Patient-tumor-specific oligonucleotides were generated for the detection of minimal residual disease (MRD) in a highly specific and sensitive clonotypic polymerase chain reaction (cPCR). The clone-specific region of highest diversity, CDR-III, was PCR amplified and sequenced. Nested CDR-III clonotypic primers were used in a semi-nested cPCR with a sensitivity of at least 1 in 10(5) cells. Patients with protocol-eligible Rai intermediate or high-risk chronic lymphocytic leukemia (CLL) received induction with fludarabine 25 mg/m(2) per day for 5 days every 4 weeks for 6 cycles, followed by consolidative high-dose cyclophosphamide (1.5, 2.25, or 3g/m(2)). cPCR was performed on peripheral blood and bone marrow mononuclear cells. All 5 patients achieving a clinical partial remission (PR) studied by cPCR were positive. Five patients achieved nodular PR (nPR) (residual nodules or suspicious lymphocytic infiltrates in a bone marrow biopsy as the sole suggestion of residual disease). Five of 5 patients with nPR were cPCR positive. In contrast, flow cytometry for CD5-CD19 dual staining and kappa--lambda clonal excess detected MRD in only 3 of the same 5 nPR patients, all of whom were cPCR positive, and immunohistochemistry detected MRD in only 1 of 4 assessable patients. Three of 7 CR patients evaluable by cPCR had MRD. Only 1 CR patient had MRD by flow cytometry; that patient was also cPCR positive. These data support the conclusions that nodular PR in CLL represents MRD and that clonotypic PCR detects MRD in CLL more frequently than flow cytometry or immunohistochemistry. (Blood. 2001;97:1929-1936)