Detection of T-cell clonality by polymerase chain reaction (PCR) and high-resolution electrophoresis facilitates differentiation of early stages of cutaneous T-cell lymphoma (CTCL) from benign T-cell-rich dermatoses. However, data regarding the sensitivity of the various electrophoresis techniques differ remarkably. In the present study, the capacity of heteroduplex (HD)-loaded temperature-gradient gel electrophoresis (TGGE) to detect clonally expanded T-cells was assessed systematically and modifications to the procedure were defined. Using our standard protocol, HD-TGGE detected clonal T-cell receptor (TCR)-gamma PCR products, generated from the Jurkat cell line, down to a total of 2 ng/muL (14 ng) DNA. However, slowly migrating single strands of the clonal PCR product reduced the amount of the clonality indicating homoduplices. To overcome this single-strand formation, thus decreasing the detection limit, the urea concentration in the gel and the temperature ramp for the HD-formation were altered, as well as the temperature gradient in the gel. Application of the modified protocol resulted in a tenfold lower detection limit of 0.15 ng/muL (1.05 ng) DNA in the clonal band. The sensitivity of the adapted HD-TGGE was investigated by dilution experiments using the well established T-cell lines Jurkat, Molt-4, MyLa and SeAx. By these approaches clonal PCR products diluted in nonclonal PCR products were detectable down to concentrations of 5-10%. Comparably, in the case of mixtures of clonal in nonclonal DNA the detection limit reached 5-10% clonal DNA. However, by dilution of clonal cells in nonclonal peripheral blood mononuclear cells, which corresponds to in vivo conditions, a lower detection limit of approximately 1-5% was observed.