Never forget.

We know from the EBCTCG meta-analysis that estrogen receptor(+) breast cancers have an insidiously high rate of distant failure at 20 years. Could that be driven by underlying molecular subtypes? The Swedish STO-3 trial is a really old adjuvant tamoxifen randomized trial from which over 700 tumor samples were maintained. In this secondary analysis, those samples were molecularly sub-typed using the PAM50 gene expression classifier. Among the ER(+) tumors, there was a roughly 3:1 split between luminal A and luminal B subtypes. At 25(!) years, the long-term patterns of failure and benefit with tamoxifen were really interesting to see. Luminal A appears to carry a low, but persistent, risk of distant failure that lingers on greater than 15 years after treatment. As a result, tamoxifen confers a long-term benefit for luminal A disease. In contrast, luminal B appears to have a much higher early risk of distant failure that confers a very short-term benefit with tamoxifen. However, both the absolute risk of metastasis and absolute benefit of tamoxifen converge with untreated patients at around 10 years. As one might expect, luminal B disease typically has intermediate to high Oncotype DX recurrence scores. TBL: Long-term patterns of distant failure and benefit from tamoxifen appear considerably different between luminal A and luminal B ER(+) breast cancer. | Yu, JAMA Oncol 2019