|
Abstract |
It is clear that myeloma (MM) is a heterogeneous disorder and that several well defined variants exist.1 The study of these genetic aberrations not only furthers our understanding of disease pathogenesis, but has allowed for their application in the clinic. This information improves our ability to prognosticate and predict likelihood of clinical benefit with specific therapeutics. Multiple classification schemes exist for MM, yet all describe one and the same thing; a heterogeneous condition best sub-classified by genetic/cytogenetic markers, with consequent variability of phenotype.1 However, as information has emerged the classification systems for the disease can be divide into three main categories; 1) Biology based classification: Mostly by identifying unique genetic subsets characterized by either chromosome translocations/structural aberrations, deletions, of whole chromosome changes, unlikely to change, usually with clinical utility, but not always. 2) Prognostic classification: A classification derived from understanding the natural course of the disease and independent of therapy administered, can possible change depending on therapy, but likely influential in outcome with most treatments. 3) Predictive classifications: A classification capable of providing estimates of clinical benefits associated with specific therapeutics, highly relevant but dynamic and subject to change.
Biology classification (enduring)
Primary genetic events. A biology classification system is one that will (should) not change with time. For instance even if MM becomes curable we will still have 15% of MM be t(11;14). This classification needs to be derived purely from biology and irrespective of clinical implications.1,2 While biologic classifiers that attain clinical significance are more convincing, important genetic lesions do not necessarily have to discriminate outcomes. For instance MM with t(11;14) is clearly a unique biologic entity, even when it is not significantly different with regards to prognosis.3-5 The two major subtypes include hyperdiploid MM and MM with IgH translocations involving multiple partners.
Secondary genetic events. A number of other genetic aberrations are thought to participate in disease pathogenesis acting as progression events. Some of them may act as catalyst factors for clonal expansion (e.g. chromosome 13 abnormalities; see abstract by T Henry et al.) while others will enhance cell survival, allow extramedullary disease and increase likelihood of therapy resistance (e.g. p53 deletions).6 Lastly others are in all likelihood contributors for clonal expansion, but with specific roles still being elucidated. (e.g. chromosome 1 abnormalities, c-myc, non-canonical NFKB activation). Regarding to chromosome 1, it should be noted that 1q gain and 1p loss are so closely related that it is hard to provide differentiation.7,8
Other categories. The fundamental observation by Bergsagel and Kuehl that MM is characterized by upregulation of any one of the three cyclin D genes,2 resulted in the observation that some patients lacking the aforementioned primary translocations will be characterized, at the gene expression level, by augmented expression of CCND2 (with or without concurrent CCND1 expression). It is possible one unifying genetic mechanism explains CCND2 elevation, but also that multiple abnormalities ultimately lead to CCND2 increased expression. Another subgroup identified by this classification is the none group. This group is characterized by minimal expression of RB1 (50%) complimenting a model consistent with hyperactivity of cyclin D or down regulation of checkpoints (RB1) mediating G1/S transition.
Prognostic classification
While we believe the major prognostic variation of MM is dictated by primary genetic categories, secondary changes can also have a profound influence in outcome by providing clonal survival/proliferation advantages. Some of the basic genetic categories have not resulted (yet) in specific clinical outcome difference, yet define unique subtypes (e.g. t(11;14)).9,10 The clinical consequences of secondary genetic changes tend not to be related to therapy administered. One possible way to define prognostic markers is that they associate with baseline features of aggressiveness (pathobiology) and they should exert their influence if patients are not treated (natural history). Since most patients will ultimately be treated, they usually will exert their influence in segregating outcome, even with therapy improvements (but may not). One could propose that for a prognostic marker to be considered meaningful minimal requirement could be; a) that the marker has shown effects when applied to patients treated with two different modalities and, b) that it has been reported in at least two statistically empowered series of patients. For instance we know for that overall patients with t(4;14) fare worse,3-5,11 even when some selected cases might fare better. Needless to say the prognostic ability of this same markers is currently challenged by the introduction of novel therapeutics(12). The same is true for t(14;16) with at least two series showing that (ECOG and UAMS).3,11 Minimal data is available regarding the other MAF variants.
Other models. Other prognostic markers exert effects across the major biology subtypes of MM. Some are well established, including the strongest cytogenetic prognostic marker; –17p1.3-5,13 Another model that has been proposed and tested in at least two series is a 70 gene signature derived from the GEP by the UAMS.14 This last model discriminates with unprecedented ability high-risk disease. While not likely to be immediately in the clinic this is an excellent benchmark with which clinical trails can be compared for inclusion of high-risk individuals. Other markers could include proliferation index by GEP, centrosome index by GEP and CTA.15 Most of these models are best considered working categories.
Predictive classification
Predictive classifications lag behind in supporting data and clinical applicability. The issue of predicting response to therapy is not as important in MM (now most patients respond), as it is to predict sustainability and quality responses. Arguably by combining these parameters (and assuming reasonable toxicity) a predictive factor will measure clinical utility. A predictive classification is more complex in that it incorporates treatment as a variable. Again, using likelihood of response alone will not be enough. For instance all preliminary data suggest that bortezomib works just as well for IgH translocated or chromosome 13 deleted cases (not necessarily better).12,16 However, this has not as of yet translated into clinical benefit. The problem with t(4;14) has not been lower responses, but rather early relapse and refractory relapse.3,4,17 The long-term outlook of t(4;14) patients treated with bortezomib is needed. Furthermore an accurate predictive model may dictate more the order (sequence) of treatments employed, since ultimately, and barring the development of curative therapy, many MM patients will be treated with all agents, just in different order. One example of prediction is our recent observation of higher likelihood of response (90%) to bortezomib amongst patients with hyperactive non-canonical NFKB pathway (using low level of expression of TRAF3 as surrogate), as opposed to those with normal level of TRAF3 expression (30%). Based on some of these considerations, despite data shortcomings, and because of the shorter benefit of autologous stem cell transplant for high-risk disease, our group has incorporated bortezomib early on in the treatment course of patients with high-risk disease (recommendations available at mSMART.org). RF is a Clinical Investigator of the Damon Runyon Cancer Research Fund.
This work is supported by the International Waldenström Macroglobulinemia Foundation, and grants R01 CA83724-01, SPORE P50 CA100707-01and P01 CA62242 from the National Cancer Institute, and the Fund to Cure Myeloma
References
1. R. Fonseca, et al. Cancer Res 64, 1546 (Feb 15, 2004).
2. P. L. Bergsagel, et al. Blood 106, 296 (Jul 1, 2005).
3. R. Fonseca, et al. Blood 101, 4569 (June 1, 2003, 2003).
4. M.A. Gertz, et al. Blood 106, 2837 (Oct 15, 2005).
5. H. Avet-Loiseau, et al. Blood (Jan 5, 2007).
6. W.J. Chng, et al. Leukemia 21, 582 (Mar, 2007).
7. J.R. Sawyer, G. Tricot, S. Mattox, S. Jagannath, B. Barlogie. Blood 91, 1732 (1998).
8. C. Debes-Marun, et al. Leukemia 17, 427 (February, 2003).
9. R. Garand, et al. Leukemia (2003).
10. R. Fonseca, et al. Blood 99, 3735 (2002).
11. F. Zhan, et al. Blood 108, 2020 (Sep 15, 2006).
12. M.V. Mateos, et al. Blood (Jun 13, 2006).
13. J. Drach, et al. Blood 92, 802 (1998).
14. J.D. Shaughnessy, Jr. et al. Blood 109, 2276 (Mar 15, 2007).
15. W.J. Chng, et al. Blood 107, 3669 (May 1, 2006).
16. S. Jagannath, et al. Leukemia 21, 151 (Jan, 2007).
17. W. Jaksic, et al. J Clin Oncol 23, 7069 (Oct 1, 2005). |
