physeal plate still

well visible. Its medial and lateral ends are wider than the

center.

R9 I The epiphysis strongly capping the metaphysis with sclerosis of

the physeal space; the growth plate is visible but blurred; the

epiphysis almost touches the metaphysis at the medial and

lateral end.

R10 I The growth plate is completely obliterated, forming a sclerotic

line, at times appearing as a broken line; the notch is visible

at the medial or lateral end of the growth plate.

R11 I Complete fusion with the lateral/medial edge of the physeal

line rounded; growth plate scar may still be visible.

U1 B The epiphysis appears at single/multiple spots.

(Continued)

318 K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

Table 1

(Continued )

DRU TW3 (RUS) Characterization of distal radius stages Illustration Radiograph

U2 C A rounded shape epiphysis.

U3 D The epiphysis is at least half the width of metaphysis.

U4 E The styloid is visible on the medial end of the epiphysis.

U5 F The head of the ulna is distinctly defined and denser than

styloid; the border adjacent to the radial epiphysis is

flattened.

U6 G The epiphysis is as wide as the metaphysis; the proximal border

of the epiphysis overlaps with the metaphysis at the center

third.

U7 H Narrowing of the medial physeal plate; the medial border of the

epiphysis and the metaphysis form a smooth curve line

(articulation with radius); fusion may be seen on the medial

half.

(Continued)

319K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

Table 1

(Continued )

DRU TW3 (RUS) Characterization of distal radius stages Illustration Radiograph

U8 H Unfused growth plate visible proximal to styloid process; in a

rotated film, the medial and lateral border appear fused but

the physeal space can be seen under the styloid.

U9 H Complete fusion.

DRU, distal radius and ulna; TW3, Tanner and Whitehouse. Note: Modified from Tanner JM, Healey MJR, Goldstein H, Cameron N. Assessment of

Skeletal Maturity and Prediction of Adult Height (TW3 Method). Third ed. London: WB Saunders; 2001.

320 K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

duration between observations. Some of these older pa-tients who had follow-up of 4 to 6 years were also included,because growth within 1 to 2 years might be too little to bedetected. The radiologic stages of the radius and ulna wereidentified, and the changes of growth parameters that weretaken place between two radiologic stages were studied. Ifone radiologic stage developed after more than one follow-up visit, the total growth between the whole period wasconsidered as the growth change between that single radio-logic stage. The DRU were studied separately.

All data were coded and entered on a spreadsheet. SPSSversion 14 (Chicago, IL, USA) was utilized to perform stat-istical analysis. Frequency and descriptive statistics wasperformed of the data set. Mean values 6 standard devia-tion, and ranges of the data were calculated.

Results

More than 1,600 scoliosis patients records from our sco-liosis anthropometry study were screened. One hundred fiftycases fulfilled our selection criteria.Mean age at presentationwas 13.161.5 years (range, 7.1–16.3). All these cases weretreated nonsurgically, with 91 cases had bracing and 54 caseshad follow-up without brace. The mean age of menarchewas12.961.3 years (range, 9.7–17.7), whereas 48 cases pre-sented before menarche and 97 cases post menarche.

As for the scoliosis curve pattern, there were 49 cases(32.7%) of double thoracolumbar curves, 2 cases (1.3%)

of double thoracic curve, 55 cases (36.7%) with singlethoracic curves, 33 cases (22%) with thoracolumbar curvesand 11 cases (7.3%) of single lumbar curve. Scolioticcurves at presentation ranged from 3� to 54�. We werenot able to assess the Risser stages because most cases trea-ted had metal bracing at the time of imaging, which pre-vented the visibility of the iliac crest apophysis.

Growth changes associated with radial stages

Chronological age and bone ageTable 2 illustrates the development of the radiologic

stage of the distal radius over the span of chronologicalage and bone age according to Greulich and Pyle [14].The lowest radius radiologic stage found in our serieswas R5, which was detected at a mean chronological ageof 9.3 years old. The mean chronological age between thetwo radiologic stages was 1.1 years (range, 0.3–1.8),whereas the bone age had a more regular interval and nar-rower range; with a mean skeletal age of 1.5 years (range,1.2–1.8) between two stages.

Standing height, sitting height, and arm spanThe standing PHV and arm span growth occurred during

radial stage R7. The sitting height, which represents thevertebral growth, also peaked at stage R7. The standingheight growth and the sitting height growth ceased at stageR9, whereas the arm span ceased at stage R11 (Fig. 1A;Table 2).

Table 2

Chronological age, bone age, and various longitudinal growth velocities according to each distal radial epiphyseal stage

Radius stage

Mean chronological age,

y (6 SD)

Mean bone age,

y (6 SD)

Longitudinal growth velocity (mm/y)

Mean arm span

(6 SD)

Mean sitting height

(6 SD)

Mean standing height

(6 SD)

R5 9.3 (1.5) 7.7 (1.0) 61.4 (25.7) 33.0 (13.2) 60.6 (31.3)

R6 11.1 (1.7) 9.5 (0.6) 53.4 (20.7) 22.0 (7.1) 53.6 (9.0)

R7 11.4 (1.1) 11.2 (0.9) 67.3 (22.2) 38.2 (12.9) 65.9 (17.2)

R8 12.7 (1.2) 13.0 (1.6) 41.2 (30.1) 26.0 (15.7) 40.4 (21.9)

R9 13.9 (1.2) 14.3 (0.8) 17.9 (16.4) 13.2 (9.3) 16.5 (13.5)

R10 15.6 (1.3) 15.7 (0.7) 5.6 (12.2) 4.6 (11.0) 4.6 (8.3)

R11 17.3 (1.1) 16.9 (0.6) 0.7 (7.2) 0.8 (6.3) 1.6 (4.4)

mm, millimeters; y, years; SD, standard deviation.

321K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

Long bone growthPeak radial length growth in relation to the distal radius

staging was at R7, decreasing thereafter up to stage R11.However, the peak of tibial length growth was during stageR6 and was less than 5 mm per year after stage R9(Fig. 1C)

Menarche and sexual maturationThe onset of menarche according to radial stage is illus-

trated in Fig. 2 (Top). The majority of patients presentedwith menarche at stage R8 and R9, which were one totwo stages after PHV at R7. All the patients in stage R5were at least breast stage 1 according to the four-stage scaledescribed by Reed and Stuart [16]. Seventy percent of caseswere in stage 2 development during R7. Ninety-one percentof patients had full development by R11.

Growth changes associated with ulna stages

Chronological age and bone ageThe lowest ulna stage in our series was stage U2. The

chronological age and bone age for each of the ulna stagesare noted in Table 3. The mean chronological age gap be-tween two stages was 1.1 years (range, 0.2–1.6), and themean bone age between two stages was 1.3 years (range,0.6–2.2).

Standing height, sitting height, and arm spanThe standing PHV and arm span growth were during ul-

na stage U4, the peak of sitting height was slightly later atstage U5 (Fig. 1B; Table 2). Cessation of these growth pa-rameters was noted at stage U9 (Fig. 1B; Table 3).

Long bone growthThe peak of long bone growth was at stage U5 for both

radius and tibia (Fig. 1D). The cessation of long bonegrowth was noted at stage U9 (Fig. 1D).

Menarche and sexual maturationOf the 150 patients, 48 (32%) attained menarche at stage

U6, one stage following after PHV. All the patients were inbreast stage 1 until stage U3. Ninety percent had full stage

4 development only at the second year of U9 (Fig. 2,Bottom).

Discussion

Pubertal growth can be divided into three stages: Prepu-bertal minimal velocity, very rapid or PHV, and decreasegrowth velocity. In girls, pubertal growth lasts for approx-imately 2 to 3 years, with PHV occurring between 11.5and 12.5 years [18,19]. Adolescent idiopathic scoliosishas been reported to have rapid curve progression duringthe peak of growth spurt [5–9]; therefore, an understandingof growth patterns and the various methods of predictingthe growth spurt of a child will provide added informationin the management of patients with scoliosis and variousgrowth-related diseases.

The menarche onset in females had been popularly usedas a reliable point in time to predict growth. However,many authors claim that it occurs at various stages of pub-erty and fewer than half of the cases that had menarche cor-responded with peak height growth [20]. In the majority ofcases, the menarche onset is 1 to 2 years after PHV [20].The Tanner staging for secondary sexual characteristics de-velopment was reported to have close relationship to peakheight, with the majority of females exhibiting budding ofthe breast approximately one year before peak growth[11–13]; however, this scheme is not commonly used indaily practice.

Direct measurement of standing and sitting heights pro-vides a direct recording of PHVor spinal growth. However,this prediction of PHV is based on a retrospective calcula-tion because growth velocity has to be in a descendingtrend so that the peak velocity can be identified. Sanderset al. [21] reported that a single measurement of height isnot informative to a child’s growth status. It takes twomeasurements to know whether a child is accelerating ordecelerating in growth, and three values are needed tounderstand growth trends.

The closure of the triradiate cartilage usually occurs be-tween ages 11 and 13 years. Ryan et al. [22] reported thatgrowth velocity was higher in cases with open triradiate

Fig. 1. Mean (A, B) longitudinal growth velocities and (C, D) bone growth of the distal radius and ulna stages of the DRU classification scheme.

322 K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

Fig. 2. Prevalence of menarche based on distal (Top) radius and (Bottom) ulna stages of the distal radius and ulna (DRU) classification scheme.

323K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

cartilage, but it could not be used as an independent indica-tor of skeletal maturity. The Risser sign had been com-monly used as an indicator of spinal growth. Based onthis sign, peak growth usually occurs approximately 2 yearsbefore its appearance (Risser 0) and growth terminateswhen the apophysis growth has completed [23]. As such,based on this sign, it is not able to predict the PHV. Further-more, studies have reported that Risser I to V is highly var-iable in relation to height velocity and that the apophysis ofthe iliac crest could remain unfused until 25 years of age

Table 3

Chronological age, bone age, and various longitudinal growth velocities accordi

Ulna stage

Mean chronological age,

y (6 SD)

Mean bone age,

y (6 SD)

Longitudinal gro

Mean arm span

U2 8.9 (1.5) 6.9 (0.6) 55.3 (8.2)

U3 9.1 (1.2) 8.3 (0.7) 53.1 (30.9)

U4 11.2 (1.4) 9.8 (0.8) 64.2 (27.7)

U5 11.0 (1.4) 10.5 (1.5) 66.4 (24.6)

U6 12.6 (1.2) 12.7 (1.2) 46.7 (30.5)

U7 13.4 (1.2) 13.8 (1.0) 24.9 (22.5)

U8 14.8 (1.3) 14.9 (0.7) 10.8 (14.3)

U9 16.3 (1.3) 16.4 (0.7) 2.5 (7.6)

mm, millimeters; y, years; SD, standard deviation.

[22]. In addition, according to Izumi [24], the use of theposteroanterior view of the spine to reduce irradiation tothe breast has led to loss of accuracy and interpretation ofthe Risser sign.

The digital skeletal age system as described by Tannerand Whitehouse [11–13] or Greulich and Pyle [14] providesa more comprehensive way of calculating bone age, whichallows more accurate prediction of bone maturation; how-ever, this is time consuming and complicated. Sanderset al. [15] made modification to the TW3 method by

ng to each distal ulna epiphyseal stage

wth velocity (mm/y)

(6 SD) Mean sitting height (6 SD) Mean standing height (6 SD)

27.4 (5.8) 44.1 (23.7)

28.0 (16.0) 54.7 (9.0)

30.2 (15.9) 70.1 (17.9)

37.4 (15.5) 67.7 (25.5)

27.3 (13.3) 44.5 (21.8)

18.3 (13.4) 24.8 (9.2)

8.5 (10.0) 9.0 (9.4)

1.3 (10.3) 3.0 (8.2)

324 K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

focusing on the finger digit epiphysis. Their study, based ona small series of 22 individuals, showed that the cappingphase of the epiphysis of the phalanges corresponded withthe PHV and deterioration of the curve. However, in thesame study, Sanders et al. [15] commented that the DRUaccording to the TW3 did not have correlation with growth.

In the new DRU classification as proposed by the au-thors, we identified additional stages for both the DRU,which were distributed more evenly throughout the pubertalphase, with a bone age gap interval of 1 year between eachstage (Tables 2 and 3). This is important as we feel the orig-inal TW3 method possessed a wide interval in betweeneach stage, and therefore this may have contributed to thenonsignificant relationship between the original radiusand ulna and PHV in previous studies.

Our study using this new staging of the DRU, known asthe DRU classification scheme, showed that the growthspurt peak occurred at radiologic stage R7 and U5, whichis characterized by medial capping of the distal radius epi-physis and appearance of the ulna styloid in the ulna epi-physis. These corresponded with a chronological age of11.4 and 11.0 years. Our mean PHV was 6.6 cm per year.These findings were similar to what Lee et al. [25] reportedin 1,139 adolescent Chinese girls, whereby the peak growthage was at 10.5 years with PHVof 7.0 cm per year. Further-more, in our study, the longitudinal growths ceased atstages R11 and U9. At these stages, the characteristic fea-tures were blurring of the growth plate of the distal radiusand complete fusion of ulna epiphysis with the metaphysis.These end stages occurred at the mean ages of 15.6 and16.3 years for the radius and ulna, respectively, whereasthe previous study by Lee et al. [25] reported that growthcessation occurred overall at 15 years of age. Most of ourpatients achieved menarche about one or two stages afterPHV, which was comparable with most studies that re-ported menarche onset having occurred around 1 to 2 yearsafter the PHV [8].

Our DRU classification scheme may have a direct clin-ical relevance to managing patients with AIS, assumingcurve deterioration is correlated with PHV. For example,to minimize the duration of bracing, bracing should pref-erably not be initiated before stage R5, but instead initiatedbetween R6 and R7/U5. Brace weaning can be consideredat R8 or 9. Stages R11 and U9, which denote skeletal ma-turity, can be useful markers to determine the need for sur-gical fusion in patients with curves in the ‘‘gray’’ zone(45�–50�) when one is not certain about the risk of furthercurve progression. This rationale can also be applied toyoung patients being managed with growing rod technol-ogy owing to early-onset scoliosis.

Although our DRU classification scheme is practicaland has direct clinical relevance, our study had certainlimitations. Not all patients were assessed at the same chro-nological age and radiographic stage, and followed longitu-dinally. Because of the age at presentation and maturitystatus, the follow-up of patients at our children’s hospital

varied. Moreover, because of the time of presentation atthe clinic, we did not have patients young enough present-ing with radius stages R1–R4 or ulna stage U1. However,the main intent of our classification scheme, which wasbased on a large sample size of 150 subjects assessed at dif-ferent time points and a sample size much larger than thatof previous reports by Greulich and Pyle [14] and Sanderset al. [15], was to focus on patients at early adolescents andbeyond and not at the juvenile or prepubertal phase. None-theless, additional and larger studies are needed to furtherverify our findings and to predict exactly the stage atwhich growth begins to accelerate and correlate withcurve progression as well as to also determine if indeeddifferent curve patterns have varied progression patterns.Furthermore, our study sample was composed of only Chi-nese subjects. The generalizability of our findings to otherethnic groups should be addressed; however, previousstudies may have been heterogeneous and more homoge-neous study subjects, as in our study, are more preferredto minimize any potential inherent biases.

Conclusions

Our DRU classification scheme is a simple and reliablemethod for the assessment of skeletal maturation, in particu-lar for scoliosis patients where understanding growth peaksand cessation provides valuable insights in clinical manage-ment options. The stages in the new DRU classification areevenly distributed throughout pubertal age. The medial cap-ping of the distal radius (R7) and early appearance of the ulnastyloid with the head of the ulna distinctly defined and denserthan the styloid (U5) signify the peak growth spurt, whereasblurring of the distal radial growth plate (R11) and fusion ofulna epiphysis (U9) indicate cessation of longitudinalgrowth. As such, our classification scheme is able to showboth the peak growth and growth cessation periods. Further-more, our classification scheme has an advantage over othermethods, such as the Risser sign and menarche age, whichhad been reported to have weak associations with PHV andno ability to predict growth cessation. As such, our classifica-tion scheme’s practicality make it a useful tool that can havetremendous utility in clinical decision making in managingscoliosis patients.

Acknowledgments

The authors thank Ms Khin Lee at the Department ofOrthopedics and Traumatology at The University of HongKong for her assistance with the data collection.

References

[1] Cheung KMC, Cheung J, Samartzis D, et al. Magnetically controlled

growing rods for severe spinal curvature in young children: a pro-

spective case series. Lancet 2012;379:1967–74.

325K.D.K. Luk et al. / The Spine Journal 14 (2014) 315–325

[2] Dominkus M, Krepler P, Schwameis E, et al. Growth prediction in

extendable tumor prostheses in children. Clin Orthop Relat Res

2001;390:212–20.

[3] Duthrie RB. The significance of growth in orthopaedic surgery. Clin

Orthop Relat Res 1959;14:7–19.

[4] Thompson GH, Akbarnia BA, Campbell RM Jr. Growing rod techni-

ques in early-onset scoliosis. J Pediatr Orthop 2007;27:354–61.

[5] Bunnell WP. The natural history of idiopathic scoliosis before skele-

tal maturity. Spine 1986;11:773–6.

[6] Charles YP, Daures JP, de Rosa V, Dimeglio A. Progression risk of

idiopathic juvenile scoliosis during pubertal growth. Spine 2006;31:

1933–42.

[7] Lonstein JE, Carlson JM. The prediction of curve progression in un-

treated idiopathic scoliosis during growth. J Bone Joint Surg Am

1984;66:1061–71.

[8] Song KM, Little DG. Peak height velocity as a maturity indicator

for males with idiopathic scoliosis. J Pediatr Orthop 2000;20:

286–8.

[9] Weinstein SL, Dolan LA, Spratt KF, et al. Health and function of pa-

tients with untreated idiopathic scoliosis: a 50-year natural history

study. JAMA 2003;289:559–67.

[10] Risser JC. The classic: the iliac apophysis: an invaluable sign in the

management of scoliosis. 1958. Clin Orthop Relat Res 2010;468:

643–53.

[11] Tanner JM, Whitehouse RH. Clinical longitudinal standards for

height, weight, height velocity, weight velocity, and stages of puberty.

Arch Dis Child 1976;51:170–9.

[12] Tanner JM, Whitehouse RH, Hughes PC, Carter BS. Relative impor-

tance of growth hormone and sex steroids for the growth at puberty of

trunk length, limb length, and muscle width in growth hormone-

deficient children. J Pediatr 1976;89:1000–8.

[13] Tanner JM, Whitehouse RH, Marubini E, Resele LF. The adolescent

growth spurt of boys and girls of the Harpenden growth study. Ann

Hum Biol 1976;3:109–26.

[14] Greulich WW, Pyle SI. Radiographic atlas of skeletal development of

the hand and wrist. Stanford, CA: Stanford University Press, 1959.

[15] Sanders JO, Browne RH, McConnell SJ, et al. Maturity assessment

and curve progression in girls with idiopathic scoliosis. J Bone Joint

Surg Am 2007;89:64–73.

[16] Reed RB, Stuart HC. Patterns of growth in height and weight from

birth to eighteen years of age. Pediatrics 1959;24:904–21.

[17] Reynolds EL, Wines JV. Physical changes associated with adoles-

cence in boys. AMA Am J Dis Child 1951;82:529–47.

[18] Little DG, Song KM, Katz D, Herring JA. Relationship of peak

height velocity to other maturity indicators in idiopathic scoliosis

in girls. J Bone Joint Surg Am 2000;82:685–93.

[19] Marshall WA, Tanner JM. Variations in pattern of pubertal changes in

girls. Arch Dis Child 1969;44:291–303.

[20] Zacharias L, Rand WM. Adolescent growth in weight and its relation

to menarche in contemporary American girls. Ann Hum Biol

1986;13:369–86.

[21] Sanders JO, Browne RH, Cooney TE, et al. Correlates of the peak height

velocity in girls with idiopathic scoliosis. Spine 2006;31:2289–95.

[22] Ryan PM, Puttler EG, Stotler WM, Ferguson RL. Role of the triradi-

ate cartilage in predicting curve progression in adolescent idiopathic

scoliosis. J Pediatr Orthop 2007;27:671–6.

[23] Risser JC. The iliac apophysis; an invaluable sign in the management

of scoliosis. Clin Orthop 1958;11:111–9.

[24] Izumi Y. The accuracy of Risser staging. Spine 1995;20:1868–71.

[25] Lee TS, Chao T, Tang RB, et al. A longitudinal study of growth

patterns in school children in Taipei area I: growth curve and height

velocity curve. J Chin Med Assoc 2004;67:67–72.