Echographic differentiation of intraocular melanomas by computer analysis

Area of Science:

• Ophthalmology
• Medical Imaging
• Oncology

Background:

• Choroidal melanomas are the most common primary intraocular malignancy.
• Accurate histological subtyping is crucial for prognosis and treatment planning.
• Distinguishing between different melanoma cell types (spindle vs. mixed/epithelioid) can be challenging with standard methods.

Purpose of the Study:

• To evaluate the efficacy of clinical and quantitative acoustic/texture parameters from echograms for differentiating histological subtypes of choroidal melanomas.
• To compare the diagnostic performance of traditional clinical echographic analysis with advanced acoustic/texture analysis.

Main Methods:

• Clinical parameters (morphological, kinetic, quantitative A- and B-mode echograms) were assessed by a diagnostician.
• Acoustic/texture parameters were derived from processed radio frequency, AM-demodulated, and FM-demodulated echograms.
• Discriminant analysis was employed to classify melanoma subtypes based on selected parameters.

Main Results:

• Clinical echographic parameters achieved 77% accuracy in retrospective classification (AUC 86%).
• Quantitative acoustic/texture parameters demonstrated superior performance, yielding 89% sensitivity and 92% specificity (AUC 96.6%).
• Prospective classification of unknown tumor types using acoustic/texture data was successful in 86% of cases.

Conclusions:

• Quantitative acoustic/texture analysis of echograms offers a highly accurate, non-invasive method for differentiating choroidal melanoma subtypes.
• This advanced echographic approach significantly improves upon traditional clinical assessment for melanoma classification.
• The findings support the potential of acoustic/texture analysis as a valuable tool in the clinical management of choroidal melanomas.