Alkylating agents for del 11q

I am sure you will recognise the Dohner et al hierarchical model for FISH results in CLL. This graph has been published hundreds of times. It shows del 17p patients doing worse that del 11q who do worse than trisomy 12, while del 13q patients do better than those with a normal karyotype


At ASH 2010 there was an interesting mini-review in the education session entitled Evidence-Based Mini-Review: The Role of Alkylating Agents in the Initial Treatment of Chronic Lymphocytic Leukemia Patients with the 11q Deletion by Wei Ding and Alessandra Ferrajoli in which they advocated the use of alkylating agents in treating CLL with del 11q.

The first thing that needs to be said is that not all patients with del 11q fall into the poor prognosis pot. Whether it is related to the integrity of the ATM gene on the intact chromosome 11 or whether the poor prognosis cases have genes other than ATM involved is not clear, but certainly some cases do rather better than expected. This means that comparisons need to be made carefully. The two arms of a clinical trial might be well matched for del 11q but poorly matched for bad-risk del 11q and we would not know about it.

I need to remind you of the survival curve discovered by Harmut Dohner

The del 11q patients survived for a much shorter period than those without this chromosomal aberration, but it is important to note that these were young patients aged under 55.


When we examine those over 55 the survival difference disappears. There are also more patients in the over-55 group. However, it must be recognized that they were all patients that needed treatment and that there is a hinterland of patients who do not come to treatment that are underrepresented.

Ding and Ferrajoli reviewed three independent studies, The GCLLSG CLL4 trial, the LRF CLL4 trial and the E2997 trial all comparing fludarabine and cyclophosphamide with fludarabine alone, all of which showed a better response rate and progression-free survival, but no difference in overall survival for the combination. The German trial is only reported in abstract form and unfortunately Ding and Ferrajoli do not give a full reference in their paper. It is Stilgenbauer et al, Blood 2008 volume 112 (ASH abstracts book 18th November, abstract 2089 page 727).

They abstracted the data referring to patients with del 11q from the papers. The German CLL4 trial was in patients under the age of 65 and in these there were 68 patients with del 11q. Ding and Ferrajoli report 100% response rate for FC and only 69% response rate for F in del 11q patients, but I am not sure where they get this information since it is not in the published abstract, nor in the 2006 paper by the same authors of an earlier look at the trial. What the abstract does say is that progression free survival (PFS) was significantly shorter for del 11q patients although they did do better with FC than with F alone. The point is that FC was statistically significantly better than F for PFS for all those with unmutated VH genes, normal karyotype, del 11q, unmutated TP53, CD38>7%, B2MG <5, while for overall survival FC was significantly better than F for those with del11q, trisomy 12 and unmutated TP53. Bear in mind that this report has not yet been subjected to full peer review, despite being more than two years old.

The US Intergroup E2997 trial enrolled 278 patients aged between 33 and 86 (median 61, but still younger than the median age for CLL). However, there were even fewer patients with del 11q, only 40. Although as Ding and Ferrajoli reported there was an indication of a longer PFS with FC (25.2 months) than with F alone (14.9) months, because the numbers were so small, this difference was not statistically significant. Nor did it become statistically significant with longer follow up. The authors conclude, “Thus there is no difference in this study between treatment arms for this group of patients with del (11q).”

The British LRF CLL4 trial was one that I was intimately involved with and my colleague David Oscier has published the impact of prognostic markets on the results. This was the largest of the three trials with 781 patients entered into a three arm comparison of chlorambucil (given at its proper dose of 70mg/sq m/month and continued for up to a year) compared with F and FC. There were 112 patients with del 11q. For PFS the adverse prognostic factors were unmutated VH genes, del 11q, B2MG >4, and treatment with either chlorambucil or fludarabine. For overall survival the adverse prognostic factors were age, B2MG >4, and unmutated VH genes but the differences between the treatment arms was not statistically significant. For patients with del 11q there is a statistically significant improvement in survival for patients treated with FC rather than F (47% v 18.5% at 2 years). What Ding and Ferrajoli fail to say is that FC is similarly superior to chlorambucil, an alkylating agent. Also the addition of cyclophosphamide to fludarabine does not lose the adverse impact of del 11q compared to other karyotypes. Since the thrust of their article is the role of alkylating agents in the initial treatment of CLL patients with the 11q deletion, this is a rather startling omission.

Ding and Ferrajoli have also looked at the addition of rituximab to the combination of purine analog and alkylating agent in some phase II trials. The trial between the Mayo Clinic and Ohio State of PCR was a rather strange one. It recruited 64 evaluable patients between the two sites, but there was some considerable heterogeneity between the sites particularly in respect of performance status, Rai stage and bone marrow diffuse infiltration. This may have accounted for quite different response times between the two sites. There were only 13 patients with del 11q in the trial but nevertheless the trial demonstrated that the treatment free survival (TFS) of del 11q patients was similar to the PFS of the whole group of patients (32 months v 36 months) . What they fail to mention is that PFS is generally shorter than TFS, since treatment is seldom started at the first sign of progression.

This trial acted as historical controls for a more recent small phase 2 trial of 33 patients treated with PR, published in Cancer. Unfortunately, I do not have access to the whole paper but they report a TFS of only 7.9 months compared to 32 months for the PCR patients in the previous trial. Since del 11q is extremely heterogeneous it is not really satisfactory to draw conclusions between such small numbers of patients treated in these trials and no statistical examination is possible. The use of historical controls is also very unsatisfactory.

The final evidence that Ding and Ferrajoli adduce comes from a retrospective analysis of the use of FCR at MDACC. There were 69 patients out of 708 with del 11q23, rather less at 9.7% than in other series, rather confirming the general impression that patients seen at MDACC are often less ill than those seen at other centers. Only 40/69 required treatment. Only 9 patients were Rai stage III or IV and only 8 had bulky lymphadenopathy. However, they generally had adverse prognostic markers with 89% having unmutated VH genes and 74% being ZAP-70 positive. The number with B2MG >4 was only 13 and this suggests to me that MDACC prefers to treat bad risk patients rather earlier in the course of their illness than some other centers. A further 8 patients who had both del 11q and del 17p were excluded from analysis.

It was also from this paper that we gather the information that the Herman CLL4 trial reported a 100% response rate for FC and only 69% response rate for F in del 11q patients, This was a personal communication from Dr Dohner to Dr Keating.

This is rather a good paper in Cancer, not least because it recognizes its limitations. “Our study is limited by the lack of testing for acquisition or presence of mutations in the second allele of the ATM gene, its retrospective nature, the short follow up period, lack of a balanced cohort not receiving chemo-immunotherapy, and lack of a uniformly treated patient group. Although 87.5% (35 of 40) of patients were treated with an FCR-containing regimen, 4 patients were treated with immunotherapy (n=3) or chemotherapy alone (n=1), and 1 patient was treated with RCVP. Caution is needed in the interpretation of the results, as some of our patients were treated on protocols based on their prognostic factors. Another weakness of our study is the lack of CT imaging for response assessment.”

Nevertheless, despite these inadequacies they are able to conclude that “an 11q22 deletion was associated with favorable clinical outcomes, despite the high incidence of other associated adverse prognostic factors, eg, ZAP-70, CD38, and unmutated IGHV gene status. Regimens that contain FCR appear to overcome the poor prognosis traditionally associated with this genetic abnormality.” The response rate for del 11q patients was 100% and at 3 years the actuarial survival was 91% and the relapse-free survival was 77%.

At the same time as this paper was being presented at ASH the results of the German CLL8 trial were becoming known. This study confirms that adding R to FC moves patients with del 11q from the high risk to the standard risk group for both PFS and overall survival. With two studies telling us this I think we can believe it.

Some patients have written to me suggesting that an alkylating agent is an essential ingredient of treatment for patients with del 11q. To my mind that is not what the data are saying. Intensification of purine analog therapy by adding an alkylating agent certainly seems to improve response and prolongs PFS for patients with del 11q, just as it does for other categories of patient with different or no chromosomal abnormalities (except for those with del 17p), but it does not move patients from poor-risk to standard risk in the way that adding rituximab does. What remains unknown is whether cyclophosphamide (with all the attendant risks of secondary MDS) is an essential component of FCR. The current trial comparing FR with FCR should give us an answer.