Introduction

Within the context of applications for international protection, “assessment of facts and circumstances” regarding the not (reliably) documented identity age-assertion ‘unaccompanied minority’, which equals the claim to far reaching, legal “guarantees” dedicated to “vulnerable” children “below the age of 18 years” [21, 23], belong to the most difficult issues administrative authorities have to deal with [22, 25, 26, 29]. Due to obvious limitations of the basic procedural means of evidence (“personal interview” [21], observation of appearance/demeanor) and to tackle the problem of intentional age-“misrepresentation” [66, 90], EU-Asylum law offers EU-Member States since 20 years the option of assigning age diagnostic expert opinion in case of doubtful minority allegation [67]. The “reliable” standard of such evidence gathering as asked for in Art. 25 para 5 Asylum Procedure Directive 2013 (APD) is summarized within the recommendations of the international and multidisciplinary “Study Group on Forensic Age Diagnostics” (AGFAD) [21, 73, 74]. Following this acknowledged, methodical “state of the art” [1, 4, 9, 84, 92], the age marker “medial clavicular ossification” (MCO) constitutes the decisive factor due to “a uniquely prolonged period of growth-related activity” [10, 83] leading well beyond the respective age limit. According to relevant reference studies comprising different imaging techniques, its sub-final and final developmental phases are solely found after the completed 18th year of life thus allowing for exclusion of age minority “beyond reasonable doubt” in individuals of both sexes [74, 77,78,79, 94, 99]. To display the anatomic site as accurately as possible, AGFAD-recommendations suggest the use of the imaging modality “sternoclavicular computed tomography” (CT) [73, 74].

Yet, the medial clavicular ossification site is prone to significant shape deformations affecting medial metaphyseal ending as well as the medial, secondary ossification process. Such characteristic not rarely renders categorization in line with the two accepted typologies of Schmeling et al. and Kellinghaus et al. difficult, which term its developmental metamorphosis towards adult state [44, 45, 72]. Accordingly, Wittschieber et al. proved that “an inexperienced examiner, who is unaware of the diversity of variants of the medial clavicular epiphysis, apparently tends to classify more cases than actually possible” coincidently being recognized as the most frequent systemic error of age-related clavicular evaluation [96]. Since a respective overview appears missing, our survey sums up not-assessable shape variants of the medial clavicle from a large sample to support expert’s distinction between assessable variations of ideal typical stage norms vs. not-assessable deformations.

Materials and methods

The results of our trial are drawn from the clavicular pairs of 2820 age-disputed, male “unaccompanied minor asylum seekers” allocated consecutively to our interdisciplinary work group for medical age assessment between July 2011 and September 2018 by the “Austrian Federal Office for Immigration and Asylum” and the “Austrian Administrative Court” based on § 2 para 25 Austrian Asylum Act and § 13 para 3 BFA-Procedural Law having completely implemented the provisions of Art. 25 para 5f APD 2013 and of Art. 22 Radiation Protection Directive 2013 [2, 3, 21, 24]. In all cases, the methodological dicta of AGFAD-recommendations were observed as demanded expressis verbis by Austrian Asylum law since 2010 [73]. As the phenomenon “doubtful unaccompanied minority assertion” within asylum procedures shows male predominance by 90% and for the sake of a homogenous sample, we did not enclose females into our study. The inclusion criteria of our test are provided in Table 1.

Table 1 Inclusion criteria of our sample in line with [21, 73]
Full size table
Table 2 Numbers (n) of included individuals (Ind) and resulting single ESC (sc), of individuals with anomalies (Ind + anom) separated into bilateral (Ind + bianom) and unilateral anomalies (Ind + unianom) plus respective ratios per main reported countries of origin (CoO)
Full size table
Table 3 Morphological structures of medial clavicular ossification relevant for stage attribution in line with [44, 45, 73, 100]
Full size table
Table 4 Typological template of not-assessable shape variants of the medial clavicle (n number, R right, L left, ax axial plane, cor coronal plane)
Full size table

In line with EU asylum statistics, over 90% of our trial reported six countries of origin [27, 28]. However, individuals from Afghanistan (1889, 66.99%), Nigeria (287, 10.18%), and Somalia (207, 7.34%) accounted for over \( \raisebox{1ex}{$5$}\!\left/ \!\raisebox{-1ex}{$6$}\right. \) of the sample (Table 2), which clearly limits the significance of results concerning the other three mentioned source countries due to low numbers of observed clavicles (Pakistan, Gambia, Algeria).

Table 5 Numbers (n) and percentages of single ESC-anomalies according to the typological template per main reported countries of origin
Full size table

Multidetector computed tomography (MDCT) studies of sternoclavicular regions were obtained using a 16-row MDCT system (SomatomSensation 16, Siemens Medical Solutions) with the technical settings: 130 kV, mAs values by using CareDose, Pitch 1, collimation 0.6, recon increment 1.0, Kernel B60s, SL 1 mm. Images were viewed as axial (ax) sections and coronal (cor) reconstructions depicting extremitas sternalis claviculae (ESC) as two frame-stripes of anterior/posterior and cranial/caudal orientation. Illustrations were gauged on screen of a diagnostic workstation with Siemens software (SyngoSomaris 5VB20B) by two examiners operating on longstanding experience of classifying MCO within the context of medical age assessment [96, 100]. We always drew our observations from all CT-slices of each case [100], since the slightly oblique position of the clavicle relative to ax and cor CT-sections together with the non-geometric character of the medial, clavicular shape [57] as well as secondary, medial calcification result on a regular basis in heterogenic ESC-profiles (Fig. 1). In addition, we made use of 3D-rendering software provided by Siemens as additional CT-data presentation in about 200 cases, because such visualization may facilitate recognition of complex surface morphologies including soft tissue calcifications, both of which are sometimes difficult to understand from ax and cor CT-sections [68].

Fig. 1
figure 1

Two examples of <MMC> with differing aspects in axial and coronal sections due to the non-geometric character of ESC

Full size image

We differentiated medial clavicular advents according to the two complementary typologies of Schmeling et al. (main stages 1 to 5) and Kellinghaus et al. (sub stages 2a–2c, 3a–3c) [44, 45, 72]. As not-assessable anatomic shape variants, we defined appearances offering obscured representations of the eponymous morphological criteria for stage attribution (Table 3) [100]. Specific features like deep rhomboid fossa usually not touching the medial ossification site [58], polygonal epiphyseal ossification reliefs being frequently encountered (Fig. 2) [95] and partly fused, yet small epiphyses [98] were not considered as such. CT-series infringed by motion artifacts plus rare cases of Allman III fractures and exostotic alterations of the medial clavicle were excluded [14, 39]. We did not see congenital malformations [52, 81], clinical testing never revealed dysfunction of shoulder/arm movements in any of our included cases.

Table 6 Thin-slice CT studies addressing medial clavicular development
Full size table
Fig. 2
figure 2

Example of an ESC-face with single, partly fused, map-like, medial, secondary ossification structure next illustrations of more than one epiphyseal kernel in cases without metaphyseal concavities as seen from 3D-renderings (type B)

Full size image

In the past, two general ESC-appearances were labeled as not-assessable shape variants, a metaphyseal trait and an epiphyseal one [96, 100]:

  • Medial metaphyseal concavities (<MMC>) equalling a substantial loss of subchondral bone structure inside the medial cortical tube of various form, extent and depth (metaphors from 2D-imaging: “fish-mouth”, “wrench”, “bowl”, “funnel”, etc.

  • More than one, medial, endochondral, secondary ossification centers unconnected to each other in all of the multiple CT-sections of ax and cor picture planes (syn. “multiple epiphyseal ossification centres”, <MEOC>).

Proceeding from that, we built a typological template with standardized descriptions of these two generic features and its varieties ranking those as “type A” (<MMC>) and “type B” (<MEOC>) (Table 4). In a subsequent step, we integrated all 2820 cases into an IT-database, whose registry contained case-information (identification-number, reported country of origin, date of investigation, body side of anomaly), furthermore linking the entries straight to the individual CT-presentations for repeated evaluation and fast comparison. Finally, we crosschecked atypical ESC-appearances of our trial with the records of the initial template to qualify and quantify respective formations. As a result, we added a “type C” comprising further anomalies.

Results

Our sample consisted of 2820 male borderline adults (Table 2), out of which 240 individuals (8.51%) presented bilateral anomalies and 326 (11.56%) unilateral ones (Σtot = 566, 20.07%). One hundred fifty-six persons (65% of 240) with bilateral anomalies showed the same typus both sided (Fig. 3) and 84 (35%) divergent ones. One hundred seventy (52.15% of 326) unilateral deformations were found on the right side and 156 (47.85%) on the left one vis-à-vis to all nine regular types of MCO-classification from stage 1 down to stage 5. Altogether, from 5640 (= 2 × 2820) observed single ESC 806 (= 2 × 240 + 326, 14.29%) anatomic shape variants were extracted, distributed as almost equal numbers between the ESC-antimeres (Table 4). Hence, we saw no lateral preference regarding formation of anomalies [69, 70].

Two generic topics dominated our overview by far (93.55%) adding up to 375 (46.53%) single clavicles with <MEOC> and 330 (40.94%) <MMC> (Table 4). In addition, we saw 49 (6.08%) combinations of these two leading features and also associations of large and small concavities within one and the same ESC, the latter being listed under A.1 (Fig. 4). Two entries of our template accounted for 60% (type B.1, 40.82%, Fig. 2; type A.1.2.1.2, 19.85%, Fig. 3). Fifty-two (6.45%) shape variants refused fitting into the two main categories A and B and were listed as “type C,” which also contains unusual examples of suspected calcified articular capsules (Figs. 5 and 6) [5, 96, 100].

Fig. 3
figure 3

Bilaterally deformed, medial collar bones with metaphyseal excavations and thin, epiphyseal-like structures (type 1.2.1.2)

Full size image
Fig. 4
figure 4

Two examples comprising combinations of large and small concavities

Full size image
Fig. 5
figure 5

Two ESC-appearances not allowing for outlining staging criteria according to [44, 45, 73, 100] (type C)

Full size image

Regarding the incidence of single, medial, clavicular deformations, the main reported source countries of our clientele offered different rates. Individuals from Algeria and Nigeria showed the relatively highest counts of aberrations from the norm (18–19%) and Somalia the lowest (9.42%) with Gambia, Pakistan, and Afghanistan ranging in between (Table 5). Furthermore, we saw differing tendencies as to which anomalies were found predominantly per reported source country. For example, individuals from Somalia and Algeria presented comparatively small numbers concerning the most prominent type B.1 of our template. On the other hand, persons from Nigeria showed counts of the second most type A.1.2.1.2 below average and those from Algeria above it. Our additional type C was found mainly in people from Nigeria and Algeria and not at all in those from Pakistan and Gambia (Table 5).

Discussion

Peculiar morphologies of ESC have “always been known to anatomists and radiologists” [63]. In particular, the extraordinary looks of medial, clavicular concavities attracted regular attention and are described in literature as geode [15], subchondral sclerosis [5] or erosion [51, 87], wrench- [56], bowl- [96], dish- [30, 79] or chalice-like [46], fossa-like [77] or fish mouth depression [63], bifid- [15, 16], funnel- [78], trumpet- [41] or V-shaped [57], caved-in sternal end [61], cupped medial clavicular end [49], or fork deformity [82]. Early referrals are found. e.g.. in Stevenson, Todd/D’Errico, Flecker, Ravelli, Fischer, McKern/Stewart, Teplick et al., Jit/Kulkarni, Owings Webb/Myers Suchey, Kumar et al., Schmidt/Freyschmidt, and Keats/Anderson 1996 [30, 31, 42, 43, 49, 53, 57, 63, 75, 83, 87, 88]. Also, <MEOC> in the region of ESC have been mentioned since long [30, 50, 65, 75, 91, 95].

Fig. 6
figure 6

Suspected calcinations of the articular capsule following in part the metaphyseal contour next <MEOC> as seen from coronal sections and “cinematic rendering”

Full size image

In recent literature concerning forensic age assessment, ESC-appearances deviating from the norm are addressed repeatedly and examples are displayed occasionally. Initially, Kreitner et al. mentioned in their CT-pilot study “the persistence of the medial epiphyseal ossification center [as] the only possible diagnostic pitfall “in evaluating MCO for age assessment purposes [30, 48]. As early as 2004, Schmeling et al. and Schulz et al. saw in “norm variants (in particular funnel-shaped clavicular epiphyses) “one of the major reasons challenging reliable ESC-classification [72, 78]. On the contrary, Schulze et al. did not refer to shape variants of the medial clavicle [80]. Butting 2008 mentioned the possibility of <MEOC> [12]. Depictions of an ossified epiphyseal kernel embedded into a medial, metaphyseal concavity as seen from different imaging modalities are shown in Vieth et al. 2010 [93]. In 2011, Garamendi et al. illustrated a “fish mouth anomaly” interestingly understood as a risk of “underestimation of the stage of development of the medial epiphyses of the clavicle, if investigators are not aware of the possible existence of this abnormality” [33]. Examples of fish mouth, bowl, and trident buildings are presented by Brinkmeier and Küppers; “concaaf” and “bifid” appearances of medial clavicles are provided by De Meulenaere [11, 16, 50]. Milenković displayed peculiar ESC as preserved specimens, and Ottow et al. MRI-based pictures of fish mouth, bowl-like, and wrench-like buildings [54, 56]. Gonsior pointed out that “alone on the grounds of barely possible identification of norm variants, the suitability of sonographic determination of ossification state of the medial clavicle has to be critically considered” [34].

Fig. 7
figure 7

Partly fused, single epiphysis embedded in metaphyseal cavity forming sometimes more or less a “trident” resemblance in certain slices (type 1.2.2.2)

Full size image
Fig. 8
figure 8

Two examples of small ESC-concavities without detectable epiphyseal-like structure sometimes pretending a cyst in certain frames (type 2.1)

Full size image

In 2012, Cameriere et al. stated that “deficient knowledge of normal morphological variants becomes crucial while differentiating between normal and abnormal shape” [13]. Wittschieber et al. proved that an inexperienced examiner is tempted to classify actually not-assessable ESC-appearances being described as “bowl- or funnel-like shape variants” [96]. Such study result is corroborated by Tangmose et al. [85, 86], El Morsi et al. [19], De Meulenaere [16], and Richel [64], all of which illustrate and stage obvious cases of <MMC>. Accordingly, Wittschieber et al. concluded that reliable ESC-scoring directly depends on “high degree of specific qualification, particularly the knowledge about the diversity of anatomic shape variants,” [96] a statement that was repeated and referred to by Franklin/Flavel and Gurses et al. [32, 36].

Concerning the incidence of atypical MCO-formations, we experienced difficulties comparing our results with most thin-slice CT-studies, since those often do not deliver the number of anomalies per observed single ESC (Table 6). Furthermore, studies like Ufuk et al., Tangmose et al., and Pattamapaspong et al. report astonishingly low counts [60, 85, 89]; others like Patil et al. and Houpert et al. avoid designating figures of “unclassifiable ossification patterns” altogether [40, 59]. Still others, like El Morsi et al., El-Gerby et al., and Bassed et al. do not even mention this important ESC-characteristic seemingly oblivious of the impediment (Milenkovic et al.: “possible inadmissible stage evaluation of anatomic shape variants“) [6,7,8, 19, 20, 55].

However, we found our outcome consistent with reliable results so far transmitted. Kellinghaus et al. reported that within their trial, an assessment of the “ossification status was not possible” in 90 (15.2%) out of 592 cases “owing to [amongst other] variants of normality (in particular, funnel-shaped clavicular epiphyses)” [45]. Wittschieber et al. and Scharte et al. both elaborating on the same sample counted 180 anomalies out of 1078 single clavicles (16.69%) [71, 97]. Vieth et al. excluded 20% of MRI-based ESC-imaging from stage classification because of “norm variants [appearing as] fish mouth form, bowl form, trident form and multiple ossification centres” (61/304 single clavicles) [94]. Schmidt et al. and Schmidt et al. had to ignore 8.1% and 14.8%, respectively, of their MRI-samples, because of (among other) “unclassifiable anatomical variants (e.g., bowl-like shape) “of ESC [76, 77].

Regarding our typology of not-assessable shape variants, we did not intend to describe the morphological variability of MCO in general. Instead, we aimed at questionable stage impressions arising from deviations of the norm, which are referred to in an additional column of the template (Table 4). To that end, we concentrated on the two primary topographies <MEOC> and <MMC> already mentioned before, which dominated our summary by over 90% appearing therefore most important for MCO-distinction as “assessable” vs. “not-assessable.” Upon encountering such features, a further morphological description of the site is not required within age assessment practice. For example, concerning <MEOC>, it is not necessary to count the number of epiphyses and to consider the metaphyseal shape. The judgment suffices that medial clavicular cartilage contains more than one to refrain from proceeding towards MCO-classification. Accordingly, we did not observe epi-/metaphyseal relationships at the sight of more than one epiphysis, apart from difficulties to distinguish small, partly, or completely fused kernels from a ragged metaphyseal ending. Likewise, in case of substantial <MMC> valuation of medial, secondary ossification seems dispensable. Beyond <MEOC> and <MMC>, Wittschieber et al. highlighted the importance of specific staging criteria to categorize a certain MCO-appearance [100].

Conclusions

The medial, clavicular ossification site shows considerable morphological diversity in borderline adults reaching sometimes extreme deformations. Depending on the source countries of observed male individuals, in about 10–20% of evaluated, single, medial clavicles not-assessable shape variants have to be expected within medical age assessment practice.

We could confirm that variations of the already denoted anomalies “more than one, medial, secondary ossification centres” and “ESC-concavities” are by far the most common ones. These two generic features plus further erratic ones carry in common the impossibility to reliably outline the leading criteria for comprehensible MCO-classification.

An inexperienced examiner is tempted to draw questionable stage impressions from anatomic shape variants constituting the major systemic error of MCO-scoring. Accordingly, reliability of ESC-evaluation within medical age assessment practice directly depends on “high degree of specific qualification”, part of which is “knowledge about the diversity of anatomic shape variants” [96], a requirement accentuated by Art. 25 para 5 APD demanding that age-related “medical examinations shall be carried out by qualified medical professionals” [21].