Recurrent ductal carcinoma in situ tumors often genetically distinct from primary lesion
Click Here to Manage Email Alerts
A substantial portion of ductal carcinoma in situ recurrences did not appear genetically related to their primary lesions, according to study findings presented at American Association for Cancer Research Annual Meeting.
De novo tumors with distinct genetic alterations accounted for nearly two in 10 recurrences, the detailed molecular analysis showed.
Results call into question whether a tumor-intrinsic biomarker alone could sufficiently predict recurrences among patients with ductal carcinoma in situ, Tanjina Kader, PhD, postdoctoral researcher at Peter MacCallum Cancer Centre in Australia, and colleagues concluded.
“One of the key findings suggests that we should test the relatedness in the clinic once patients walk in the door with a recurrent tumor,” Kader told Healio. “These patients are highly susceptible to developing new tumors; therefore, preventive therapies should be considered — for example, bilateral mastectomy or endocrine therapy, or even genetic counseling.”
Background
Up to one-quarter of patients with DCIS develop recurrence, and half of recurrences are in the form of invasive breast cancer. Standard treatment consists of surgery and/or radiotherapy.
It had been assumed that all recurrences are related to the primary DCIS lesion and that identification of a biomarker to predict which cases are most likely to recur could allow for treatment de-escalation for those at lower risk for recurrence.
However, prior research had not confirmed that all secondary ipsilateral tumors are related to the original tumor.
Results of a prior study from their group, published in 2015 in Modern Pathology, suggested two of eight evaluated cases were nonclonal.
“[That] study was based on very small sample size and only investigated copy number changes,” Kader said. “Therefore, we have assembled a larger cohort with a high-depth sequencing so that we can be confident about our findings.”
Methodology
Researchers used a large recurrence cohort to evaluate the clonal relatedness of primary-recurrence tumor pairs.
Investigators microdissected and extracted DNA from 67 pairs of primary DCIS and recurrences, half of which were ipsilateral breast cancer.
They used targeted sequencing or low-coverage whole-genome sequencing to analyze 23 pairs and whole-exome sequencing to analyze 44 pairs.
They also analyzed 32 nonrecurrent DCIS cases treated with wide local excision, all of which had a minimum 7 years of follow-up.
Kader and colleagues used several statistical approaches to assess clonal relatedness through copy number alterations and mutations. These included the clonality package, manual breakpoint inspection and clonality indexes, and, more importantly, phylogenetic analysis by MEDICC2. This allowed them to determine if primary DCIS and recurrent tumors shared cancer-driving genetic events, confirming derivation from a common ancestor cell.
Kader compared the approach to “building a family tree” of tumors based on genetic events.
“We can estimate how distantly related the primary DCIS and recurrence pairs are to each other and to a potential original ancestor cell, and we can track how the tumor pairs have evolved over time,” Kader said in a press release.
Key findings
Researchers characterized the majority of recurrent cases as clonal and sharing the same genetic events as the original tumor.
However, they characterized 18% of recurrent tumors as new primary lesions that developed de novo.
“The previous study from our lab suggested 25%; however, that study was limited to only one type of analysis — chromosomal changes,” Kader told Healio. “I expected that the number of nonclonal cases would go down once we added mutation data. Therefore, I am not surprised to see nonclonal tumors. I was expecting the number would be between 10% and 15%.”
The findings can guide how patients are managed in the clinic, Kader said.
“The occurrence of a new primary lesion in the same patient suggests a high-risk breast environment in which new tumors may develop over the years,” she said in the release. “Therefore, such patients would be candidates for [preventive] breast removal surgery even if the tumor is small, and they might be referred to genetic testing to ascertain whether they have any genetic predisposition.”
Results showed no significant differences in clonal relatedness regardless of whether recurrences were DCIS or invasive breast cancer.
In addition, clonal primary DCIS exhibited a significantly higher number of TP53 mutations than nonclonal primaries and nonrecurrent cases (P < .001), as well as a higher level of copy number events overall.
Researchers frequently detected specific variations in TP53 genes in recurrences related to the primary lesion; however, these gene variations did not occur frequently among primary DCIS cases that did not recur or those that had nonclonal recurrence.
Kader highlighted another finding she described as intriguing.
“We found that the genetic profiles of nonclonal tumors are very similar to nonrecurrent tumors,” Kader told Healio. “We still do not understand why this is so, and it emphasizes that we need to know more about the biology of both nonclonal and clonal tumors, as well as nonrecurrent DCIS.”
Kader and colleagues acknowledged study limitations. For example, they based the phylogenetic analysis only on chromosomal changes, as a full phylogenetic analysis based on all tumor-specific mutations could not be performed due to the lack of baseline genetic information from normal cells.
However, researchers emphasized they could detect the most relevant breast cancer driver mutations, such as those in the PIK3CA or TP53 genes.
Implications
A determination of whether the recurrent lesion is related to the original tumor should guide treatment decisions for patients who have apparent recurrence of DCIS, Kader and colleagues concluded. That’s because patients who develop a second independent DCIS or cancer are more likely to develop subsequent tumors.
Additional studies in this area are needed, Kader said.
“[Because] we don’t know how and why these independent tumors are being developed for a very small subset of patients, we should investigate tumor microenvironment and immune and stromal cell cross-talk,” Kader told Healio. “We should try to understand the biology of DCIS a lot more than we do now.”
Tracking cancer evolution is challenging, Kader said. Sequencing technologies — such as targeted sequencing panels or low-coverage genome sequencing — only allow total copy numbers to be analyzed due to their lower depth.
“This indicates we will miss out on allele-specific copy number changes,” Kader told Healio. “The available statistical tools for clonality analysis utilizing these sequencing methods also do not follow the evolutionary concept. As a result, you may overcall clonality (ie, almost all cases will be clonal).”
“Therefore, phylogenetic analysis such as MEDICC/MEDICC2/Pyclone to date is the most accurate and commonly used analysis that can help track tumor development,” Kader added. “I think it is essential to keep this in mind — sequencing depth and analysis — before utilizing these findings either in the clinic or for further research.”
Perspective
Back to Top
Larry Norton, MD
This paper adds to the expanding literature supporting the key role for evolutionary analysis in understanding cancer and in providing clinically-useful insights.
Because cancers in an individual all have their origin in normal cells, they all have the same fundamental root. Once the nascent molecular changes occur, a process of subsequent changes leads to the full clinical genotypic/phenotypic presentation, which continues to evolve. The question with multiple occurrences of cancer is: Do all share those nascent changes or can some arise new from the original germline cells, as is demonstrated in a significant minority of cases here?
In clinical oncology, we often are obligated to manage disease as if the biology of the nascent lesions is maintained in further episodes, but the reality of de novo lesions changes the paradigm regarding prognostication and even prediction of response to therapy. It also calls into question the concept of “normality,” because perhaps the so-called normal cells preceding the nascent event might be predisposed to carcinogenic transformation — and hence not “normal” — opening opportunities for targeted prevention.
Cancer is a moving target and, hence, cannot be captured in its full complexity with the analysis of a single specimen at one point in time. We may anticipate that both the technology for time-sequence data acquisition (eg, needle biopsies or cell-free DNA) and data analysis (more sophisticated mathematics) will advance, and that this will have a direct and important impact on patient care.
Collapse
Kader T, et al. Abstract 43. Presented at: American Association for Cancer Research Annual Meeting; April 8-13, 2022; New Orleans.
Read more about
Click Here to Manage Email Alerts