Orbital Complications and Cost Analysis of Enucleation by Myoconjunctival versus Conventional Techniques for Retinoblastoma

Abstract

Introduction

The aim of the study was to compare complication rates and hospital costs of myoconjunctival versus conventional enucleation techniques in retinoblastoma.

Methods

This retrospective cohort and cost analysis reviewed patients with retinoblastoma treated by primary or secondary enucleation between 2003 and 2021 and a minimum 6-month follow-up. Cases were reviewed for three postsurgical complications: chronic conjunctivitis, implant exposure/extrusion, and cellulitis. Cases were excluded if surgery was performed elsewhere or documentation was incomplete. Treatment costs were estimated based on two sample cases billed in 2021 that manifested the studied complications and represented each of the two surgical techniques. Univariate and multivariate analyses were applied to compare complication rates and treatment costs.

Results

Included were 180 eyes (179 patients); 239 eyes (227 patients) were excluded. Patients had median age of 18.9 (0–104.4) months at diagnosis, the majority were male (94, 52%), with unilateral (115, 64%) group D or E (163, 91%) eyes. Enucleation was performed by conventional techniques in 107 eyes (59%) and by myoconjunctival approach in 73 (41%). Orbital complications occurred in 61 eyes (34%) during a median follow-up of 7.9 (0.5–33.7) years, more frequently in the conventional technique group (p = 0.014). The myoconjunctival technique had significantly lower costs for implant price (p < 0.001) and estimated treatment cost, including complication management (p < 0.001).

Conclusion

Enucleation by myoconjunctival technique showed significantly less complication burden and treatment cost, indicating advantages over conventional approaches. Study limitations include the retrospective nature, confounders’ complexity, and follow-up time variations.

Introduction

Progress in retinoblastoma care in the latest decades has improved children’s survival and quality of life, notably through eye-sparing therapies [1–4]. However, enucleation remains the most curative and accessible intervention for disease management, being especially important in advanced stages [3–5]. Approximately 50% of retinoblastoma eyes staged cT2 or cT3 (groups D or E, by the International Intraocular Retinoblastoma Classification, IIRC [6]) may require primary enucleation at diagnosis [4, 6–8]. In addition, many children benefit from secondary globe removal when eye salvage fails or results in a painful, blind eye [4, 5]. In underserved locations, enucleation may be the single modality available to treat retinoblastoma when medical resources are limited [9, 10]. Eye removal can also offer prompt disease control, sparing children the burden of multiple examinations under anesthesia.

Historically, various enucleation techniques and orbital implants have been used to reform the orbital volume with few complications [5, 11]. The now common approach to enucleation (conventional techniques, CTs), consists of suturing the extraocular muscles to the front of an orbital implant manufactured with porous material, originally hydroxyapatite [12–15]. The porous implant allows internal fibrovascular ingrowth, which creates adherence to the orbital tissues, to theoretically result in good prosthetic motility. While no study has consistently confirmed these benefits, complications related to porous designs include conjunctival erosion and consequent implant exposure/extrusion [11, 13–17]. Variations of CTs include implant wrapping with donor sclera or Vicryl mesh to protect overlying conjunctiva, and various orbital porous and nonporous implants. Outcomes have not been determined by comparative studies [17].

The “myoconjunctival technique” (MT) was developed to address some of the issues related to porous implants and muscle placement and to enhance later prosthetic motility [18–21]. By the MT approach, extraocular muscles are attached to the conjunctival fornices and a nonporous implant is placed deep behind the posterior Tenon’s capsule, creating deep conjunctival fornices for external prosthesis centration, and movement [11]. Benefits of the MT technique were suggested in two randomized trials [11, 18]. Nonintegrated (nonporous) implants used with MT are less costly than porous implants [11, 17, 21]. The retinoblastoma team at SickKids has employed the MT in the standard of care for enucleation for retinoblastoma since 2010.

No study, to our knowledge, has compared complication rates between CTs and MT in retinoblastoma. However, a global survey on enucleation trends for retinoblastoma reported a general preference for CTs with porous implants, which is endorsed by other studies [19, 22, 23]. Porous/integrated implants have been chosen in approximately 70% of enucleations performed in developed countries and 30% in underdeveloped nations, with no worldwide consensus for children with retinoblastoma [19]. In countries where orbital implants are not readily available, dermal-fat grafts continue as a viable option for prompt anophthalmic socket reconstruction [24, 25].

In 2018, a systematic review by the American Academy of Ophthalmology [17] found that adverse events were considered low with CTs and MT. However, this report noted a “paucity of data on motility and absence of direct, objective comparisons of porous and nonporous implants.” While implant exposure was the most frequent complication, few studies focused on prevalence of complications in children with retinoblastoma or compared these surgical techniques. We present comparison of orbital complications and treatment costs following enucleation by MT versus CT in a cohort of retinoblastoma patients.

Methods

Study Design

This retrospective single-institution cohort study reviewed clinical and institutional costs for patients diagnosed with retinoblastoma and treated with primary or secondary enucleation between January 2003 and November 2021 at The Hospital for Sick Children (SickKids), Toronto, Canada. The study protocol was approved by the Institutional Research Ethics Board and followed the tenets of the Declaration of Helsinki. Patient and parental informed consent was waived for this retrospective study.

Eligibility Criteria

Eligible cases were identified from the in-hospital electronic database (eCancer Care – Retinoblastoma) used in routine patient care since 2003. All retinoblastoma cases treated with enucleation at any point of care with a minimum follow-up of 6 months after the surgical procedure were eligible, subsequently confirmed by a complete review of medical records. Cases were excluded if medical records were incomplete or had inconsistent clinical information, were inaccessible or archived, or if the enucleation surgery was performed elsewhere.

Data Collection

Medical records were reviewed for the following clinical and demographic features: age of retinoblastoma diagnosis; laterality of enucleated eye; eye disease group by IIRC [6]; date of enucleation; surgical technique (MT vs. CTs); type, size, and wrapping of the orbital implant; indication for enucleation (primary vs. secondary); and follow-up duration. Three post-enucleation common orbital complications were selected as outcomes: (a) chronic conjunctivitis or ocular discharge reported for at least 4 weeks, requiring medical management with close surveillance or topical antibiotic eyedrops, with/without culture; (b) periorbital or orbital cellulitis diagnosed by clinical exam and/or imaging requiring specific management such as systemic antibiotics, hospitalization, steroid therapy, or surgical intervention; and (c) implant exposure or extrusion detected on clinical exam. Surgery in all cases was performed by the same group of retinoblastoma specialists led by the senior author (B.G.). The surgical enucleation steps followed previous reported descriptions [11].

Cost comparison between CTs and MTs was based on the price of orbital implants and total treatment costs using SickKids hospital data from one MT case and one CT case sampled from 2021. Information was extracted from the records submitted to the Ontario Health Insurance Plan, Canada, and included the baseline enucleation surgery cost and hospital costs to manage the two most clinically relevant enucleation complications, implant exposure/extrusion, and cellulitis. Costs included hospital admission for short- or long-term care, surgical and examination procedures, medications, laboratory, and imaging tests. Ambulatory visits and standard follow-up procedures related to routine retinoblastoma care were disregarded in the cost evaluations.

Data Analysis

Analyses comprised descriptive statistics and univariate comparisons based on Student’s t test (continuous variables) and logistic regression (categorical variables). Complication rates were analyzed by univariate and multivariate modeling adjusted for five clinically and statistically relevant confounders: retinoblastoma staging by IIRC [6], indication for enucleation (primary/secondary), type of implant, use of implant wrapping, and follow-up duration. Time to complication by treatment group was calculated using Kaplan-Meier curves considering the two most clinically relevant complications, cellulitis and implant exposure/extrusion. For cost analyses, values were converted from Canadian (CAD) to American dollar (USD) currencies using the annual average exchange rates (1 USD = 1.2535 CAD) by the Bank of Canada for 2021 [26]. Analyses were performed on STATA (version 17.0; StataCorp LP, College Station, TX, USA), with p < 0.05.

Results

Patient Characteristics

A total of 180 enucleated retinoblastoma eyes (179 patients) met the inclusion criteria, with 227 patients (239 eyes) excluded due to incomplete or inconsistent clinical information in the chart (145), inaccessible or archived chart (45), or enucleation performed elsewhere (37). Retinoblastoma was diagnosed at a median of 18.9 months (0–104.4) of age and was most commonly unilateral (115, 64%). The enucleated eye was IIRC group B (5, 3%), C (12, 7%), D (93, 52%), and E (70, 39%), with no group A eye. Enucleation was primary in 143 (80%) and secondary in 37 (21%). One patient with bilateral retinoblastoma had one eye enucleated primarily and the other eye secondarily. Six eyes secondarily enucleated had a history of prior radiotherapy, one external beam, and five plaque brachytherapy.

Enucleation and Implant Specifics

Enucleation was performed by CT in 107 eyes (59%) and by MT in 73 (41%). Of the eyes with prior radiotherapy, three were enucleated by CT (including the case with previous external-beam radiotherapy) and two by MT. The most common type of orbital implant used in the CT group was bioceramic (85, 79%), followed by polymethyl methacrylate (PMMA) (17, 16%), acrylic (3, 3%), and silicone (2, 2%). All eyes enucleated by MT received an unwrapped PMMA implant (73, 100%). The median implant size was 18 mm (14–20). Of the eyes enucleated by CTs, 22 (21%) had implant wrap at the time of surgery. Comparative details between MT and CT enucleation are summarized in Table 1.

Table 1.

Patient baseline characteristics, enucleation complication details, and surgical costs for MTs and CTs by univariate and multivariate analysis

	MT (n = 73 eyes)	CTs (n = 107 eyes)	p value
Patient characteristics
Age of diagnosis, months (median, range)	18.2, 0–104.4	19.6, 0.3–82.5	0.72A
Sex, n (%)			0.45B
Male	41 (56)	53 (50)
Female	32 (44)	54 (50)
Enucleated eye, n (%)			0.46B
Right	33 (45)	52 (49)
Left	40 (55)	55 (51)
IIRC stage, n (%)			0.26B
B	4 (5)	1 (1)
C	6 (8)	6 (6)
D	37 (51)	56 (52)
E	26 (36)	44 (41)
Enucleation indication, n (%)			0.06B
Primary	53 (73)	90 (84)
Secondary	20 (27)	17 (16)
Follow-up
Duration, years (median, range)	4.5, 0.6–11.3	12.0, 0.5–33.7	<0.001 A
Implant features
Implant type and material, n (%)			<0.001 B
Integrated
Bioceramic	0	85 (79)
Nonintegrated
PMMA	73 (100)	17 (16)
Acrylic	0	3 (3)
Silicone	0	2 (2)
Type of wrap, n (%)			<0.001 B
No wrapping	73 (100)	84 (78.5)
Vicryl mesh	0	11 (10.3)
Mersilene	0	5 (4.7)
Donor sclera	0	7 (6.5)
Size of implant, mm (median, range)	20, 16–20	18, 14–20	<0.001 A
Enucleation complications
Orbital complication, n (%)
Yes	12 (16)	49 (46)	<0.001 B
No	61 (84)	58 (54)	0.014 C
Type of complication, n (%)
Conjunctivitis	11 (15)	34 (32)	0.01 B
Implant extrusion/exposure	2 (3)	24 (22)	<0.001 B
Cellulitis	1 (1)	13 (12)	0.008 B
Number of complications, n (%)			<0.001 B
0	61 (84)	55 (51)
1	11 (15)	36 (34)
2	0	13 (12)
3	1 (1)	3 (3)
Time to complication,D years (median, range)	3.7, 0.6–11.3	7.3, 0.6–33.7	0.03 A
Surgical costs
Implant price, US$ (median, range)	44.4 (n/a)	611.8 (44.4–611.8)	<0.001 A
Treatment cost,B US$ (mean, range)	5,224.7 (4,969.4–15,812.2)	7,977.9 (4,969.4–33,560.3)	<0.001 A

^AStudent’s t test.

^BUnivariate logistic regression.

^CMultivariate logistic regression adjusting for retinoblastoma staging (International Intraocular Retinoblastoma Classification), enucleation indication (primary/secondary), type of implant, use of implant wrapping, and follow-up duration.

^DOnly cases with implant exposure/extrusion or cellulitis.

Enucleation Complications

Complications were identified in 61/180 eyes (34%) at a median postoperative follow-up of 7.9 years (0.5–33.7). Complications occurred in 12 (16%) eyes treated by MT and 49 (46%) by CT. Chronic conjunctivitis was the most common complication in both groups (45, 25%), followed by implant exposure/extrusion (26, 16%) and cellulitis (14, 7.8%). The child who received external-beam radiotherapy later developed both cellulitis and implant exposure/extrusion.

By univariate analysis, all three complications were significantly more prevalent within the CT group (p < 0.001; p = 0.01) (also shown in Table 1). Ten of 17 eyes treated by CTs with PMMA implant had complications: 6/17 chronic conjunctivitis, 1/17 implant exposure/extrusion, and 3/17 cellulitis (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000531871). Within the CTs, no difference was found between eyes treated by integrated (bioceramic) versus nonintegrated (PMMA, acrylic, or silicone) implants (p = 0.35). Univariate analysis for any of the three complications between the MT and CTs showed a significantly lower number of complications among the MT group (p < 0.001). The MT technique also had a lower number of complications which required simple (chronic conjunctivitis) and complex (implant exposure/extrusion plus orbital cellulitis) management compared to CTs (p = 0.02 and p < 0.001, respectively).

A model adjusted for the five most clinically and statistically relevant parameters confirmed that complications were less prevalent among eyes managed by MT compared to CTs (p = 0.014). Univariate analysis of complications showed no differences between primarily versus secondarily enucleated eyes (online suppl Table 2). The MT enucleations had a shorter time to complication (implant exposure/extrusion or cellulitis) compared to CT enucleations (p = 0.03) (Fig. 1).

Fig. 1.

Kaplan-Meier comparison between conventional (CTs) versus myoconjunctival (MC) techniques for time to cellulitis or implant exposure/extrusion.

Surgical Costs

The median cost of nonintegrated orbital implants (PMMA) used in MT was USD 44.4 versus USD 611.8 (44.4–611.8) in the CT group (p < 0.001). The total cost of treatment considering the two most clinically relevant complications (implant exposure/extrusion and cellulitis) was lower for MT enucleations [USD 5,224.7 (4,969.4–15,812.2)] compared to CTs [USD 7,977.9 (4,969.4–33,560.3)] (p < 0.001). The costs of CTs were likely explained by the predominant use of porous implants and higher complication burden.

Discussion

This study confirms a significant advantage of enucleation using MT for retinoblastoma compared to CTs, which mainly involve integrated/porous implants. In this almost two decades of retinoblastoma management, the complication rate of MTs (16%) was significantly lower compared to CTs (46%). MT was also significantly less expensive than CTs, mostly attributable to the lower cost of nonintegrated (PMMA) implants and the higher orbital complication burden of CTs. Although MT has been used in our practice only for about a decade, no significant differences in patient characteristics were identified between the two surgical groups. In addition, no difference was observed in complication rates by enucleation indication (online suppl. Table 2).

MT enucleation was proposed in the 1970s by Coston, who speculated that prosthetic movement was optimized by anchoring the extraocular muscles to the conjunctival fornices [27]. Other authors supported this principle with satisfactory surgical results, which evolved into the MT as currently known [11, 18]. In 2010, 150 patients randomized to three treatment groups (“MT plus PMMA,” “muscle imbrication plus PMMA,” and “enucleation plus porous polyethylene”) and followed for approximately 15 months confirmed that MT was equivalent to or better than the traditional techniques (p = 0.001–0.002) [11]. In that study, there was no case of implant displacement/extrusion in the MT group versus 40% in the “CT plus PMMA” group, reinforcing the safety of MT.

Different factors have been proposed to explain the increased rate of complications after enucleation for retinoblastoma, including the nature of the infant orbit, healing specificities of the pediatric population, and immunosuppression induced by systemic chemotherapy, which raises the risk of orbital and skin infection [1, 3–5, 10, 13, 16, 19, 28, 29]. Our study identified chronic conjunctivitis and cellulitis as relatively frequent adverse events after enucleations by CTs, suggesting that porous/integrated implants could play a role in promoting or facilitating these processes. Among the cases studied, only 1 patient received external-beam radiation prior to enucleation, a now abandoned treatment routine, making radiation a less probable explanation for chronic orbital inflammation in this particular cohort. The nature of prostheses (customized vs. stock) could not be analyzed in our study, since all patients had customized protheses.

In a retrospective series of 347 adults enucleated with unwrapped hydroxyapatite (porous), recurrent conjunctival infection was identified in 8 (2.3%) eyes over a mean of 3.5 years, with no mention of cellulitis [15]. A study of 26 patients receiving different types of unwrapped integrated and nonintegrated implants [30] reported only 2 complications – persistent chemosis and giant papillary conjunctivitis – in those who received porous polyethylene implants. Jongman et al. [31] found a 0.6% rate of post-enucleation cellulitis in 179 enucleated eyes receiving sclera-wrapped acrylic implants at mean 12-month follow-up. Comparatively, in our study, rates of conjunctivitis and cellulitis among CTs were close to 20% and 10%, respectively.

Implant exposure or extrusion is a frequently reported complication following enucleation, estimated at approximately 10% for MT and up to 50% for techniques with porous implants, despite the more superficially placed orbital implant in the MT [5, 11, 13, 15–21, 32, 33]. This has been of particular focus among CTs, not only in the setting of porous implants but also related to muscle imbrication, which may not be stable enough to prevent implant displacement [11, 16, 17, 19, 33–36]. Potential contributing factors include the grinding of porous spheres on the conjunctiva, contributing to abnormal postsurgical scarring, neovascular ingrowth into the implant, and developmental changes in pediatric sockets [13, 29]. Implant wrapping may address these risks, especially for hydroxyapatite implants, but is less commonly performed with porous polyethylene implants, which do not require wrapping [11, 13, 15, 17, 30, 35, 37].

While the 22% rate of exposure/extrusion here identified for CTs should be considered in light of the variety of implant types analyzed together, a sub-analysis of cases treated with PMMA implants had only 1 case of implant exposure/extrusion, suggesting the benefit of nonintegrated implants. Overall, the 3% exposure/extrusion rate in the MT group highlights the advantages of this implant/technique. Similar to our results, a series of 321 unilateral retinoblastoma enucleations managed by MT reported a 7% rate of exposure/extrusion at a mean follow-up of 40 months [21]. In our study, the difference in implant size between MTs and CTs could be attributable to the use of wrapping coupled with the implant or other unassessed surgeon’s decision. Finally, this study also confirms that enucleation by MT is more affordable, as conveyed in previous analyses, whereby significantly lower implant and complication-related costs were noted [11, 19].

Conclusion

Enucleation by MTs showed significantly lower complications and treatment costs than by CTs. In addition, as almost all MT sockets received nonporous implants, conclusions also support benefits from this type of implant. A prospective, randomized, controlled trial would be ideal to further delineate the role of implant specifics versus enucleation technique on complications. Limitations of this study include its retrospective nature, the considerable number of excluded cases for data quality, complexity of confounders, role of retinoblastoma therapies before enucleation, factors underlying improvements in hospital care or surgeon’s experience over time, histologic and genetic specificities of the tumors, patient’s socioeconomic background, and differences in follow-up duration.

Statement of Ethics

This study was approved by the research committee at the Hospital for Sick Children (SickKids), University of Toronto, Toronto, Canada, and abided by the tenets of the Declaration of Helsinki. Full informed consent was waived for all participants given the retrospective nature of the study. Approval number: 1000041604.

Conflict of Interest Statement

None of the authors report any conflict of interest related to this study. Leonardo Lando reports past grants by the Pan-American Association of Ophthalmology (Sear Scholarship, 2019–2020) and the International Society of Ocular Oncology (ISOO, 2022).

Funding Sources

The study was not funded by any sponsor or institution.

Author Contributions

Leonardo Lando: concept, design, literature search, data acquisition, data analysis, statistical analysis, manuscript preparation, manuscript editing, and manuscript review. Ashwin Mallipatna and Stephanie N. Kletke: concept, design, data analysis, manuscript editing, and manuscript review. Brenda Gallie: concept, design, data analysis, manuscript preparation, manuscript editing, and manuscript review. All authors have reviewed and approved this manuscript. All authors meet the authorship requirements. All authors acknowledge that this manuscript represents original work not previously published.

Funding Statement

The study was not funded by any sponsor or institution.

Data Availability Statement

Data are not available for this study given its sensitive nature, population, and potentially identifiable content related to a rare medical condition. Further inquiries can be directed to the corresponding author.