Bilateral familial retinoblastoma in a 2-month-old girl. A, Subtle macular retinoblastoma inferotemporal to foveola in left eye. B, After thermotherapy, tumor regressed to an atrophic scar at 1-year follow-up. Foveola remains intact, and she has not needed chemotherapy and external beam radiotherapy, despite a total of 16 tumors in both eyes, all treated with thermotherapy alone.
Bilateral familial retinoblastoma in a 4-month-old girl. A, Small retinoblastoma in papillomacular region of right eye. B, After thermotherapy coupled with chemotherapy, tumor completely regressed to a flat scar. C, On fluorescein angiography, papillomacular retinal vessels continue to perfuse over the scar into the fovea, allowing retained visual acuity in eye at more than 3 years of follow-up.
Bilateral familial retinoblastoma in a 2-week-old boy. A, Right eye demonstrating juxtapapillary and larger juxtafoveal retinoblastomas. B, Left eye demonstrating large macular retinoblastoma overhanging the optic disc. C, Right eye after chemoreduction and coupled thermotherapy. Tumors regressed, and retinal vessels perfuse over the scars. D, Left eye after chemoreduction and coupled thermotherapy. Tumor is a calcified, regressed mass and remains stable at 2-year follow-up.
Focal iris atrophy and paraxial lens opacities in an eye that required 70 minutes of thermotherapy for a large juxtapapillary retinoblastoma.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Shields CL, Santos MCM, Diniz W, et al. Thermotherapy for Retinoblastoma. Arch Ophthalmol. 1999;117(7):885–893. doi:10.1001/archopht.117.7.885
To evaluate the results of thermotherapy for retinoblastoma.
Prospective, nonrandomized analysis of the treatment method.
A total of 188 retinoblastomas in 80 eyes of 58 patients who were treated with thermotherapy.
Main Outcome Measures
Tumor response and ocular adverse effects.
Of 188 retinoblastomas treated with thermotherapy, mean tumor base was 3.0 mm and tumor thickness was 2.0 mm. Complete tumor regression was achieved in 161 tumors (85.6%), and 27 tumors (14.4%) developed recurrence. Using univariate analysis, the predictors of local tumor recurrence were male sex (P=.005), no color change ("no visible take") in tumor after treatment (P=.01), increasing number of treatment sessions (P=.002), and previous use of chemoreduction (P=.02). By multivariate analysis, the most important predictors of local tumor recurrence were male sex (P=.01) and previous use of chemoreduction (P=.03), the latter likely reflecting the fact that these tumors were initially larger with more ominous findings, and required chemoreduction therapy to reduce them to a size amenable to focal treatment with thermotherapy. When evaluating thermotherapy variables as a function of tumor size, it was apparent that larger tumors (≥3.0-mm base) required greater energy and time than did smaller tumors (<3.0-mm base). Comparison of treatment variables for larger vs smaller tumors was as follows: number of treatment sessions, 3.3 vs 2.3; spot size, 1.7 vs 1.3 mm; power, 540 vs 370 mW; treatment duration, 49 vs 14 minutes; and coupling of thermotherapy with chemotherapy, 79% vs 48% of cases (P≤.001 for each variable). Complications of thermotherapy in the 80 eyes included focal iris atrophy in 29 eyes (36%), peripheral focal lens opacity in 19 eyes (24%), retinal traction in 4 eyes (5%), retinal vascular obstruction in 2 eyes (2%), and transient localized serous retinal detachment in 2 eyes (2%). There were no cases of corneal scarring, central lens opacity, iris or retinal neovascularization, or rhegmatogenous retinal detachment. All eyes with focal lens opacity demonstrated adjacent focal iris atrophy. By multivariate analysis, the predictors of thermotherapy-induced focal iris atrophy were increasing number of treatment sessions (P=.001) and increasing tumor base (P=.02).
Thermotherapy is used for relatively small retinoblastomas without associated vitreous or subretinal seeds. This treatment provides satisfactory control for selected retinoblastomas, with 86% of tumors demonstrating lasting regression. Tumors that measure 3.0 mm or larger in base at the time of thermotherapy require more intense treatment than smaller tumors and are at greatest risk for ocular complications such as focal iris atrophy and focal paraxial lens opacity.
MANAGEMENT of retinoblastoma is complex and involves many issues.1-4 Treatment selection is determined by patient age; tumor size and location; laterality; and associated findings such as retinal detachment, subretinal tumor seeds, vitreous tumor seeds, and status of the opposite eye. Management options for retinoblastoma include enucleation, external beam radiotherapy, plaque radiotherapy, laser photocoagulation, cryotherapy, chemoreduction, and thermotherapy.5-11 There has been a continuing trend toward more conservative focal treatment for retinoblastoma, avoiding enucleation and external beam radiotherapy.5,12
This evolution in treatment philosophy is reflected in a recent editorial, "A New Era for the Treatment of Retinoblastoma."13 The authors remarked that recently published treatment strategies14-17 using chemotherapy (termed chemoreduction) "provide new hope for alternative therapies for retinoblastoma." One of the most exciting of the new alternative treatments is thermotherapy, a method of delivering heat to tissues to raise the local temperature to cytotoxic levels, leading to tumor cell death.
Thermotherapy is an established technique for use in various solid tumors, especially when in combination with chemotherapy.18,19 When used in combination with radiotherapy or chemotherapy or as a sole treatment modality, it is an important treatment modality for small- and medium-sized malignant melanoma of the skin and uvea.19-26
The recent application of thermotherapy for retinoblastoma was popularized using a specific treatment protocol that combines intravenous carboplatin with thermotherapy delivered through the operating microscope.27 Later, other observers3,17,28,29 reported modifying the technique by coupling it with a 3-agent chemoreduction protocol, using it alone, and delivering the heat via an indirect ophthalmoscope system. However, there remains a need for a more comprehensive understanding of the clinical effects of thermotherapy on retinoblastoma. A detailed, comprehensive analysis evaluating the treatment variables and risks for treatment failure and ocular adverse effects using this technique has not yet been performed, despite its multicenter use. In this article, we review our initial experience with 188 retinoblastomas managed with thermotherapy and analyze treatment variables and other clinical features as they affect 2 major end points—tumor recurrence and ocular complications.
A review of computer-coded medical charts of the Ocular Oncology Service at Wills Eye Hospital, Philadelphia, Pa, was performed. Patients with a diagnosis of retinoblastoma managed with thermotherapy were selected. The following clinical data were recorded prospectively: patient age, race (white, African-American, Hispanic, or Asian), sex, inheritance (sporadic or familial), tumor laterality (unilateral or bilateral), eye involved, and initial visual acuity. Eyes were assessed for Reese-Ellsworth group (I, II, III, IV, or V)1 and estimated visual acuity. Tumors were analyzed for previous treatment (chemotherapy, radiotherapy, laser photocoagulation, cryotherapy, or thermotherapy), anatomic location (intraretinal, subretinal, or vitreous), meridional location (superior, superonasal, nasal, inferonasal, inferior, inferotemporal, temporal, supertemporal, or macula), basal dimension (in millimeters), thickness (in millimeters), proximity to optic nerve (in millimeters), and proximity to foveola (in millimeters). The presence of associated subretinal fluid (yes or no), associated active subretinal tumor seeds (yes or no), extent of active subretinal tumor seeds (number of affected quadrants), and associated active vitreous seeds (yes or no) was recorded.
Inclusion criteria for treatment with thermotherapy were the presence of viable retinoblastoma within the retina or subretinal space, with less than 1.0 mm of overlying subretinal fluid, in an eye with clear media. Further criteria were based on the use of systemic chemotherapy.6 If the patient was not undergoing systemic chemotherapy, we included tumors 3.0 mm or less in base and 3.0 mm or less in thickness. Larger tumors and the presence of active vitreous seeds emanating from the selected tumor were criteria for exclusion. If the patient was undergoing systemic chemotherapy, tumor base or thickness and overlying vitreous seeds did not play into the inclusion criteria. Complex decision making regarding treatment of retinoblastoma exists because of the multiplicity of tumors of different sizes and features, often in both eyes. Therefore, use of chemotherapy and, thus, thermotherapy is based on multiple overlapping variables.
Treatment variables were evaluated for the number of sessions, delivery system (indirect ophthalmoscope, operating microscope, or transscleral probe), spot size (in millimeters), power (in milliwatts), duration (in minutes), end point (no change, light gray "take," or white take), and coupling of chemoreduction with thermotherapy (yes or no).
At the most recent visit, the recorded data included estimated final visual acuity and reason for decreased vision. Tumors were assessed for basal dimension (in millimeters), thickness (in millimeters), regression type (0, no visible tumor scar; I, completely calcified tumor scar; II, completely noncalcified tumor scar; III, partially calcified and partially noncalcified tumor scar; or IV, flat, atrophic chorioretinal atrophy with no residual tumor),1 status (regressed, recurrence, or no change), and treatment of recurrence (chemotherapy, radiotherapy, laser photocoagulation, cryotherapy, or thermotherapy). Ocular complications were recorded, and systemic status was noted.
Under general anesthesia, thermotherapy was delivered using infrared radiation (810 nm) delivered by a specially modified diode system. Delivery was either transpupillary or transscleral. The transpupillary method involved infrared radiation delivered through an adaptor on the operating microscope and a wide-angle contact lens, or through an adaptor on an indirect ophthalmoscope and a 20-diopter (D) lens. In both instances, wide pupillary dilation was necessary. The operating microscope allowed for a spot size of 0.8, 1.2, and 2.0 mm, whereas the indirect ophthalmoscope provided a 1.2-mm spot size. The largest spot size available was used to cover, but not extend onto, normal tissue. The transscleral method involved a special hand-held adaptor, which delivered a 1-mm spot of infrared radiation via scleral depression and indirect ophthalmoscopy and a 20-D lens. The end point of all methods was a gentle, light gray color change ("take") within the tumor during a 1- to 5-minute period without causing vascular spasm or rapid tumor whitening. In general, the power was started at 200 mW and was increased or decreased at 50-mW increments until an adequate, slow-onset take was observed in the mass.
Tumors nearest the optic disc and fovea and large in size (>6 mm) were treated with the operating microscope system. Tumors that were smaller or situated peripheral to the macula were treated with the indirect ophthalmoscope system. Small tumors at the ora serrata were treated with the indirect ophthalmoscope system or transscleral probe.
The clinical and treatment variables were analyzed as they affected 2 end points—local retinoblastoma recurrence and iris atrophy (the most common complication). Retinoblastoma recurrence was defined as the development of a viable tumor within a previous thermotherapy site.
A series of univariate Cox proportional hazards regression analyses were performed to assess the individual variables as they affected the major end point of recurrence. An SAS statistical program30,31 was used to account for the correlated data, wherein there were multiple observations per patient (ie, tumors). For the major end point of iris atrophy, the Global Evaluation of Efficacy32,33 approach to analyzing clustered binomial data was used for a series of univariable logistic regressions. Age at presentation, tumor base, tumor thickness, proximity to optic disc, proximity to foveola, number of treatment sessions, spot size, power, and treatment duration were analyzed as continuous variables. All others were analyzed as discrete variables. A preliminary stepwise model included all variables that were significant on a univariate level to determine an independent set of predictors for each outcome. Subsequent multivariate models simultaneously fitted the set of independent predictors and tested other biologically important variables for inclusion in the final multivariable model.
We treated 188 retinoblastomas in 80 eyes of 58 patients. Mean patient age was 29 months (median, 28 months; range, 6-76 months); 51 were white (88%), 3 were African American (5%), 3 were Asian (5%), and 1 was Indian (2%). Thirty-three patients (57%) were boys and 25 (43%) were girls. The retinoblastoma was bilateral in 49 patients (84%) and unilateral in 9 (16%). The disease was sporadic in 39 patients (67%) and familial in 19 (33%). The right eye was treated in 41 cases (51%) and the left eye was treated in 39 (49%). The initial visual acuity was fix and follow in 21 eyes (26%), no fix or follow in 25 (31%), and unable to obtain visual acuity because of a lack of patient cooperation in 34 (42%).
The Reese-Ellsworth classification was type I in 12 eyes (15%), type II in 14 (18%), type III in 34 (42%), type IV in 11 (14%), and type V in 9 (11%). Chemoreduction using vincristine, etoposide, and carboplatin, as previously described,6,15 was delivered before thermotherapy to 71 tumors (37.8%). Mean number of chemoreduction cycles before initiation of thermotherapy was 3 (median, 3).
At the time of thermotherapy, mean tumor base was 3.0 mm (median, 2.0 mm; range, 0.2-18.0 mm) and tumor thickness was 2.0 mm (median, 2.0 mm; range, 0.2-6.3 mm) (Figure 1). The closest tumor margin was a mean of 6.0 mm (median, 6.0 mm; range, 0-23 mm) to the optic disc and 7.0 mm (median, 6.0 mm; range, 0-24 mm) to the foveola. At the time of thermotherapy, subretinal fluid was present with the tumor in 11 cases (14%). No active subretinal or vitreous seeds were associated with the retinoblastoma.
Thermotherapy was coupled with chemotherapy in 108 cases (57.4%) (Figure 2 and Figure 3). In 79 of them (73.1%), the chemotherapy coupling was within 6 hours of thermotherapy, and in 29 (26.9%), the coupling was within 1 to 2 days. The delivery system of treatment included indirect ophthalmoscope in 109 cases (58.0%), operating microscope in 72 (38.3%), and transscleral probe in 7 (3.7%). The thermotherapy spot size was 0.8 mm in 11 cases (5.8%), 1.2 mm in 106 (56.4%), 2.0 mm in 68 (36.2%), and not recorded in 3 (1.4%). Mean thermotherapy power was 437 mW (median, 400 mW; range, 100-1000 mW). The number of treatment sessions for tumor control was 1 in 57 cases (30.3%), 2 in 52 (27.6%), 3 in 31 (16.5%), 4 in 23 (12.2%), 5 in 10 (5.3%), 6 in 8 (4.2%), 7 in 5 (2.6%), and 8 in 2 (1.1%). Mean number of sessions per tumor was 3 (median, 2). Mean total cumulated thermotherapy time for each individual tumor was 27 minutes (median, 15 minutes; range, 1-210 minutes). A comparison of treatment variables for larger vs smaller tumors was as follows: number of treatment sessions, 3.3 vs 2.3; spot size, 1.7 vs 1.3 mm; power, 540 vs 370 mW; treatment duration, 49 vs 14 minutes; and coupling of thermotherapy with chemotherapy, 79% vs 48% of cases (P<.001 for each variable) (Table 1 and Table 2).
Mean follow-up was 12 months (median, 10 months; maximum, 45 months). Final visual acuity was measured at fix and follow in 40 eyes (50%), no fix or follow in 1 (1%), enucleation in 1 (1%), and undetermined due to patient cooperation in 38 (48%). Macular findings that accounted for ultimate poor vision included regressed retinoblastoma involving the fovea in 36 eyes (67%), foveal traction in 6 (11%), amblyopia in 4 (7%), residual subretinal fluid in 3 (6%), previous subretinal fluid in 3 (6%), cataract from previous radiotherapy in 1 (2%), and enucleation in 1 (2%).
At the last examination, mean tumor base was 2.0 mm (median, 1.0 mm; range, 0-20.0 mm) and tumor thickness was 1.0 mm (median, 0 mm; range, 0-6.0 mm). The entire treatment scar measured a mean of 4.0 mm in base (median, 3.0 mm; range, 0-20.0 mm). The final tumor regression after completion of thermotherapy was judged to be type I in 34 cases (21.1%), type II in 11 (6.8%), type III in 22 (13.6%), and type IV in 94 (56.5%). No patients developed systemic metastasis. The tumor remained regressed in 161 cases (85.6%) and demonstrated recurrence in 27 (14.4%).
Using univariate analysis, the predictors of tumor recurrence included sex (male) (P=.005), end point (no color change) (P=.01), greater number of treatment sessions (P=.02), and chemoreduction preceding thermotherapy (P=.02) (Table 3 and Table 4). By multivariate analysis, the most important combination of variables was chemoreduction preceding thermotherapy (P=.03) and sex (male) (P=.01) (Table 5).
Complications of thermotherapy included transient corneal edema in 1 eye (1%), focal iris atrophy in 29 (36%), focal paraxial lens opacity in 19 (24%), retinal traction in 4 (5%), retinal vascular occlusion in 2 (2%), serous retinal detachment in 2 (2%), optic disc edema in 4 (5%), and eventual sector optic disc atrophy in 10 (12%) (Table 6). Optic atrophy was judged to be retrograde in the case of an immediately adjacent large regressed retinoblastoma.
By univariate analysis, the predictors of focal iris atrophy (Figure 4) were greater number of treatment sessions (P=.003), increasing treatment power (P=.004), proximity of tumor to the optic disc (P=.006), proximity of tumor to the foveola (P=.008), chemotherapy coupled with thermotherapy (P=.01), increasing tumor thickness and base dimensions (P=.01), larger spot size (P=.02), longer treatment duration (P=.03), and delivery system (operating microscope) (P=.04). By multivariate analysis, the most predictive variables of thermotherapy-induced focal iris atrophy were increasing number of treatment sessions (P=.001) and increasing tumor base (P=.02) (Table 7).
Conservative management of retinoblastoma has evolved during the past decades from external beam radiotherapy to chemoreduction plus focal consolidation treatment.3,12 An important focal treatment method is thermotherapy, which applies focused heat to tissue at subphotocoagulation levels to induce cell necrosis. Thermotherapy can be delivered either as a single modality or in synergism with chemotherapy or radiotherapy. The ideal tissue temperature for single-modality thermotherapy is 45°C to 60°C; this is slightly lower, at 42°C to 44°C, for thermotherapy coupled with chemotherapy or radiotherapy in order to avoid local heat-related complications.34
Thermotherapy has been used to treat systemic cancers such as cutaneous melanoma, superficial head and neck cancers, and breast carcinoma.18-23 In most instances, thermotherapy is coupled with lower dose radiotherapy or chemotherapy as a sensitizing agent and to allow for fewer adverse effects. Thermotherapy is also effective for selected eye cancers, including choroidal melanoma.24,25 With regard to choroidal melanoma, short-term results26 at 14-month follow-up show that properly applied thermotherapy offers 94% tumor control for selected small choroidal melanomas less than 4.0 mm thick. For medium and large choroidal melanoma, thermotherapy has been coupled with iodine plaque radiotherapy, and preliminary results35 indicate that complete tumor flattening was achieved in 86% of patients, many of whom did not respond adequately to radiotherapy alone.
Thermotherapy for retinoblastoma is not a new concept. In 1982, Lagendijk36,37 designed a microwave applicator to deliver whole eye hyperthermia for retinoblastoma.34 With this device, the temperature distribution reached 43°C to 45°C at the retinal surface and less than 39°C at the lens. This gradient was planned in an effort to minimize heat-induced cataract. The device was designed for combination with lower dose external beam radiotherapy to provide enhanced radiosensitivity at a low dose and spare radiotherapeutic complications. Kaneko38 later reported a few successful cases using the Lagendijk microwave applicator with local chemotherapy.
Recent enthusiasm for thermotherapy has developed with the introduction of improved heat delivery devices. Murphree et al17 and Murphree and Munier27 used an infrared radiation device delivered through an operating microscope to heat individual tumors, and combined it with intravenous single-agent carboplatin and later 3-agent carboplatin, etoposide, and vincristine therapy. Shields and associates3,28,39 found similar good control with 3-agent chemotherapy combined with thermotherapy. The adaptation of various heat delivery instruments now allows focused heat to be applied by either a transpupillary or transscleral route.39
The decision to use chemoreduction or thermotherapy in a child with retinoblastoma should be made carefully.6 If a tumor is of a size and location that is treatable without needing chemoreduction, then chemotherapy should be avoided. In our practice of ocular oncology, approximately one third of patients with retinoblastoma are selected for chemoreduction; two thirds are treated with nonchemotherapy methods.6 Furthermore, the decision to use chemoreduction, before or coupled with thermotherapy, is complex and will be the subject of future investigations. With regard to infrared thermotherapy, most heat absorption is by chromophores in the posterior segment of the eye—the retinal pigment epithelium and the choroidal melanocytes. Small retinoblastomas receive enough heat transfer from the adjacent pigmented tissue to induce visible color change (take), thereby indicating a direct cellular effect. Larger retinoblastomas show little take with thermotherapy alone, and in these cases, we prefer to first use chemoreduction so that the smaller residual mass may be more susceptible to heat treatment. In addition, the combination of chemotherapy with thermotherapy is synergistic so that, even if there is less absorption of heat and subsequent lower tissue temperature, the synergism is sufficient to control the tumor. In general, we use chemoreduction if a tumor is 4 mm or greater, then apply thermotherapy after reduction. Tumors less than 3 to 4 mm are treated initially with thermotherapy alone. However, the decision to use chemoreduction is complicated and based on many variables, including age of the child; status of the opposite eye; size and location of other tumors; presence of vitreous seeds, subretinal seeds, and retinal detachment; and a host of other variables.1,6
Results of recent studies of thermotherapy combined with chemotherapy in the transgenic murine retinoblastoma cell line confirmed the synergistic action of heat and carboplatin. Tumor cells treated with no heat required 542 ng of carboplatin for 50% of cells to reach a lethal dose, whereas cells treated with 43°C heat for 30 minutes needed only 327 ng of carboplatin to reach the lethal dose for 50% of cells.40 Results of further studies41 confirmed the synergistic effects of thermotherapy using ferromagnetic seeds and external beam radiotherapy in a murine retinoblastoma model.
In the present study, we report our experience with thermotherapy for retinoblastoma in the clinical setting. In 188 retinoblastomas, local tumor control was achieved in 85.6% and recurrence developed in 14.4%. Variables leading to tumor recurrence included male sex (P=.005), no color change (no visible take) in tumor after treatment (P=.01), increasing number of treatment sessions (P=.002), and previous use of chemoreduction (P=.02). No color change in the tumor and increasing number of treatment sessions likely reflect an inability to achieve adequate temperature levels in the retinoblastoma. The main predictors of recurrence from the multivariate standpoint included male sex (P=.01) and preceding chemoreduction before thermotherapy (P=.03). It is difficult to understand how patient sex could affect tumor recurrence. Preceding chemoreduction is easier to understand. Larger tumors are reduced in size by chemoreduction to allow for conservative treatments such as thermotherapy.6,15 The reduced mass is then treated with thermotherapy, but all margins where the larger mass previously existed may be at risk for recurrence. In addition, chemoreduction nearly completely eradicates macroscopic and microscopic subretinal and vitreous seeds, but at least 1 seed can recur in up to 67% and 75% of cases, respectively, based on previous reports.6
Thermotherapy via infrared radiation can be delivered through an operating microscope, indirect ophthalmoscope, or transscleral probe.3 The delivery is time-intensive and tedious; it involves a continuous period of tumor monitoring by the ocular oncologist as the temperature in the tumor is elevated and maintained. Often, a gray-white discoloration in the tumor is seen, indicating a successful take. Retinal vessels generally maintain their caliber during treatment, but retinal hemorrhage can occur. In our group of 188 tumors, the average time to adequately heat a single retinoblastoma was 27 minutes per tumor divided over a mean of 3 sessions. The duration varied with tumor size, and larger tumors measuring 4 mm or more in thickness required a mean of 68 minutes of thermotherapy per tumor, divided over a mean of 3 sessions. This time-intensive treatment required continuous concentration on the part of the ocular oncologist to maintain alignment of the infrared beam on the tumor for the entire period. This is especially important with juxtapapillary and macular tumors because misalignment could lead to maltreatment and immediate visual loss. When considering that the mean number of tumors in a child with bilateral retinoblastoma is 5,42 it is obvious that thermotherapy involves a substantial time commitment for the patient and surgeon.
Complications of thermotherapy for retinoblastoma included focal iris atrophy (36%), peripheral focal lens opacity (24%), retinal traction (5%), retinal vascular obstruction (2%), and transient serous retinal detachment (2%). There were no cases of corneal scarring, central lens opacity, iris or retinal neovascularization, or rhegmatogenous retinal detachment. All eyes with focal lens opacity demonstrated adjacent iris atrophy. We suspect that the infrared radiation transmitted through the anterior chamber is absorbed by the pigmented iris, leading to focal iris atrophy and adjacent heat damage to the lens in the form of focal opacity. In addition, the heat gradient generated in the anterior chamber may be a problem during treatment because it causes pupillary spasm and transient lens edema. In an effort to avoid pupillary spasm, topical mydriatic agents were used and cold balance salt solution was dripped onto the corneal surface.
By multivariate analysis, the most predictive variables of thermotherapy-induced focal iris atrophy were increasing number of treatment sessions (P=.001) and increasing tumor base (P=.02). Both variables likely reflect more difficult treatment with greater chance for heat gradient in the anterior chamber, leading to pupillary spasm and inadvertent iris heating. With mean follow-up of 1 year (maximum, 4 years), the focal paraxial lens opacity did not progress or obstruct vision in any child.
Our results should be interpreted with caution. This is a nonrandomized study, and patients treated with thermotherapy were carefully selected. Our goals were to assess tumor control provided by thermotherapy and its associated ocular complications. We did not compare a chemotherapy/thermotherapy combination with thermotherapy alone; this would have been unrealistic because most tumors in the former group were large with prominent associated features, whereas those in the latter group had less serious features. We chose to assess all patients as a whole to provide a comprehensive overview of this technique. As more patients are treated in the future, we will evaluate each individual group. In addition, our patients were not randomized to chemotherapy but carefully selected as candidates for chemotherapy based on many clinical variables, as outlined in previous studies.6,15 Our overview did not address the timing of chemotherapy coupling, specific chemotherapy agents used, and other related variables. Finally, these results are preliminary, with a maximum of only 4 years of follow-up. Longer follow-up may reveal a greater incidence of tumor recurrence or ocular complications.
Management of retinoblastoma is complex, and decision making for pediatric patients requires extensive experience with the disease; familiarity with treatment techniques; and understanding of expected tumor, vision, and systemic outcomes. Thermotherapy is among many important techniques for retinoblastoma treatment. We continue to use plaque radiotherapy, generally for tumors that fail thermotherapy. Cryotherapy remains important for treating small peripheral tumors. Laser photocoagulation has become less important because it has been largely replaced by thermotherapy, which leaves the retina with patent vessels and a smaller scar. In this study, we found that thermotherapy is most effective for selected small retinoblastomas. Tumors measuring less than 3 mm in base and 2 mm in thickness can sometimes be treated with thermotherapy alone without the need for systemic chemotherapy. Larger tumors, 3 mm or more in base and 4 mm or more in thickness, require higher power, longer duration, increased number of treatment sessions, and greater need for previous chemoreduction, leading to greater risk for local tumor recurrence and ocular complications. Although thermotherapy plays a major role in the management of retinoblastoma, advanced retinoblastoma often requires enucleation or chemoreduction plus external beam radiotherapy.6,16,43 Future developments of thermotherapy for retinoblastoma may include local chemotherapy delivery44 and improved heat delivery devices, providing promise for this technique.
Accepted for publication February 12, 1999.
Support was provided by the Paul Kayser International Award of Merit in Retina Research, Houston, Tex (Dr J. A. Shields); the Macula Foundation, New York, NY (Dr C. L. Shields); Research to Prevent Blindness, New York, NY; The Andres Soriano Cancer Research Foundation, Manila, the Philippines (Dr Mercado); CAPES, Brasilia, Brazil (Dr Santos); and the Eye Tumor Research Foundation, Philadelphia.
We thank Jacqueline Cater, PhD, Philadelphia, for statistical consultation.
Presented as a poster at the American Academy of Ophthalmology, New Orleans, La, November 10-11, 1998, and as a paper at the International Congress of Ocular Oncology, Philadelphia, Pa, May 3, 1999.
Reprints: Carol L. Shields, MD, Ocular Oncology Service, Wills Eye Hospital, 900 Walnut St, Philadelphia, PA 19107.