Article Text

Download PDFPDF

Original article
Orbital development in survivors of retinoblastoma treated by enucleation with hydroxyapatite implant
  1. Hsiu-Yi Lin,
  2. Shu-Lang Liao
  1. Department of Ophthalmology, National Taiwan University Hospital, Taipei, Taiwan
  1. Correspondence to Dr Shu-Lang Liao, Department of Ophthalmology, National Taiwan University Hospital, 7, Chung Shan South Road, Taipei, 100, Taiwan; liaosl89{at}


Aims To determine the impact of enucleation with hydroxyapatite implant on bony orbital development in survivors of retinoblastoma (RB) by measuring orbital volume based on CT imaging.

Methods The authors used CT images obtained at a median age of 6 years to measure orbital volume of RB and contralateral orbits in 18 patients who underwent enucleation with hydroxyapatite implant for RB. Comparison of the orbital volume of RB and contralateral orbits was done using the Wilcoxon rank sum test.

Results The mean age at diagnosis and operation was 29±23 months, and the mean follow-up was 49±31 months. The mean volume difference between RB and contralateral orbits was 0.93±1.13 cm3. RB orbits with hydroxyapatite implant were statistically significantly smaller than contralateral orbits (p=0.002). The age at operation was significantly negatively correlated with orbital volume difference (p=0.033). Orbital volume differences for children treated by enucleation before the age of 12 months were also statistically significantly larger than those treated later. (p=0.03).

Conclusion Significant orbital growth retardation remained after enucleation, even with a hydroxyapatite implant for the RB orbit. Orbital growth retardation was correlated with operation age and also more prominent in children treated in the first year of life.

  • Retinoblastoma
  • enucleation
  • hydroxyapatite implant
  • orbital development
  • orbit
  • embryology and development
  • neoplasia

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Retinoblastoma (RB) is the most common intraocular malignant tumour in children. The disease is diagnosed in infancy or early childhood at a median age of 2 years. The tumour is generally managed with enucleation in unilateral cases and external-beam irradiation in bilateral cases. Since the prognosis is good, with cure rates of 75–90%,1 cosmetic problems and quality of life are of great concern.

Orbital growth is still in progress in children. The orbital volume increases rapidly in the initial 3 years after birth.2 The medial and lateral orbital walls enlarge significantly in the first year2 and continue with remodelling until age 5–7.3 4 By 12 years, the orbital volume is near adult size. Several studies suggest that orbital growth is retarded after enucleation for RB.5–8 To improve orbital growth after enucleation, orbital implants of various materials have been used.5 6 9 10 The mechanism of preventing orbital growth retardation is not fully understood.

Porous implants have been widely used in recent years to reconstruct anophthalmic sockets.11 Their ability to maintain orbital size in rapidly growing young children has not been established. In this study, we sought to measure orbital volumes using CT images of patients with RB who had undergone enucleation with hydroxyapatite implants by comparing orbital volumes of enucleated and contralateral orbits.

Materials and methods

Eighteen consecutive patients who had undergone enucleation for unilateral RB with hydroxyapatite implant (18 mm) at the National Taiwan University Hospital between 1995 and 2002 were studied retrospectively. The Ethics Committee of the National Taiwan University Hospital approved the protocol of this study, and parents of each subject signed written informed consent to participate. None of the patients received radiotherapy after enucleation.

Orbital volumes were measured on sequential axial images with a slice thickness of 2.0 mm on CT using Mirror DPS software (Canfield Clinical Systems, Fairfield, New Jersey). The area of the orbital cross-section of each CT image slice was determined by marking the margin of the orbital cavity and having it automatically calculated by Mirror DPS software (figure 1). Total orbital volume was calculated by summing the orbital cross-section area of all 2 mm CT image slices and multiplying by 0.2 (2 mm equals 0.2 cm). All measurements were performed by the same examiner. Orbital volumes of enucleated and contralateral orbits were compared by subject. The Wilcoxon rank sum test was used with p<0.05 considered statistically significant. Pearson correlation and generalised linear regression were used to find the correlation between age at operation and orbital volume difference. All analyses were conducted using SPSS version 12.0.

Figure 1

Orbital volume was measured on sequential axial images with a slice thickness of 2.0 mm of CT. A cross-section area of each orbital slice was automatically calculated by marking the margin of orbital cavity using the software Mirror DPS.


The 18 survivors of RB consisted of 13 males and 5 females, whose malignancy was diagnosed and treated at a mean age of 29±23 months (range 3–84 months) (table 1). The mean follow-up was 49±31 months, with a median age at follow-up imaging of 6 years (table 2). At the time of enucleation, 5 of 18 patients were younger than 1 year.

Table 1

Orbital volume, volume difference, operation age and follow-up time for survivors of retinoblastoma

Table 2

Demographics for survivors of retinoblastoma

The mean difference in orbital volume between the enucleated and contralateral orbits was 0.93±1.13 cm3. The volume of the enucleated orbits with hydroxyapatite implant was statistically significantly smaller than the contralateral orbits (p=0.002). Age at operation was significantly negatively correlated with orbital volume difference (p=0.033) (figure 2). The equation using orbital volume differences as a dependent variable and operation age as an independent variable was: orbital volume difference=−0.024×operation age (months)+1.64 (p=0.033 for operation age) (table 3). Orbital volume differences in children treated before the age of 12 months (1.83±1.47 cm3) were statistically significantly larger than those of children treated after that age (0.58±0.58 cm3) (p=0.03) (table 4). The follow-up time after surgery did not correlate significantly with orbital volume difference (p=0.578).

Figure 2

Correlation plot of volume differences and operation age for survivors of retinoblastoma. Operation age was significantly negatively correlated with orbital volume differences (p=0.033).

Table 3

Multivariate analyses of the influence of operation age to orbital volume difference by linear regression for survivors of retinoblastoma (N=18)

Table 4

Mean difference in orbital volume according to age at enucleation


Enucleation in paediatric patients involves removing an eye that has not reached adult size. Studies found that an eye achieved 85% or more of its axial length by the age of 2 years and continued to grow 1% per year until it reached its final size.10 The rapid growth presents certain cosmetic challenges in this age group. Early studies described orbital contraction resulting from enucleation in children who did not receive an implant. Orbital growth retardation was reported to be the result of reduced volume stimulus.8 9

Several methods are used to treat the anophthalmic socket in congenital microphthalmos and anophthalmos.12–14 To promote facial symmetry, spherical implants, orbital expanders, orbital osteotomies and, more recently, hydrophilic osmotic expanders have been used.12–15 The drawback of traditional soft-tissue expanders is extrusion and rupture. Low predictability of the orbital volume after expansion is observed, and further surgeries are needed to replace them with a solid implant.13 Orbital osteotomies are reserved for the child whose orbital growth is complete. Cicatricial retraction of the conjunctival sac occurs, and total eyelid reconstruction is always required.13 Self-expanding hydrogel both improves the bony orbit and encourages soft-tissue growth.15 The expander provides a good alternative for the anophthalmic orbit, increasing volume up to 10-fold. A newer injectable self-inflating hydrogel pellet to treat microphthalmos children also shows promise.16 No report yet exists for using hydrophilic osmotic expanders or injectable self-inflating hydrogel pellet in an anophthalmic socket of RB children.

According to Wolff's law published in the 19th century, pressure on the long endochondral bones is responsible for bone transformation and leads to bone growth.17 However, the exact mechanism of orbital bone growth is still unknown. Several animal studies have been conducted to determine the factors related to orbit size. In an anophthalmic feline model, orbital development after enucleation without an implant was 47.6% of the normal side orbit, and orbital growth with 8 mm and 12 mm implants was 46.5% and 57.0% of the contralateral orbit, respectively.18 The use of expandable orbital implants showed socket size proportional to volume implanted.19 A neonatal swine model was used to measure intraorbital pressure.20 When orbital expansion was near normal pressure (20 mm Hg), radial dimension increased 8%. In high-pressure orbits (60 mm Hg), the radius increased 16%, a statistically significant amount. This study demonstrated not only that orbital volume increased with intraorbital pressure, but also that it was possible to maintain normal orbit size if near-normal intraorbital pressure could be achieved.20

Since the overall mechanism controlling orbital development remains unclear, we may presume that several factors determine the orbit size in an anophthalmic socket with orbital implant. Prior studies noted that the younger the child at the time of enucleation, the greater the resulting orbital contraction.21 This is probably due to the rapid facial growth that occurs in the first 3 years of life. Orbital development improved in RB survivors treated with implants.9

The introduction of porous implants, such as porous polyethylene or hydroxyapatite implants, led some to hypothesise that they would improve orbital growth and cosmetic outcomes. Hydroxyapatite orbital implants are commonly used during enucleation, evisceration and secondary orbital implant surgeries. These porous orbital implants permit better fibrovascular ingrowth of the microporous structure and offer less implant migration, extrusion and infection.11 22 Owing to the potentially better vascularised environment created by the porous implant, orbital growth retardation after enucleation could be reduced by orbital stimulation. Besides porous implants, a hypervascularised socket has also been developed using dermis fat. However, animal studies using free-fat and dermis grafts for orbital reconstruction were disappointing and showed no significant effect on orbital bone growth.23 Instead, porous polyethylene implants displayed orbital volumes not significantly smaller than normal orbits.23 In the present study, statistically significant orbital growth retardation after enucleation was observed even with porous hydroxyapatite implants. Another study compared hydroxyapatite (median, 18 mm) with silicone (median, 16 mm) implants in enucleated patients with RB.24 While the median orbital volume in the hydroxyapatite group was slightly larger than in the silicone group, the hydroxyapatite implants showed no better symmetry of orbital volume than did traditional silicone implants. These findings also concluded that the type of orbital implant alone may not determine orbital growth in young children with anophthalmos.

Kaste et al showed no orbital volume differences for children treated at less than 1 year of age and those treated later.8 In the present study, we found significantly greater volume differences in patients treated at less than 1 year of age. The result was similar to Peylan's report that orbital growth retardation was most prominent in children treated in the first year of life.6 In addition, we found age at operation to be significantly negatively correlated with orbital volume difference (p=0.033). The equation using orbital volume differences as a dependent variable and operation age as an independent variable was: orbital volume differences=−0.024×operation age (months)+1.64 with p=0.033 for operation age. We assumed that the impact of enucleation, even stimulated by the mechanical force exerted by hydroxapatite implants, was also much more significant when the operation was performed in younger children. The discrepancies in speed of orbital development between enucleated and contralateral orbits result in a more asymmetrical orbital volume for children less than 1 year of age.

There are some limitations to this study. The power of this study was restricted by the small sample size. The CT scans to calculate orbital volume were performed at different ages for each patient. Measurements taken at different time points during orbital development may affect the result. Also, the volume of the ocular prosthesis was not taken into account. The prosthesis itself had been suggested to be important in orbital growth of the anophthalmic socket.25 The lack of data on the ocular prosthesis may have affected the results. Additionally, orbital volume calculated from CT images may not reflect actual orbital size. First, the calculation depended on manual marking of the orbital cavity margin, and slight errors could have been made in marking. Second, orbital cross-section area of CT image slices could be overestimated or underestimated if the head was tilted during CT imaging. The results could be less accurate as a result.

In sum, orbital growth retardation after enucleation in survivors of RB was significant, even with a porous hydroxyapatite implant. The impact of enucleation on orbital development was significantly correlated with age at operation and was more prominent in children who underwent surgeries before the age of 1 year. Since other factors likely impact orbital development, further studies are needed to determine which factors affect anophthalmic orbital growth retardation.



  • See Editorial, p 601

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by the National Taiwan University Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles