The fact that cancer is heterogeneous has been known for centuries. The fact that some specific types of cancer are microscopically heterogeneous has been known for over a century. Chromosomal and molecular heterogeneity was discovered in the latter half of the 20th century as a result of the technological advances in cellular and molecular biology. Further studies showed that cancer in a given individual may have started as a clone, but mutations are common due to the pressures of therapy or other evolutionary factors that can lead to metastases or therapeutic resistance.
However, a recent study reported by a large international group (primarily in the U.K.) has demonstrated how cancer in any one patient can be far more heterogeneous than I ever imagined (Marco Gerlinger et al: Intratumor Heterogeneity and Branched Evolution Revealed by Multiregion Sequencing: NEJM 2012;366:883-892); also see Dr. Dan Longo's editorial in the same issue.
The authors evaluated multiple tumor-biopsy samples from a variety of primary and metastatic sites before and after cytoreductive nephrectomy in four consecutive patients with metastatic renal-cell carcinoma who were enrolled in a targeted therapeutics clinical trial of everolimus. Biopsy samples were obtained before the initiation of six weeks of treatment with everolimus. After a one-week washout period in which patients did not receive everolimus, a nephrectomy was performed. Everolimus treatment was continued after recovery from surgery until tumor progression. Extremely detailed cytogenetic and genetic analyses allowed them to reconstruct the evolutionary growth of tumors in each of the samples. The key tactic was to take multiple biopsies of the same tumor sample, some only a centimeter apart.
A summary of their findings:
1. Phylogenetic reconstruction revealed branched evolutionary tumor growth, with 63 to 69 percent of all somatic mutations not detectable across every tumor region (italics mine).
2. Intratumor heterogeneity was observed for a mutation within an autoinhibitory domain of the mammalian target of rapamycin (mTOR) kinase, correlating other molecular control points such as other kinase activity.
3. Mutational intratumor heterogeneity was seen for multiple tumor-suppressor genes converging on loss of function with multiple distinct and spatially separated inactivating mutations within a single tumor, suggesting convergent phenotypic evolution.
4. Gene-expression signatures of good and poor prognosis were detected in different regions of the same tumor.
5. Allelic composition and ploidy profiling analysis revealed extensive intratumor heterogeneity, with 26 of 30 tumor samples from four tumors harboring divergent allelic-imbalance profiles.
If this proves to be a general finding for other solid cancers, the effects on our diagnostic confidence would be profound. Our usual diagnostic routine involves a needle biopsy of the accessible tumor with the sample subdivided for various diagnostic tests, including a part frozen for some types of molecular analysis. “Personalized medicine” or “personalized therapy” in this case means the selection of specific therapeutic approaches based on a positive historical response in a patient with the same genomic or cytogenetic profile. This approach depends on identifying an appropriate genomic or cytogenetic profile.
But one must now ask, “Which genomic or cytogenetic profile?” When one is using more general use therapy, such precision is not required and we accept the fact that some tumors will respond and some won't. When one has a single dominant genomic aberration, such as in CML, personalized therapy works very well. But if even a handful of integral genomic determinants are as variable as demonstrated in this study, personalized therapy becomes logarithmically more difficult. In addition, this study found that the heterogeneity problem also applied to favorable vs. unfavorable prognostic features.
As the authors calmly state, “Intratumor heterogeneity can lead to underestimation of the tumor genomics landscape portrayed from single tumor-biopsy samples and may present major challenges to personalized-medicine and biomarker development. Intratumor heterogeneity, associated with heterogeneous protein function, may foster tumor adaptation and therapeutic failure through Darwinian selection.”
Cancer researchers have known about mutational evolution of cancer for some time but at least in my own simple mind, I thought it was a more deliberate process. But the life expectancy of a cancer is no longer than the human host's, so its evolutionary time frame is short; cancer is one of the most resourceful parasites on earth.
What this means for cancer therapy in general and personalized therapy specifically is not yet clear. It depends on whether this is a general phenomenon that applies, more or less, to all cancers. It also depends on the clinical significance of the mutations.
Are they all deadly? I would guess that is unlikely. Are some of the mutants sensitive to conventional therapy? I would guess that is likely. Finally, is the extent of the observed phenomenon limited to the widely metastatic cancer stage of the four participating patients? Would we see such extreme variation in a clinically localized tumor?
What I believe to be certain if this degree of heterogeneity is found in all or many common tumors is th`e following: Clinical trials will become even more difficult or even impossible to complete using modern molecular diagnostics with current trial methods; combination therapy must be considered earlier in the drug development process.
Cancer never ceases to amaze and frustrate, and these findings continue and reinforce that effect.