Intrascleral invasion. Undifferentiated cells, scarce cytoplasm, and hyperchromatic nuclei infiltrate the sclera without surpassing it (hematoxylin-eosin, original magnification ×4).
Transscleral invasion. Tumoral cells surpass the sclera, invading the orbit (hematoxylin-eosin, original magnification ×2).
Cuenca A, Giron F, Castro D, Fandiño A, Guitter M, de Dávila MTG, Chantada G. Microscopic Scleral Invasion in RetinoblastomaClinicopathological Features and Outcome. Arch Ophthalmol. 2009;127(8):1006-1010. doi:10.1001/archophthalmol.2009.174
To describe the clinical and pathological features of patients with retinoblastoma and microscopic scleral invasion.
We reviewed all pathology slides of patients with microscopic scleral invasion who were included in 3 prospective treatment protocols (1988-2007). All patients received adjuvant chemotherapy (moderately intensive chemotherapy in the first 2 protocols or a more intensive combination in the third one). Only patients with cut-end invasion received orbital radiotherapy.
Thirty-two of 386 patients had enucleated eyes with intrascleral (21 cases) and transscleral (11 cases) invasion. Of these cases, 20 had tumor invading the optic nerve beyond the lamina cribrosa, with 6 of these having tumor at the surgical margin. Sixteen were treated with moderately intensive chemotherapy and 16 received a higher-intensity regimen. Five-year overall survival was 0.77. Seven patients had an extraocular relapse (central nervous system metastasis, n = 4; systemic metastasis, n = 2; and involving the orbit, n = 3, isolated in 1 and combined with central nervous system disease in 2). All patients who had a relapse died. Patients receiving the intensive regimen had a significantly better outcome (P = .007).
Microscopic scleral invasion might be a risk factor for extraocular relapse, and more intensive chemotherapy results in improved survival for these patients.
Retinoblastoma is the most common intraocular malignancy in childhood.1 Most cases in affluent nations are diagnosed when the disease is in the intraocular stage and the disease-free survival is more than 90%.2 On the other hand, in less developed countries, it is usually diagnosed when extraocular dissemination has occurred and the survival is much lower.3 In this setting, patients usually present with overt extraocular disease massively involving the orbit or with systemic or central nervous system (CNS) metastasis. They may also present with disease apparently limited to the eye but with microscopic extraocular extension or intraocular disease with high-risk pathological features. Some of these patients may be at risk for extraocular relapse and adjuvant therapy is usually given in hope of reducing this risk.4 Postlaminar optic nerve (PLONI), choroid, and scleral invasion have been postulated as risk factors for extraocular relapse in these cases.4- 8 Scleral invasion may be evident after massive orbital invasion or it may be detected only microscopically. Microscopic scleral invasion has been identified as a significant risk factor for extraocular relapse in retinoblastoma by some studies; however, its prognostic significance has been less well studied than other pathological risk factors, such as choroidal or optic nerve involvement.5,7,9- 12 Scleral invasion of retinoblastoma occurs when the tumor extends beyond the choroid and true scleral invasion should be differentiated on histopathological grounds from “floaters,” which are free neoplastic cells that are dragged passively to the sclera during tissue processing thereby simulating scleral invasion. Relatively little information is available on the outcome of patients with microscopic scleral invasion when treated with current therapy. In addition, scleral invasion is usually the consequence of advanced disease, so it may occur simultaneously with other pathological risk factors, such as PLONI, making the evaluation of independent risk factors more complicated.9,13 Patients with this feature have been treated under prospective studies in our center since 1987 and our aim over the years was to try to identify patients with significantly higher risk of relapse based on a careful histopathological examination to tailor adjuvant therapy in a risk-adapted fashion.14,15 However, because of its rarity, each study included a relatively small number of patients with scleral invasion. Therefore, to collect information in a larger cohort, we sought to assess the outcome of patients with microscopic scleral invasion treated over 3 prospective treatment protocols done at our institution.
The clinical records of all patients with microscopic scleral invasion seen at our institution and included in 3 successive prospective protocols from January 1988 to December 2007 were analyzed. The first protocol took part from 1987 to 1993 and included a total of 101 patients; the second one ran from 1994 to 2001 and included a total of 169 patients.14,15 The third one was opened in 2002 and closed in 2007 and included a total of 162 patients; its final results have not been published yet. We evaluated patients with newly diagnosed disease and those in whom therapy failed. Only patients whose eyes were enucleated and who were treated at our institution were included in this analysis. Histologic slides were analyzed by a pathologist (M.T.G.D.) at diagnosis, retrieved for this review, and analyzed by 3 pathologists (M.T.G.D., F.G., and D.C.). From 1990, all enucleated eyes underwent a uniform, previously reported, comprehensive procedure for pathological examination.13 In case of discrepancy between the pathologists, the decision was made by consensus. Choroidal invasion was categorized as full when tumor cells invaded more than 50% of the choroid or when more than 1 tumor cluster was evident.15 We also considered the number of tumor clusters in the sclera as an indirect estimation of tumor burden and we divided our population into 2 groups: those with 3 or fewer clusters with presumably less tumor burden and those with more than 3 tumor clusters reflecting a higher tumor burden.
For the purpose of this study, scleral invasion was categorized as (1) intrascleral, which included cases when tumor cells invaded the sclera without surpassing the episclera (outer layer contiguous to periorbital adipose tissue) (Figure 1) or (2) transscleral, which included cases when tumor cells invaded the whole width of the sclera, invading the periorbital tissue (Figure 2).
All patients were included in 3 successive prospective treatment protocols with different recommendations for adjuvant therapy after institutional review board approval and written informed consent. Adjuvant chemotherapy was given to all patients with any degree of scleral invasion, but the regimen used varied among the protocols and depended on the occurrence of other concomitant pathological risk factors, such as PLONI. In the first protocol, the chemotherapy regimen included standard doses of cyclophosphamide, doxorubicin, and vincristine sulfate, along with intrathecal therapy (Table 1, regimen 1).14 In the second protocol, patients with isolated scleral invasion or those with associated PLONI received 8 cycles of chemotherapy alternating cyclophosphamide and vincristine with carboplatin and etoposide (Table 1, regimen 2).15 From this point, all patients with cut-end invasion of the optic nerve received a higher-dose regimen including cyclophosphamide, idarubicin hydrochloride, and vincristine alternating with higher-dose carboplatin and etoposide (Table 1, regimen 3) along with orbital irradiation (45 Gy). Orbital radiotherapy was not used for any patient without invasion of the optic nerve resection margin. In the third protocol, all patients with any degree of scleral invasion, with or without associated PLONI, received this intensive chemotherapy regimen (Table 1, regimen 3).13 From 1995, chemoreduction with carboplatin and vincristine with or without etoposide was offered for patients in whom eye preservation was possible.16 For the purpose of this study, we classified the adjuvant chemotherapy regimens into 2 groups: lower intensity (regimens 1 and 2) and higher intensity (regimen 3).
Contingency tables were constructed and χ2 or Fisher exact tests were used for categorical variables, with 2-tailed calculation of the P value when necessary. The Mann-Whitney test was used for continuous variables. “Event” was defined as any extraocular relapse (including orbital recurrence in the treated or the fellow eye and distant or CNS relapse) and event-free survival curves were calculated according to Kaplan-Meier. Survival status was updated to June 2008.
Three hundred eighty-six evaluable patients with nonmetastatic retinoblastoma were admitted to our hospital in the study period. Of these, 34 consecutive patients (8.8%) with microscopic scleral invasion were analyzed. Two patients were excluded after pathological review because scleral invasion was doubtful. Archival material was not available for review in 2 patients. However, these 2 patients were analyzed for the other parameters and considered for survival since scleral invasion did not change the staging and further management of these patients because they had associated invasion of the optic nerve resection margin. Twenty-three (71.8%) had unilateral retinoblastoma and 9 (28.2%) had bilateral retinoblastoma. Median age at diagnosis was 27 months (range, 1-74 months) and the male to female ratio was 1.1:1. Median follow-up was 76 months (range, 6-185 months).
For the 9 patients with bilateral retinoblastoma, 2 had both eyes enucleated at diagnosis. In 1 of them, invasion of the optic nerve resection margin was also evident in the contralateral eye and there was an extraocular relapse (CNS metastasis). The second patient had isolated choroidal invasion in the contralateral eye and is now disease free. Three patients had 1 eye with scleral invasion enucleated and the remaining eye preserved and all are now disease free. The remaining 4 patients had both eyes enucleated at different times during follow-up. The contralateral eye showed invasion of the optic nerve resection margin in 1 patient who had an extraocular relapse (systemic metastasis), choroidal invasion in 2 patients (1 had an extraocular relapse, described later), and no extraretinal invasion in 1 patient who is disease free.
Twenty-one patients (66%) had intrascleral invasion and 11 (34%) had transscleral invasion. Optic nerve invasion was present in 23 patients (72%). Invasion was prelaminar in 3 (9.3%), PLONI in 14 (43%), and involving the resection line in 6 (19%) of the patients.
Seven patients had events during follow-up. The median time from diagnosis to the event was 9 months (range, 5-20 months). The CNS was involved in 4 cases of relapse (2 with concomitant orbital relapse). There were 2 cases of systemic metastasis involving bone metastasis without CNS or orbital involvement. Isolated orbital relapse occurred in 1 patient. In that patient with bilateral disease, resistance to chemotherapy and radiotherapy was evident when conservative treatment was attempted and an orbital relapse occurred soon after enucleation while the patient was still receiving adjuvant therapy. Orbital exenteration was done, but it was followed by lymph node and CNS metastasis and the patient died of retinoblastoma. All patients who had a relapse died of progressive disease after second-line treatment. Scleral invasion of patients who had a relapse was categorized as transscleral in 4 and intrascleral in 3. Postlaminar optic nerve invasion was present in 3 patients with events (2 had transscleral and 1 intrascleral invasion).
There were 11 patients with microscopic scleral invasion and either no or prelaminar optic nerve involvement (10 had intrascleral and 1 had transscleral invasion). Two had a relapse (1 isolated orbital relapse, described earlier, and 1 systemic metastasis). Both had intrascleral involvement.
All patients who had events had received lower-intensity protocols for adjuvant therapy, and 3 received adjuvant orbital radiotherapy because of concomitant invasion of the optic nerve resection margin. The comparative analysis between the cohorts with and without events is shown in Table 2. The probability of event-free survival and overall survival at 5 years from diagnosis was 0.77 (95% confidence interval, 0.58-0.92).
Patients receiving high-intensity adjuvant chemotherapy had a better outcome compared with those receiving lower-intensity regimens (Table 2). Both populations were comparable in terms of optic nerve involvement or degree of scleral compromise, except that a higher number of patients who received higher-intensity chemotherapy had PLONI (Table 3).
There is little published information about scleral invasion in retinoblastoma. Most recent studies included fewer than 20 patients. We were able to find only 3 studies including more than 10 patients who provided survival data.5,12,17 Twenty-one of 32 patients were disease free in the Stannard et al study12 dealing with patients treated from 1952 to 1975 and 7 of 14 were disease free in the Khelfaoui et al study5 including patients treated from 1977 to 1990. More recently, Antoneli et al17 reported 14 patients with microscopic scleral invasion and the probability of 3-year survival was 58%. However, their survival results were analyzed together with patients with overt orbital extension. Therefore, with a survival rate of 77%, our results compare favorably with those series. Given the lack of information on this population, many questions about their management, such as the need for adjuvant therapy for those with only intrascleral invasion, the type of chemotherapy to be given, and the need for orbital radiotherapy, remain controversial.4,9,10,18 Our series, including 32 patients with microscopic scleral invasion, uniformly staged and treated under prospective treatment protocols, provides information on these issues, but only a multicentric study including more patients could provide a definitive answer to some of these questions. Our study showed that microscopic scleral invasion is an uncommon feature in patients with retinoblastoma in our setting. As reported previously, it tended to occur together with other major pathological risk factors, such as PLONI, reflecting aggressive disease probably due to late diagnosis.9,13 However, since only a limited area of the enucleated eye is sampled for pathological examination, microscopic scleral invasion or other pathological risk factors may be missed in some cases, even when a comprehensive pathological examination is done; the true incidence of these features may be higher.
Interestingly, even though all patients received adjuvant chemotherapy, some of them still had relapses. Subgroup analysis was difficult because of the relatively low patient number and no multivariate analysis could be done because of that, but despite these limitations, our results seem to suggest that patients receiving higher-intensity regimens for adjuvant therapy do better. However, although the decision of giving adjuvant therapy depended also on the presence of other risk factors, such as optic nerve invasion, and this evaluation was retrospectively designed, both subgroups of patients receiving lower- and higher-intensity treatments were comparable. The adequate chemotherapy regimen for adjuvant therapy in these patients is yet to be determined, but according to our study, it should include drugs at a dose that enables an adequate passage to the CNS to avoid the occurrence of CNS-involved relapse, which was the most frequent event. A similar conclusion was reached by Antoneli et al17 in their series of 30 patients with orbital involvement. In their series, also including patients with overt orbital disease, the addition of ifosfamide and etoposide to the chemotherapy regimen resulted in better disease-free survival rates.17 The occurrence of microscopic scleral invasion may suggest the possibility of an increased risk of orbital relapse because of possible seeding of tumor cells in the orbit. To prevent this event, some authors recommend orbital radiotherapy in addition to adjuvant chemotherapy.18 However, orbital radiotherapy after enucleation of an eye with retinoblastoma is associated with severe cosmetic sequelae and endocrinological disorders when the chiasm is included in the radiation field.19 In our study, only patients with concomitant optic nerve invasion beyond the resection margin received orbital radiotherapy. There was no case of orbital relapse in children receiving the higher-intensity regimen, suggesting that an adequate local/regional control could be achieved without radiotherapy if intensive chemotherapy regimens are used. In our series, orbital relapse occurred in 3 patients, of whom 2 had other metastatic sites, again suggesting that scleral invasion is indeed a marker for aggressive disease with the potential for systemic dissemination. Systemic metastases were found concomitantly to orbital relapse in 62% of the cases in a recently published series, suggesting that most patients with orbital recurrence have systemic disease.20 Therefore, our limited data do not seem to support the use of orbital radiotherapy in patients with microscopic scleral invasion when there is no concomitant invasion of the resection margin of the optic nerve and an intensive chemotherapy is given in an adjuvant setting.
We also examined other putative risk factors for extraocular relapse associated with scleral invasion. We speculated that the extent of scleral invasion may play a role, so we analyzed separately the outcome of patients who had only intrascleral invasion with those in whom the tumor extended beyond the sclera (transscleral invasion). Both subgroups received the same adjuvant therapy but we were not able to detect any difference in outcome. Extraocular relapse was seen also in 2 of 11 patients with less than PLONI, suggesting that scleral invasion per se may be a marker for aggressive disease. We also compared patients with more than 3 tumor foci of scleral invasion, possibly reflecting more extensive disease, and there was a trend for poorer prognosis, although it did not reach statistical significance. The impact of other concomitant pathological risk factors, such as PLONI, was also analyzed, but we were not able to find a significant difference. However, the occurrence of relapse could be determined for the presence of invasion of other ocular coats in some of our patients and we could not evaluate the impact of scleral invasion as an independent factor by multivariate analysis because of the relatively low patient number and the low number of events. The importance of optic nerve invasion may be critical since, for example, all the CNS relapses occurred in patients with PLONI with or without tumor at the resection margin in our cohort. However, the relapse pattern is not clearly associated with globe pathological findings as reported by other studies so it is difficult to determine the role of invasion at a given ocular coat for a specific relapse site.4,21 Therefore, a larger cohort is needed to assess the role of scleral invasion as an independent risk factor for extraocular relapse. We also speculated that patients in whom preoperative chemotherapy or radiotherapy had been given for a preservation attempt could be at higher risk because of presumed chemotherapy or radiotherapy resistance. However, we found no statistically significant increased risk of metastasis in that cohort, but the population receiving preenucleation treatment was small. In some of these cases, we speculate that scleral invasion would not be a consequence of advanced, aggressive intraocular disease but it may have occurred because intensive local treatment may have caused a disruption of the anatomy of the eye, leading to a narrowing of the choroidal layer allowing tumor cells to reach the sclera more easily. However, since only a limited area of the enucleated eye is sampled for pathological examination, microscopic scleral invasion or other pathological risk factors may be missed in some cases, even when a comprehensive pathological examination is done.
To conclude, microscopic scleral invasion is an uncommon feature in retinoblastoma and is a marker for aggressive disease with the potential for systemic dissemination. Since it occurred often in combination with other pathological risk factors, such as PLONI, and most events included relapse at the CNS or distant metastatic sites, patients presenting with these features may benefit from an intensive systemic adjuvant therapy, including drugs with good penetration to the CNS. Orbital radiotherapy may not be needed in this population.
Correspondence: Guillermo Chantada, MD, Department of Hemato-Oncology, Hospital JP Garrahan, Combate de los Pozos 1881, C1245AAL, Buenos Aires, Argentina (firstname.lastname@example.org).
Submitted for Publication: January 23, 2009; final revision received April 27, 2009; accepted April 28, 2009.
Author Contributions: Dr Chantada had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: Supported in part by a grant from the Fund for Ophthalmic Knowledge (New York, New York) (Drs Chantada and Guitter) and St. Jude Children's Research Hospital (International Outreach Program) (Memphis, Tennessee) (Drs Giron and Castro).
Role of the Sponsor: These organizations had no role in the design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.