Seeking to optimize the immunocytochemical assay of P-glycoprotein, a 170-kilodalton (P-170) molecule associated with multidrug resistance, we experimented with a variety of antibodies (JSB-1, C219, and MRK-16), fixation conditions, and titers using both human myeloma cell lines and clinical myeloma specimens. Under optimized conditions, using all three antibodies and the cell lines as standards and controls, the ICC method proved sensitive, specific, reliable, rapid, and within the realm of everyday hospital laboratory expertise. The 3 anti-P-glycoprotein antibodies revealed different reactivities with P-170. Both C219 and JSB1 were optimized by fixation in cold acetone. With MRK-16 optimal results were obtained on unfixed or formalin fixed specimens. Under optimal fixation and titering conditions, low level (DOX 4) detection was possible. Given that the three antibodies differ in reactivity and recognize different P-170 epitopes, it follows that using the antibodies in a small panel is a useful strategy in increasing the likelihood of detecting true P-glycoprotein expression by the immunocytochemical method. In dilution experiments, the immunocytochemical method was as sensitive as RNase protection assay and more sensitive than Western blot detection. Immunocytochemistry coupled to computer-assisted single-cell densitometry, showed a strong correlation (R = 0.98) between cellular P-170 density and in vitro resistance to doxorubicin. Multidrug-resistant specific probes for RNA expression and Western blot assays confirmed the specificity of P-170 expression in both cell lines and clinical samples. Thus, a small panel of antibodies, under optimized immunocytochemical conditions, appears to have potential as a rapid, sensitive, clinically useful assay for multidrug resistance in myeloma.