Article Text


Transpupillary thermotherapy (TTT) for the treatment of choroidal neovascularisation
  1. R S B Newsom,
  2. J C McAlister,
  3. M Saeed,
  4. J D A McHugh
  1. King's College Hospital, Denmark Hill, London SE5 9RS, UK
  1. Mr R S B Newsom, Department of Vitreoretinal Surgery, Moorfields Eye Hospital, City Road, London EC1V 2PD, UKrbnewsom{at}


AIM To assess the effectiveness of transpupillary thermotherapy (TTT) for the treatment of classic and occult choroidal neovascularisation (CNV).

METHOD In a retrospective, case selected, open label trial 44 eyes of 42 patients with CNV secondary to age related macular degeneration (ARMD) were studied. 44 eyes with angiographically defined CNV were treated with diode laser (810 nm) TTT. Laser beam sizes ranged between 0.8 and 3.0 mm and power settings between 250–750 mW. Treatment was given in one area for 1 minute, the end point being no visible change, or a slight greying of the retina. Outcome was assessed with Snellen visual acuity and clinical examination; in 24/44 patients angiographic follow up was available.

RESULTS 12 predominantly classic CNV and 32 predominantly occult membranes were followed up for a mean of 6.1 months (range 2–19). Mean change in vision for classic membranes was −0.75 (SD 1.75) Snellen lines and occult membranes was −0.66 Snellen lines (2.1) (p>0.05). Predominantly classic membranes were closed in 75% (95% CI: 62.5–87.5) of eyes, remained persistent in 25% (95% CI: 12.5–37.5); no recurrences occurred. Predominantly occult membranes were closed in 78% (95% CI: 70.1–85.3) of eyes, remained persistent in 12.5% (95% CI: 6.6–18.5), and were recurrent in 5.1% (95% CI: 4.2–14.3).

CONCLUSIONS Transpupillary thermotherapy is a potential treatment for CNV. It is able to close choroidal neovascularisation while maintaining visual function in patients with classic and occult disease. Further trials of TTT are needed to compare this intervention with the natural history and other treatment modalities.

  • age related macular degeneration
  • choroidal neovascularisation
  • transpupillary thermotherapy
  • diode laser

Statistics from

Age related macular degeneration (ARMD) is the leading cause of visual loss in the developed world and choroidal neovascularisation (CNV) is the main cause of severe visual loss in this group.1-4 Laser photocoagulation treatment for classic CNV reduces the incidence of severe visual loss in patients with extrafoveal and juxtafoveal membranes. However, because photocoagulation damages neurosensory mechanisms, severe visual loss is often experienced in patients with subfoveal lesions.1-10The benefits of laser treatment are also reduced by recurrence in many patients.9 In addition, 85–90% of patients with CNV are not eligible for argon laser treatment because the CNV is occult, or subfoveal and large.1-3 The limitations of photocoagulation have led to a proliferation of new treatments (photodynamic therapy,11-14 radiotherapy,15interferon α2,16 macular rotation,1718and submacular membrane excision for occult and subfoveal lesions19).

The diode laser (emission wavelength λ 810 nm, near infrared) has a number of biophysical advantages in the treatment of macular disease. It has low absorption in xanthophyll, minimising nerve fibre layer damage, it is poorly absorbed by haemoglobin, allowing treatment through preretinal and subretinal blood but is mainly absorbed in the choroid, enabling effective treatment of choroidal lesions.20 Initial studies using photocoagulation have shown it to be effective in closing CNV2122 and in treating macular oedema.2324 Low irradiance, long pulse diode laser irradiation (“transpupillary thermotherapy; TTT”) has been usefully employed for the treatment of choroidal melanomas.2526

A recent report by Reichel and colleagues27 has further demonstrated the benefit of diode laser TTT for the treatment of subfoveal CNV. In a non-randomised case selected pilot study, 94% of CNV showed clinical or angiographic improvement and vision was stabilised or improved in 75%. Optical coherence tomography demonstrated membrane resolution with a restoration of the retinal anatomy.

We aimed to assess the efficacy of transpupillary thermotherapy for the treatment of both classic and occult CNV.

Materials and methods

Forty two consecutive patients with CNV secondary to ARMD, referred to the retinal clinic at King's College Hospital, were recruited to a non-randomised pilot trial. Baseline characteristics and eligibility criteria are shown in Table 1. The patients were over 55, had signs of ARMD, and had symptoms of visual loss or metamorphopsia within the previous 4 months. There was angiographic evidence of CNV with or without pigment epithelial detachment. The Snellen visual acuity ranged between 6/9 and counting fingers. Patients who had received previous photocoagulation or were on alternative treatments for AMD were excluded from the trial, as were patients with other serious ocular disease or medical conditions that would prevent them from completing treatment.

Table 1

Baseline characteristics of patients


Initial clinical assessment included Snellen visual acuity, anterior segment examination, IOP measurement, mydriasis with 1% tropicamide, and funduscopy with a 78D or 60D Volk lens (Volk Optical, Mentor, OH, USA). Following this, patients were assessed with digital or photographic fundus colour (Topcon, Tokyo, Japan) and fluorescein angiography (sodium fluorescein 5 ml, 10%). Patients were counselled as to the prognosis for their condition and their treatment options. Patients not willing to undergo treatment or to complete the counselling were excluded from the trial. The patients consented to treatment and the investigation took place with the recognition of the King's College Hospital ethics committee.


Treatment was scheduled between 1 hour and 4 days of the initial angiogram. The membranes were classified in accordance with their angiographic appearance. Transpupillary thermotherapy with an 810 nm diode laser (Iris Medical OcuLight SLx, Iridex Corporation, Mountain View, CA, USA), was delivered through a dedicated slit lamp mounted delivery system. Laser settings were adjusted to give faint retinal greying following 1 minute of treatment. Beam diameters varied between 800 and 3000 μm.

The goal of TTT is to create and maintain tissue hyperthermia. It has been found that higher irradiance is needed for smaller laser spot diameters owing to the efficient heat conductance of the choroidal circulation. Hence, laser irradiance must be inversely proportional to the spot diameter, which determines the size of the thermally affected tissue volume.28 For a 3 mm beam the settings were between 500–700 mW (irradiance 6–9 W/cm2); for a 2 mm beam between 400–650 (irradiance 11–18 W/cm2); and for a 1.2 mm beam 300–600 mW (irradiance 24–47 W/cm2). A Volk (Mentor, OH, USA) area centralis contact lens was used in every case giving a magnification of 1.06× of the laser spot at the retina. Patients were anaesthetised with topical benoxinate 1%. Laser targeting was achieved with a diode red beam, the lesion being covered and extending 100 μm beyond the border. If whitening of the lesion was seen the laser power was reduced by 1–200 mW if further treatment was necessary. In patients with large lesions, confluent overlapping treatments were given.


Follow up was arranged at 2, 4, 6 weeks and 3, 6, 9, 12, and 18 months. At each visit Snellen visual acuity and retinal examination were performed with fundus photography and fluorescein angiography, if clinically indicated. Retreatment was given to patients who retained active membranes for more than 4 weeks following TTT.


Outcomes measured were clinical or angiographic resolution and visual acuity. The results are presented in terms of percentage of membranes resolved and number of lines of vision lost within each group. The amount of visual deterioration was assessed using Wilcoxon's rank sum test on the Snellen data. Correlation between the visual acuity and the size of the CNV was performed using the Fisher's R-Z correlation.



The patient characteristics are summarised in Table 1; 44 eyes of 42 patients were treated. The mean patient age was 77.9 years (SD 8.38 years); 12 (27%) membranes were classic and 32 (73%) were occult with a mean area of 1.7 DA (range of 0.5–4.0 DA). Of the classic membranes one was extrafoveal, two were juxtafoveal, and nine were subfoveal. There was no correlation between visual acuity and size of CNV Z = −1.23 p=0.2. Haemorrhage was present in 7/12 of the classic membranes and 22/32 of the occult membranes. Seven patients were pseudophakic and two had glaucoma treated with timolol 0.25% BD. Table 1 shows the baseline data.


The average number of treatments sessions was 1.80 (SD 0.87) spread over 1–6 months. The average power was 486 mW (140), with a beam diameter of 1.69 mm (0.63). Eyes received 4.25 (1.9) laser spots per treatment. Two patients were noted to have a significant post-treatment haemorrhage, no patients complained of pain, although some sensation was common. No anterior segment damage was noted or reported by the patients. Clinical examples of the treatment of occult and classic membranes are shown in Figures 1, 2, 3.


Predominantly classic membranes

Predominantly classic membranes were closed in 9/12 (75%, CI: 62.5–87.5) eyes and remained persistent in 3/12 (25%, 95% CI: 12.5–37.5) with no recurrences; with an average follow up of 6.7 months (SD 4.33, range 3–19) and 1.75 treatments. The initial visual acuity was 6/32.4 with a post-treatment acuity of 6/40.5 (p=0.17), an average of 0.75 (SD 1.75) Snellen lines lost (Fig 4A). Vision improved (>2 lines) in 0/12 patients, was stabilised (0 to + 1 line) in 7/12 (58.1%, 95% CI: 44.1–72.5) patients, there was mild loss (–1 line) in 1/12 patients (8.1% 95% CI: 3.5–16.3), and four patients lost vision (2–4 lines) 33.3% (95% CI: 19.7–46.9) (Table 2).

Figure 1

(Top left and right) Colour and fluorescein images showing a small classic juxtafoveal membrane. (Bottom left and right) Colour and fluorescein images following TTT with resolution of the membrane; the patient's initial vision improved from 6/12 to 6/9 3 months after treatment.

Figure 2

(Top left and right) Colour and fluorescein images showing an occult subfoveal membrane. (Bottom left and right) Colour and fluorescein images following TTT, with complete closure of the CNV. The patient's vision was stabilised 3 months following treatment.

Figure 3

(Top left and right) Colour and fluorescein images showing an occult subfoveal membrane with extensive subretinal haemorrhage. (Bottom left and right) Colour and fluorescein images showing resolution of the CNV and absorption of subretinal haemorrhage 6 weeks following TTT treatment. The patient's vision improved from 6/60 to 6/24.    

Figure 4

(A) Change in visual acuity following transpupillary thermotherapy for classic CNV. There was no significant deterioration in vision between the pretreatment group (VA = 6/32.4) and the post-treatment group (6/40) (p=0.1). (B) Change in visual acuity following transpupillary thermotherapy for occult CNV. There was no significant deterioration in vision between the pretreatment group (VA = 6/24.0) and the post-treatment group (6/26.4) (p=0.1).

Table 2

Pre and post-TTT visual acuity classic membranes

Predominantly occult membranes

Predominantly occult membranes were closed in 25/32 (78%, CI: 70.1–85.3) eyes, remained persistent in 4/32 (12.5%, CI: 6.6–18.5), and were recurrent in 3/32 (5.1%, CI: 4.2–14.3) eyes; at an average follow up of 7.2 months (SD 4.33 range 3–18) and 1.8 treatments. The initial visual acuity was 6/24 with a post-treatment acuity of 6/26.4 (p=0.68), an average of 0.66 (SD 2.1) Snellen lines lost (Fig 4B). The vision improved (2+ lines) in 4/32 (12.5%, CI: 6.6–18.3) patients, 14/32 (43.8%; CI: 34.9–52.5) were stable (0 to +1 line), there was mild loss (–1 line) in 5/32 patient (15.6% 95% CI: 9.2–25.6), 6/32 (19.2% CI: 11.8–25.6) had moderate loss (2–4 lines), and 3/32 eyes (9.5% CI: 4.2–14.5) eyes had severe visual loss (Table 3). A Kaplan–Meyer estimate of a three line change in vision is shown in Figure 5.

Table 3

Pre and post-TTT visual acuity occult membranes

Figure 5

Kaplan–Meyer estimates of the cumulative proportion of eyes treated with TTT with moderate visual acuity loss (⩾ 3 Snellen lines).


The results from this study show that TTT is effective in closing membranes in patients with CNV. Seventy seven per cent of membranes were closed at 6 months and only 7.1% developed recurrent CNV. This high closure rate was associated with a stabilisation or improvement in vision in 71% of patients with occult disease and 67% of patients with classic disease. These results compare well with the natural history of occult CNV where 63% of patients suffered 3 or more lines of visual loss in the first 6 months28 and to the MPS study of classic membranes, where recurrence rates of 59% were reported.10

Early clinical work recently published by Reichelet al have shown encouraging results for patients with occult subfoveal CNV.27 They followed 16 eyes of 15 patients for 13 months following TTT for occult subfoveal CNV: 3/16 eyes experienced a 2+ line improvement, 9/16 stabilisation of vision (+/− 1 line) and 4/16 showed a significant decrease (>1 line). Optical coherence tomography and fluorescein angiography demonstrated resolution of the membranes. This compares with our findings that moderate loss of >2 lines occurred in 33% and severe loss >5 lines occurred in 0% of our patients with occult disease.

Results from this study compare well with the natural history of occult disease29 and the recently reported results from the TAP study group, which evaluated photodynamic therapy (PDT) for predominantly classic membranes.14 Their results show that, at 12 months, PDT stabilised vision in 61.2% (<3 lines lost) of the treatment group compared with 46% of the control group. The visual results of our study are comparable; 75% of patients, with predominantly classic membranes, having <3 lines of visual loss. However, TTT closed a higher proportion of CNV at 6 months (75%) compared with PDT (19%), and the control group (9%).

The MPS demonstrated that argon laser photocoagulation was effective in closing classic membranes (72% at 6 months) but that this was associated with visual loss for subfoveal membranes. TTT may therefore combine the high rate of closure found with argon photocoagulation and the low recurrence rate with PDT, by closing CNV without damaging the neuroretina. Table 4 shows a comparison of the results between argon laser photocoagulation, photodynamic therapy, and transpupillary thermotherapy.

Table 4

Chart of closure rate and visual acuity results at 6 months for TTT, PDT14 and argon laser photocoagulation of subfoveal membranes1

Transpupillary thermotherapy uses hyperthermia to disrupt the neovascular process and has been widely used for the treatment of small choroidal melanoma.2526 The laser is set to a low power with long pulse duration, the aim being to produce a moderate intralesional temperature rise (4–9°C), compared with 40°C for conventional laser photocoagulation. The lower temperature rise confines the thermal damage profile in the target tissue, thereby reducing damage to the middle and inner retina. Conventional photocoagulation may cause dense scotomata and recurrence following argon laser photocoagulation, particularly when treating subfoveal CNV. A large spot (0.8–3 mm) size, and a long low energy pulse enables complete treatment of the lesion and allows dose titration. The mechanism of action with TTT is still unclear. However, histological sections of treated human melanoma have demonstrated intravascular thrombosis,3031 which may account for the small amount of leakage, or haemorrhage following CNV treatment. Vascular closure may be precipitated by release of cytotoxic free radicals released from irradiated tissue. Certainly the relatively rapid involution of the CNV following TTT prevents further subretinal scarring and permanent photoreceptor damage, thereby limiting central scotomata in patients with subfoveal membranes.

The treatment variables have not been completely elucidated. In contrast with Reichel et al27we used several spots during one treatment session (average 4.25). It is clear that the end point is clinically difficult to determine (as there is little retinal reaction) and that treatment parameters may need to be adjusted for the degree of neurosensory retinal detachment and the amount of subretinal haemorrhage, as well as for the location of the CNV.

Several complications have been reported following TTT for uveal tumours, including branch retinal vein and artery occlusions and retinal and choroidal haemorrhage. There also have been reports of ciliary nerve damage and corneal ulceration following panretinal photocoagulation.32 However, low power settings used in TTT for CNV have meant that complications are rare and patient tolerance is good.

The precise role of TTT in the treatment of CNV needs to be further defined. However, the clinical improvement found in several of our patients and the high closure rate of occult CNV membranes, with little recurrence, suggest that this treatment will have a role in the management of some patient groups. Prospective, randomised controlled trials will further clarify the role of TTT in the treatment of CNV.


We thank Dr Lyn Jenkins for coordination of angiography and image processing; Iris Medical for the loan of an OcuLight 810 nm diode laser; and Colin Clements, medical photographer, King's College Hospital, London.


View Abstract

    In the paper published in the February issue of the BJO (2001;85:173–8) by Newsom et al, there was an error in Figure 4 (p 176). The axes are wrongly labelled. The y axis should read Post-TTT VA and the x axis should be Pre-TTT VA. We apologise for the error.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd.BMA House, Tavistock Square, London, WC1H 9JR