ASH 1 - The German CLL8 trial

For CLL patients and doctors alike, the main event at ASH was the German CLL8 trial comparing FC with FCR. We already knew from the Roche website that the trial had shown that FCR was sufficiently better than FC for the trial to be stopped after the first interim analysis, but this was our first opportunity to examine the details.

817 patients from several countries were randomized between FC and FCR in standard dosages. The patients were younger than most with a median age of 61 (though some patients were as old as 80) and they were fit with good kidney function. 61% were Binet stage B and 32% stage C, so it is unlikely that this trial took the easy way out by treating patients that didn't really need treating. The overall incidences of trisomy 12 and abnormalities of 13q, 11q23, and 17p13 detected by FISH were 12%, 57%, 25%, and 8%, respectively, with no statistically significant differences between treatment arms. On average patients managed to get 5 of the recommended 6 courses in.

Analysis was in June 2008, with a median of just over 2 years follow-up. Overall response rate was significantly higher for the FCR arm (95% v 88%; p=0.001). The CR rate (1996 NCI criteria) was significantly higher for the FCR arm (52% v 27%; p<0.0001). Progression-free survival was was 76.6% at 2 years in the FCR arm and 62.3% in the FC arm (p<0.0001). But the difference for overall survival was not significant (91% v 88% at 2 years).

Adding the rituximab resulted in more toxicity. Grade 3 and 4 events occurred in 62% v 47% and severe hematologic toxicity in 55% v 39% (the significant difference was for neutropenia), but there was no significant difference in the incidence of serious infections (18.8% v 14.8%) nor of treatment related mortality (2% v 1.5%).

Minimal residual disease (MRD) was assessed in this trial in both blood and bone marrow by 4-color flow - capable of detecting one residual cell among 10,000 tested (for comparison a CR just means that there are fewer than 30% lymphocytes in the bone marrow). Assessment was after 3 courses of treatment, again one month after finishing treatment and again three months after treatment. Whether tested on blood or bone marrow and at whatever time tested there was significantly less residual disease in the FCR arm than in the FC arm. For example, fewer than one in 10,000 cells were detected in 33.6% vs. 6.4% at the first assessment of blood MRD, 67.6% vs. 36.6% at the second and 66.4% vs. 34.2% at the final assessment (for all comparisons FCR is first and p < 0.0001). The results from the marrow showed slightly less clearance of MRD (eg 47.6% v 66.4% for the final assessment of the FCR arm).

Progression-free survival (PFS) was greater if lower levels of MRD were achieved. Although this was more likely when rituximab was included, if it could be achieved it didn't matter which arm of the trial was involved. If FC can give fewer than 1 cell in 10,000 at the final assessment, progression-free survival will be no longer if rituximab is added.

Incidentally, this trial compared two methods of detecting MRD, the 4 color flow method described above and an RQ-PCR method. The results were virtually identical where more than 1 in 10,000 cells were detectable, but below that level RQ-PCR was more sensitive.

Any residual cells present up to 180 days after the last rituximab infusion lacked CD20 - evidence that persisting antibody continued to modulate the antigen.

The trial also looked at the role of prognostic markers in this trial. The incidences of the most common genomic aberrations were 13q- 56.7%, 13q- single 36.4%, 11q- 24.6%, +12 12.0%, and 17p- 8.2%. IGHV was unmutated in 63.4% and V3-21 was rearranged in 4.9%. AS expected, particularly poor outcomes were observed for 17p- in both arms (FC and FCR): CR (4.5% and 19.0%), CR+PR (45.5% and 71.4%), PFS (at 24 months: 0.0% and 29.6%), and OS (at 24 months: 41.0% and 53.3%). However, even these poor results do not suggest that rituximab is completely inactive in p53 deleted disease. Unmutated IGHV gene status was associated with shorter PFS in both arms combined and individually (all p<.001), shorter overall survival in the FC arm (p=.006), and probably in the FCR arm though follow-up is not long enough for this to be statistically significant (p=.092).

Multivariate analysis was applied to the trial. Regarding PFS, independent prognostic factors were 17p- (Hazard Ratio (HR) 6.76, p<.001), unmutated IGHV(HR 1.97, p<.001), FCR (HR 0.51, p<.001) and +12 (HR 0.58, p=.020). Regarding overall survival, only 17p- (HR 7.47, p<.001) and unmutated IGHV (HR 2.09, p=.018) were identified as significant independent factors, though FCR might be with longer follow-up (HR 0.66, p=.085).

In conclusion, genetic parameters remain powerful prognostic markers after 1st line FC and FCR treatment. The overall improvements by FCR result from specific treatment effects in distinct genetic subgroups. It seems particularly important that patients with 11q- appears to benefit most from the addition of rituximab. However, 17p- and unmutated IGHV status remain predictors for shorter PFS and OS independently of the overall improvement by FCR.