Defining High Risk Myeloma Using Co-Segregating FISH Variables; Results of MRC Myeloma IX

Basing treatment decisions on the clinical variability known to exist in myeloma represents a way to improve patient outcome, however, a robust way of describing this variability is needed. We have investigated the association of tumour-specific genetic variables detected by FISH with survival in the MRC Myeloma IX trial with the aim of identifying high risk subgroups suitable for alternate treatment strategies.

Patients followed an intensive or non-intensive pathway determined by a combination of performance status, patient and physician preference. The intensive pathway randomised patients to CVAD (cyclophosphamide, vincristine, adriamycin and dexamethasone) vs CTD (cyclophosphamide, thalidomide and dexamethasone), followed by HDM and ASCT. The non-intensive pathway randomised patients to MP (melphalan and prednisolone) vs attenuated CTD. The trial recruited 1960 patients with a median age of 59 in the intensive pathway and 73 in the non-intensive, with a median follow-up of 3.7 years. Plasma cells were purified from presenting diagnostic bone marrow samples using CD138 magnetic beads. FISH was performed to detect an IgH split, the common IgH translocation partners (4, 6, 11, 16, 20), hyperdiploidy, del(1p), del(13q), del(16q), del(17p) and gain of 1q. FISH results were available from 1180 patients. The lesions often co-segregated, and consequently we investigated their impact as single and combined lesions.

t(4;14), t(14;16), t(14;20), del(17p) and 1q+ were all associated with adverse overall survival (OS) in multivariate analysis, so these ‘adverse FISH lesions’ were examined in more detail.

t(4;14) was identified in 11% of patients and was positively associated with 1q+ (p<0.001), with 62% of t(4;14) cases co-segregating with 1q+, and a further 12% co-segregating with both 1q+ and del(17p). Interestingly the 25% of patients in which t(4;14) occurred independently were not associated with impaired OS (median 45 vs 49 months, p=0.76). However, patients with t(4;14) and 1q+ combined were associated with an impaired OS of 26 months, whilst the combination of t(4;14), 1q+ and del(17p) had a very poor OS of 9 months(p<0.001).

del(17p) was identified in 8% of patients, and in 38 patients where del(17p) occurred without other adverse lesions there was a trend towards adverse OS which did not reach significance (38 months vs 49 months, p=0.42). However, when del(17p) occurred in combination with 1 other adverse FISH lesion the median OS was 20 months, which shortened to 9 months with 2 lesions (p<0.001).

1q+ was found in 39% of cases and as an isolated lesion it was associated with adverse OS (42 vs 53 months, p=0.015), but again, the association was stronger when it co-segregated with another adverse lesion (median OS 25 months) or 2 adverse lesions (median OS 9 months) (p<0.001).

These data suggest that the accumulation of adverse FISH lesions was associated with progressive impairment of OS, and used this observation to define a group with no adverse lesions associated with a median OS of 61 months, an intermediate group with one adverse lesion (median OS 42 months) and a high risk group defined by the co-segregation of ≥2 adverse lesions (median OS 22 months). These groups retained association with adverse prognosis in the context of thalidomide or conventional induction chemotherapy, intensive or non-intensive therapy. In comparison, the ISS stratified the same group to give a median OS of 68 months (ISS1) vs 48 months (ISS2) vs 36 months (ISS3). Combining the genetic grouping with the ISS identified patients associated with very poor survival that constitute a high risk subgroup.

We show that when genetic lesions such as t(4;14) and del(17p) occur in isolation, they are not strongly associated with short survival. However, combining the results of t(4;14), t(14;16), t(14;20), del(17p) and 1q+ can be used to define high risk myeloma. The true value of these lesions lies in interpreting them together, with the worst performing patients defined by the co-segregation of these lesions. Combining these tumour genetic groups with patient-specific variables can robustly identify patients suitable for alternative therapeutic strategies. More detailed statistical modelling incorporating patient-specific variables will be presented at the meeting.

View largeDownload slide

View largeDownload slide

Close modal

Boyd:Celgene: Honoraria. Davies:Ortho Biotech: Honoraria, Membership on an entity's Board of Directors or advisory committees; Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees; Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees.