Previous results with individualised tumour response testing (ITRT) in vitro in chronic lymphocytic leukaemia (CLL) have consistently shown good correlation with patient response and survival. We describe here an improved test and report its use with samples from the Leukaemia Research Fund CLL4 randomised clinical trial and previously treated patients. ITRT was performed by the tumour response to anti-neoplastic compounds (TRAC) assay, a modification of the differential staining cytotoxicity (DiSC) assay. Improvements included drying drugs into wells before assay and using the Octospot system to cytocentrifuge eight spots of cells onto one microscope slide. We successfully tested 765/782 (98%) cellular blood samples received within 48 h of phlebotomy. Cross-resistance (Pearson's r > 0.7) in untreated CLL was found between similar drugs. Mitoxantrone (r = 0.31), cyclophosphamide (r = 0.35) and pentostatin (r = 0.29) had low cross-resistance with fludarabine. Treatment resulted in increased resistance to chlorambucil, cyclophosphamide, doxorubicin, mitoxantrone, corticosteroids, cladribine and fludarabine (P < 0.01) but not to pentostatin. These results provide further rationale for standard drug combinations such as fludarabine-mitoxantrone and fludarabine-mitoxantrone-cyclophosphamide and suggest possible pentostatin salvage in fludarabine-resistant patients. ITRT results could assist both in determining the best treatment for individual patients and in the design and rationale of future clinical trials.