G.Valet, F.Otto1)
Malignant melanomas are usually surgically removed upon histological confirmation of the diagnosis from bioptic material. This intervention either cures the malignant affection or the disease further progresses. There is a high need at this point to establish individual patient prognosis to rationally decide e.g. on chemotherapy after surgery.
It was investigated whether several routinely determined clinical parameters like: diameter, infiltration depth, presence of an ulcerative surface and localisation of the tumor in conjunction with flow cytometric DNA-aneuploidy and determination of S-phase cells contained prognostic information at the individual patient level. Deeply infiltrated, large and ulcerated tumors are statistically accepted signs for bad patient prognosis, the predictive value of any one of the parameters alone is not sufficient to reliably foresee the ultimate disease outcome for individual patients.
Data were classified in a standardized and automated way with the CLASSIF1 multiparameter data analysis program.
It was found(1) that the resulting triple matrix pattern classifier permits to predict disease outcome substantially better than any single one of the three most discriminatory parameters: tumor diameter, infiltration depth and ulceration.
1. G Valet, H Kahle, F Otto, E Bräutigam, L Kestens:
Prediction and precise diagnosis of diseases by data pattern analysis
in multiparameter flow cytometry: Melanoma, Juvenile Asthma, HIV Infection.
in: Cytometry (3rd edition), eds: Z Darzynkiewicz, JP Robinson,
HA Crissman, Academic Press, San Diego (2000) in press
2. G Valet, F Otto: 10 year survival prognosis for melanoma patients
by automated classification of clinical and cytometric parameters.
Cytometry Suppl.8 65, (1996)
3. G Valet, M Valet, D Tschöpe, H Gabriel, G Rothe,
W Kellermann, H Kahle: White cell and thrombocyte disorders:
Standardized, self-learning flow cytometric list mode data classification
with the CLASSIF1 program system. Ann.NY Acad.Sci. 677,233-251(1993)