Precision Medicine Needs 'Precision Genomics'

Pam Harrison

April 15, 2015

When tumor-only gene sequencing is used to identify potential genetic alterations, nearly half of all patients are at risk of receiving treatment that is inappropriate for their tumor type, new research shows.

"Our analyses have highlighted the limitations of current tumor-only sequencing approaches and the benefit of tumor-normal analyses," said principal investigator Victor Velculescu, MD, PhD, professor of oncology and pathology at Johns Hopkins University School of Medicine in Baltimore, who is a cofounder of Personal Genome Diagnostics.

"We all want precision medicine, but one conclusion from these analyses is that we cannot have precision medicine without precision genomics," he explained.

"Inaccurate genetic information can have substantial consequences, ranging from serious side effects related to inappropriate therapies, to a lack of useful targeted therapies, to increased cost of patient care because of misguided medicines," he said at a press briefing.

"Using accurate versions of these sequencing tests will be essential to make appropriate decisions and to better understand the effect of new therapies," he added.

The study results were published online April 15 in Science Translational Medicine.

Concept of Genotyping Oversold?

These findings are important for several reasons, said George Demetri, MD, senior vice president for experimental therapeutics at the Dana-Farber Cancer Institute in Boston.

"I think it'll help physicians and scientists recognize the complexity of the fact that there's not one type of genotyping. If you only do one type of genotyping and think you understand cancer, you're fooling yourself," he told Medscape Medical News.

We cannot have precision medicine without precision genomics.

He said he believes that the medical community has over-represented the importance of tumor sequencing to the public. In fact, he noted, results can be anywhere from promissory to not helpful or even dangerous, which is what this research effectively conveys.

"Part of the problem with cancer is that it does not have high-fidelity copying mechanisms," Dr Demetri explained.

"Cancer has lots of mistakes. What's just random noise in those mistakes and what's a true pathogenic causative or contributing factor in that cancer is not clear for the vast majority of human cancers," he said.

Indeed, with the exception of certain mutations — such as EGFR-mutant non-small cell lung cancer and KIT-mutant gastrointestinal sarcoma (GIST) — the clinical relevance of alterations in most cancers, even when identified, is still not known.

"It's just like a ding in the finish of a car," Dr Demetri said.

"Is there a ding in the paint, an abnormality?" he asked. "Yes there is. Is that going to hurt the way the car runs? Absolutely not. So yes, the abnormality's there, but what does it mean?"

"I think that's the problem with all this genetic testing," he noted.

Comparing the Genomes of Tumor and Normal Tissue

In their study, Dr Velculescu and colleagues compared tumor-tissue genomes with normal-tissue genomes in a cohort of 815 patients with 15 different tumor types.

In tumor-only analyses, after common germline and inherited variants were excluded, 382 candidate mutations per tumor were identified.

When the full germline genome of a patient was compared with the corresponding tumor genome, an average of 241 of the 382 (63%) alterations in actionable genes were part of the normal inherited genetic variation, and were not specific to the tumor.

Actionable genes are targets or modifiers for the new generation of precision cancer therapies.

"Even after filtering the data with well-known germline alterations, approximately two-thirds of the remaining mutations were part of the patient's normal inherited genetic variation," Dr Velculescu reported. "In other words, they were false positive."

For genes that are likely to be actionable (or potential targets for precision-based oncology therapies), about one-third of the changes on tumor-only analyses were false positive.

"Because each patient, on average, had multiple such alterations, these false-positive changes affected roughly one in two patients analyzed," Dr Velculescu said.

"In other words, one in two patients receiving tumor-only panel gene sequencing is potentially at risk of receiving a treatment that may be inappropriate," he noted.

However, the study confirmed that cancer genome analyses are likely to be useful for physicians and patients because about three-quarters of the patients had alterations in genes associated with current therapies or clinical trials and, as such, could be potential candidates for such therapies.

Analyses of normal or germline DNA revealed that there were alterations in cancer-predisposing genes in a small but significant proportion of patients who had no known family history of cancer.

This points to the possibility that genomic sequencing could be a simple way to identify at-risk individuals and affected family members.

More Accurate Diagnostic Yield

Without comparison to normal samples, DNA analysis of the patient's tumor, as shown in this study, frequently misidentifies genetic changes in the tumor as targetable when, in reality, the genetic changes are present throughout the patient's normal cells, said Angela Colmone, PhD, associate editor of Science Translational Medicine.

At the same time, the tumor-only analyses "failed to identify mutations that were not present in normal cells that were contributing to tumor progression," she said during the press briefing.

"These finding suggest that sequencing paratumor and normal samples would yield more accurate diagnosis and treatment plans than the standard methods," Dr Colmone added.

However, the cancer genomics company Foundation Medicine disagrees with this conclusion. The company offers genotyping of tumor tissue and has sophisticated algorithms based on 35,000 clinical cases that it uses to identify driver mutations in tumor tissue that could be targets for therapy. "With our sophisticated approach, we do not need to compare tumor tissue with normal tissue," Vince Miller, MD, chief medical officer of Foundation Medicine, told Medscape Medical News.

A blog post on the company website states that it "strongly disagrees that matched tumor and normal DNA from the same patient significantly improves the accuracy of identification of actionable mutations enabling better targeting for potential treatments in the vast majority of patients with cancer."

"We believe the only genomic alterations that should be considered for targeted treatment decisions are alterations with strong evidence that they are important in transforming a normal cell into cancer or maintaining a cancer cell's viability once transformed, regardless of whether they are somatically acquired or are germline alterations; examples include activating point mutations (e.g., EGFR L858R), amplification of known oncogenes (e.g., HER2), loss of function alterations in tumor suppressor genes (e.g., PTEN), and gene fusions (e.g., ALK)."

"Foundation Medicine's comprehensive genomic profiling assays are designed and validated [Nat Biotechnol. 2013;31:1023-1031] to accurately detect these types of genomic change in real-world clinical specimens. Through our proprietary sample preparation and computational biology algorithms, we identify the full complement of genomic alterations in all cancer-related genes, and we then remove the variants that should not be considered for treatment decisions. Such variants, whether somatically acquired or germline, are 'variants of unknown significance' (VUS), and we believe they should not be used for treatment decisions because the overwhelming majority will prove to be nonpathogenic and could indeed expose the patient to costly and ineffective therapies," the company adds.

"In an ideal world, routinely obtaining a matched normal would allow more rapid discernment of the minority of VUS that may be pathogenic, which could be valuable for research efforts, but not for evidence-based treatment decisions," the company points out. "In addition, a matched normal can more quickly discriminate germline pathogenic mutations, which were identified in ~3% of patients in the research study, but reporting such findings to physicians (and patients) is still a matter of intense debate and, if reported, should come with appropriate genetic counseling."

The study was funded by the Dr Miriam and Sheldon G. Adelson Medical Research Foundation, the Commonwealth Foundation, the American Association for Cancer Research Stand Up to Cancer dream team translational cancer research grant, the John G. Ballenger Trust, a FasterCures Research Acceleration Award, Swim Across American, and the National Institutes of Health. Dr Velculescu and one of his coauthors are cofounders of Personal Genome Diagnostics, are members of its scientific advisory board and board of directors, and own stock in the company. Dr Demetri is on the scientific advisory board of Caris Life Sciences (and receives minor equity grants as part of that) and several other scientific advisory boards of biopharma companies interested in molecular therapeutics, and serves on the board of directors of Blueprint Medicines.

Sci Transl Med. Published online April 15, 2015. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.