- Email: [email protected]
The Classification of Vitreous Seeds in Retinoblastoma and Response to Intravitreal Melphalan
The Classification of Vitreous Seeds in Retinoblastoma and Response to Intravitreal Melphalan Jasmine H. Francis, MD,1,2 David H. Abramson, MD,1,2 Marie-Claire Gaillard, MD,3 Brian P. Marr, MD,1,2 Maja Beck-Popovic, MD,4 Francis L. Munier, MD3 Purpose: To evaluate the clinical characteristics of the 3 classifications of vitreous seeds in retinoblastomaddust (class 1), spheres (class 2), and clouds (class 3)dand their responses to intravitreal melphalan. Design: Retrospective, bi-institutional cohort study. Participants: A total of 87 patient eyes received 475 intravitreal injections of melphalan (median dose, 30 mg) given weekly, a median of 5 times (range, 1e12 times). Methods: At presentation, the vitreous seeds were classified into 3 groups: dust, spheres, and clouds. Indirect ophthalmoscopy, fundus photography, ultrasonography, and ultrasonic biomicroscopy were used to evaluate clinical response to weekly intravitreal melphalan injections and time to regression of vitreous seeds. KaplaneMeier estimates of time to regression and ocular survival, patient survival, and event-free survival (EFS) were calculated and then compared using the ManteleCox test of curve. Main Outcome Measures: Time to regression of vitreous seeds, patient survival, ocular survival, and EFS. Results: The difference in time to regression was significantly different for the 3 seed classes (P < 0.0001): the median time to regression was 0.6, 1.7, and 7.7 months for dust, spheres, and clouds, respectively. Eyes with dust received significantly fewer injections and a lower median and cumulative dose of melphalan, whereas eyes with clouds received significantly more injections and a higher median and cumulative dose of melphalan. Overall, the 2-year KaplaneMeier estimates for ocular survival, patient survival, and EFS (related to target seeds) were 90.4% (95% confidence interval [CI], 79.7e95.6), 100%, and 98.5% (95% CI, 90e99.7), respectively. Conclusions: The regression and response of vitreous seeds to intravitreal melphalan are different for each seed classification. The vitreous seed classification can be predictive of time to regression, number, median dose, and cumulative dose of intravitreal melphalan injections required. Ophthalmology 2015;122:1173-1179 ª 2015 by the American Academy of Ophthalmology.
The current literature on retinoblastoma emphasizes vitreous seeding as the primary reason for treatment failure and loss of the eye.1 In fact, vitreous seeds have been recognized as the defining feature for failure by the Reese and Ellsworth classification group (Vb)2 and the International Classification of Retinoblastoma group (D).3 With the increased adoption of intravitreal melphalan, salvage rates for eyes with vitreous seeds are surpassing all historical data.4 However, describing vitreous disease with the blanket statement “seeds” does not capture the clinical heterogeneity that is seen; furthermore, we have noted that there is a spectrum of responses to intravitreal melphalan. Thus, our group believes differentiating vitreous seeds into 3 categories is a strong clinical tool. We have proposed a classification scheme5 as a means of distinguishing among vitreous seeds to aid in the interpretation of disease and enhance reporting in the literature. Our classification system is based on morphologic features of seeds and divides vitreous seeds into 3 groups: dust (class 1), spheres (class 2), and clouds (class 3).5 In this study, we evaluate the clinical response of vitreous seeds 2015 by the American Academy of Ophthalmology Published by Elsevier Inc.
to intravitreal melphalan to establish and define the clinical characteristics of each seed classification.
Methods This institutional review boardeapproved study included all eyes that received intravitreal melphalan for vitreous disease at Jules-Gonin Hospital and Memorial Sloan-Kettering Cancer Center between September 2009 and April 2014. Informed consent was obtained for each patient from their guardian, caregiver, or parent. The study was Health Insurance Portability and Accountability Act compliant. Research adhered to the tenets of the Declaration of Helsinki. After induction of anesthesia, the intraocular pressure was lowered with an anterior chamber paracentesis or by digital massage. Intravitreal melphalan (20e40 mg in 0.05 to 0.15 ml) was injected through the conjunctiva, sclera, and pars plana with a 32or 33-gauge needle. Upon needle withdrawal, the injection site was sealed and sterilized with cryotherapy and the eye was shaken in all directions during cryo-application, as previously described.4 In cases performed at Memorial Sloan-Kettering Cancer Center, the ocular surface was submerged in irrigating sterile water for 3 minutes.6 http://dx.doi.org/10.1016/j.ophtha.2015.01.017 ISSN 0161-6420/15
1173
Ophthalmology Volume 122, Number 6, June 2015 Table 1. Summary of Vitreous Seed Classification and Clinical Findings Class
Type
Type I
Dust
Type 2
Spheres
Type 3
Cloud
Description Small granules of vitreous opacities Can be seen as a vitreous haze overlying tumor Spherically shaped opacities within vitreous Dust may be present around spheres Can be homogenously opaque or have a translucent outer shell with relatively transparent or whitish center, or vice versa Dense collection of punctate vitreous opacities Can appear as a sheet or globule of seed granules and often with wispy edge Dust and spheres are sometimes also visible
Regression Characteristics
Median Time to Regression (wks)
Median No. of Injections
Median Melphalan Dose (mg)
Typically regress to type 0 (not visible)
2e3
3
20
Initially disperse (pseudogrowth) and then disappear, but can become calcific (type I), amorphous (type II), or a mixture of types I and II (type III)
6e7
5
30
Initially disperse (pseudogrowth), become calcific, or disappear, but can remain calcific (type I) or amorphous (type II)
30e32
8
33
The clinical status was evaluated under anesthesia with indirect ophthalmoscopy, RetCam fundus photography (Clarity, Pleasanton, CA), B-scan ultrasonography (OTI Scan 2000; Ophthalmic Technologies, North York, Ontario, Canada), and ultrasonic biomicroscopy (OTI Scan 2000; Ophthalmic Technologies). At each subsequent examination, the burden of residual disease was reevaluated and intravitreal melphalan given every 7 to 10 days up to 12 injections. Patient data included age, sex, laterality, age at start of injection course, eye status (salvaged or enucleated), life status (alive or dead), treatment status (naïve vs. prior treatment with systemic chemotherapy or external beam radiation), and follow-up time from beginning of injection course. Treatment data included time from initial injection to final regression, number of injections, cumulative/mean dose of melphalan, prior treatment with systemic chemotherapy, ophthalmic artery chemosurgery (OAC) or radiation (plaque brachytherapy or external beam), concomitant OAC defined as occurring within 1 month before initial injection or 1 month after final injection completion, concomitant focal treatment (laser or cryotherapy) performed at the time of injection but exclusive of the injection site cryotherapy, additional treatment related to the target vitreous disease for which the eye was receiving injections, and additional treatment unrelated to the target vitreous disease for which the eye was receiving injections (e.g., retinal tumor recurrence or a different focus of vitreous disease). Tumor data included the International Classification,3 final seed regression pattern (type 0 ¼ not visible, type 1 ¼ calcific, type 2 ¼ amorphous, type 3 ¼ types 1 and 2),5 seed classification at presentation (class 1 ¼ dust, class 2 ¼ spheres dust, or class 3 ¼ clouds spheres or dust), and extent of disease (localized [1 quadrant] or diffuse [>1 quadrant]). Outcome measurements were compared for all 3 seed classifications and included time to final regression, ocular survival, patient survival, and ocular event-free survival (EFS) related and unrelated to target vitreous seeds. Time to regression was calculated as the time from the initial injection to the first date of examination when regression was noted. For ocular EFS, an event was defined as recurrent (or new) disease that required additional focal treatment, OAC, radiation, intravitreal melphalan, or enucleation. Time to regression was compared for extent of disease, concomitant OAC,
1174
radiation and focal treatments, treatment status, number of injections (5 vs. >5), cumulative dose of melphalan (<160 vs. 160 mg), and mean dose of melphalan (<30 vs. 30 mg). Statistical analysis was performed using Prism (GraphPad Software, Inc, La Jolla, CA). KaplaneMeier survival data with log-rank test was used to evaluate ocular and progression-free survival, and the ManteleCox test was used to compare curves. Statistical analysis was performed with linear regression analysis, 2-tailed Student t test, and analysis of variance using GraphPad software (GraphPad Software Inc.) and NCSS software (NCSS, Kaysville, UT).
Results A total of 35 patients had unilateral disease, and 52 patients had bilateral disease, of whom 30 were monocular. The median followup was 20.3 months (range, 2e56 months), and the median age at
Figure 1. Illustration summarizing vitreous seed classification and response to intravitreal melphalan: number of injections received, time to response, and median dose of melphalan per injection.
Francis et al
Classification of Vitreous Seeds
Figure 2. Representative cases of seed classification and response to intravitreal melphalan. (A) Dust before intravitreal melphalan and (B) after 3 injections; the dust responds with a type 0 regression pattern. (C) Sphere surrounded by dust before intravitreal melphalan and (D) after 6 injections spheres and dust respond with a type 0 regression pattern. (E) Cloud before intravitreal melphalan and (F) after 8 injections; the cloud responds with a type 1 regression pattern.
treatment was 43.6 months (range, 11e216 months). Seeds were classified into the 3 groups of the vitreous seed classification system based on the morphologic features outlined in Table 1 and Figure 1. Examples are shown in Figure 2. A total of 475 injections in 87 eyes (International Classification A ¼ 2 eyes, B ¼ 10 eyes, C ¼ 17 eyes, D ¼ 53 eyes, and E ¼ 5 eyes) were included in this study.
Table 2 depicts patient, treatment, and tumor data, along with outcome measurements by seed classification. As shown in Figure 3, there was a significant difference in the median time to regression for the 3 seed classifications (P < 0.0001): class 1 (dust) regressed in the shortest interval, and class 3 (clouds) took significantly longer. Each seed class received a different mean
1175
Ophthalmology Volume 122, Number 6, June 2015 Table 2. Patient, Treatment, and Tumor Data and Outcome Measurements by Seed Classification Characteristic
Class I: Dust
Class II: Spheres
Class III: Cloud
Overall
No. of eyes Median time to regression (mos) Regression pattern Type 0: not visible Type I: calcific Type II: amorphous Type III: calcific and amorphous Configuration Localized Diffuse No. of injections Mean Median Cumulative melphalan (mg) Mean Median Melphalan dose (mg) Mean Median EFS (related to target seeds) at 2 yrs EFS (unrelated to target seeds) at 2 yrs Percentage enucleated Ocular survival at 2 yrs Percentage dead Patient survival at 2 yrs
11 0.6*
60 1.7*
16 7.7*
87 2.0
11 0 0 0
(100%) (0%) (0%) (0%)
9 (82%) 2 (18%)
47 6 5 2
(78%) (10%) (8%) (3%)
11 3 2 0
27 (45%) 33 (55%)
(69%) (19%) (13%) (0%)
69 9 7 2
2 (12%) 14 (88%)
(79%) (10%) (8%) (2%)
38 (44%) 49 (56%)
3.0 3.0
5.5 5
7.2 8
5.5 5
75.5 60
156.4 161
232.3 229
160.1 162
23.4 20 100.0% 70% (32.8e89.2) 0.0% 100.0% 0.0% 100%
28.2 30 98% (86e100) 66.4% (50.6e78.2) 10.0% 85.8% (70.8e93.4) 3.4% 100.0%
31.4 32.5 100.0% 56.3% (25.4e78.6) 0.0% 100.0% 0.0% 100%
28.2 30 98.5% (90e99.7) 65.6% (53.1e75.5) 6.9% 90.4% (79.7e95.6) 2.3% 100%
Bold numbers are statistically significant. 95% Confidence intervals are shown in parentheses. EFS ¼ event-free survival. *P < 0.0001.
number of injections and amount of melphalan: Compared with the other classes, class 3 received a significantly increased number of injections (P < 0.01) and cumulative (P < 0.01) and mean melphalan dose (P < 0.05). Table 3 shows factors related to seed regression. Localized seeding took significantly less time to regress compared with diffuse seeding (1.1 vs. 3.2 months; P < 0.001). Concomitant OAC was associated with a significantly longer time to regression (2.5 vs. 1.5 months; P ¼ 0.04); however, more eyes receiving concomitant OAC had class 3 seeds compared with eyes that did not have concomitant OAC (35% vs. 12%). Concomitant radiation and focal treatment had no influence on
time to regression. Eyes receiving more than 5 injections or more than 160 mg of cumulative melphalan took significantly longer to regress (both P < 0.0001), which reflects the higher number of class 3 seeds in this group. As shown in Figure 4, the overall 2-year KaplaneMeier estimates were 90.4% (95% confidence interval [CI], 79.7e95.6) for ocular survival and 100% for patient survival. Overall, the 2-year KaplaneMeier estimates for ocular EFS related to the target seeds were 98.5% (95% CI, 90e99.7) and 65.6% (95% CI, 53.1e75.5) for EFS unrelated to the target seeds (Fig 4). None of Table 3. Factors Related to Seed Regression (KaplaneMeier Estimates of Time to Final Seed Regression with ManteleCox Test of Curve Comparison) Comparison Localized vs. diffuse disease Concomitant radiation: yes vs. no Concomitant OAC: yes vs. no. Concomitant focal treatment: yes vs. no <5 vs. 5 Injections <160 vs. 160 mg Cumulative melphalan dose <30 vs. 30 mg Mean melphalan dose New York vs. Swiss cohorts
Figure 3. KaplaneMeier survival curves for time to seed regression after intravitreal melphalan. Note that each seed classification has a significantly different time to regression: Dust takes the least time to regress, and cloud takes the most prolonged course to regression.
1176
Median Time (mos)
P Value
Variable 1
Variable 2
1.1 1.4 2.5 1.5
3.2 2.1 1.5 2.4
<0.001 0.44 0.04 0.68
0.9 0.9
3.5 3.3
<0.0001 <0.0001
1.4
2.5
0.08
1.7
2.2
0.18
Bold numbers are statistically significant. OAC ¼ ophthalmic artery chemosurgery.
Francis et al
Classification of Vitreous Seeds
Figure 4. (A) KaplaneMeier survival curves for ocular survival of all eyes, (B) patient survival of all patients, (C) event-free survival of all eyes related to target seeds, and (D) event-free survival of all eyes unrelated to target seeds. Note that all curves show no significant difference between seed classification.
these estimates were significantly different among the 3 seed classes, although comparisons for ocular and patient survival may be limited by the sample size. There were 2 deaths among patients; 1 patient, who remained disease-free until lost of follow-up, underwent secondary enucleation for late relapse and died under poorly defined conditions 27 months after intravitreal injections, and 1 patient with trilateral retinoblastoma displayed no ocular recurrence for 30 months after intravitreal injections when she died of disseminated intravascular coagulation, a complication of treatment for pinealoblastoma.
Discussion Reese and Ellsworth2 recognized vitreous seeding as clinically unfavorable, giving eyes with this feature the last grouping in their classification scheme. Historically, response rates of these eyes confirmed Reese and Ellsworth’s assessment with globe salvage rates for International Classification group D ranging from 10% to 47% for eyes treated with systemic chemotherapy and focal therapies.7,8 External beam radiation also had dismal salvage rates for eyes with vitreous seeding: Of group V eyes that progressed with an initial course of external beam radiation, only 2.2% could be salvaged with a second course.9 By comparison, vitreous seeds have demonstrated improved response after OAC, with 2-year KaplaneMeier estimates of 64% and 76% for naïve and previously treated eyes, respectively.10 But OAC requires a comprehensive team of specialists and is inaccessible to many treatment centers. Now that safety-enhanced techniques have calmed fears
regarding extraocular extension through the needle tract, the increased adoption of intravitreal melphalan has proven it to be an efficacious method of treating vitreous seeds, with some groups reporting salvage rates as high as 100% (at a modest follow-up of only 9 months).11 The efficacy of intravitreal melphalan has eradicated the use of external beam radiotherapy in the management of vitreous seeding. In treating these eyes, we recognize that not all vitreous seeds are equal, and, instead, they can be distinguished into 3 separate categories based on morphology: dust, spheres, and clouds. This study summarizes our collective experience of treating 87 eyes with 475 intravitreal melphalan injections. We learn that our proposed classification system for vitreous seeds not only guides identification of each seed type based on morphologic features but also provides powerful information regarding the distinct clinical characteristics and response to intravitreal melphalan. Each seed type has a significantly distinct time to regression, number of injections, and cumulative and mean dose of melphalan. These findings are summarized in Tables 2 and 3. Dust results from a displacement of tumor cells into the vitreous,5 appears as small granules of vitreous opacities, and can be seen as a vitreous haze overlying tumor (Fig 2A). They responded in the least amount of time to intravitreal melphalan (P < 0.0001), with their median time to regression calculated at 0.6 months after the first weekly injection. They required the significantly fewest injections (median, 3; P < 0.01), least amount of cumulative melphalan (median, 60 mg; P < 0.01), and lowest mean melphalan dose (median, 20 mg; P < 0.05)
1177
Ophthalmology Volume 122, Number 6, June 2015 before regression was noted. Dust typically regressed into a type 0 pattern, meaning it was not detectable via ophthalmoscopy (Fig 2B). Spheres result from translocation of tumor cells into the vitreous, which then undergo clonal expansion into spheres.5 They are spherically shaped opacities within the vitreous, and dust may be present around them (Fig 2C). They can be homogeneously opaque and have a translucent outer shell with a relatively transparent or whitish center. They responded to intravitreal melphalan in a time period that was significantly more rapid than a cloud but more prolonged than dust (P < 0.0001), with their median time to regression calculated at 1.7 months. Likewise, their regression was apparent after significantly fewer injections and less melphalan than a cloud, but more than dust: The median number of injections was 5, cumulative melphalan was 161 mg, and mean melphalan was 30 mg. As spheres responded to intravitreal melphalan, they initially dispersed and eventually disappeared on ophthalmoscopy or became calcific (regression type I) or amorphous in shape (regression type II) or even a mixture of these 2 (regression type III) (Fig 2D). Clouds result from massive transference of tumor cells into the vitreous5 and are typically visualized as a dense collection of punctate vitreous opacities. They can appear as a sheet or globule of seed granules and often with wispy edges, and dust and spheres are sometimes also visible (Fig 2E). They responded to intravitreal melphalan in a significantly prolonged fashion (P < 0.000), with median time to regression being calculated at 7.7 months. Probably because of this prolonged response, they were treated with significantly more injections (median, 8; P < 0.01), highest amount of cumulative melphalan (median, 229 mg; P < 0.01), and highest mean melphalan dose (median, 32.5 mg; P < 0.05). They typically regressed to calcific or amorphous granules, and over time, some slowly became undetectable to ophthalmoscopy (Fig 2F). Table 3 summarizes other factors related to time to seed regression. Overall, diffuse disease took significantly longer to regress compared with localized disease (3.2 vs. 1.1 months; P < 0.001) and can be explained by a higher burden of disease requiring more time to respond. Although concomitant radiation and focal treatment had no significant impact on time to regression of seeds, concomitant OAC resulted in a significantly longer time to regression compared with seeds without OAC (P ¼ 0.04). This is likely a result of eyes with a heavier burden of disease and higher percentage of class 3 seeds in this group that required treatment with 2 delivery methods of chemotherapy (OAC and intravitreal). Eyes receiving <5 injections of melphalan and <160 mg of cumulative melphalan took significantly less time to regress (P < 0.0001), and this is most likely reflective of the seed classification (with more “dust” falling into these groups). Perhaps future visualization techniques will allow for a keener understanding of the diseased vitreous architecture (identification of the seeded vitreous sinus, confluence of this diseased vitreous sinus with other sinuses, proximity and placement of the injection) and provide further understanding regarding response to treatment. Two-year KaplaneMeier estimates for ocular and patient survival were calculated at 90.4% and 100%, respectively,
1178
and are concordant with other published studies.4,11e13 It may be worth noting that all eyes that were enucleated were in patients with vitreous seeds classified as spheres: this may be a coincidence or may hint at an ominous feature of spheres. The 2-year EFS related to the target seeds was calculated at 98.5% and suggests that intravitreal melphalan is efficacious at treating the intended vitreous seeds with few supplementary therapies needed. However, the markedly lower EFS unrelated to the seeds (65.6% at 2 years) demonstrates that eyes with vitreous seeds may require additional treatments for other foci of disease, whether new or recurrent retinal tumors or a new distinct crop of vitreous seeds. In conclusion, distinguishing vitreous seeds into 3 classes based on morphologic features can provide valuable clinical information. This classification seems to be predictive of the time it will take seeds to regress and the number of injections and amount of melphalan that are used. This information would be particularly useful to clinicians as they formulate an individualized treatment plan for patients and their parents. We encourage other groups to use the proposed classification scheme and further confirm our findings with additional cohorts of patients. Acknowledgments. The authors thank Susan Houghton for data management.
References 1. Francis JH, Schaiquevich P, Buitrago E, et al. Local and systemic toxicity of intravitreal melphalan for vitreous seeding in retinoblastoma: a preclinical and clinical study. Ophthalmology 2014;121:1810–7. 2. Reese AB, Ellsworth RM. The evaluation and current concept of retinoblastoma therapy. Trans Am Acad Ophthalmol Otolaryngol 1963;67:164–72. 3. Linn Murphree A. Intraocular retinoblastoma: the case for a new group classification. Ophthalmol Clin North Am 2005;18: 41–53. viii. 4. Munier FL, Gaillard M-C, Balmer A, et al. Intravitreal chemotherapy for vitreous disease in retinoblastoma revisited: from prohibition to conditional indications. Br J Ophthalmol 2012;96:1078–83. 5. Munier FL. Classification and management of seeds in retinoblastoma. Ellsworth Lecture Ghent August 24, 2013. Ophthalmic Genet 2014;35:193–207. 6. Francis JH, Xu XL, Gobin YP, et al. Death by water: precautionary water submersion for intravitreal injection of retinoblastoma eyes. Open Ophthalmol J 2014;8:7–11. 7. Bartuma K, Pal N, Kosek S, et al. A 10-year experience of outcome in chemotherapy-treated hereditary retinoblastoma. Acta Ophthalmol 2014;92:404–11. 8. Berry JL, Jubran R, Kim JW, et al. Long-term outcomes of Group D eyes in bilateral retinoblastoma patients treated with chemoreduction and low-dose IMRT salvage. Pediatr Blood Cancer 2013;60:688–93. 9. Abramson DH, Ellsworth RM, Rosenblatt M, et al. Retreatment of retinoblastoma with external beam irradiation. Arch Ophthalmol 1982;100:1257–60. 10. Abramson DH, Marr BP, Dunkel IJ, et al. Intra-arterial chemotherapy for retinoblastoma in eyes with vitreous and/or subretinal seeding: 2-year results. Br J Ophthalmol 2012;96: 499–502.
Francis et al
Classification of Vitreous Seeds
11. Shields CL, Manjandavida FP, Arepalli S, et al. Intravitreal melphalan for persistent or recurrent retinoblastoma vitreous seeds: preliminary results. JAMA Ophthalmol 2014;132: 319–25. 12. Ghassemi F, Shields CL, Ghadimi H, et al. Combined intravitreal melphalan and topotecan for refractory or recurrent
vitreous seeding from retinoblastoma. JAMA Ophthalmol 2014;132:936–41. 13. Suzuki S, Kaneko A. Vitreous injection therapy of melphalan for retinoblastoma [abstract]. XVth Biannual Meeting, International Society of Ocular Oncology, November 14e17, 2011, Buenos Aires, Argentina.
Footnotes and Financial Disclosures Originally received: December 2, 2014. Final revision: January 15, 2015. Accepted: January 19, 2015. Available online: March 19, 2015.
Author Contributions: Conception and design: Francis, Abramson, Munier Manuscript no. 2014-1939.
1
Ophthalmic Oncology Service, Memorial Sloan-Kettering Cancer Center, New York, New York.
2
Weill-Cornell Medical Center, New York, New York.
3
Jules-Gonin Eye Hospital, Lausanne, Switzerland.
4
Pediatric Hematology Oncology Unit, University Hospital CHUV, Lausanne, Switzerland. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Supported by The Fund for Ophthalmic Knowledge, The New York Community Trust, and Research to Prevent Blindness. The sponsors or funding organizations had no role in the design or conduct of this research.
Analysis and interpretation: Francis, Abramson, Munier Data collection: Francis, Abramson, Munier Obtained funding: Not applicable Overall responsibility: Francis, Abramson, Gaillard, Marr, Beck-Popovic, Munier Abbreviations and Acronyms: CI ¼ confidence interval; EFS ¼ event-free survival; OAC ¼ ophthalmic artery chemosurgery. Correspondence: Jasmine H. Francis, MD, Memorial Sloan Kettering Cancer Center, 1275 York Ave, New York, NY 10065. E-mail: [email protected].
1179
The current literature on retinoblastoma emphasizes vitreous seeding as the primary reason for treatment failure and loss of the eye.1 In fact, vitreous seeds have been recognized as the defining feature for failure by the Reese and Ellsworth classification group (Vb)2 and the International Classification of Retinoblastoma group (D).3 With the increased adoption of intravitreal melphalan, salvage rates for eyes with vitreous seeds are surpassing all historical data.4 However, describing vitreous disease with the blanket statement “seeds” does not capture the clinical heterogeneity that is seen; furthermore, we have noted that there is a spectrum of responses to intravitreal melphalan. Thus, our group believes differentiating vitreous seeds into 3 categories is a strong clinical tool. We have proposed a classification scheme5 as a means of distinguishing among vitreous seeds to aid in the interpretation of disease and enhance reporting in the literature. Our classification system is based on morphologic features of seeds and divides vitreous seeds into 3 groups: dust (class 1), spheres (class 2), and clouds (class 3).5 In this study, we evaluate the clinical response of vitreous seeds 2015 by the American Academy of Ophthalmology Published by Elsevier Inc.
to intravitreal melphalan to establish and define the clinical characteristics of each seed classification.
Methods This institutional review boardeapproved study included all eyes that received intravitreal melphalan for vitreous disease at Jules-Gonin Hospital and Memorial Sloan-Kettering Cancer Center between September 2009 and April 2014. Informed consent was obtained for each patient from their guardian, caregiver, or parent. The study was Health Insurance Portability and Accountability Act compliant. Research adhered to the tenets of the Declaration of Helsinki. After induction of anesthesia, the intraocular pressure was lowered with an anterior chamber paracentesis or by digital massage. Intravitreal melphalan (20e40 mg in 0.05 to 0.15 ml) was injected through the conjunctiva, sclera, and pars plana with a 32or 33-gauge needle. Upon needle withdrawal, the injection site was sealed and sterilized with cryotherapy and the eye was shaken in all directions during cryo-application, as previously described.4 In cases performed at Memorial Sloan-Kettering Cancer Center, the ocular surface was submerged in irrigating sterile water for 3 minutes.6 http://dx.doi.org/10.1016/j.ophtha.2015.01.017 ISSN 0161-6420/15
1173
Ophthalmology Volume 122, Number 6, June 2015 Table 1. Summary of Vitreous Seed Classification and Clinical Findings Class
Type
Type I
Dust
Type 2
Spheres
Type 3
Cloud
Description Small granules of vitreous opacities Can be seen as a vitreous haze overlying tumor Spherically shaped opacities within vitreous Dust may be present around spheres Can be homogenously opaque or have a translucent outer shell with relatively transparent or whitish center, or vice versa Dense collection of punctate vitreous opacities Can appear as a sheet or globule of seed granules and often with wispy edge Dust and spheres are sometimes also visible
Regression Characteristics
Median Time to Regression (wks)
Median No. of Injections
Median Melphalan Dose (mg)
Typically regress to type 0 (not visible)
2e3
3
20
Initially disperse (pseudogrowth) and then disappear, but can become calcific (type I), amorphous (type II), or a mixture of types I and II (type III)
6e7
5
30
Initially disperse (pseudogrowth), become calcific, or disappear, but can remain calcific (type I) or amorphous (type II)
30e32
8
33
The clinical status was evaluated under anesthesia with indirect ophthalmoscopy, RetCam fundus photography (Clarity, Pleasanton, CA), B-scan ultrasonography (OTI Scan 2000; Ophthalmic Technologies, North York, Ontario, Canada), and ultrasonic biomicroscopy (OTI Scan 2000; Ophthalmic Technologies). At each subsequent examination, the burden of residual disease was reevaluated and intravitreal melphalan given every 7 to 10 days up to 12 injections. Patient data included age, sex, laterality, age at start of injection course, eye status (salvaged or enucleated), life status (alive or dead), treatment status (naïve vs. prior treatment with systemic chemotherapy or external beam radiation), and follow-up time from beginning of injection course. Treatment data included time from initial injection to final regression, number of injections, cumulative/mean dose of melphalan, prior treatment with systemic chemotherapy, ophthalmic artery chemosurgery (OAC) or radiation (plaque brachytherapy or external beam), concomitant OAC defined as occurring within 1 month before initial injection or 1 month after final injection completion, concomitant focal treatment (laser or cryotherapy) performed at the time of injection but exclusive of the injection site cryotherapy, additional treatment related to the target vitreous disease for which the eye was receiving injections, and additional treatment unrelated to the target vitreous disease for which the eye was receiving injections (e.g., retinal tumor recurrence or a different focus of vitreous disease). Tumor data included the International Classification,3 final seed regression pattern (type 0 ¼ not visible, type 1 ¼ calcific, type 2 ¼ amorphous, type 3 ¼ types 1 and 2),5 seed classification at presentation (class 1 ¼ dust, class 2 ¼ spheres dust, or class 3 ¼ clouds spheres or dust), and extent of disease (localized [1 quadrant] or diffuse [>1 quadrant]). Outcome measurements were compared for all 3 seed classifications and included time to final regression, ocular survival, patient survival, and ocular event-free survival (EFS) related and unrelated to target vitreous seeds. Time to regression was calculated as the time from the initial injection to the first date of examination when regression was noted. For ocular EFS, an event was defined as recurrent (or new) disease that required additional focal treatment, OAC, radiation, intravitreal melphalan, or enucleation. Time to regression was compared for extent of disease, concomitant OAC,
1174
radiation and focal treatments, treatment status, number of injections (5 vs. >5), cumulative dose of melphalan (<160 vs. 160 mg), and mean dose of melphalan (<30 vs. 30 mg). Statistical analysis was performed using Prism (GraphPad Software, Inc, La Jolla, CA). KaplaneMeier survival data with log-rank test was used to evaluate ocular and progression-free survival, and the ManteleCox test was used to compare curves. Statistical analysis was performed with linear regression analysis, 2-tailed Student t test, and analysis of variance using GraphPad software (GraphPad Software Inc.) and NCSS software (NCSS, Kaysville, UT).
Results A total of 35 patients had unilateral disease, and 52 patients had bilateral disease, of whom 30 were monocular. The median followup was 20.3 months (range, 2e56 months), and the median age at
Figure 1. Illustration summarizing vitreous seed classification and response to intravitreal melphalan: number of injections received, time to response, and median dose of melphalan per injection.
Francis et al
Classification of Vitreous Seeds
Figure 2. Representative cases of seed classification and response to intravitreal melphalan. (A) Dust before intravitreal melphalan and (B) after 3 injections; the dust responds with a type 0 regression pattern. (C) Sphere surrounded by dust before intravitreal melphalan and (D) after 6 injections spheres and dust respond with a type 0 regression pattern. (E) Cloud before intravitreal melphalan and (F) after 8 injections; the cloud responds with a type 1 regression pattern.
treatment was 43.6 months (range, 11e216 months). Seeds were classified into the 3 groups of the vitreous seed classification system based on the morphologic features outlined in Table 1 and Figure 1. Examples are shown in Figure 2. A total of 475 injections in 87 eyes (International Classification A ¼ 2 eyes, B ¼ 10 eyes, C ¼ 17 eyes, D ¼ 53 eyes, and E ¼ 5 eyes) were included in this study.
Table 2 depicts patient, treatment, and tumor data, along with outcome measurements by seed classification. As shown in Figure 3, there was a significant difference in the median time to regression for the 3 seed classifications (P < 0.0001): class 1 (dust) regressed in the shortest interval, and class 3 (clouds) took significantly longer. Each seed class received a different mean
1175
Ophthalmology Volume 122, Number 6, June 2015 Table 2. Patient, Treatment, and Tumor Data and Outcome Measurements by Seed Classification Characteristic
Class I: Dust
Class II: Spheres
Class III: Cloud
Overall
No. of eyes Median time to regression (mos) Regression pattern Type 0: not visible Type I: calcific Type II: amorphous Type III: calcific and amorphous Configuration Localized Diffuse No. of injections Mean Median Cumulative melphalan (mg) Mean Median Melphalan dose (mg) Mean Median EFS (related to target seeds) at 2 yrs EFS (unrelated to target seeds) at 2 yrs Percentage enucleated Ocular survival at 2 yrs Percentage dead Patient survival at 2 yrs
11 0.6*
60 1.7*
16 7.7*
87 2.0
11 0 0 0
(100%) (0%) (0%) (0%)
9 (82%) 2 (18%)
47 6 5 2
(78%) (10%) (8%) (3%)
11 3 2 0
27 (45%) 33 (55%)
(69%) (19%) (13%) (0%)
69 9 7 2
2 (12%) 14 (88%)
(79%) (10%) (8%) (2%)
38 (44%) 49 (56%)
3.0 3.0
5.5 5
7.2 8
5.5 5
75.5 60
156.4 161
232.3 229
160.1 162
23.4 20 100.0% 70% (32.8e89.2) 0.0% 100.0% 0.0% 100%
28.2 30 98% (86e100) 66.4% (50.6e78.2) 10.0% 85.8% (70.8e93.4) 3.4% 100.0%
31.4 32.5 100.0% 56.3% (25.4e78.6) 0.0% 100.0% 0.0% 100%
28.2 30 98.5% (90e99.7) 65.6% (53.1e75.5) 6.9% 90.4% (79.7e95.6) 2.3% 100%
Bold numbers are statistically significant. 95% Confidence intervals are shown in parentheses. EFS ¼ event-free survival. *P < 0.0001.
number of injections and amount of melphalan: Compared with the other classes, class 3 received a significantly increased number of injections (P < 0.01) and cumulative (P < 0.01) and mean melphalan dose (P < 0.05). Table 3 shows factors related to seed regression. Localized seeding took significantly less time to regress compared with diffuse seeding (1.1 vs. 3.2 months; P < 0.001). Concomitant OAC was associated with a significantly longer time to regression (2.5 vs. 1.5 months; P ¼ 0.04); however, more eyes receiving concomitant OAC had class 3 seeds compared with eyes that did not have concomitant OAC (35% vs. 12%). Concomitant radiation and focal treatment had no influence on
time to regression. Eyes receiving more than 5 injections or more than 160 mg of cumulative melphalan took significantly longer to regress (both P < 0.0001), which reflects the higher number of class 3 seeds in this group. As shown in Figure 4, the overall 2-year KaplaneMeier estimates were 90.4% (95% confidence interval [CI], 79.7e95.6) for ocular survival and 100% for patient survival. Overall, the 2-year KaplaneMeier estimates for ocular EFS related to the target seeds were 98.5% (95% CI, 90e99.7) and 65.6% (95% CI, 53.1e75.5) for EFS unrelated to the target seeds (Fig 4). None of Table 3. Factors Related to Seed Regression (KaplaneMeier Estimates of Time to Final Seed Regression with ManteleCox Test of Curve Comparison) Comparison Localized vs. diffuse disease Concomitant radiation: yes vs. no Concomitant OAC: yes vs. no. Concomitant focal treatment: yes vs. no <5 vs. 5 Injections <160 vs. 160 mg Cumulative melphalan dose <30 vs. 30 mg Mean melphalan dose New York vs. Swiss cohorts
Figure 3. KaplaneMeier survival curves for time to seed regression after intravitreal melphalan. Note that each seed classification has a significantly different time to regression: Dust takes the least time to regress, and cloud takes the most prolonged course to regression.
1176
Median Time (mos)
P Value
Variable 1
Variable 2
1.1 1.4 2.5 1.5
3.2 2.1 1.5 2.4
<0.001 0.44 0.04 0.68
0.9 0.9
3.5 3.3
<0.0001 <0.0001
1.4
2.5
0.08
1.7
2.2
0.18
Bold numbers are statistically significant. OAC ¼ ophthalmic artery chemosurgery.
Francis et al
Classification of Vitreous Seeds
Figure 4. (A) KaplaneMeier survival curves for ocular survival of all eyes, (B) patient survival of all patients, (C) event-free survival of all eyes related to target seeds, and (D) event-free survival of all eyes unrelated to target seeds. Note that all curves show no significant difference between seed classification.
these estimates were significantly different among the 3 seed classes, although comparisons for ocular and patient survival may be limited by the sample size. There were 2 deaths among patients; 1 patient, who remained disease-free until lost of follow-up, underwent secondary enucleation for late relapse and died under poorly defined conditions 27 months after intravitreal injections, and 1 patient with trilateral retinoblastoma displayed no ocular recurrence for 30 months after intravitreal injections when she died of disseminated intravascular coagulation, a complication of treatment for pinealoblastoma.
Discussion Reese and Ellsworth2 recognized vitreous seeding as clinically unfavorable, giving eyes with this feature the last grouping in their classification scheme. Historically, response rates of these eyes confirmed Reese and Ellsworth’s assessment with globe salvage rates for International Classification group D ranging from 10% to 47% for eyes treated with systemic chemotherapy and focal therapies.7,8 External beam radiation also had dismal salvage rates for eyes with vitreous seeding: Of group V eyes that progressed with an initial course of external beam radiation, only 2.2% could be salvaged with a second course.9 By comparison, vitreous seeds have demonstrated improved response after OAC, with 2-year KaplaneMeier estimates of 64% and 76% for naïve and previously treated eyes, respectively.10 But OAC requires a comprehensive team of specialists and is inaccessible to many treatment centers. Now that safety-enhanced techniques have calmed fears
regarding extraocular extension through the needle tract, the increased adoption of intravitreal melphalan has proven it to be an efficacious method of treating vitreous seeds, with some groups reporting salvage rates as high as 100% (at a modest follow-up of only 9 months).11 The efficacy of intravitreal melphalan has eradicated the use of external beam radiotherapy in the management of vitreous seeding. In treating these eyes, we recognize that not all vitreous seeds are equal, and, instead, they can be distinguished into 3 separate categories based on morphology: dust, spheres, and clouds. This study summarizes our collective experience of treating 87 eyes with 475 intravitreal melphalan injections. We learn that our proposed classification system for vitreous seeds not only guides identification of each seed type based on morphologic features but also provides powerful information regarding the distinct clinical characteristics and response to intravitreal melphalan. Each seed type has a significantly distinct time to regression, number of injections, and cumulative and mean dose of melphalan. These findings are summarized in Tables 2 and 3. Dust results from a displacement of tumor cells into the vitreous,5 appears as small granules of vitreous opacities, and can be seen as a vitreous haze overlying tumor (Fig 2A). They responded in the least amount of time to intravitreal melphalan (P < 0.0001), with their median time to regression calculated at 0.6 months after the first weekly injection. They required the significantly fewest injections (median, 3; P < 0.01), least amount of cumulative melphalan (median, 60 mg; P < 0.01), and lowest mean melphalan dose (median, 20 mg; P < 0.05)
1177
Ophthalmology Volume 122, Number 6, June 2015 before regression was noted. Dust typically regressed into a type 0 pattern, meaning it was not detectable via ophthalmoscopy (Fig 2B). Spheres result from translocation of tumor cells into the vitreous, which then undergo clonal expansion into spheres.5 They are spherically shaped opacities within the vitreous, and dust may be present around them (Fig 2C). They can be homogeneously opaque and have a translucent outer shell with a relatively transparent or whitish center. They responded to intravitreal melphalan in a time period that was significantly more rapid than a cloud but more prolonged than dust (P < 0.0001), with their median time to regression calculated at 1.7 months. Likewise, their regression was apparent after significantly fewer injections and less melphalan than a cloud, but more than dust: The median number of injections was 5, cumulative melphalan was 161 mg, and mean melphalan was 30 mg. As spheres responded to intravitreal melphalan, they initially dispersed and eventually disappeared on ophthalmoscopy or became calcific (regression type I) or amorphous in shape (regression type II) or even a mixture of these 2 (regression type III) (Fig 2D). Clouds result from massive transference of tumor cells into the vitreous5 and are typically visualized as a dense collection of punctate vitreous opacities. They can appear as a sheet or globule of seed granules and often with wispy edges, and dust and spheres are sometimes also visible (Fig 2E). They responded to intravitreal melphalan in a significantly prolonged fashion (P < 0.000), with median time to regression being calculated at 7.7 months. Probably because of this prolonged response, they were treated with significantly more injections (median, 8; P < 0.01), highest amount of cumulative melphalan (median, 229 mg; P < 0.01), and highest mean melphalan dose (median, 32.5 mg; P < 0.05). They typically regressed to calcific or amorphous granules, and over time, some slowly became undetectable to ophthalmoscopy (Fig 2F). Table 3 summarizes other factors related to time to seed regression. Overall, diffuse disease took significantly longer to regress compared with localized disease (3.2 vs. 1.1 months; P < 0.001) and can be explained by a higher burden of disease requiring more time to respond. Although concomitant radiation and focal treatment had no significant impact on time to regression of seeds, concomitant OAC resulted in a significantly longer time to regression compared with seeds without OAC (P ¼ 0.04). This is likely a result of eyes with a heavier burden of disease and higher percentage of class 3 seeds in this group that required treatment with 2 delivery methods of chemotherapy (OAC and intravitreal). Eyes receiving <5 injections of melphalan and <160 mg of cumulative melphalan took significantly less time to regress (P < 0.0001), and this is most likely reflective of the seed classification (with more “dust” falling into these groups). Perhaps future visualization techniques will allow for a keener understanding of the diseased vitreous architecture (identification of the seeded vitreous sinus, confluence of this diseased vitreous sinus with other sinuses, proximity and placement of the injection) and provide further understanding regarding response to treatment. Two-year KaplaneMeier estimates for ocular and patient survival were calculated at 90.4% and 100%, respectively,
1178
and are concordant with other published studies.4,11e13 It may be worth noting that all eyes that were enucleated were in patients with vitreous seeds classified as spheres: this may be a coincidence or may hint at an ominous feature of spheres. The 2-year EFS related to the target seeds was calculated at 98.5% and suggests that intravitreal melphalan is efficacious at treating the intended vitreous seeds with few supplementary therapies needed. However, the markedly lower EFS unrelated to the seeds (65.6% at 2 years) demonstrates that eyes with vitreous seeds may require additional treatments for other foci of disease, whether new or recurrent retinal tumors or a new distinct crop of vitreous seeds. In conclusion, distinguishing vitreous seeds into 3 classes based on morphologic features can provide valuable clinical information. This classification seems to be predictive of the time it will take seeds to regress and the number of injections and amount of melphalan that are used. This information would be particularly useful to clinicians as they formulate an individualized treatment plan for patients and their parents. We encourage other groups to use the proposed classification scheme and further confirm our findings with additional cohorts of patients. Acknowledgments. The authors thank Susan Houghton for data management.
References 1. Francis JH, Schaiquevich P, Buitrago E, et al. Local and systemic toxicity of intravitreal melphalan for vitreous seeding in retinoblastoma: a preclinical and clinical study. Ophthalmology 2014;121:1810–7. 2. Reese AB, Ellsworth RM. The evaluation and current concept of retinoblastoma therapy. Trans Am Acad Ophthalmol Otolaryngol 1963;67:164–72. 3. Linn Murphree A. Intraocular retinoblastoma: the case for a new group classification. Ophthalmol Clin North Am 2005;18: 41–53. viii. 4. Munier FL, Gaillard M-C, Balmer A, et al. Intravitreal chemotherapy for vitreous disease in retinoblastoma revisited: from prohibition to conditional indications. Br J Ophthalmol 2012;96:1078–83. 5. Munier FL. Classification and management of seeds in retinoblastoma. Ellsworth Lecture Ghent August 24, 2013. Ophthalmic Genet 2014;35:193–207. 6. Francis JH, Xu XL, Gobin YP, et al. Death by water: precautionary water submersion for intravitreal injection of retinoblastoma eyes. Open Ophthalmol J 2014;8:7–11. 7. Bartuma K, Pal N, Kosek S, et al. A 10-year experience of outcome in chemotherapy-treated hereditary retinoblastoma. Acta Ophthalmol 2014;92:404–11. 8. Berry JL, Jubran R, Kim JW, et al. Long-term outcomes of Group D eyes in bilateral retinoblastoma patients treated with chemoreduction and low-dose IMRT salvage. Pediatr Blood Cancer 2013;60:688–93. 9. Abramson DH, Ellsworth RM, Rosenblatt M, et al. Retreatment of retinoblastoma with external beam irradiation. Arch Ophthalmol 1982;100:1257–60. 10. Abramson DH, Marr BP, Dunkel IJ, et al. Intra-arterial chemotherapy for retinoblastoma in eyes with vitreous and/or subretinal seeding: 2-year results. Br J Ophthalmol 2012;96: 499–502.
Francis et al
Classification of Vitreous Seeds
11. Shields CL, Manjandavida FP, Arepalli S, et al. Intravitreal melphalan for persistent or recurrent retinoblastoma vitreous seeds: preliminary results. JAMA Ophthalmol 2014;132: 319–25. 12. Ghassemi F, Shields CL, Ghadimi H, et al. Combined intravitreal melphalan and topotecan for refractory or recurrent
vitreous seeding from retinoblastoma. JAMA Ophthalmol 2014;132:936–41. 13. Suzuki S, Kaneko A. Vitreous injection therapy of melphalan for retinoblastoma [abstract]. XVth Biannual Meeting, International Society of Ocular Oncology, November 14e17, 2011, Buenos Aires, Argentina.
Footnotes and Financial Disclosures Originally received: December 2, 2014. Final revision: January 15, 2015. Accepted: January 19, 2015. Available online: March 19, 2015.
Author Contributions: Conception and design: Francis, Abramson, Munier Manuscript no. 2014-1939.
1
Ophthalmic Oncology Service, Memorial Sloan-Kettering Cancer Center, New York, New York.
2
Weill-Cornell Medical Center, New York, New York.
3
Jules-Gonin Eye Hospital, Lausanne, Switzerland.
4
Pediatric Hematology Oncology Unit, University Hospital CHUV, Lausanne, Switzerland. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Supported by The Fund for Ophthalmic Knowledge, The New York Community Trust, and Research to Prevent Blindness. The sponsors or funding organizations had no role in the design or conduct of this research.
Analysis and interpretation: Francis, Abramson, Munier Data collection: Francis, Abramson, Munier Obtained funding: Not applicable Overall responsibility: Francis, Abramson, Gaillard, Marr, Beck-Popovic, Munier Abbreviations and Acronyms: CI ¼ confidence interval; EFS ¼ event-free survival; OAC ¼ ophthalmic artery chemosurgery. Correspondence: Jasmine H. Francis, MD, Memorial Sloan Kettering Cancer Center, 1275 York Ave, New York, NY 10065. E-mail: [email protected].
1179