乌鸦传媒

NEW ORLEANS -- A search for biomarkers to predict recurrence of ductal carcinoma in situ (DCIS) instead showed that a substantial fraction of recurrences had genetically distinct features from the primary DCIS, questioning the potential value of recurrence markers.

Almost 20% of ipsilateral recurrent DCIS had a different genetic makeup from the primary tumor. The findings dispelled the widely held assumption that ipsilateral recurrence, whether DCIS or invasive breast cancer, shared clonality with the original tumor, said Tanjina Kader, PhD, of the Peter MacCallum Cancer Center in Melbourne, Australia, during the American Association for Cancer Research (AACR) meeting.

The finding has implications for breast cancer treatment and for research aimed at identifying biomarkers for recurrence, she said. In about a fourth of cases, patients with DCIS have recurrences, which are invasive breast cancer about half the time.

"At the moment, we assume that recurrent tumors are genetically related to the primary DCIS," Kader said during an AACR press briefing. "Therefore, the treatment options for all of these patients are exactly the same. All of the patients with recurrent DCIS will be treated with further surgery or surgery and radiotherapy. But the question remains, do we really know the second tumor is actually related to the primary tumor?"

The findings related to clonality differences came from a study to examine the genetics of primary and recurrent DCIS. The goal was to identify potential biomarkers that might distinguish DCIS with a high risk of recurrence from tumors with a low likelihood of recurrence.

Investigators performed DNA sequencing on 67 paired specimens of DCIS (primary vs recurrent tumor). They compared mutations and changes in copy number in the primary and recurrent tumor and performed phylogenetic analyses to gain insight into the origin of the paired specimens. Paired specimens with shared genetic events indicated common ancestry and clonality, whereas absence of shared events indicated distinct origins for the primary and recurrent tumor.

Additionally, they performed sequencing studies of specimens from 29 cases of nonrecurrent DCIS. Comparing the genetics of recurrent and nonrecurrent DCIS might reveal potential biomarkers to distinguish high-risk from normal-risk DCIS, said Kader.

New Primary Tumors

The comparison of paired specimens showed that 12 of 67 (18%) recurrent tumors did not have shared clonality with the primary DCIS. The likelihood of clonality was not significantly related to the grade of primary DCIS, hormone receptor or HER2 status of the primary tumor, recurrence type (DCIS vs invasive breast cancer), or whether the patient received radiotherapy.

"These [12] tumors are new primaries, and independent tumors suggest there is a high-risk environment or genetic predisposition that make patients highly susceptible to develop new tumors," said Kader. "Therefore, we are proposing that in the future, instead of treating all the recurrent patients the same, we have to find out first whether the [recurrences] are nonclonal."

"Surgery alone is not going to be sufficient for patients who are highly susceptible to develop new tumors," she continued. "We should consider preventive therapies, which could include bilateral mastectomy or endocrine therapy or genetic counseling."

Comparison of the genetics of recurrent and nonrecurrent tumors identified five potential markers to distinguish DCIS with a high likelihood of recurrence from normal-risk DCIS. Recurrent DCIS was associated with changes in chromosomes 5, 11, 17, and 20, as well as TP53 mutation.

However, investigators found that nonclonal recurrent DCIS had a genetic profile similar to nonrecurrent DCIS. Consequently, the "predictive biomarkers" have questionable clinical value.

"Even though we can get excited about these biomarkers, this particular finding [about the genetic making of nonclonal recurrent DCIS] raises a question as to whether we should actually utilize predictive markers for DCIS," said Kader.

During a discussion that followed her presentation, Kader said no information is available on the natural history or response to treatment of the nonclonal recurrences and reiterated that the study is the first to document the different genetic makeup and clonality of recurrent DCIS.

Press briefing moderator Timothy Yap, MD, PhD, of the University of Texas MD Anderson Cancer Center in Houston, asked about the next step of investigation, given the apparent lack of utility for the biomarkers identified during the study.

Kader suggested a need for more studies of the interaction between tumor cells and the tumor microenvironment to determine how tumors may change as a result of the interaction, as well as what molecular and genetic factors are involved.

'On the Right Track'

The study "reinforces my belief that we’re on the right track, that we can find groups of people who are low enough risk that we can afford to watch them and that we can get to them in time to intervene," said Laura Esserman, MD, of the University of California San Francisco. "It reinforces my belief that we have to do a better job of characterizing people fully at the beginning, really understanding all the risk factors and start improving the system."

"Women do not like the [treatment] choices they are being given, and I don't blame them. It's adequate and we can do better," she added.

For more than a decade, Esserman has overseen an active surveillance program for women who have newly diagnosed breast cancer and do not want surgery. Breast MRI represents a key decision-making tool in the program. Early results suggest that a patient who has diffuse enhancement on MRI in the absence of focal lesions probably does not need surgery. A patient with a focal lesion that does not improve is probably a good surgical candidate.

"The people who we think are good candidates for active surveillance had about a 10% chance of getting invasive cancer over the next seven years," said Esserman. "That the same as for people who get surgery and radiation. There's something wrong here. Let's start figuring out how to right-size the interventions and give people better choices. Using the tools and the information that we already have in hand, just start teasing this out."

A multicenter randomized trial will begin later this year with the goal of validating what has been learned thus far from active surveillance studies.

Comment