We have analysed the clinical and laboratory features in 544 patients with chronic lymphocytic leukaemia (CLL) with available cytogenetics and fluorescence in-situ hybridization (FISH) analysis for trisomy 12 in half of them, to examine the correlation between chromosome abnormalities and clinical or laboratory parameters. Five chromosome groups were defined: (1) trisomy 12 (18%), detected as the sole abnormality or associated with other changes; (2) del(13)(q12-14) (7%); (3) other abnormal karyotypes (20%); (4) normal karyotype (41%); and (5) no divisions (14%). There were no differences in the age distribution between the five groups. Clinical stages (Binet) were: A (74%), B (12%) and C (14%). Stage A was common in cases with del(13q)(82%), normal (84%) and other abnormal karyotypes (74%), whereas it was less common in trisomy 12 cases (64%) and those with no divisions (48%). Typical CLL morphology was found in 83% of cases; 10% had more than 10% prolymphocytes (CLL/PL) and 7% had other atypical features. CLL with trisomy 12 was the only group with a high frequency of either CLL/PL (31%) or atypical morphology (24%). Atypical morphology and CLL/PL were even more frequent when trisomy 12 was associated with other chromosomal abnormalities (70% v 46%). The incidence of cases with CLL/PL and other atypical morphology was significantly lower in the other chromosome groups (P < 0.001). There were no differences in immunophenotype among the various groups except for a higher frequency of stronger Smlg and FMC7 expression in cases with trisomy 12, particularly those with CLL/PL and other atypical morphology. Our findings confirm that trisomy 12 defines a subgroup of CLL with more frequent atypical morphology, including CLL/PL, stronger SmIg and FMC7 expression, more advanced stages (B and C in 18%) and possibly worse prognosis.

Summary.We have analysed the prognostic influence of cytogenetic findings at diagnosis in a group of 502 children with acute lymphoblastic leukaemia (ALL), treated on MRC UKALL X, in whom clonal cytogenetic abnormalities were detected at diagnosis. Despite the overall improvement in outcome in children treated on this protocol compared with previous trials, some cytogenetically-defined groups were still associated with a poor outcome and ploidy retained some prognostic significance. Patients with high hyperdiploid ALL (39% of those with clonal abnormalities) had a favourable outcome with event free survival of 71% at 5 years. Those with near haploidy (1%), hypodiploidy (9%) and low hyperdiploidy (16 . 5%) had a relatively poor prognosis with event-free survival at 5 years of 17%, 42% and 49% respectively. Only two of 12 children with Ph-positive leukaemia are alive in remission and abnormalities of chromosome 11q23 were also associated with a high risk of treatment failure. In contrast, the t(1;19) was associated with improved event-free survival of 87 . 5% at 5 years. A number of other non-random abnormalities were identified with no clear prognostic significance.We conclude that identification of certain genetic changes remains important in the management of acute lymphoblastic leukaemia, although whether molecular diagnosis of clinically relevant abnormalities can now supplant cytogenetics in the clinical trials context remains to be determined.