Relevance of SNP 88 study

Hi Casey4:

Here is a live link to the paper:

Cuzick (2016): "Impact of a Panel of 88 Single Nucleotide Polymorphisms on the Risk of Breast Cancer in High-Risk Women: Results From Two Randomized Tamoxifen Prevention Trials"

http://ascopubs.org/doi/full/10.1200/JCO.2016.69.8944

And here is a link to the accompanying commentary, which may address some of your questions:

Chlebowski (2016): "Improving Breast Cancer Risk Assessment Versus Implementing Breast Cancer Prevention"

http://ascopubs.org/doi/full/10.1200/JCO.2016.70.9386?utm_medium=cpc&utm_campaign=J_Clin_Oncol_TrendMD_0&utm_source=TrendMD

BarredOwl

I agree that the comparison to statins is not a good one. Similar to those seeking even broader-based use of statins, his focus seems to be on statistical risk / benefit of chemo-prevention at the population level, whereas individual patients may take quite a different and more personalized view. Also, as you note, understanding of risks continues to evolve and the risks can fairly be seen to weigh heavier against benefit in the preventive versus therapeutic setting.

While he clearly thinks (with his population focus) that current risk prediction models ought to support greater uptake of chemo-prevention, I don't think he is saying improved prediction is not needed. He seems both surprised and disappointed by what he considers to be the modest improvement achieved in this study:

"In the current analyses, the predicted power of the SNP88 model (interquartile range odds ratio [IQ-OR], 1.37; 95% CI, 1.14 to 1.66) was similar to the TC model (IQ-OR, 1.45; 95% CI, 1.21 to 1.73), and in the combined model, a modest 13% improvement (IQ-OR 1.64; 95% CI, 1.36-1.97) was seen.[1] The limited improvement in the combined analysis is surprising given the lack of correlation between the TC model and the SNP score (P = .7)."

He also notes with disappointment that similar genetic approaches so far have not enhanced risk prediction to the degree it was hoped:

"When the two decades of discovery began[3] the expectation and hope was that genetic analysis would identify new high-penetrance mutations; that integration of an increasing number of high-frequency, low-penetrance mutations would substantially improve breast cancer risk prediction; or that consideration of the function of newly identified mutations associated with breast cancer would suggest potentially modifiable approaches not identified through the nearly half-century of observational study efforts. These expectations have largely not been met. . .

Although the available body of evidence suggests that breast cancer risk prediction is a mature area, there seems to be room for at least some improvement."

I'd say more than some.

BarredOwl

Hi Casey4:

You may be interested in this NCCN guideline entitled "Breast Cancer Risk Reduction", if you have not already seen it. The "algorithms" in the charts up front were recently updated. However, the "Discussion" section (with extensive bibliography) may not be up-to-date in all aspects and is currently under revision (Discussion Section - "Last updated 06/30/15").

NCCN Guideline "Breast Cancer Risk Reduction" (Version 1.2017): https://www.nccn.org/professionals/physician_gls/pdf/breast_risk.pdf

(Available with free registration)

Note: There are separate NCCN guidelines available for "Breast Cancer Screening and Diagnosis" and "Genetic/Familial High-Risk Assessment: Breast and Ovarian".

Guidelines represent snapshots in time and may not be totally current in all aspects. In addition, they address what is provided in the general case, and there may be appropriate individual exceptions.

As such guidelines are technical documents geared at professionals, if they influence decision-making in any way, patients should be sure to discuss their thinking with their doctor to ensure accurate understanding and applicability to their particular case.

BarredOwl