Photodynamic therapy: a 5-year study of its effectiveness in the treatment of posterior uveal melanoma, and evaluation of haematoporphyrin uptake and photocytotoxicity of melanoma cells in tissue culture.

A human melanoma cell line RVH-421 which selectively demonstrates melanin synthesis when cultured in L15 Leibowitz medium but not in RPMI medium was used as a model to examine haematoporphyrin derivative (HPD) uptake and the photocytotoxicity of photodynamic therapy (PDT). Confocal scanning microscopy and extraction fluorometry showed HPD uptake in both non-pigmented and pigmented melanoma cells. Phototoxicity was determined by incubating pigmented and non-pigmented monolayer cells with HPD, exposing them to variable periods of white fluorescent light and calculating the number of viable cells in the samples relative to the controls. Both the non-pigmented and pigmented melanoma cells were affected by light-dependent cytotoxicity which was greater in the non-pigmented cells. Melanin or other substances may reduce the photo-oxidative effects of PDT. Posterior uveal melanomas in 36 patients were treated with PDT with the longest duration of tumour control being 6.5 years. Kaplan-Meier survival analysis showed that 76% of melanomas were not growing at the end of the first year, declining to 62% at the end of the second year, with 38% showing no signs of growth at the end of the fifth year. No eyes were lost as a result of PDT. Cox's hazards analysis showed that the degree of tumour pigmentation and patient age at therapy significantly influence the tumour response to PDT.

Phototherapy of posterior uveal melanomas.

The results of phototherapy on a small series of 19 patients with uveal melanomas are presented. Laser radiation at a wavelength of 620 to 630 nm was used in conjunction with a 5 to 7.5 mg/kg dose of haematoporphyrin derivative administered 24 hours before treatment, with total doses/treatment averaging 960 J/cm2. Eleven patients received two treatments, and one received a third. The longest duration of tumour control to 30 September 1990 was 31 months. Of the 19 patients treated six had complete regression of the tumour, while another five had minor to significant regression. A strong correlation between degree of regression and degree of tumour pigmentation was found, the lighter tumours responding much better than darker tumours. There were also strong indications that if a tumour did not respond to the initial phototherapy it was very unlikely that subsequent phototherapy would provide any further benefit.

Volume measurement of intraocular tumours by cross-sectional ultrasonographic scans.

The volume of intraocular tumours can be calculated from tumour height and cross-sectional areas measured in maximal orthogonal B-mode ultrasonograms, on the assumption that the tumour morphology is hemi-ellipsoidal. Comparison of this method with an accurate but more complex computer-aided B-mode ultrasonic tomographic method demonstrated a highly linear relationship with a correlation coefficient of 0.93. Furthermore, the comparison of cross-sectional areas measured from B-scan images and histological cross-sections in a small series of enucleated eyes demonstrated that tumour cross-sectional area measured from ultrasound images provides an accurate representation of the true physical size of the tumour in the globe. This method can be easily implemented with readily available B-mode ultrasound equipment. Unlike previously reported volume measurement methods, customised hardware or software is not required. The method is rapid and easy to perform, and may offer a useful standard method of measuring and monitoring intraocular tumours.

Computer-assisted analysis of the electro-oculogram.

A method of acquiring and analysing electro-oculogram signals was developed using computer-aided techniques. The repetitive nature, characteristic shape and spectral properties of the electro-oculogram are used in the automated analysis, enabling the retinal resting potential to be investigated. Online computer-aided automatic analysis of the electro-oculogram is efficient, reliable and accurate. Automated analysis produced results that have a highly significant degree of correlation (P less than 0.001) with the results obtained manually from strip chart recording of the electro-oculogram signals. The increased accuracy of the automatic analysis system demonstrates a significantly reduced (McNemar's Test) false positive rate in classifying abnormal results on the basis of Arden ratio. The computer-aided analysis system offers a useful clinical means of measuring the electro-oculogram for diagnostic and research purposes.

Computer-aided area measurement of ocular images.

Using a combination of computer-aided measurement techniques and fundus camera images, the area of regions of interest on the fundus can be measured. If these area measurements are corrected for the magnification effects of the patient's visual system, and the fundus camera optics, an estimate of the true anatomical size of the measured object can be determined in physical units. While the results of measuring models indicate a maximum error of 6% for the system, the results of clinical trials demonstrate an average error of 10% for well-defined areas of interest, with increasing error of up to 30% for poorly defined areas of interest.

Computer-aided volume measurement of choroidal melanomas.

Using a combination of conventional B-scan ultrasonography, standard video techniques, computer image digitization and area analysis, it is possible to measure the volume of choroidal melanomas by area calculation of successive ultrasound image slices, allowing the monitoring of such tumors. The design of a combined water-bath and linear tracking B-scan transducer is presented, together with a description of the instrumentation and computer software required. The results of volume measurements performed upon eye models demonstrated that melanomas of volume greater than 100 mm3 could be measured with an average error of 6%. The problems associated with linear calibration of the system from scan-head to analysed image, resolution of the digitised image, reproducibility of the measurements, sources of error and assumptions made in formulating the measurement procedure are discussed.

Video and digital fluorescein angiography.

A video recording system, which provides real-time imaging and instant reporting of retinal fluorescein angiograms, with improved patient comfort, is described. Digital image-processing methods can be applied to video angiograms to improve image quality, quantitate retinal blood flow, measure areas, improve storage and access, and send images via phone lines or other electronic communication channels. Video image quality is somewhat inferior to conventional photographic techniques, but is more than adequate for diagnostic interpretation.