Postlaminar Invasion Research Articles

Magnetic Resonance Imaging Can Reliably Differentiate Optic Nerve Inflammation from Tumor Invasion in Retinoblastoma with Orbital Cellulitis

To investigate the prevalence and magnetic resonance imaging (MRI) phenotype of retinoblastoma-associated orbital cellulitis. Additionally, this study aimed to identify postlaminar optic nerve enhancement (PLONE) patterns differentiating between inflammation and tumor invasion. A monocenter cohort study assessed the prevalence of orbital cellulitis features on MRI in retinoblastoma patients. A multicenter case-control study compared MRI features of the retinoblastoma-associated orbital cellulitis cases with retinoblastoma controls. A consecutive retinoblastoma patient cohort of 236 patients (311 eyes) was retrospectively investigated. Subsequently, 30 retinoblastoma cases with orbital cellulitis were compared with 30 matched retinoblastoma controls without cellulitis. In the cohort study, retinoblastoma MRI scans were scored on presence of inflammatory features. In the case-control study, MRI scans were scored on intraocular features and PLONE patterns. Postlaminar enhancement patterns were compared with histopathologic assessment of postlaminar tumor invasion. Interreader agreement was assessed, and exact tests with Bonferroni correction were adopted for statistical comparisons. Prevalence of retinoblastoma-associated orbital cellulitis on MRI was calculated. Frequency of intraocular MRI features was compared between cases and controls. Sensitivity and specificity of postlaminar optic nerve patterns for detection of postlaminar tumor invasion were assessed. The MRI prevalence of retinoblastoma-associated orbital cellulitis was 6.8% (16/236). Retinoblastoma with orbital cellulitis showed significantly more tumor necrosis, uveal abnormalities (inflammation, hemorrhage, and necrosis), lens luxation (all P < 0.001), and a larger eye size (P= 0.012). The inflammatory pattern of optic nerve enhancement (strong enhancement similar to adjacent choroid) was solely found in orbital cellulitis cases, of which none (0/16) showed tumor invasion on histopathology. Invasive pattern enhancement was found in both cases and controls, of which 50% (5/10) showed tumor invasion on histopathology. Considering these different enhancement patterns suggestive for either inflammation or tumor invasion increased specificity for detection of postlaminar tumor invasion in orbital cellulitis cases from 32% (95% confidence interval [CI], 16-52) to 89% (95% CI, 72-98). Retinoblastoma cases presenting with orbital cellulitis show MRI findings of a larger eye size, extensive tumor necrosis, uveal abnormalities, and lens luxation. Magnetic resonance imaging contrast-enhancement patterns within the postlaminar optic nerve can differentiate between tumor invasion and inflammatory changes.

Histopathological assessment of optic nerve invasion guided by radiological findings in enucleated globes with retinoblastoma

BackgroundOptic nerve (ON) invasion is an important high-risk feature, and an indicator for neoadjuvant chemotherapy and prognosis. We aim through this study to correlate the detected-ON invasion by Magnetic resonance imaging (MRI) with the corresponding confirmed histopathological level of invasion.MethodsA retrospective study of enucleated globes with the diagnosis of retinoblastoma received in the histopathology department(s) from January 2015 to December 2016 (2 years). Slides were reviewed for ON invasion assessment, charts were reviewed for basic demographic data. All patients underwent MRI under sedation upon diagnosis and MRI findings were collected for the above correlation.ResultsA total of 38 patients were included: 21 males and 17 females. 29 (77.3%) had unilateral involvement, 7 (18.4%) had bilateral involvement and 2 cases had trilateral disease. The overall mean age at diagnosis was 22.63 ± 15.15 months. Histopathological examination revealed ON invasion in 28 cases (74%) distributed as follows: prelaminar (31.6%), laminar (18.4%), and post-laminar (23.7%). MRI confirmed post-laminar ON invasion in 8 cases (true positive) but failed to detect this in 1 case. Additionally, MRI detected another 8 cases of ON invasion that were false positive on histopathology (accuracy: 63.3%; sensitivity: 88.9%; specificity: 72.4%; Positive predictive value (PPV): 50%; Negative predictive value (NPV): 95.5%).ConclusionsMRI is found to be less sensitive in evaluating prelaminar and laminar ON invasion (0.0 and 42.9%) compared to post-laminar invasion (88.9%). MRI has generally better specificity in detecting ON invasion irrespective of the invasion level. In our study, obtaining deeper and/or additional histologic sections from the other surface of the tissue block in cases where a post-laminar ON invasion by MRI is found but not confirmed histopathologically in routine sections is essential to avoid missing such an important high-risk feature.

Evaluation the risk of optic nerve invasion associated with optic nerve obscuration in advanced retinoblastoma

Objective: To evaluate the potential association between optic nerve invasion and optic nerve obscuration during treatment of advanced retinoblastoma. Methods: Retrospective case series study. Medical records of 77 patients (77 eyes) with advanced retinoblastoma (Group D/E) who were treated with primary or secondary enucleation in the Ophthalmology Department of Peking University People's Hospital from January 1st 2012 to December 31th 2015 were retrospectively reviewed. RetCam photographs under general anesthesia at diagnosis and each subsequent follow-up were evaluated for complete obscuration of the optic nerve. The primary endpoints included prelaminar invasion, postlaminar invasion and optic nerve transection invasion. Group difference was calculated with chi-square. Results: There were 46 boys and 31 girls in the study. The mean age at the first diagnosis was (27.1±22.1) months. The optic nerve was obscured in 62 eyes (80.5%) at the first diagnosis and 61 eyes (79.2%) at the last ocular examination prior to enucleation. Twenty-nine eyes (37.7%) underwent primary enucleation. Forty-eight eyes (62.3%) were treated with eye-preserving therapy, followed by enucleation. Fourteen eyes (18.2%) were in Group D and 63 eyes (81.8%) were in Group E. Histopathologic analysis of enucleated eyes without optic nerve obscuration (16 eyes) showed prelaminar invasion in 7 eyes, postlaminar invasion in 2 eyes and optic nerve transection invasion in 0 eyes. Histopathologic analysis of enucleated eyes with optic nerve obscuration (61 eyes) showed prelaminar invasion in 26 eyes, postlaminar invasion in 9 eyes and optic nerve transection invasion in 4 eyes. The difference between two groups did not achieve statistical significance (P=0.935, 1.000, 0.296). Histopathologic analysis of enucleated eyes with persistent complete obscuration of the optic nerve showed a high risk factor in 10 eyes (10/40), while in 1 eye (1/8) the optic nerve was visible at the initial presentation and obscured before secondary enucleation (P=0.529). Conclusion: Optic nerve obscuration at the last examination prior to enucleation may not be associated with postlaminar optic nerve invasion in advanced retinoblastoma. (Chin J Ophthalmol, 2020, 56: 681-687).

Clinicopathological Findings of Retinoblastoma: A 10-Year Experience from a Tertiary Hospital in Kampala, Uganda.

Background Retinoblastoma (RB) is one of the most common cancers occurring in young children in sub-Saharan Africa. The incidence rate reported in the literature is 9,000 new cases per year, which corresponds to 1 in 15,000 births. This study aimed at analyzing the clinicopathological findings in children with RB in Uganda. Purpose The purpose of this study was to describe and analyze the clinicopathological findings in the patients with RB. Design This was a cross-sectional analytical study involving 234 eyeball surgical specimens from 214 patients with RB diagnosed between January 2006 and December 2015. Results The mean age of the patients was 27.8 months (SD = 21.413, range: 1–132 months). More than half of the cases, 50.9%, presented with leucokoria. Invasion of choroid, sclera, anterior chamber, and optic nerve was found in 26.5% (n = 58), 51.2% (n = 88), 26.2% (n = 45), and 29.2% (n = 49), respectively. Twenty-six percent (n = 56) of the cases with intraocular tumour were at stage I and all patients with metastasis 4.7% (n = 11) had stage IV. The correlation between postlaminar optic nerve invasion and massive choroidal invasion was statistically significant (P = 0.002). Also, there was a statistical significance difference between metastasis and postlaminar invasion (P = 0.004). Conclusion The majority of children with RB in Uganda present clinically with leucokoria, and their parents or guardians seek medical intervention at a later stage. Moreover, there was a noticeably significant lag period for the patients to begin treatment after the diagnosis was done.

The predictive value of magnetic resonance imaging of retinoblastoma for the likelihood of high-risk pathologic features.

To evaluate the predictive value of magnetic resonance imaging in retinoblastoma for the likelihood of high-risk pathologic features. A retrospective study of 64 eyes enucleated from 60 retinoblastoma patients. Contrast-enhanced magnetic resonance imaging was performed before enucleation. Main outcome measures included demographics, laterality, accuracy, sensitivity, and specificity of magnetic resonance imaging in detecting high-risk pathologic features. Optic nerve invasion and choroidal invasion were seen microscopically in 34 (53%) and 28 (44%) eyes, respectively, while they were detected in magnetic resonance imaging in 22 (34%) and 15 (23%) eyes, respectively. The accuracy of magnetic resonance imaging in detecting prelaminar invasion was 77% (sensitivity 89%, specificity 98%), 56% for laminar invasion (sensitivity 27%, specificity 94%), 84% for postlaminar invasion (sensitivity 42%, specificity 98%), and 100% for optic cut edge invasion (sensitivity100%, specificity 100%). The accuracy of magnetic resonance imaging in detecting focal choroidal invasion was 48% (sensitivity 33%, specificity 97%), and 84% for massive choroidal invasion (sensitivity 53%, specificity 98%), and the accuracy in detecting extrascleral extension was 96% (sensitivity 67%, specificity 98%). Magnetic resonance imaging should not be the only method to stratify patients at high risk from those who are not, eventhough it can predict with high accuracy extensive postlaminar optic nerve invasion, massive choroidal invasion, and extrascleral tumor extension.

Assessment of early-stage optic nerve invasion in retinoblastoma using high-resolution 1.5 Tesla MRI with surface coils: a multicentre, prospective accuracy study with histopathological correlation.

To assess the accuracy of high-resolution (HR) magnetic resonance imaging (MRI) in diagnosing early-stage optic nerve (ON) invasion in a retinoblastoma cohort. This IRB-approved, prospective multicenter study included 95 patients (55 boys, 40 girls; mean age, 29 months). 1.5-T MRI was performed using surface coils before enucleation, including spin-echo unenhanced and contrast-enhanced (CE) T1-weighted sequences (slice thickness, 2 mm; pixel size <0.3 × 0.3 mm(2)). Images were read by five neuroradiologists blinded to histopathologic findings. ROC curves were constructed with AUC assessment using a bootstrap method. Histopathology identified 41 eyes without ON invasion and 25 with prelaminar, 18 with intralaminar and 12 with postlaminar invasion. All but one were postoperatively classified as stage I by the International Retinoblastoma Staging System. The accuracy of CE-T1 sequences in identifying ON invasion was limited (AUC = 0.64; 95 % CI, 0.55 - 0.72) and not confirmed for postlaminar invasion diagnosis (AUC = 0.64; 95 % CI, 0.47 - 0.82); high specificities (range, 0.64 - 1) and negative predictive values (range, 0.81 - 0.97) were confirmed. HR-MRI with surface coils is recommended to appropriately select retinoblastoma patients eligible for primary enucleation without the risk of IRSS stage II but cannot substitute for pathology in differentiating the first degrees of ON invasion. • HR-MRI excludes advanced optic nerve invasion with high negative predictive value. • HR-MRI accurately selects patients eligible for primary enucleation. • Diagnosis of early stages of optic nerve invasion still relies on pathology. • Several physiological MR patterns may mimic optic nerve invasion.

The validity of routine brain MRI in detecting post-laminar optic nerve involvement in retinoblastoma

ObjectiveTo describe the diagnostic validity of a gadolinium-enhanced routine brain MRI scan with 5 mm section thickness for the detection of optic nerve involvement in patients with retinoblastoma.MethodsMedical records of...

Correlation between Clinical Features, Magnetic Resonance Imaging, and Histopathologic Findings in Retinoblastoma: A Prospective Study

To correlate clinical features with histopathology findings in advanced intraocular retinoblastoma and to determine the diagnostic accuracy of magnetic resonance imaging (MRI) in detecting tumor invasion. Prospective, nonrandomized case series. We included 75 patients with group E retinoblastoma. Demographic and clinical features were recorded at presentation. Contrast-enhanced MRI was performed to study tumor characteristics and extent of invasion. Primary enucleation was performed and histopathologic features noted. Statistical analysis was done using the Kruskal-Wallis test to determine correlation between clinical features and histopathology. Sensitivity, specificity, and accuracy of MRI in detecting tumor invasion were determined. Significant associations between clinical findings at presentation and high-risk histopathology, and correlation between MRI results and histopathologic evidence of tumor invasion. A significant association was found between iris neovascularization and choroidal invasion (P = 0.032), intraocular pressure and optic nerve invasion (P = 0.034), and shallow anterior chamber and iris invasion (P = 0.021). Corneal diameter did not show any significant correlation with high-risk histopathology. On MRI, tumor volume showed a significant association with optic nerve invasion (P = 0.023). The accuracy of MRI in detecting choroidal invasion was 68% (sensitivity, 60%; specificity, 80%). Prelaminar invasion was correctly identified in 9 out of 15 eyes (accuracy, 84%; sensitivity, 60%; specificity, 90%), whereas the accuracy of MRI in detecting postlaminar invasion was 76% (sensitivity, 61.9%; specificity, 81.5%). Ciliary body invasion was correctly identified in 5 out of 7 eyes (accuracy, 93.3%; specificity, 95.6%) and scleral invasion in 5 out of 6 eyes (accuracy, 98.7%; specificity, 100%). As far as we are aware, this is the first prospective study on the correlation of clinical features and MRI findings with histopathologic risk factors in eyes primarily enucleated for retinoblastoma. Neovascularization of iris, intraocular pressure, shallow anterior chamber, and tumor volume correlated well with high-risk histopathology. Because MRI has limitations in reliably predicting microscopic infiltration of the choroid and optic nerve, decision in favor of neoadjuvant chemotherapy on the basis of suspected postlaminar invasion on MRI is not justified in the absence of histopathologic evidence of disease.

Lack of Correlation between the Histologic and Magnetic Resonance Imaging Results of Optic Nerve Involvement in Eyes Primarily Enucleated for Retinoblastoma

To correlate the histologic and magnetic resonance imaging results of the optic nerve in eyes primarily enucleated for retinoblastoma. Retrospective, clinicopathologic correlation. Sixty-seven consecutive patients with retinoblastoma who underwent primary enucleation. The histologic results of 67 eyes from 67 patients with retinoblastoma who underwent primary enucleation between March 1997 and January 2008 were studied for evidence of optic nerve invasion. Two neuroradiologists independently reviewed available preoperative magnetic resonance imaging studies with special emphasis on nonenhanced T2-weighted and gadolinium-enhanced T1-weighted imaging for evidence of optic nerve invasion. A weighted kappa statistic was used to assess agreement between observers. Correlation between neuroradiologists and histologic results. Of the 67 eyes studied, 60 had preoperative magnetic resonance images, 58 of which were deemed appropriate for review by both neuroradiologists. Review of the histologic results showed optic nerve involvement in 62 (93%) of 67 eyes: 28 prelaminar (42%), 24 laminar (36%), and 10 postlaminar (15%). On review of the magnetic resonance scans, the first neuroradiologist identified optic nerve involvement in 57 (95%) of 60 eyes: 26 prelaminar (43%), 10 laminar (17%), and 11 postlaminar (18%). The second neuroradiologist identified optic nerve involvement in 46 (77%) of 60 eyes: 33 prelaminar (55%), 9 laminar (15%), and 4 postlaminar (7%). Moderate agreement existed between neuroradiologists (kappa, 0.55). Poor and fair agreement existed between each of the 2 neuroradiologists and histologic results, respectively (kappa, 0.29 and 0.17). Exophytic tumors showed the greatest disparity (kappa, -0.20 and -0.13) between magnetic resonance imaging and histologic results. Limited correlation was found between magnetic resonance imaging and histologic results in assessing optic nerve invasion in eyes with retinoblastoma. Magnetic resonance imaging using routine imaging technologies, although useful in the evaluation of retinoblastoma, has limited usefulness in assessing the exact extent of optic nerve invasion; high-risk features of retinoblastoma such as postlaminar invasion remain best defined by histologic analysis. This study demonstrates that the interpretation of optic nerve involvement by a radiologist should not be the determining factor to defer enucleation in favor of neoadjuvant therapy.

Relevance of CT and MRI in retinoblastoma for the diagnosis of postlaminar invasion with normal-size optic nerve: a retrospective study of 150 patients with histological comparison

Detection of optic nerve invasion is mandatory in children primarily enucleated for retinoblastoma to ensure a free resection margin. To assess the accuracy of CT and MRI for the detection of postlaminar invasion in normal-size nerves. A total of 150 patients enucleated for retinoblastoma were included. Imaging data (119 CT and 46 MRI) were retrospectively reviewed and compared with histological findings. Abnormal contrast enhancement of the optic nerve was used as diagnostic criterion for invasion. The associations between postlaminar invasion and several indirect signs were also assessed. Statistical analysis was performed with the Kruskal-Wallis and Fisher exact tests. Postlaminar invasion on histology was observed in 8% (12/150). The sensitivity, specificity, accuracy and negative and positive predictive values were 60%, 95%, 91%, 95% and 60% for MRI, and 0%, 100%, 94% and 94% (PPV not assessable) for CT, respectively. Tumour diameter was the only indirect radiological sign significantly associated with postlaminar optic nerve invasion (P=0.002). Our results suggest that MRI is more relevant than CT for preoperative detection of optic nerve invasion in patients with retinoblastoma. Tumour diameter is the only indirect sign significantly associated with postlaminar invasion.

Primary and secondary tumours of the optic nerve, with emphasis on invasion of uveal malignant melanoma

Tumours of the optic nerve may originate in the optic nerve itself (i.e. primary tumours) or in structures adjacent to the optic nerve (i.e. secondary tumours). Furthermore, leukaemic infiltration and tumour metastases from distant solid tumours are also found in the optic nerve (i.e. secondary tumours). Large surveys have described primary optic nerve tumours and their treatment modalities whereas secondary tumours have received little attention. Most reported work-ups have been based on material from referral centres and specialized departments. The correlation between invasion of the optic nerve in uveal melanoma and the risk of dying from the disease is not fully elucidated. The impact on prognosis may be correlated with tumour cells left in the proximal nerve after enucleation. In addition, melanomas with secondary invasion of the optic nerve may be a highly malignant invasive subcategory of uveal melanomas. Clinical factors that should make the ophthalmologist aware of optic nerve invasion when investigating a uveal melanoma would be most important. Melanomas with invasion of the optic nerve may histopathologically display different invasive patterns and represent different types of uveal melanoma. A subtype of uveal melanoma with preponderance for neural invasion may exist. Such a subtype has been identified for cutaneous melanoma (desmoplastic/neurotropic). In order to investigate tumours of the optic nerve in a known population, we studied all surgically removed tumours of the optic nerve in Denmark during a study period of 25 years. Furthermore, in order to elucidate questions of secondary invasion of the optic nerve in uveal melanoma, all eyes enucleated with a uveal melanoma and optic nerve invasion in Denmark between 1942 and 2001 (n = 157) were investigated. Tumours of the optic nerve in children were dominated by optic glioma and invasion from retinoblastoma, whereas tumours in adults mainly comprised optic nerve sheath meningioma and invasion from uveal melanoma. Optic nerve invasion of uveal melanoma was found in 1 in 20 patients. Increased intraocular pressure and juxtapapillary location were associated with both prelaminar/laminar and postlaminar invasion of the optic nerve. Furthermore, age older than 70 years, reduced vision to light perception, non-visible fundus and large tumour size were associated with postlaminar invasion. Only patients with optic nerve invasion had metastases to the CNS, kidney and heart. Histopathological features associated with optic nerve invasion were focal retinal invasion, neovascularization of the chamber angle and scleral invasion. Furthermore, non-spindle cell type and rupture of the inner limiting membrane were both associated with postlaminar invasion. The optic nerve was invaded in four different ways: 1) by tumour extension from the neuroretina through the lamina cribrosa 2) by direct extension into the optic nerve head between Bruch's membrane and the border tissue of Elschnig 3) by direct invasion through the border tissue of Elschnig and 4) by spreading along the inner limiting membrane to the optic nerve head and through the lamina cribrosa. Most melanomas with invasion of the optic nerve were large juxtapapillary tumours invading the nerve because of their close location to the nerve, but a subcategory of neurotropic uveal melanomas with a diffuse invasive growth pattern was also found. These tumours invade the optic nerve and retina regardless of tumour size and location. Invasion of the optic nerve had no impact on the prognosis per se, but was seen in association with a series of other prognostic factors (e.g. cell type, extrascleral extension). I wish to express my sincere gratitude to my invaluable mentors Associate Professor Steffen Heegaard MD and Professor Jan Ulrik Prause MD for introducing me to the fascinating world of ophthalmology. Their optimism and encouraging guidance in the fields of research and pathology has been a great inspiration. This study was made possible by grants from The Danish Cancer Society, Værn om Synet, Synoptikfonden, Købmand M. Kristian Kjær og hustru Margrethe Kjær født la Cour-Holmen's Fond, Overlærer Svend Hansen's Fond, Savværksejer Jeppe Juhl og Hustru Ovita Juhls Mlndelegat, Alcon and H:S Generelle Forskningspulje.

Optic nerve invasion of uveal melanoma.

The aim of the study was to identify the histopathological characteristics associated with the invasion of the optic nerve of uveal melanoma and to evaluate the association between invasion of the optic nerve and survival. In order to achieve this, all uveal melanomas with optic nerve invasion in Denmark between 1942 and 2001 were reviewed (n=157). Histopathological characteristics and depth of optic nerve invasion were recorded. The material was compared with a control material from the same period consisting of 85 cases randomly drawn from all choroidal/ciliary body melanomas without optic nerve invasion. Prelaminar/laminar optic nerve invasion was in multivariate analysis associated with focal retinal invasion, neovascularization of the chamber angle, and scleral invasion. Postlaminar invasion was further associated with non-spindle cell type and rupture of the inner limiting membrane of the retina. The optic nerve was invaded in four different ways: 1) by tumor extension from the neuroretina through the lamina cribrosa; 2) by direct extension into the optic nerve head between Bruch's membrane and the border tissue of Elschnig; 3) by direct invasion through the border tissue of Elschnig; and 4) in one case a tumor spread along the inner limiting membrane to the optic nerve through the lamina cribrosa. Invasion of the optic nerve had no impact on all-cause mortality or melanoma-related mortality in multivariate analyses. The majority of melanomas invading the optic nerve are large juxtapapillary tumors invading the optic nerve because of simple proximity to the nerve. A neurotropic subtype invades the optic nerve and retina in a diffuse fashion unrelated to tumor size or location.

Optic Nerve Invasion of Uveal Melanoma: Clinical Characteristics and Metastatic Pattern

To determine the frequency of optic nerve invasion in uveal melanoma, to identify clinical factors associated with optic nerve invasion, and to analyze the metastatic pattern and the association with survival. All iris, ciliary body, and choroidal melanomas (N = 2758) examined between 1942 and 2001 at the Eye Pathology Institute, University of Copenhagen, Denmark, and the Institute of Pathology, Aarhus University Hospital, Aarhus, Denmark, were reviewed. Cases with optic nerve invasion were identified and subdivided into prelaminar or laminar invasion and postlaminar invasion. Clinical characteristics were compared with those from 85 cases randomly drawn from all ciliary body and choroidal melanomas without optic nerve invasion from the same period. Survival data were obtained by the Kaplan-Meier method, and the Mantel-Cox log-rank test was used to test differences in survival among the three patient groups. Optic nerve invasion was found in 157 uveal melanomas (5.7%; 95% confidence interval [CI], 4.8%-6.6%). Frequency varied during the observation period between 5% and 7%. Only choroidal and ciliary body melanomas were found to invade the optic nerve. Eighty-five (54%) were confined to the prelaminar or laminar part, and 72 (46%) were confined to the postlaminar part. Increased intraocular pressure (IOP) and juxtapapillary location were associated with prelaminar or laminar invasion and postlaminar invasion. Age older than 70 years, reduced vision to light perception or worse, nonvisible fundus, and large (>15 mm) tumor size were associated with postlaminar spread. In univariate analysis, patients with postlaminar invasion had significantly higher all-cause and melanoma-related mortality than the other patients. Optic nerve invasion in uveal melanoma is found in 1 in 20 patients. Visible juxtapapillary melanoma or loss of light perception should make the clinician suspicious of melanoma with optic nerve invasion, and special awareness of postlaminar spread should be addressed when increased IOP is present independently of decreased visual acuity and tumor location.

Results of a stage-based protocol for the treatment of retinoblastoma.

To describe the treatment of retinoblastoma at a single institution using a prospective protocol based on histopathologic staging. We included 116 consecutive patients (101 eligible, 46 bilateral) from August 1987 to December 1993. Treatment was enucleation or conservative therapy for intraocular disease (stage I patients). Stage II patients (orbital or postlaminar invasion) received vincristine, cyclophosphamide, and doxorubicin for 57 weeks. Patients with orbital mass and extension beyond the cut end of the optic nerve also received orbital radiotherapy (45 Gy). The latter received intrathecal therapy. In those with CNS (stage III) or hematogenous metastasis (stage IV), cisplatin and etoposide were added along with cranial (in patients with a CNS mass and prophylactically in stage IV) or craniospinal (in patients with positive CSF) radiotherapy. The median follow-up time was 39 months (range, 12 to 84). The overall survival rate was 0.84. Survival rates according to stage were as follows: stage I probability of overall survival [pOS] = 0.97) (alive/total), 59 of 60; stage II (pOS = 0.85) including patients with scattered episcleral cells, three of three; orbital mass, one of one; postlaminar invasion up to and beyond the cut end of optic nerve, 10 of 11 and 11 of 14, respectively; of stage III (pOS = 0), zero of six; and stage IV (pOS = 0.50), three of six. Only those patients with preauricular adenopathy as the only metastatic site survived in the latter group. Acute toxicity was mild. Chemotherapy is not warranted to prevent systemic metastasis for intraocular disease. Patients with extraocular orbital disease and had a good outcome with this therapy. Patients with metastatic disease fared poorly, except for those with isolated malignant preauricular adenopathy.