While the embryological development of the eyelid and the lacrimal system has been thoroughly studied recently,1 , 2 the literature is singularly outdated with respect to the developmental anatomy of the orbit.
Orbital development is a protracted process that begins in the third week of intrauterine life, with a strong influence played by the developing eye, the brain, and even the face. Some of the complicated configurations of the early embryonic and fetal orbit result from the fact that almost all positional reference points in the developing orbit alter as growth advances. Nevertheless, the orbit grows pari passu with all 3 structures, although not exactly establishing a linear growth curve with any of them.3–5 This association still maintains stable reference points, which makes it impractical to discuss the developing orbit without alluding to the development of the eye, face, and brain.
The authors have reviewed developmental anatomy of the orbit in an orderly chronological fashion dividing the prenatal period into an embryonic period, which spans the first 8 weeks postfertilization, and the fetal period, which spans the rest from postfertilization week 9 until birth. Intense organogenetic activity characterizes the embryonic period, whereas the fetal period is hallmarked by less intense changes. Because of significant variations in the methodology used by different authors to assess fetal and embryonic ages, the authors strictly divided development into weeks and avoided use of the Carnegie staging system because the fetal period is less amenable to staging. Since gestational age (postmenstrual age) is an ambiguous term of convenience used by obstetricians, we opted for the more precise term postfertilization age because technically, at a gestational age of 1 week the embryo does not even exist yet.6 Normative values for crown-rump length (CRL), and hence approximate postfertilization age in weeks are used.7 Although this article discusses prenatal orbital development and ends the discussion at birth, birth does not mark a precise landmark in the development of the orbit as the structure continues to grow in its physical dimensions during the first few years of life.4 , 8
Orbital morphogenesis is multifaceted and complex, and the literature is full of deeply entrenched misconceptions and uncertainties about the timing and nature of specific events, and the origin or fate of several structures.9 In the ensuing discussion, therefore, we highlight only the most significant developmental processes occurring week by week with a particular emphasis on fundamental and controversial issues. The molecular bases and the clinical implications will be discussed in a separate article.
THE EMBRYONIC ORBIT
Week 3 (CRL 1–2.5 mm).
In the third week postfertilization and specifically on day 19, the neural plate appears.10 This represents the first step in the genesis of the entire nervous system including the eyes and orbit. Once the neural plate has formed, both of its lateral borders elevate forming the neural folds, and these move toward each other in the midline.10 When these come into contact, the central part of the neural folds begins to fuse at the neck region, and this process proceeds bidirectionally cranially and caudally,10–12 forming a structure that looks like an elongated deflated balloon. While the neural tube and the facing surfaces of the large neural folds are made up of neural ectoderm, they are covered by surface ectoderm, which also covers the embryo as a whole.11 , 12 The cranial end of the evolving neural tube dilates to form 3 distinct brain segments, the prosencephalon (forebrain), the mesencephalon (midbrain), and the rhombencephalon (hindbrain) (Fig. 1A). The forebrain later further differentiates into 2 parts, the telencephalon and the diencephalon.11 , 12 Even before the neural folds have completed their fusion, the initial sign of the developing eye is observed during week 3 when the optic sulci make their first appearance as an invagination on the inner surface of diencephalon. Progressive deepening of the sulci produces the optic evaginations, which later develop into full-fledged optic pits.13 There is still no indication of an orbit proper except for a loose network of highly undifferentiated mesenchymal cells.14
At this early stage in embryonic life, pluripotential progenitor cells in the neural folds have the ability to form multiple derivative cells including the cranial neural crest cells (CNCC), which populates the embryonic head, and together with another cell type, cells of the paraxial mesoderm play a pivotal role in mammalian head development.15 CNCC are neuroectodermal cells that originate from the edge of the neural folds at the neurosomatic junction, and lose their epithelial affinities as they form a flattened mass between the neural ectoderm and the overlying surface ectoderm. The CNCC undergo a process of epithelial-mesenchymal transition and acquire a unique migratory capacity that contributes heavily to orbital and ocular development.11 , 16 Three distinct populations of CNCC (the prosencephalon, mesencephalon, and rhombencephalon) are already distinguishable by week 3 (Fig. 1A) and differentiate into a broad range of cell types.17 , 18 These include cranial bone, connective tissue, cartilage, nerves and sensory ganglia, and sclera, choroid, iris, and ciliary body.
The other important landmark event that also starts before complete closure (zipping) of the neural tube is the segmentation or rounding of the paraxial mesoderm (mesoderm immediately adjacent to both sides of the neural tube), into transient blocks of cells or segments called somites.11 , 19 Somitic segments caudal to the developing inner ear (post-otic somites) are clear cut, well-defined mesodermal blocks and are called somites.11 Mesoderm rostral to this level (pre-otic somites) forms ill-defined, incomplete rounded blocks with indistinct borders called somitomeres (Fig. 1A).11 , 19 As we shall see later, these somitomeres will be committed to the myogenic lineage in the orbit. The unsegmented somitomeres and the somites are characterized by strikingly different upstream and downstream gene expression patterns, which could explain their differential growth patterns.20
Because there are no well-defined paraxial mesodermal somites in the head, the CNCC must take up the job of providing many of the connective tissue phenotypes of the developing skull and orbit.21 Since it replaces those traditionally mesodermal elements seen elsewhere in the body, the CNCC is a major protagonist in orbital development, and we strongly believe the orbit is fundamentally a byproduct of the CNCC. When CNCC and mesoderm migrate to their destinations in the orbit, the mesoderm becomes augmented with CNCC, forming the mesenchyme,15 also referred to as ectomesenchyme, mesectoderm, or more accurately periocular mesenchyme.21 , 22 It is important to realize that whereas mesoderm specifically relates to the middle embryonic layer, mesenchyme is a loose term used to describe any embryonic connective tissue.11 The periocular mesenchyme contributes heavily to the development of the anterior segment of the eye, the eyelids, and the orbit.1 , 22
Week 4 (CRL 4–5 mm).
In week 4, the optic pits are directed outward away from the brain, and toward the surface ectoderm. The 2 lateral bulges, caused by the outward extension of the growing optic pits, become pouch-shaped outpocketings called the optic vesicles.13 At this stage, the angle between the imaginary lines passing through the optical axis of OU is 180 degrees,23 but despite the expansion of the optic vesicles, there is still no indication of an orbit proper. A constriction in the optic vesicle at its attachment to the wall of the forebrain forms the optic stalk.24 Initially short, the optic stalk will form the future optic nerve (ON).25 Originally a hollow structure with walls that are composed of a single layer of undifferentiated neuroectodermal epithelial cells,26 the optic stalk differs from the ON proper in that it possesses a wide central lumen that is freely continuous with the cavities of the optic vesicle and the forebrain on either end.26
The 3 CNCC populations that initially appeared during week 3 migrate and merge with the cephalic mesoderm to form the future face.27 , 28 In humans, CNCC contributing to the developing face arise from the diencephalon, mesencephalon, and rhombencephalon, but not from the telencephalon.12 , 18 The most rostral stream of CNCC covers the forebrain, a medial stream migrates toward the frontonasal process, while the most caudal stream (rhombencephalic CNCC) migrates toward the paired first pharyngeal arches.12 , 27 , 29 , 30 Because rhombomeric CNCC contribute only to the mandibular processes in the developing face, maxillary elements are exclusively derived from midbrain CNCC (Fig. 1B).30 Together with mesoderm, these migratory neural crest waves form a series of small buds of tissue called the facial primordia, which form around a central depressed area, the stomodeum (primitive mouth).27 The face is formed from the merging of 5 buds or facial prominences that appear in the 4th week (a single frontonasal process, and the paired maxillary and mandibular processes).27 , 31 , 32 Contrary to some textbook illustrations where they are shown to be separated by furrows, these 5 facial prominences are confluent with no clear boundary between them.27 Mesencephalic and diencephalic waves of CNCC migrate toward the optic vesicles, and later contribute heavily to the development of the eye and orbit (Fig. 1B).17 , 22 , 33 The optic vesicle plays a crucial role in guiding the migration of CNCC that are destined for the orbit.22 , 27
The very first indication of blood supply to the orbital region is observed at day 22 or 23 as a plexus of undifferentiated vessels containing nucleated blood cells that can be recognized in the periocular mesenchyme.34–37 At day 24, the internal carotid artery (ICA) emerges from a combination of the distal aorta and the branchial arch arteries. It divides into 2 major divisions near the summit of the optic vesicle: a caudal division, which forms the posterior communicating artery, and a cranial division, which curves around the base of the optic vesicle and gives off several important branches to the developing brain and face.36–38 A primitive maxillary artery originates from the future cavernous segment of the ICA.39 Several short branches appear in the vascular plexus around the optic vesicle derived from the primitive maxillary artery,40 the primitive ICA,36 , 37 , 39 , 41 and possibly the primitive olfactory artery.40
Development of orbital veins is a complicated and under-researched subject. Classical teaching indicates that small blood spaces appear within the orbital mesenchyme contemporaneous with the appearance of the arterial plexus around the optic vesicle, and that these later coalesce to form venous channels.34 However, at this stage, there is no visibly discernible difference between arteries and veins throughout the nervous system. They remain indistinguishable until dynamics of venous flow starts.42 By the end of the 4th week, a well-established primitive venous plexus encases the developing brain and is initially connected to a nonpaired, nondraining transient midline vessel within the hindbrain called the vena capitis medialis.43 , 44
Week 5 (CRL 5–7 mm Stage).
The optic vesicle undergoes active invagination onto itself producing a double-layered goblet-shaped structure called the optic cup. This invagination is circumferentially asymmetrical or unequal, which simultaneously creates a groove called the optic fissure or embryonic fissure.24 Starting at week 5, the process of frontalization of the eyes starts, whereby the eyes migrate from the initial position of extreme hypertelorism observed at week 4 (1800) toward a more frontal position. This process of lateral-to-frontal reorientation continues at an intense pace until week 9, and then slows down, but does not stop except after birth.45
At day 29, the rhombencephalon (hindbrain) is divided into distinct segments called rhombomeres (Fig. 1A,B). This is a crucial first step toward the highly ordered and quite complex genesis of cranial nerves. Each rhombomere has its own attendant CNCC, and its own set of genes and molecular pathways, which usually do not cross rhombomere boundaries, thereby directing the respective axons to a predetermined stereotypical location.46
A unique developmental pattern that sets the sensory nerves within the orbit apart from motor nerves is that they originate from 2 distinct embryonic sources, rhombomere CNCC and the ectodermal placodes, which are plate-like condensations derived from surface ectoderm. In contrast to CNCC, these placodes almost exclusively generate neural tissue, except for the adenohypophysial and lens placodes.46–48 While very complex, in general rhombomeres 2 through 6, CNCC, and the ectodermal placode variously contribute to formation of the trigeminal and facial ganglia and nerves.
The picture is not as clear-cut for other cranial nerves. The origin of the trochlear nucleus with its unusual nerve fiber exit trajectory is still a matter of debate. Most authors maintain it arises from the mesencephalon just caudal to the oculomotor nucleus,49–51 whereas others argue for a rhombomere origin from the hindbrain,52–54 or from the isthmus rhombencephali in between.46 Unlike other cranial nerves, the oculomotor nerve and the ciliary ganglion have traditionally been assumed to originate embryologically within mesencephalic CNCC,46 , 55 but recent evidence suggests a possible additional origin for the ciliary ganglion and possibly the oculomotor nerve from rhombomere 1.55
Week 5 also witnesses the earliest indication of the tripartite division of the trigeminal ganglion, and the ophthalmic division is the first to appear.17 At this stage, the ciliary ganglion also makes its first appearance in the developing orbit.56
At this stage, the blood supply to the orbit is still plexiform.41 The ICA gives off a primitive dorsal ophthalmic artery (DOA), which appears as a long branch along the dorso-caudal margin of the developing optic cup. The primitive DOA gives off 2 distinct branches, the hyaloid artery (future central retinal artery) that passes through the embryonic fissure in the optic cup, and the common temporal ciliary artery (future lateral posterior ciliary artery).36 , 37 , 39 , 41 , 57 , 58 The vena capitis medialis disappears, and the primitive cerebral venous system is now divided into paired anterior, middle, and posterior plexi.43 They are now connected to a definitive draining primitive head vein (venae capitis lateralis or primary head sinus).43 This vein has a major tributary, the primitive maxillary vein that is closely related to the primitive optic vesicle and trigeminal ganglion. Initially, this vein drains most of the primordial orbit and the olfactory region.42 , 59–61
Week 6 (CRL 8–14 mm).
At week 6, the first indication of osteogenesis of bones is observed around the developing major neural and visceral contents of the orbit.27 The 7 bones comprising the bony orbit, like most bones of skull vault and base, undisputedly originate from CNCC.27 This is in contrast to bones in the posterior portion of the skull (parietal and occipital bones), which are postulated to be of a mesodermal origin.27 , 62 , 63 Orbital bones initially form as a network of tiny bony ossification zones centrally, separated by large nonossified areas. Each is composed of a thin membranous mesenchymal capsule that acts as a template over which individual bones develop from a complex series of ossifications involving both intramembranous (direct) and endochondral (indirect) ossification patterns. When osteoblasts are laid directly in these mesenchymal membranes, membranous ossification is realized without any cartilaginous intermediary. The slower endochondral ossification involves an intermediary cartilaginous template stage followed by secondary ossification.27 , 34 Either way, as ossification proceeds, separate ossification zones gradually approach each other leaving unossified sutures in between. These sutures are true fibrous connective tissue structures developing from periocular mesenchyme, usually of CNCC origin.27 , 34 , 62 , 64
The very first bone to appear around the orbit is the maxillary bone, which is seen initially as a single ossification center or a thin strip of bone above the dental lamina of the future canine tooth.34 , 65–67 With the exception of the sphenoid and the ethmoid bones, bones in the orbit develop by direct intramembranous ossification and not by endochondral ossification,19 , 34 but the greater wing of the sphenoid bone is the most controversial and still eludes researchers (Professor Hiroki Otani, personal communication), with contradictory evidence pointing either to an endochondral68 , 69 or intramembranous origin,70 or even a dual origin (Fig. 2).71
The exact events that lead to the development of extraocular muscles (EOM) are conflicting at best, and the signals that promote or retard myogenesis of EOMs have not yet been fully identified. The earliest indication of EOMs is 2 populations of mesodermal cells. First are the rostral somitomeres of the unsegmented or poorly segmented paraxial head mesoderm, which likely contribute to the lateral rectus (LR) and superior oblique (SO) muscles (innervated by the abducens and trochlear nerves, respectively). The second population is the prechordal head mesoderm (Fig. 1A), which gives rise to the medial (MR), inferior rectus (IR), and superior rectus (SR) muscles and the inferior oblique muscle (muscles innervated by the oculomotor nerve).72 , 73 This is a thickening of endoderm at the rostral end of the notochord, some cells of which undergo transition to head mesenchyme (Fig. 1A). The role of the prechordal plate in human EOM ontogenesis remains unclear (Professor Shahragim Tajbakhsh, personal communication), but vertebrate fate mapping studies do demonstrate early migration of prechordal cells to join the paraxial mesoderm that participate in EOM formation.20 Some authors provide an alternative view claiming that the MR, LR, and IR together with the inferior oblique and the levator palpebrae muscles (LPS) arise from early myoblasts migrating from the first and second somitomeres, the SO from the third somitomere myoblasts, and the LR muscle from the fifth somitomere.11 , 15 , 20
Regardless of their exact origin, these transient early myoblasts appear to originate away from the eye field.73 Later, they migrate en masse anteriorly toward an already CNCC-enriched orbit to initiate muscle morphogenesis and rapidly give rise to myoblasts.74 Reciprocal interactions between the eye and EOMs are compulsory for later formation of primary and secondary myotubes, and mature muscle fibers,73 although it seems that these migratory waves may already contain some multinucleated myotubes before arrival into periocular territories.20 It is important to point out that the EOMs, the motor nerves that innervate them, and the CNCC that will form the connective tissue elements of these muscles all arise at separate axial locations, and do not establish stable relationships until all have reached their sites of terminal differentiation in the orbit.15 This contrasts with most other head and neck muscles where myogenic primordia, neural crest progenitors, and their respective nerves all arise at the same axial level, and move in concert maintaining this close relation throughout their journey.15 , 20 The significance of this unusual behavior in unknown,15 but the developing optic vesicle is a required prerequisite for CNCC migration to the orbit,22 and the latter is required for subsequent orbital migration of mesodermal cells. Therefore, it appears that the developing eye serves a nonvisual role during development as an organizer of orbital and craniofacial activity.75
Once inside the primitive orbit, myoblasts form superior and inferior mesodermal complexes at the apex of the orbit.74 This is followed by segregation and breakaway of individual muscles groups from these common progenitor pools.20 The superior complex forms the SR, LPS, and the SO. All 3 initially share a common epimysium.74 The inferior complex forms the IR and the inferior oblique. The MR and the LR arise from both complexes.74 Although all 6 EOMs are distinguishable during week 6, their tendons are not yet discernible.34 , 74 , 76
By the end of the sixth week, the growing edges of the optic fissure fuse; a process that starts centrally and progresses in both directions,24 enclosing the hyaloid artery and associated mesenchyme in the center of the optic stalk. Thin, elongated mesenchymal cells presumably derived from the CNCC start to surround the optic stalk toward the end of the 6th week. These cells will form the future ON sheaths.25 While the pia and arachnoid are entirely derived from CNCC, the dura appears to be exclusively mesodermal or has a dual mesodermal/CNCC origin.77
Due to the precocious growth of the cranial nerves, the rhombencephalon occupies almost one-half of the developing brain, and all cranial nerves from III to XII are now easily recognizable. The trigeminal fibers form 3 noticeably thick bundles.46 , 78 Toward the end of week 6, all the main branches of the facial nerve can also be discerned, but the main facial nucleus is not yet apparent.46
Development of the adult ophthalmic artery (OA) is rather delayed and extremely complex. The adult OA results from the simultaneous regression, anastomosis, and fusion of 3, and possibly 4 or even 5 different transient arteries (DOA, ventral ophthalmic artery [VOA], stapedial artery (SA), primitive maxillary artery, and the primitive olfactory artery.40 During week 6, the plexiform arterial plexus investing the optic cup and supplying the orbital region becomes extensive.41 This rich plexus is supplied by 2 ophthalmic arteries that can be clearly identified in the orbit, the larger DOA that already appeared at week 5, and a smaller VOA that arises later and at a higher level.36 , 37 , 40 , 41 , 79 The VOA provides the common nasal ciliary artery (future medial posterior ciliary artery). The primitive maxillary artery regresses during week 6, but its lateral branch may persist throughout adult life as an anastomotic branch between the OA and the ICA.40 Another significant vasculogenic event during week 6 is that the SA makes its first appearance.41 The SA arises from the primitive hyoid branch of the petrous ICA and is a major supplier of arterial blood to the primordial orbit. 80–82
The evolution of orbital vasculature is not only complex but heavily contested as well.36 , 37 , 39 , 41 , 79 , 80 Padget’s formidable anatomic study showed that the development of the DOA precedes the VOA, both are derivatives of the ICA, both access the primordial orbit through the optic canal, and both exclusively supply the eye and not the orbit.41 An alternative interpretation maintains that both the DOA and VOA develop simultaneously, but that the VOA arises from the anterior cerebral artery, and not from the ICA. It also proposes that the VOA enters through the optic canal while the DOA enters through the superior orbital fissure (SOF), and that the DOA contributes to the orbital blood supply, while the VOA supplies the eye.79 A more recent view, however, has corroborated Padget’s original findings.82 That study maintains that the VOA and the DOA both pass through the optic canal and are more concerned with vision, while the blood supply to the primordial orbit is primarily derived from the superior ramus of the SA (SRSA), which enters the orbit through the SOF.82 We believe that in light of the extensive variations of arterial supply to the orbit observed in clinical practice, both theories may be correct, and the discrepancy may simply be explained by anatomical variations.
A small vein, the supraoptic vein (also referred to as the primitive supraorbital or superior optic vein) arises from the superficial tissues cranial and dorsal to the eye and is closely related to the developing ophthalmic division of the trigeminal nerve.59–61 Like the maxillary vein, it also drains into the primitive head vein through the anterior dural plexus,42 , 59–61 but the orbit still predominantly drains into the maxillary vein.
Week 7 (CRL 16–18 mm).
Except for a single article, no detailed anthropomorphic data are available for the face and orbits at this very early stage in embryonic development.45 According to its authors, the width of the central-most facial zone (interorbital distance) occupies a huge percentage of the overall size of the face, contributing about 60% to the facial width.45
Paradoxically, as the optic stalk lengthens, it initially becomes thinner with a diameter of 0.5 mm.83 The previously hollow tubal lumen is now progressively occupied by axons that originate from ganglion cells of the retina (RGC) and gradually penetrate the optic stalk through a defect in the pigment epithelium to surround the hyaloid artery.25 , 34 , 56 At the molecular level, several molecules repulsive to axonal growth, and axon pathfinding mechanisms work in tandem to ensure proper projection of axons away from the eye, through the optic stalk en route toward the brain.84 , 85 A chondroitin sulfate proteoglycan pathway acts as an inhibitory barrier to prevent axon misdirection toward the eye. An axon guidance molecule (netrin-1) has been identified on neuroepithelial cells surrounding exiting RGC axons at the disk, associated with netrin receptors on RGC axons.84 Concomitant with this process, the neuroectodermal cells that initially occupied the optic stalk start a process of natural cell death and disintegration.56 The elongated mesenchymal cells that started to encircle the optic stalk at week 6 become a single compact layer that will subsequently form the future pia, arachnoid, and dura mater surrounding the ON.13 , 25 , 34
Through a complicated anastomotic process whereby the proximal portions of the DOA and the VOA regress, the permanent stem of the OA finally appears,36 , 37 , 85 and gradually annexes the ocular branches of the primitive DOA. Later, annexation of the VOA by the permanent OA establishes all the adult ocular branches of the OA, but no orbital branches appear yet.36 , 37 Besides the SRSA, which at this stage is still the major supplier of blood to the embryologic orbit, the SA gives off an inferior branch, called the ramus inferior or the maxillo-mandibular branch.81 , 82 The SRSA is also sometimes referred to in the literature as the supraorbital branch of the SA and should not be confused with the adult supraorbital artery.81 , 82 The SRSA can enter the orbit either through the SOF or through the meningo-lacrimal foramen (Hyrtl’s foramen) in the greater wing of the sphenoid.40 , 86 , 87 The greater sphenoid wing does not ossify until week 10, so that the position of this foramen is variable.87 The hyaloid arterial system is now fully developed.57
A branch of the primitive maxillary vein located within the optic stalk is a likely precursor to the central retinal vein.60 , 61 This vessel has occasionally been referred to as the hyaloid vein, but it is not homologous to the true hyaloid vein of lower vertebrates.
Toward the end of week 7, the muscle belly of the SO becomes clearly visible, initially running a straight course along the lateral wall of the nasal capsule without any change in direction.88 The most anterior portion of the SO belly is still barely discernible from the sclera, but the trochlea cannot be identified.88 There is still no evidence of orbital fat, connective tissue, fascia, or muscle sheaths.69
There is controversy regarding the temporal and cellular origin of the human lacrimal gland. In the mouse, a single bud-like invagination of the conjunctival epithelium adjacent to the temporal extremity of the eye is the earliest hint of the development of the future lacrimal gland.9 , 89 In humans, this event begins with epithelial-mesenchymal thickening at the level of the superior conjunctival fornix at week 7 (16–18 mm, 46th day).90 Although the mesenchyme is clearly of CNCC origin, the cell of origin of epithelial components of the lacrimal gland remains controversial.90 Earlier authors91 placed the origin of these epithelial components from the CNCC, but more recent evidence confirms that they derive from surface ectoderm and not from CNCC.13 , 90 , 92 Development of the lacrimal gland is an archetypal example of a morphogenetic epithelial-mesenchymal interaction and has been studied extensively in the mouse.93 Epithelial differentiation into acinar, ductal, and myoepithelial cells is controlled by fibroblast growth factor (FGF10), and although their exact role in lacrimal gland morphogenesis is not yet clear, evidence suggests that myoepithelial cells play a pivotal role,94 , 95 and that this central role is not limited to the fetus alone. It seems that these enigmatic cells hold the key for maintenance and repair of lacrimal gland function throughout adult life as well.94
Week 8 (CRL 23–31 mm).
This is the most critical week in orbital and facial development, and while 1 week may not seem much in relation to the whole human life span, significant day-to-day changes occur throughout this defining week.45 Reorientation of both orbits to a more medial position continues at a rapid pace, but the width of interorbital distance is still sizable.45 Predictably, there is a concomitant increase in all dimensions and zones of the growing face.45 Toward the end of the 8th week, the formation of most of the embryonic face is complete, but because the orbital axes have not yet stabilized, the face may not become fully recognizable except a week later.16 , 31 , 69
It is also during this week that intramembranous ossification of the frontal bone begins anteriorly close to the supraorbital convexity and progresses posteriorly concurrently with the rest of the face.65 , 69 , 96 The lesser wing of the sphenoid (aka, ala orbitalis or orbitosphenoid) initially appears in week 8 as a well-differentiated cartilaginous center lateral to the ON.96 Between the frontal bone and the lesser wing of the sphenoid, a small cartilaginous structure is seen extending from the ethmoid precursor medially toward the ala orbitalis. Although de Haan and Willekens65 hypothesized that this cartilaginous precursor is the forerunner of the lesser wing of the sphenoid and the ethmoids, a more recent study suggested that this sphenoethmoidal cartilage is a transient structure and progressively disappears without ossifying (Fig. 2).96 Intramembranous ossification of the zygomatic and palatine bones is also observed during this week.65 The floor of the orbit is separated from the pterygopalatine fossa by the primordial Müller orbital muscle, an enigmatic structure with an uncertain cell of origin.97
As the optic stalk becomes occupied fully with axons, it reaches the diencephalon enclosing the hyaloid artery. The definitive ON has now formed, which is a relatively late development in the embryonic period.25 The estimated number of axons at this time is around 1.9 million.98 Faint suggestions of the dura, arachnoid, and pia mater are just beginning to take shape.34 En route toward the brain, the developing ON picks up glial cells derived from the neuroectoderm. Two glial cell lineages populate the nerve, the astrocyte precursor cells, and the oligodendrocyte precursor cells. These in turn produce 3 different macroglial cell populations. The astrocyte precursor cells develop locally from primitive neuroepithelial cells that initially colonize the outer layer of the optic stalk and produce type I astrocytes.34 , 99 The oligodendrocyte precursor cells originate from the ventral forebrain and produce both the oligodendrocyte and the astrocyte type II glial cells.26
Week 8 also witnesses the appearance of undifferentiated microglial cells interspersed within bundles of ON axons, and macrophages in the meningeal layers of the orbital ON.100 Unlike the macroglial astrocytic precursors, microglia, and macrophages derive from a hemopoietic origin and not from a neuroectodermal origin.101
According to Mann,35 week 8 witnesses the first appearance of the oculomotor nucleus, which consists of 2 groups of cells on either side of the midline in the basal plate of the mesencephalon.
Once inside the orbit, the SRSA accompanies the ophthalmic nerve (V1), and gives off corresponding arterial branches that accompany the lacrimal, frontal, and nasociliary nerves.40 , 81 It then anastomoses with the stem of the OA forming an arterial ring around the ON.40 , 87 The inferior ramus of the SA (maxillo-mandibular artery) further branches into the infraorbital artery and the inferior alveolar artery that accompany the maxillary (V2), and mandibular divisions (V3) of the trigeminal nerve, respectively.36 , 37 , 82
The predominant drainage role of the primitive maxillary vein is supplemented with other veins as it starts to cede its territorial orbital tributaries. Concurrently, the supraoptic vein, which will become the superior ophthalmic vein (SOV) in adult life, enlarges and elongates anteriorly and slightly downward. Through a complicated anastomotic process that reroutes and links the supraoptic vein to the receding main trunk of the primitive maxillary vein, the supraoptic vein assumes the dominant role in orbital venous drainage.42 , 59–61 This rerouting may explain the unusual trajectory of the adult SOV.59 Remnants of the maxillary vein become the inferior ophthalmic vein (IOV) of adult life.60 , 61 Most portions of the primitive head vein recede and disappear. The only part that persists is the pro-otic sinus, which continues into adult life as the cavernous sinus. Both the SOV and the IOV drain into the cavernous sinus either separately, or more commonly together after the IOV joins the SOV.42 , 59–61 A recent cadaveric study102 corroborated Padget’s findings that the IOV more commonly joins the SOV and drains into the cavernous sinus with a common trunk, which Padget termed the common orbital or orbito-ophthalmic vein. What this means is that the cavernous sinus only drains orbital extracranial veins (SOV and IOV). This exclusive relation lends embryological support to the modern anatomical treatise that refers to the orbital apex and the cavernous sinus as a single anatomic transition zone, the so-called lateral sellar orbital junction.103
The epithelial-mesenchymal thickening at the superior conjunctival fornix now takes on a nodular shape, and 5 or 6 rounded epithelial buds subsequently invaginate into the surrounding mesenchyme forming the primordial lacrimal gland.90 Toward the end of week 8, the primordial orbital lobe of the lacrimal gland receives the lacrimal artery and lacrimal nerve. The nerve enters the gland through its posteromedial end.90 This vascularization and innervation process precedes the development of acinar lumina, which appear at the very end of week 8, coinciding with the conclusion of the process of eyelid fusion.90
Histological sections from Leo Koornneef’s milestone article about the development of orbital connective tissue clearly show that all EOM bellies are already clearly seen by week 8 (Fig. 2).69 Toward the end of week 8, the tendons of the rectus muscles are also observed microscopically to be attached to the apical perichondrium (future periosteum).76 Similar to the muscle bellies, Sevel76 favored a mesodermal origin for the tendons of the EOMs and not a CNCC origin. Near its origin, the MR muscle fuses with the dura of the ON, and even at this early stage, it is stronger and more fully developed than the LR.69 , 76 The trochlea develops from the mesenchymal tissue at the superomedial corner of the orbit as a small plate-like mesenchymal condensation, which quickly becomes semilunar in shape.56 , 76 , 88 Initially muscular in nature, the origin of the inferior oblique is attached to the periorbita adjacent to the bony opening of the nasolacrimal duct.76 The LPS muscle, which is the last muscle to appear in the orbit,72 , 104 originally develops from the superior mesodermal complex and is initially indistinct from the SR and the SO.76 , 105 The LPS first appears as a mesenchymal condensation that begins to separate from the medial aspect of the SR by a method of differential growth (Fig. 2). Throughout most of the fetal period, the LPS shares a common epimysium with the SR, which is initially quite dense and well demarcated.
THE FETAL ORBIT
Week 9 (CRL 40 ± 2 mm).
As evidenced from anthropomorphic measurements, it is at week 9 that the rate of the lateral-to-frontal reorientation of the orbits reaches its peak, and then starts to slow down. This leads to a proportional decrease in the interorbital or inner canthal distance in comparison with the overall width of the face. This brings together and consolidates the major facial primordia, giving rise to an anthropologically recognizable human face for the first time. The interorbital distance now constitutes only 20% of the overall facial width, an abrupt decline from a value of 60% at week 7.45 Second and third trimester ultrasonographic studies show that frontalization continues and reaches its postnatal value of 68° to 71° by the time of birth.92 , 106 , 107 This rotational process continues even after birth particularly during the first year of life.108 During week 9, a median nucleus appears, which connects the 2 lateral nuclei of the oculomotor nerve.35
Near the end of week 9, the origin of the LR straddles the SOF with 2 heads. The superior and inferior heads arise from the superior and inferior muscle complexes, respectively.76 Whether this dual head origin is a constant finding or not remains controversial, but recent anatomic studies have validated this embryologic view and demonstrated that in adult cadavers the LR arises by 2 separate heads, which delineate the oculomotor foramen.103 , 109 , 110
By now, the OA has attained approximate adult configurations.39 The SRSA starts to regress and the stem of the OA annexes the SRSA acquiring its orbital branches to become the adult OA.36 , 37 , 80 , 81 Thus, all the orbital branches of the adult OA are derived from the SRSA. The arterial ring around the ON, which appeared during week 8, also involutes. The course of the OA, whether above or below the ON, depends on whether the superior or the inferior part of this ring regresses.41 , 87 Subsequently, the connection between the intracranial and intraorbital portions of the SA is lost and the intracranial segment is annexed by the external carotid artery. This persists into adult life as the middle meningeal artery.36 , 37 , 41 , 111 Failure of the SRSA to regress results in its persistence into adult life as the orbital branch of middle meningeal artery (the meningolacrimal artery), or the recurrent meningeal branch of the OA, both probably embryologically the same vessel.86 , 87
Week 10 (CRL 55–68 mm).
The orbital plate of the frontal bone is now clearly ossified particularly medially.69 , 96 Ossification of the lacrimal and orbital plates of the greater wing of the sphenoid begins at this time.27 , 65 , 69 The bones of the orbit are not yet completely ossified, nor have they made contact with each other, so that the sphenofrontal suture separating the frontal bone from the lesser and greater sphenoid wings is still very wide (Fig. 3). It occupies a vast area contributing most of the roof and lateral wall of the orbit.69 , 96 The uniqueness of this suture lies in the fact that it is a chondromembranous junction between the frontal bone (membranous ossification), and the greater and lesser wings (probably both arise through endochondral ossification).96 By now, the transient sphenoethmoidal cartilage has almost completely regressed (Fig. 4). The significance of this cartilage is unknown, but it seems to act like a supportive framework of the orbital roof before maturation of the sphenofrontal suture96 akin to the role played by the Müller muscle in the floor of the orbit.
A well-defined pia matter surrounding the ON is clearly identifiable as a separate layer.25 While the pia mater and arachnoid investing the ON are homologous to the choroid, the sclera is homologous to, and indeed continuous with, the dura mater surrounding the ON.24 The parasympathetic Edinger-Westphal nucleus differentiates from the median oculomotor nucleus as paired masses of neuroblasts dorsolateral to the main nucleus.35 , 112 , 113
Most EOM mesenchymal cells have differentiated into myoblasts,69 and EOMs have started to disperse away from each other, resulting in an orbital mesenchyme that appears less compact.69 The tendon of the SO, which until now was still parallel to the cornea, begins to curve posterolaterally immediately medial to the primitive trochlea. The LPS acquires its adult shape, and although it still lies medial to the SR it is now gradually moving laterally, slightly overlapping the medial edge of the SR.72 , 104 With the exception of the LPS, all other EOMs have attained their adult position and architecture.
A milestone in orbital development during the 10th week is that the Müller orbital muscle acquires the appearance of a well-developed muscle plate that occupies more than half of the orbital floor.69 At this early phase of fetal life, the inferior orbital fissure is initially very wide since intramembranous ossification of its delimiting bones is still far from complete. Thus, the Müller muscle essentially constitutes the orbital floor and apparently affords a protective role for the developing eye and orbit.114 , 115 The orbital aspect of the muscle is slightly concave and is closely related to the IR muscle and the inferior division of the oculomotor nerve.97 , 114 The inferior surface of the muscle is in close proximity to infraorbital nerve, while the maxillary nerve courses through the body of the muscle.97 , 115 The IOV runs along the posterior half of superior surface of the muscle, and some muscle fibers even follow the vein as it empties into the anterior confluence of the cavernous sinus (where the IOV and the SOVs join). This intimate relation with the IOV and the cavernous sinus suggests a possible influence on autonomically mediated vascular dynamics, specifically, constriction of venous flow.34 , 116 Müller muscle fibers extend posteriorly toward the SOF, and laterally along the lateral orbital wall as is also clearly seen in the adult orbit.34
The lacrimal gland starts to show morphological characteristics similar to the adult lacrimal gland.90 , 117 For the first time, it is divided into palpebral and orbital lobes by the emerging lateral horn of the levator aponeurosis.90
As the hyaloid artery becomes fully developed, it contemporaneously starts to show its very first signs of regression.57 The maxillary vein continues to recede its facial and pharyngeal drainage, while its inferior drainage is captured by the external jugular vein.42 , 43 Abundant vessel-like structures are already prominent in the primordial orbit, which are considered a precursor step toward organogenesis of orbital fat, but actual fat is still lacking.69 , 115
Week 11 (CRL 87 ± 8 mm).
Thanks to the pioneering work of Haas and associates in 1993, orbitometric measurements are available from the 11th week until full-term.118 If observed individually, their measurements show a linear growth curve for the orbital width, height, depth, and volume. However, taken together, these changes in the orbital measurements are not proportional. This causes the orbital index (height/width × 100) to express a nonlinear increase as the fetus grows. The shape of the orbital entrance therefore evolves from a chamaeconchal shape (low or rectangular orbit) at week 11 to a hypsiconchal (higher, rounded shape) at full term because the height is approximately equal to the width at birth.118
After weeks of extensive remodeling with formation of new vessels and resorption of old ones, the orbital and cerebral venous systems assume a definitive adult-like configuration.42 , 43 , 59–61
Week 12 (CRL 98 ± 9 mm).
By week 12, both orbits are getting closer together, but the angle between them is still 105°,31 and the mean transverse fetal orbital diameter is 5.2 mm.119 From the 12th week onward, a linear growth curve is observed between the age in weeks and orbital diameter.119–121 Another valuable spatial dimension that closely correlates with the growth of orbital bones is the interlens distance. Kivilevitch et al.122 found a parallel increase of interlens distance and bony orbital diameter, with a fixed growth ratio of 1.5 from 10 to 35 weeks postfertilization. They proposed that the interlens distance is an important marker for ocular and orbital growth in the fetal period, similar to the interpupillary distance in the postnatal period for the evaluation of ocular abnormalities.
The LPS has changed its position and is now moving superolaterally but is still firmly connected to the SR.69 The levator aponeurosis is observed expanding downward in the anterior orbit, dividing the lacrimal gland into orbital and palpebral lobes.90 Connective tissue condensations extend from the SR/LPS complex to the LR. The spaces between the EOMs have broadened significantly. Each muscle is surrounded by a layer of collagenous fibers that run parallel to the muscle, and which represent the primordial muscle fascia.69
Starting from the 12th week onward, there is a dramatic increase in vascularity of the ON.123 This process of intense vasculogenic activity continues until the 14th week and then gradually tapers. Within the orbit, mesenchyme, capillaries, and thin vessels are present medially and inferiorly,69 and this complex probably represents the future fat (Fig. 3).
Week 13 (CRL 109 ± 10 mm).
At 13 weeks, the mean fetal orbital diameter is 6.0 mm.119 Ossification proceeds significantly, and the unossified cartilaginous precursor of the ethmoid bone develops 3 outgrowths representing the future conchae.69
Müller orbital muscle directly separates the orbit from the rest of the developing face, but the sphenopalatine fossa is still continuous with the infratemporal fossa and the parasellar area (cranial cavity) without delineation.115 The SO tendon continues curving laterally until it initially becomes attached to the superomedial part of the sclera. It is possible that frontalization of the eyes, or possibly rotation of the eyeball along its own axis may play a role in this unusual insertion of the SO tendon during early fetal life.88 In the frontal bone, a fossa for the trochlea is not yet identified, but fascial bands are seen connecting the trochlea to both the frontal bone and the sclera.88 Thus, although in adults the trochlea is firmly attached to the frontal bone, formation of the trochlea seems to be independent of frontal bone development.88 Instead, because the trochlea is indistinguishable from the sclera prior to week 8, Katori et al.88 hypothesized that the trochlea might develop from a common analge within the sclera. The LPS finally reaches its adult position above the SR.69 Mesenchymal condensations around the EOMs are much more prominent, with strands connecting the SR/LPS complex to the SO, the MR, and the LR. Similarly, the LR is connected to the IR.69 These strands are likely the precursors of the intermuscular septum that is seen with some discontinuities in adults.
The thin mesenchymal vessels that were first seen during week 10 are more prominent and are now immersed in loose relatively homogenous tissue suggesting the initiation of fatty tissue differentiation.115 , 124 As we mentioned earlier, adipogenesis is closely coupled to blood vessels development, and the very first indication of orbital adipogenesis is the proliferation of primitive blood vessels followed by the aggregation of a dense mass of mesenchymal cells (preadipocytes) (Fig. 3).124 Both findings are early signs of fatty tissue differentiation throughout the body and not just in the orbit.124 For decades, immunohistochemical and electron microscopic studies have indeed repeatedly confirmed a perivascular origin of preadipocytes.125 More interestingly, the expression of preadipocyte markers by capillary endothelial cells in embryonic, fetal and even adult adipose tissue, suggests a possible role for capillary endothelial cells themselves as the “birth place” for adipocyte precursors.125
The exact mesenchymal progenitors of adipocyte cell lineages in the orbit are largely unknown. Although in the trunk and limbs, adipocytes are traditionally regarded as having a mesodermal origin,126 in vivo and in vitro studies have demonstrated recently that in the cephalic region, preadipocytes display an overrepresentation of CNCC-associated genes, suggesting a conceivable neural origin for cephalic fat from CNCC, and refuting the long held dictum that mesoderm is the sole precursor of adipocytes throughout the body.127 , 128 A likely dual origin for orbital and eyelid fat has also been demonstrated recently by Korn et al.,129 who demonstrated that intraconal and nasal upper eyelid fat pads showed a double-fold more immune staining with CNCC markers than central eyelid fat.129 , 130 Despite some paradoxical over-expression of the central fat pad with few neuronal and glial antigens, the authors suggested that intraconal fat and the nasal upper eyelid fat pad derive from CNCC similar to head fat, while central eyelid fat originates from a mesodermal lineage like visceral and subcutaneous fat elsewhere in the body. Either way, their results do confirm the contribution of CNCC to orbital fat ontogenesis. Up until this point, adipogenesis of orbital fat runs a similar course to development of fat elsewhere. Hereafter, and as we shall see later, orbital fat morphogenesis will run a different course.
A connective tissue band is seen extending from the orbital (outer) surface of the MR and LR muscles toward the medial and lateral periphery of the conjunctiva. These bands represent the primordial medial and lateral check ligaments.131 The MR band is much thicker and better defined than the LR band.
Three separate events occur in the lacrimal gland during week 13: arborization or branching of the glandular parenchyma,132 anastomosis inside the gland between the lacrimal and the zygomatic nerves, and a significant increase in glandular vascularization.90 The branching process in the lacrimal gland which appears to be strongly influenced by the NOTCH pathway starts off with the formation of clefts within a single epithelial bud. These clefts invaginate cleaving the single bud into smaller buds, each of which grows to become a lobule. Repetitive looping of this process leads to the formation of the lacrimal gland.132
Week 14 (CRL 121 ± 11 mm).
The canthal index (inner canthal/outer canthal distance × 100) is around 40%.107 The mean fetal orbital diameter is 6.6 mm. The lesser wing of the sphenoid now possesses a lateral and a medial ossification center, but ossification in the lateral center usually starts earlier than the medial.71 Nevertheless, there are considerable individual variations in the ossification patterns of the orbitosphenoid.71 Similarly, the alisphenoid (greater wing of the sphenoid) possess 2 ossification centers, a medial and lateral, but in contrast to the lesser wing, ossification starts with the medial center and gradually extends laterally.71
The estimated number of ON axons reaches a peak of 3.7 to 3.9 million and then starts a process of decline.98 Vascularization within the lacrimal gland increases significantly during this week, but the stroma is still sparse.90
Week 15 (CRL 133 ± 11 mm).
The common muscle connective tissue sheath can be distinguished, and the IR is surrounded by circularly arranged connective tissue fibers.69 Lacrimal gland lobules are now observed, and the stroma condenses around them.90
Week 16 (CRL 145 ± 12 mm).
The medial ossification center of the orbitosphenoid is now well developed, consequently increasing the size of the lesser wing.71 Effectively, this narrows the opening of the infratemporal fossa to the orbital space, but the Müller orbital muscle still separates the orbit from the sphenopalatine-infratemporal space complex.115 The ethmoidal cartilaginous chondrocranium has not yet ossified,69 but perivascular osteoprogenitor cells, probably arising from the orbitosphenoid and ethmoid, start to colonize the Müller muscle. This initiates a process whereby its smooth muscle fibers are gradually replaced by collagen fibers. This process seems to provide guidance for ossification of the inferior orbital wall.115
Progression of fatty tissue differentiation continues during this stage, and although it is more concentrated in the retrobulbar area, fat is now seen as a definite structure throughout the orbit.115 , 124 , 133 , 134 Numerous connective tissue septa are seen between fatty tissue islands.69 Orbital connective tissue septa between EOMs are now extensive. Circular septae surround the SO and are connected to the septae surrounding the SR/LPS complex by means of a septum that encloses the SOV. Septa also extend from the MR to the medial periorbita and to the IR.69 Each lobule of the lacrimal gland receives its own arterial vessel.90
Week 17 (CRL 158 ± 13 mm).
By the beginning of week 17, the tendons of the EOMs become macroscopically identifiable for the first time, but all insertions are still equidistant from the limbus and remain so until the 22nd week.76 The EOMs have generally arrived at their adult topographical positions and are surrounded by an extensive, almost complete system of connective tissue. The LR is in direct contact with the periorbita of the sphenoid.69
Week 18 (CRL 171 ± 14 mm).
The canthal index has changed relatively little and is now around 39%.107 The mean fetal orbital diameter is 9.8 mm.119 The medial and lateral ossification centers of the orbitosphenoid are now fully fused to form the lesser wing of the sphenoid,71 and for the first time in fetal life the orbit is now separated from the sphenopalatine and infratemporal fossae by a layer of bone.115 The medial and lateral ossification centers of the greater wing now appear to be of equal density.71 According to a recent morphometric study based on the Kyoto embryo collection, the lesser sphenoid wing gradually surrounds the ON to form the optic canal around week 18.71 The space between the orbitosphenoid (lesser wing) and the alisphenoid (greater wing), which later creates the SOF, has not yet fully formed and is still only observed as a dehiscence.
The process of replacement of smooth muscle fibers by connective tissue in the Müller muscle is accelerating, resulting in a sharp reduction in actual muscle size as evidenced by the reduction in smooth muscle actin reactivity.115 By now, the absolute and relative size of Müller muscle has decreased.115
The amount of adipose tissue increases significantly but is least concentrated around the ON.69 The unique architecture of the orbital fat with its lobule-like arrangement and numerous connective tissue septae is already seen, and from this point forwards, no histologic differences are noted between fetal and adult orbital fat.115 , 134 This is in sharp contrast to nearby fat in the sphenopalatine-infratemporal fossae where it still maintains a homogenous honeycomb appearance devoid of septa. Osanai et al.115 proposed that mechanical stresses from the developing eye and EOM contractions might be the critical differentiating factors.
Week 19 (CRL 184 ± 14 mm).
The process of vascularization of the ON is now complete, and the ongoing process of ON axonal degeneration has reduced the total number of axons to an estimated 1.8 million.98
Week 20 (CRL 195 ± 15 mm)`.
The orbital index (height/width × 100) reaches 75%, and the orbit changes its shape from a more flattened chamaeconchal shape at week 11, to a more rounded mesoconchal shape.118 The relative increase in orbital height as the fetus matures is due to development of the facial skeleton and growth of the lateral wall of the nasal cavity related to development of the paranasal sinuses.118 , 135 Ossification centers appear in the ethmoidal chondrocranium and in its conchal offshoots.69
A well-defined dura mater is now observed.25 , 34 The oculomotor and the parasympathetic Edinger-Westphal nuclei approach their approximate adult configurations.35 , 112 , 113
The lateral horn of the levator aponeurosis continues its expansion underneath the orbital roof, above the globe, and divides the lacrimal gland into orbital and palpebral lobes.13 , 117 The SR/LPS complex is now firmly in contact with periorbita of the roof.69
All EOM insertions are still equidistant from the limbus.76 Müller muscle is reduced in thickness from an initial plate-like structure to a membrane-like structure.69 , 115
Week 22 (CRL 213 ± 16 mm).
The diameter of the ON is now 1.5 mm.83 Up until week 22, the tendinous insertions of the EOMs are equidistant from the limbus, and still have wide attachments to the sclera, extending from the equator of the eye to the limbus. From this point on the insertions start to recede posteriorly, and start to vary their distance from the limbus,76 to ultimately form the Spiral of Tillaux.34
Week 23 (CRL 223 ± 17 mm).
Ossification of the frontal, maxillary, sphenoid, and ethmoid bones accelerates.69 Müller muscle is further reduced in size as most of the smooth muscle fibers are now replaced by collagenous fiber bundles.115 Osanai et al.115 cautiously hypothesized that this newly formed periosteal-like tissue later undergoes direct membranous ossification, ultimately laying down bony laminae along the inferior orbital fissure. To the best of our knowledge, this muscle-to-bone transformation is the only example of a physiologic transition of muscle to bone throughout the body. Considering that the adult inferior orbital fissure is bordered by 4 different bones and measures 25 to 35 mm in length (mean, 29 mm),116 the authors do not specify where these muscle-derived bony laminae are deposited, although their vimentin and actin-stained sections do show unequivocal replacement of smooth muscle cells by osteoprogenitor cells.115
Week 24 (CRL 232 ± 18 mm).
The canthal index is around 37.6 %.107 By week 24, the mean fetal orbital diameter reaches 12.7 mm.119 As the cartilage at the apex of the orbit develops into bone and the perichondrium becomes periosteum, a ring of periosteal condensation forms the annulus of Zinn, which is located over the optic foramen and the medial central aspect of the SOF.25 , 26 , 76
Week 25 (CRL 242 ± 19 mm).
The orbital connective tissue system reaches a near adult configuration (Fig. 4).69 An extensive system of septae is already established within and outside the muscle cone between the EOMs, Tenon capsule, and the ON.69 The septum, which encircles the SOV and closely connects it to the SR/LPS complex, now sends several strands to the Tenon capsule.69 All EOMs send firm septal attachments to their adjacent periorbital regions establishing firm attachments and close apposition to the periorbita. As the fetus matures, the distance between each EOM and the periorbita broadens, but these septal attachments persist.69 The LR is the first muscle to distance itself from the periorbita.69
Week 26 (CRL 250 ± 21 mm).
Retrobulbar fat increases progressively in size and continues this exponential growth until the 30th week (Fig. 4).136 A well-defined arachnoid sheath is observed around the ON.25 , 34
Week 28 (CRL 267 ± 27 mm).
The total number of ON axons has stabilized to a number compatible with the adult ON.98 It is not exactly clear when the adult number is reached, but this process of apoptosis appears to peak between 14 and 18 weeks postfertilization and slows down later until between 50% and 70% of those superfluous axons are eliminated by week 28.98 , 137 It appears that the processes of retinal ganglion cell proliferation, differentiation, and cell death are complexly intertwined and coordinated. Morphogenetic cell death is not unique to the developing RGC as it is a widespread phenomenon throughout the entire nervous system including the cranial nerves supplying the orbit,137 , 138 and outside the nervous system in the developing eyelid.1 The evolutionary advantages and functional implications of this phenomenon remain relatively unknown. Myelin derived from oligodendrocytes begins centrally at the optic chiasm and progresses in a centrifugal direction toward the eye. The exact fetal age of myelinization of ON axons is controversial, variously given as weeks 21 to 30.11 , 34 , 83
The MR and LR check ligaments, initially observed during week 13, remain attached to the peripheral conjunctiva. However, they have not yet attained their adult attachment to bone, which only occurs after birth.131 , 139
Week 30 (CRL 284 ± 32 mm).
The diameter of the ON is 2.1 mm.83 The orbital index now exceeds 85 and the orbital entrance becomes hypsiconcal.118 , 135
Week 31 (CRL 292 ± 33 mm).
The orbital dimensions have increased substantially, and the fat/connective tissue ratio increases significantly. The general adult features of connective tissue are now easily discernible.69
Week 32 (CRL 301 ± 33 mm).
The canthal index is around 37%.107 The mean fetal orbital diameter is 15.2 mm.119 The primordial maxillary sinus appears.34 As ossification of the lesser wing of the sphenoid and the frontal bone near maturity, both bones approach each other and overlap, each with its own ossification front, its own basement membrane, and with a cartilaginous center in between them forming a transient 5-layered junctional structure that is termed the orbitosphenofrontal suture.96 Because of this overlap, the junction between both ossification centers appears obliquely located in the sagittal plane between the dura mater and the periosteum. Hypothetically, this overlap gives way to an unimpeded growth of the anterior cranial fossa (orbital roof) in length and height.96 This transient suture does not completely disappear except near birth when the lesser wing is completely ossified.96 In contrast, the ossification centers of the frontal bone (nearly horizontal), and the great wing of the sphenoid (nearly vertical) progress directly toward each other near the lateral wall of the orbit (Fig. 5). Here, they form the lateral sphenofrontal suture, which by now already looks like an adult end-to-end mature suture.96 This suture should not be confused with the transient and more medial orbitosphenofrontal suture described above (Professor Guillaume Captier, personal communication). Ossification of the maxillary bone also progresses significantly, but the ethmoid bone is still only partially ossified.69
Week 36 (CRL 336 ± 32 mm).
The mean fetal orbital diameter is 16.5 mm,121 and the eye-orbit index (eye volume/orbital volume) is now 75%.3
Full Term (CRL 367 ± 25 mm).
The canthal index is around 40 in neonates.140 The eye volume is about 38% of the average value in adults,3 while the volume of the orbit averages about 17.3% of the average adult value.141 The eye-orbit index at birth is 45, while in adulthood the index is around 32. Taking into consideration a value of 75 at 8 months postfertilization,3 it appears that the rate of increase in the size of the eye initially outpaces growth of the orbit during fetal life, but relatively falls behind as the infant approaches full term.3 , 4 , 142 This clearly indicates that the growth of the eye and orbit do not follow a linear growth curve.143 That the eye and orbit follow separate growth trajectories may lend support to a contentious novel view that the eye volume may only explain a small amount of variance (15%) in orbital volume, the latter apparently correlating more with the cranium and frontal lobe size.142 This unequal growth also explains why at birth the eyeball appears relatively proptotic compared with adults, and projects markedly beyond the orbital rim.4 This unequal growth pattern continues into adulthood and well into adult life, where orbital volume increases to a greater extent than ocular volume as the overall body size increases.142
As we explained earlier (weeks 11 and 20), the orbit becomes hypsiconchal in shape at full term and onwards because the orbital height keeps increasing throughout fetal life, and even later. This process continues until the orbital index reaches a peak of 100 at the age of 6 months of life or so.118 , 141 Here again, orbital growth mimics the growth of the face as a whole, because vertical growth of the face outpaces horizontal growth during fetal life.106 , 107 After the age of 6 months, the pattern reverses and in adults the orbit becomes relatively mesoconchal again.118 , 144 This lends further credence to the concept discussed in the preceding paragraph which is often underappreciated when discussing congenital anophthalmia, that growth of the orbit is not solely dependent on the development of the eye, but the developing face and cranium also play a role in determining the eventual shape and size of the orbit.106 , 142
According to De Haan et al.,145 even at term the orbit should still be considered primordial because nonossified connective tissue is still present in the orbit, especially in the apex, the ethmoid bone is not yet completely ossified, and almost 50% of the floor is still formed by the orbital muscle of Müller, which may have reduced in thickness but not in extent.145
Myelinization of the ON is now complete and terminates at the level of the lamina cribrosa at full term or 1-month postpartum.13 Transient overexpression of netrin-1 and tenascin-C in the developing ON act as molecular stop signals that prevent the migration of oligodendrocytes into the retina in 99% of humans.26
At birth, the EOM tendons are located between the adult site of insertion and the limbus.76 That is, the distance from the rectus muscle insertion to the limbus is roughly 2 mm less than in adults, while at the age of 6 to 9 months of age this distance is 1 mm less than in adults. They reach their adult distances at about 20 months of age.141 , 146 With enlargement of the orbit the tendinous origin of the SO sequestrates from the annulus of Zinn and acquires a separate origin from the region of the frontoethmoidal suture, immediately superior and medial to the origin of the MR.76 The insertion of the SO also changes position and now inserts with a flattened tendon in a position corresponding to its adult position into the superolateral quadrant of the globe.76
Despite conventional teaching that the secretory function of the lacrimal apparatus in infancy is either absent or minimally present at birth,147 more than 80% of infants have normal basal tear flow within the first 2 days of life.148
Development of the orbit is a collective enterprise necessitating interactions between, and contributions from different cell populations both within and beyond the realm of the orbit. Understanding the morphogenesis of all the definitive structures in the orbit in a methodical and timely fashion poses a significant challenge because the orbit represents a complex set of structures and tissues with a multitude of cellular raw material required to build the final product. Our goal while preparing this exhaustive review was to clarify this impossible chaos of those seemingly incomprehensible embryological events. Only through a strict sequentially ordered chronological narrative could we unfold those complexly intertwined morphogenetic processes, together with their attendant unfathomable embryological jargon.
Although a single cell type of neural cell origin, the CNCC could be singled out as the fundamental cell in orbital morphogenesis, the processes that trigger and contribute to formation of the orbit are far more complex and seem to be intricately regulated not just by the CNCC, but by mesodermal-CNCC interaction, and by epithelial mesenchymal bidirectional cross-talk.18
Orbital morphogenesis is a relatively unexplored area for orbital experimentation. Identifying and linking signaling cascades and regulatory genes to existing orbital diseases may seem like a tedious and painstaking processes at the moment, and the use of progenitor/stem cell therapy132 and even regenerative therapy149 to treat intractable genetic or even acquired orbital pathology, may appear pointless or even counterintuitive, until we realize that fetal tissues and their adult counterparts share similar molecular and genetic pathways. The story is just starting to unfold,132 but these shared signaling pathways may be involved both in morphogenetic events during fetal life, and physiologic maintenance of the same tissue in adults.132 Therefore, a basic understanding of the processes underlying orbital ontogenesis which we hope we have achieved in this study, coupled with research tools that are already available at hand today,149 is a crucial first step toward making this far-reaching goal easily attainable. But we need to move forward with caution because as we have seen throughout the discussion, the literature is plagued with several controversial hypotheses about the true nature of events during the embryologic or fetal period, and to complicate matters further the exact embryologic origin of several major orbital structures remains heavily contested or largely unknown up to this day. Those fundamental issues need to be resolved first, but with more thorough research, and more data accumulating, we believe that for the first time in history, there is now tangible optimism that in the not-so-distant future we may be able to offer patients with chronic or incurable orbital diseases some form of effective therapy. Finally, there is no better way to conclude this review than to borrow Drew Noden’s concluding remarks in his article about the differentiation of craniofacial muscles, “If this review provokes some to engage in the adventure, we will happily consider it a worthwhile investment.”20
1. Tawfik HA, Abdulhafez MH, Fouad YA, et al. Embryologic and fetal development of the human eyelid. Ophthalmic Plast Reconstr Surg 2016;32:407–14.
2. Ali MJ, Kakizaki H. Ali MJ. Embryology of the lacrimal drainage system. In: Principles and Practice of Lacrimal Surgery. 2015:New Delhi, India: Springer, 9–15.
3. Schultz AH. The size of the orbit and of the eye in primates. Am J Phys Anthropol 1940;26:389–408.
4. Sarnat BG, Shanedling PD. Postnatal growth of the orbit and upper face in rabbits after exenteration of the orbit. Arch Ophthalmol 1965;73:829–37.
5. Lesshaft P. Sarnat BG, Shanedling PD; Grundlagen der theoretischen Anatomie, Erster Teil, Leipzig: W. Hartmann, 1892. Postnatal growth of the orbit and upper face in rabbits after exenteration of the orbit. In. Arch Ophthalmol 1965;73:829–37.
6. O’Rahilly R, Müller F. Developmental stages in human embryos: revised and new measurements. Cells Tissues Organs 2010;192:73–84.
7. Archie JG, Collins JS, Lebel RR. Quantitative standards for fetal and neonatal autopsy. Am J Clin Pathol 2006;126:256–65.
8. Escaravage GK Jr, Dutton JJ. Age-related changes in the pediatric human orbit on CT. Ophthalmic Plast Reconstr Surg 2013;29:150–6.
9. Som PM, Naidich TP. Illustrated review of the embryology and development of the facial region, part 1: early face and lateral nasal cavities. AJNR Am J Neuroradiol 2013;34:2233–40.
10. Sadler TW. Embryology of neural tube development. Am J Med Genet C Semin Med Genet 2005;135C:2–8.
11. Cook CS, Sulik KK, Wright KW. Wright KW, Spiegel PH. Embryology. In: Pediatric Ophthalmology and Strabismus. 2013:New York: Springer Science & Business Media, 3–38.
14. Fink WH. The development of the orbital fascia. Am J Ophthalmol 1956;42(4 Part 2):269–77.
15. Noden DM, Trainor PA. Relations and interactions between cranial mesoderm and neural crest populations. J Anat 2005;207:575–601.
16. Piest KL. Katowitz JA. Embryology and anatomy of the developing face. In: Pediatric Oculoplastic Surgery. 2002:New York, NY: Springer, 11–29.
17. O’Rahilly R, Müller F. The development of the neural crest in the human. J Anat 2007;211:335–51.
18. Santagati F, Rijli FM. Cranial neural crest and the building of the vertebrate head. Nat Rev Neurosci 2003;4:806–18.
20. Noden DM, Francis-West P. The differentiation and morphogenesis of craniofacial muscles. Dev Dyn 2006;235:1194–218.
21. Kahana A, Levin LA. Albert D, Miller J, Azar D. Peripheral nerve sheath tumors of the orbit. In: Albert & Jakobiec’s Principles and Practice of Ophthalmology. 2008:3rd ed. Philadelphia, PA: Saunders/Elsevier, 3019–41.
22. Langenberg T, Kahana A, Wszalek JA, et al. The eye organizes neural crest cell migration. Dev Dyn 2008;237:1645–52.
24. Forrester J, Dick A, McMenamin P, et al. Forrester J, Dick A, McMenamin P, et al. Embryology and early development of the eye and adnexa. In The eye. 2016:Amsterdam, Netherlands: Elsevier, 103–29.
25. Zhang J, Rubin RM, Rao NA. Tasman W, Jaeger EA. Anatomy and embryology of the optic nerve. In: Duane’s Foundations of Clinical Ophthalmology. 2006:Philadelphia, PA: LWW, 1–10.
26. Rizzo JF III.. Miller NR, Newman NJ, Biousse V, Kerrison JB. Embryology, anatomy, and physiology of the afferent visual pathway. In: Walsh and Hoyt’s Clinical Neuro-ophthalmology. 2005:Philadelphia, PA: Lippincott Williams & Wilkins, 3–82.
27. Catala M. [Embryology of the sphenoid bone]. J Neuroradiol 2003;30:196–200.
28. Cordero DR, Brugmann S, Chu Y, et al. Cranial neural crest cells on the move: their roles in craniofacial development. Am J Med Genet A 2011;155A:270–9.
29. O’Rahilly R, Müller F. The meninges in human development. J Neuropathol Exp Neurol 1986;45:588–608.
30. Köntges G, Lumsden A. Rhombencephalic neural crest segmentation is preserved throughout craniofacial ontogeny. Development 1996;122:3229–42.
32. Streeter GL. Developmental horizons in human embryos. Description of age group XIII, embryos about 4 or 5 millimeters long, and age group XIV, period of indentation of the lens vesicle. Contrib Embryol Carnegie Inst Wash 1945;31:27–63.
33. Williams AL, Bohnsack BL. Neural crest derivatives in ocular development: discerning the eye of the storm. Birth Defects Res C Embryo Today 2015;105:87–95.
34. Dutton JJ. Atlas of Clinical and Surgical Orbital Anatomy. 1994:Philadelphia, PA: W.B. Saunders Company, 1–240.
35. Mann IC. The developing third nerve nucleus in human embryos. J Anat 1927;61(Pt 4):424–38.
36. Hayreh SS. Arteries of the orbit in the human being. Br J Surg 1963;50:938–53.
37. Hayreh SS. Orbital vascular anatomy. Eye (Lond) 2006;20:1130–44.
38. Menshawi K, Mohr JP, Gutierrez J. A functional perspective on the embryology and anatomy of the cerebral blood supply. J Stroke 2015;17:144–58.
39. Toma N. Anatomy of the ophthalmic artery: embryological consideration. Neurol Med Chir (Tokyo) 2016;56:585–91.
40. Bertelli E, Regoli M, Bracco S. An update on the variations of the orbital blood supply and hemodynamic. Surg Radiol Anat 2017;39:485–96.
41. Padget DH. The development of the cranial arteries in the human embryo. Contrib Embryol 1948;32:205–61.
42. Mullan S, Mojtahedi S, Johnson DL, et al. Embryological basis of some aspects of cerebral vascular fistulas and malformations. J Neurosurg 1996;85:1–8.
43. Mullan S. Reflections upon the nature and management of intracranial and intraspinal vascular malformations and fistulae. J Neurosurg 1994;80:606–16.
44. Scher MS. Normal and abnormal cerebrovascular development: gene-environment interactions during early life with later life consequences. Handb Clin Neurol 2013;112:1021–42.
45. Burdi AR, Lawton TJ, Grosslight J. Prenatal pattern emergence in early human facial development. Cleft Palate J 1988;25:8–15.
46. Müller F, O’Rahilly R. The initial appearance of the cranial nerves and related neuronal migration in staged human embryos. Cells Tissues Organs 2011;193:215–38.
47. Steventon B, Mayor R, Streit A. Neural crest and placode interaction during the development of the cranial sensory system. Dev Biol 2014;389:28–38.
48. Barlow LA. Cranial nerve development: placodal neurons ride the crest. Curr Biol 2002;12:R171–3.
49. Pearson AA. The trochlear nerve in human fetuses. J Comp Neurol 1943;78:29–43.
50. Cooper ERA. The development of the nuclei of the oculomotor and trochlear nerves (somatic efferent column). Brain 1946;119:50–7.
51. Yamaguchi K, Honma K. Development of the human abducens nucleus: a morphometric study. Brain Dev 2012;34:712–8.
52. Cordes SP. Molecular genetics of cranial nerve development in mouse. Nat Rev Neurosci 2001;2:611–23.
53. Carlson BM. Human Embryology and Developmental Biology (Fifth Edition). 2014:Philadelphia, PA: Saunders, Elservier, 1–472.
54. Lumsden A, Keynes R. Segmental patterns of neuronal development in the chick hindbrain. Nature 1989;337:424–8.
55. Lee VM, Sechrist JW, Luetolf S, et al. Both neural crest and placode contribute to the ciliary ganglion and oculomotor nerve. Dev Biol 2003;263:176–90.
56. Barishak YR. Embryology of the eye and its adnexae. Dev Ophthalmol 1992;24:1–142.
57. Saint-Geniez M, D’Amore PA. Development and pathology of the hyaloid, choroidal and retinal vasculature. Int J Dev Biol 2004;48:1045–58.
58. O’Rahilly R. The timing and sequence of events in the development of the human eye and ear during the embryonic period proper. Anat Embryol (Berl) 1983;168:87–99.
59. Keller JT, Leach JL, van Loveren HR, et al. Dolenc VV, Rogers L. Venous anatomy of the lateral sellar compartment. In: Cavernous Sinus. 2009:Vienna, Austria: Springer. 34–51.
60. Padget DH. The cranial venous system in man in reference to development, adult configuration, and relation to the arteries. Am J Anat 1956;98:307–55.
61. Padget DH. The development of the cranial venous system in man from the viewpoint of comparative anatomy. Contrib Embryol 1957;36:81–139.
62. Mishina Y, Snider TN. Neural crest cell signaling pathways critical to cranial bone development and pathology. Exp Cell Res 2014;325:138–47.
63. Gross JB, Hanken J. Review of fate-mapping studies of osteogenic cranial neural crest in vertebrates. Dev Biol 2008;317:389–400.
64. Jiang X, Iseki S, Maxson RE, et al. Tissue origins and interactions in the mammalian skull vault. Dev Biol 2002;241:106–16.
65. de Haan AB, Willekens BL. Embryology of the orbital walls. A preliminary report. Mod Probl Ophthalmol 1975;14:57–64.
66. Mall FP. On ossification centers in human embryos less than one hundred days old. Am J Anat 1906;5:433–58.
67. Vacher C, Copin H, Sakka M. Maxillary ossification in a series of six human embryos and fetuses aged from 9 to 12 weeks of amenorrhea: clinical implications. Surg Radiol Anat 1999;21:261–6.
68. O’Rahilly R, Müller F. O’Rahilly R, Müller F. The skull. In: Human Embryology & Teratology. 2001:3rd ed. New York, NY: Wiley-Liss, 374–81.
69. Koornneef L. The development of the connective tissue in the human orbit. Acta Morphol Neerl Scand 1976;14:263–90.
70. Presely R, Steel FL. On the homology of the alisphenoid. J Anat 1976;121(Pt 3):441–59.
71. Zhang Q, Wang H, Udagawa J, et al. Morphological and morphometric study on sphenoid and basioccipital ossification in normal human fetuses. Congenit Anom (Kyoto) 2011;51:138–48.
72. Gilbert PW. The origin and development of the human extrinsic ocular muscles. Contrib Embryol Carneg Instn 1957;36:5978.
73. Bohnsack BL, Gallina D, Thompson H, et al. Development of extraocular muscles requires early signals from periocular neural crest and the developing eye. Arch Ophthalmol 2011;129:1030–41.
74. Sevel D. A reappraisal of the origin of human extraocular muscles. Ophthalmology 1981;88:1330–8.
75. Kish PE, Bohnsack BL, Gallina D, et al. The eye as an organizer of craniofacial development. Genesis 2011;49:222–30.
76. Sevel D. The origins and insertions of the extraocular muscles: development, histologic features, and clinical significance. Trans Am Ophthalmol Soc 1986;84:488–526.
77. Adeeb N, Mortazavi MM, Tubbs RS, et al. The cranial dura mater: a review of its history, embryology, and anatomy. Childs Nerv Syst 2012;28:827–37.
78. Streeter GL. The nuclei of origin of the cranial nerves in the 10 mm human embryo. Am J Anat 1908;2:111–15.
79. Lasjaunias P, Moret J, Mink J. The anatomy of the inferolateral trunk (ILT) of the internal carotid artery. Neuroradiology 1977;13:215–20.
80. Diamond MK. Homologies of the meningeal-orbital arteries of humans: a reappraisal. J Anat 1991;178:223–41.
81. Louw L. Different ophthalmic artery origins: embryology and clinical significance. Clin Anat 2015;28:576–83.
82. Komiyama M. Embryology of the ophthalmic artery: a revived concept letter to the editor. Interv Neuroradiol 2009;15:363–8.
83. Takayama S, Yamamoto M, Hashimoto K, et al. Immunohistochemical study on the developing optic nerves in human embryos and fetuses. Brain Dev 1991;13:307–12.
84. Deiner MS, Kennedy TE, Fazeli A, et al. Netrin-1 and DCC mediate axon guidance locally at the optic disc: loss of function leads to optic nerve hypoplasia. Neuron 1997;19:575–89.
85. Brittis PA, Canning DR, Silver J. Chondroitin sulfate as a regulator of neuronal patterning in the retina. Science 1992;255:733–6.
86. Tawfik HA, Habib MA. Intraorbital false (thrombosed) aneurysm of the meningolacrimal branch of the middle meningeal artery: a case report and an appraisal of anatomy. Ophthalmic Plast Reconstr Surg 2011;27:387–90.
87. Georgiou C, Cassell MD. The foramen meningo-orbitale and its relationship to the development of the ophthalmic artery. J Anat 1992;180 (Pt 1):119–25.
88. Katori Y, Rodríguez-Vázquez JF, Kawase T, et al. Early fetal development of hard tissue pulleys for the human superior oblique and tensor veli palatini muscles. Ann Anat 2011;193:127–33.
89. Kammandel B, Chowdhury K, Stoykova A, et al. Distinct cis-essential modules direct the time-space pattern of the Pax6 gene activity. Dev Biol 1999;205:79–97.
90. de la Cuadra-Blanco C, Peces-Peña MD, MeÂrida-Velasco JR. Morphogenesis of the human lacrimal gland. J Anat 2003;203:531–6.
91. Tripathi BJ, Tripathi RC. Evidence for the neuroectodermal origin of the human lacrimal gland. Invest Ophthalmol Vis Sci 1990;31:393–5.
92. Duke-Elder S, Cook C. Duke-Elder S. Normal and abnormal development, part 1: embryology. In: System of Ophthalmology. 1963:St. Louis: CV Mosby, 215–6.
93. Makarenkova HP, Ito M, Govindarajan V, et al. FGF10 is an inducer and Pax6 a competence factor for lacrimal gland development. Development 2000;127:2563–72.
94. Makarenkova HP, Dartt DA. Myoepithelial cells: their origin and function in lacrimal gland morphogenesis, homeostasis, and repair. Curr Mol Biol Rep 2015;1:115–23.
95. Edman MC, Marchelletta RR, Hamm-Alvarez SF. Dartt DA, Bex P, D’Amore P, et al. Lacrimal gland overview. In: Ocular Periphery and Disorders. 2011:San Diego, CA: Elsevier Ltd, 68–73.
96. Captier G, Cristol R, Montoya P, et al. Prenatal organization and morphogenesis of the sphenofrontal suture in humans. Cells Tissues Organs 2003;175:98–104.
97. Rodríguez Vázquez JF, Mérida Velasco JR, Jiménez Collado J. Orbital muscle of Müller: observations on human fetuses measuring 35-150 mm. Acta Anat (Basel) 1990;139:300–3.
98. Provis JM, van Driel D, Billson FA, et al. Human fetal optic nerve: overproduction and elimination of retinal axons during development. J Comp Neurol 1985;238:92–100.
99. Soukkarieh C, Agius E, Soula C, et al. Pax2 regulates neuronal-glial cell fate choice in the embryonic optic nerve. Dev Biol 2007;303:800–13.
100. Sturrock RR. An electron microscopic study of macrophages in the meninges of the human embryonic optic nerve. J Anat 1988;157:145–51.
101. Moujahid A, Navascués J, Marín-Teva JL, et al. Macrophages during avian optic nerve development: relationship to cell death and differentiation into microglia. Anat Embryol (Berl) 1996;193:131–44.
102. Cheung N, McNab AA. Venous anatomy of the orbit. Invest Ophthalmol Vis Sci 2003;44:988–95.
103. Froelich S, Aziz KMA, van, Loveren HR, et al. Dolenc VV, Rogers L. The transition between the cavernous sinus and orbit. In: Cavernous Sinus. 2009:Vienna, Austria: Springer, 28–33.
104. Reuter K. Ucber die Entwicklung der Bugenmuskcln beim Schwein. Anat Hefte. 1897;9:365–88. In: Johnson CE. The development of the prootic head somites and eye muscles in chelydra serpentina. Am J Anat
105. Sevel D. A reappraisal of the development of the eyelids. Eye (Lond) 1988;2 (Pt 2):123–9.
106. Denis D, Burguiere O, Oudahi F, et al. Measurement of facial growth in the human fetus. Graefes Arch Clin Exp Ophthalmol 1995;233:756–65.
107. Denis D, Burguière O, Burillon C. A biometric study of the eye, orbit, and face in 205 normal human fetuses. Invest Ophthalmol Vis Sci 1998;390:232–8.
108. Laestadius ND, Aase JM, Smith DW. Normal inner canthal and outer orbital dimensions. J Pediatr 1969;74:465–8.
109. Morard M, Tcherekayev V, de Tribolet N. The superior orbital fissure: a microanatomical study. Neurosurgery 1994;35:1087–93.
110. Govsa F, Kayalioglu G, Erturk M, et al. The superior orbital fissure and its contents. Surg Radiol Anat 1999;21:181–5.
111. Vignaud J, Hasso AN, Lasjaunias P, et al. Orbital vascular anatomy and embryology. Radiology 1974;111:617–26.
112. Pearson AA. The oculomotor nucleus in the human fetus. J Comp Neurol 1994;80:47–63.
113. Yamaguchi K. Development of the human oculomotor nuclear complex: Centrally-projecting Edinger-Westphal nucleus. Neurosci Lett 2017;646:8–14.
114. Rodríguez-Vázquez JF, Mérida-Velasco JR, Arráez-Aybar LA, et al. Anatomic relationships of the orbital muscle of Müller in human fetuses. Surg Radiol Anat 1998;20:341–4.
115. Osanai H, Abe S, Rodríguez-Vázquez J, et al. Human orbital muscle: a new point of view from the fetal development of extraocular connective tissues. Invest Ophthalmol Vis Sci 2011;52:1501–6.
116. De Battista JC, Zimmer LA, Theodosopoulos PV, et al. Anatomy of the inferior orbital fissure: implications for endoscopic cranial base surgery. J Neurol Surg B Skull Base 2012;73:132–8.
118. Haas A, Weiglein A, Faschinger C, et al. Fetal development of the human orbit. Graefes Arch Clin Exp Ophthalmol 1993;231:217–20.
119. Goldstein I, Tamir A, Zimmer EZ, et al. Growth of the fetal orbit and lens in normal pregnancies. Ultrasound Obstet Gynecol 1998;12:175–9.
120. Dilmen G, Köktener A, Turhan NO, et al. Growth of the fetal lens and orbit. Int J Gynaecol Obstet 2002;76:267–71.
121. Sukonpan K, Phupong V. A biometric study of the fetal orbit and lens in normal pregnancies. J Clin Ultrasound 2009;37:69–74.
122. Kivilevitch Z, Salomon LJ, Benoit B, et al. Fetal interlens distance: normal values during pregnancy. Ultrasound Obstet Gynecol 2010;36:186–90.
123. Sturrock RR. A light and electron microscopic study of proliferation and maturation of fibrous astrocytes in the optic nerve of the human embryo. J Anat 1975;119(Pt 2):223–34.
124. Poissonnet CM, Burdi AR, Garn SM. The chronology of adipose tissue appearance and distribution in the human fetus. Early Hum Dev 1984;10:1–11.
125. Gupta RK, Mepani RJ, Kleiner S, et al. Zfp423 expression identifies committed preadipocytes and localizes to adipose endothelial and perivascular cells. Cell Metab 2012;15:230–9.
126. Tseng YH, Cypess AM, Kahn CR. Cellular bioenergetics as a target for obesity therapy. Nat Rev Drug Discov 2010;9:465–82.
127. Billon N, Iannarelli P, Monteiro MC, et al. The generation of adipocytes by the neural crest. Development 2007;134:2283–92.
128. Billon N, Dani C. Developmental origins of the adipocyte lineage: new insights from genetics and genomics studies. Stem Cell Rev 2012;8:55–66.
129. Korn BS, Kikkawa DO, Hicok KC. Identification and characterization of adult stem cells from human orbital adipose tissue. Ophthalmic Plast Reconstr Surg 2009;25:27–32.
130. Oh SR, Chokthaweesak W, Annunziata CC, et al. Analysis of eyelid fat pad changes with aging. Ophthalmic Plast Reconstr Surg 2011;27:348–51.
131. Osanai H, Rodríguez-Vázquez JF, Abe H, et al. Fetal check ligament connected between the conjunctiva and the medial and lateral recti. Invest Ophthalmol Vis Sci 2011;52:7175–9.
132. Dvoriantchikova G, Tao W, Pappas S, et al. Molecular profiling of the developing lacrimal gland reveals putative role of notch signaling in branching morphogenesis. Invest Ophthalmol Vis Sci 2017;58:1098–109.
133. Byun TH, Kim JT, Park HW, et al. Timetable for upper eyelid development in staged human embryos and fetuses. Anat Rec (Hoboken) 2011;294:789–96.
134. Wolfram-Gabel R, Kahn JL. Adipose body of the orbit. Clin Anat 2002;15:186–92.
136. Birnholz JC. Ultrasonic fetal ophthalmology. Early Hum Dev 1985;12:199–209.
137. Chavarría T, Baleriola J, Mayordomo R, et al. Early neural cell death is an extensive, dynamic process in the embryonic chick and mouse retina. ScientificWorld J 2013;2013:627240.
138. Yamaguchi K, Honma K. Development of the human trochlear nucleus: a morphometric study. Ann Anat 2011;193:106–11.
139. Plock J, Contaldo C, Von Lüdinghausen M. Extraocular eye muscles in human fetuses with craniofacial malformations: anatomical findings and clinical relevance. Clin Anat 2007;20:239–45.
140. Ozkağnici A, Büyükmumcu M, Zengin N, et al. Ocular and periorbital anthropometric measurements in term Turkish newborns. Surg Radiol Anat 2001;23:321–4.
141. Eustis HS, Guthrie MA. Wright KW, Spiegel PH. Postnatal development. In: Pediatric Ophthalmology and Strabismus. 2013:New York, NY: Springer Science & Business Media, 39–53.
142. Masters M, Bruner E, Queer S, et al. Analysis of the volumetric relationship among human ocular, orbital and fronto-occipital cortical morphology. J Anat 2015;227:460–73.
143. Chau A, Fung K, Pak K, et al. Is eye size related to orbit size in human subjects? Ophthalmic Physiol Opt 2004;24:35–40.
144. Bron AJ, Tripathi RC, Tripathi BJ. Wolff’s Anatomy of the Eye & Orbit. 1997.8th ed. London, United Kingdom: Chapman & Hall Medical.
145. de Haan AB, Willekens B, Klooster J, et al. The prenatal development of the human orbit. Strabismus 2006;14:51–6.
146. Swan KC, Wilkins JH. Extraocular muscle surgery in early infancy—anatomical factors. J Pediatr Ophthalmol Strabismus 1984;21:44–9.
147. Morin JD. Crawford JS, Morin JD. Glaucoma. In: The Eye in Childhood, New York: Grune & Stratton, 1983.1-616. 2013:New York, NY: Springer Science & Business Media, 40.In: Eustis HS, Guthrie MA. Postnatal development. In: Wright KW, Spiegel PH, eds. Pediatric Ophthalmology and Strabismus
148. Apt L, Cullen BF. Newborns do secrete tears. JAMA 1964;189:951–3.
149. Hirayama M, Kawakita T, Tsubota K, et al. Challenges and strategies for regenerating the lacrimal gland. Ocul Surf 2016;14:135–43.
© 2018 by The American Society of Ophthalmic Plastic and Reconstructive Surgery, Inc., All rights reserved.
150. Sambasivan R, Kuratani S, Tajbakhsh S. An eye on the head: the development and evolution of craniofacial muscles. Development 2011;138:2401–15.