Intratumor heterogeneity drives need for multiple biopsies

Major finding: Significant intratumor heterogeneity occurred in the genetic-profile results from three individual biopsies for 27% of primary breast cancers tested.

Data source: Single-center study of 26 patients with either early- or advanced-stage breast cancer.

Disclosures: Dr. Reis-Fiho, Dr. Lau, Dr. Jarzab, and Dr. Di Leo had no disclosures.



BRUSSELS – When biopsying a primary breast cancer, once is probably not enough based on accumulating evidence of primary-tumor heterogeneity, but despite this evidence, taking multiple biopsies has not yet become routine practice.

"Intra-tumor heterogeneity is one of the causes of variation during genetic profiling" of tumors, Dr. Jorge S. Reis-Fiho said. "Taking multiple biopsies from a primary tumor and putting them all together may mitigate the effect of intra-tumor heterogeneity," but taking multiple biopsies has not yet become standard of care, he said at the conference, which was sponsored by the European Society for Medical Oncology.

Mitchel L. Zoler/IMNG Medical Media

Dr. Jorge S. Reis-Fiho

"I think we should start to seriously consider taking more than one biopsy. Quite a few places already do it, and our radiologists will routinely take two or three biopsies," said Dr. Fiho, a surgical pathologist at Memorial Sloan-Kettering Cancer Center in New York. "With only a single, small biopsy of the primary tumor we may not have sufficient information to predict what the metastasis will respond to," he said in an interview.

Results from two studies reported at the meeting added to the evidence favoring multiple biopsies as a way to more precisely characterize primary tumors and target the best therapy.

In one study, Dr. Michal Jarzab and his associates took three core biopsies from the primary tumors of 26 patients and assessed them genetically for estrogen receptor, progesterone receptor, and HER2-receptor mutations. Their sample included some patients with early-stage disease and others with advanced-stage cancer. The results of the three individual biopsy tests showed significant heterogeneity in seven of the 26 tumors (27%), reported Dr. Jarzab, a researcher at the Maria Sklodowska-Curie Cancer Center and Institute of Oncology in Gliwice, Poland.

The intratumor heterogeneity in these seven patients "could negatively impact genomic assessment if done with one specimen," he said.

"If other studies report similar results" on intratumor heterogeneity, "then it could become common practice to evaluate prognostic or predictive breast cancer markers from more than one primary tumour area," Dr. Angelo Di Leo, chairman of oncology at Prato (Italy) Hospital, said in a written statement.

The second study looked at the impact of assessing one, two, or three biopsies on the variance in results from four different single- or multiple-genetic tests in biopsies collected from 51 breast cancer patients. The amount of variance depended on the gene or genes tested, but, overall, using three biopsy specimens rather than one substantially reduced the variance, reported Dr. Rosanna Lau, a researcher at the University of Texas M.D. Anderson Cancer Center in Houston.

"Pooling biopsies from a single tumor reduced outlier results and reduced variance more generally for certain genes," Dr. Lau said. "Pooling biopsies is usually preferable to a single biopsy." The results showed that "we can get a comprehensive picture of the genes being expressed in the tumor by sampling multiple areas of the tumor and pooling the samples together. This increases the precision of the assay and allows us to make more reliable predictions related to the disease," she said in a written statement.

Her study also looked at the scope of analytic variance – variance caused by technical issues – in the biopsies taken from a 20-patient subgroup and found that the extent of analytic variance also varied by gene and was generally of the same magnitude as intratumor heterogeneity. The analytic variance wasn’t helped by pooling multiple biopsies, but data-processing solutions such as normalizing and scaling reduced this part of the variance, Dr. Lau said.

Dr. Reis-Fiho, Dr. Lau, Dr. Jarzab, and Dr. Di Leo had no disclosures.

Next Article:

   Comments ()