We have done a patient level meta-analysis regarding a new signature called POLAR and POLAR can identify patients benefitting from radiotherapy. POLAR was developed within the SweBCG 91 RT trial, a trial where patients after breast-conserving surgery were randomised to radiotherapy or no radiotherapy to the breast. We collected all the blocks in that trial and we also have long-term follow-up and in the training part of that cohort and we did define the POLAR signature. Now we have made a patient level meta-analysis including three different trials in the meta-analysis: the validation part of the SweBCG 91 RT and the Princess Margaret trial in Toronto and also the Scottish conservation RT trials. So we have tested the performance of the POLAR signature in all these three studies.

The methodology was that the objective was to look at the performance of the POLAR signature to prognosticate local regional failures and also to examine the performance of the POLAR signature to predict RT benefits. We combined all these three trials and we could see that in the POLAR high patients there was a great benefit of radiation and in POLAR low it was no benefit. In the POLAR low there was in the no RT arm a very low risk for local regional recurrences. So POLAR both prognosticated the risk for local regional recurrences in the group without RT and can also predict the benefit from radiation.

It is an interesting finding that we have found a genomic classifier that can predict RT benefit. This classifier must be further validated in contemporary new studies to go into the clinic but, as far as we know, this is the first predictive classifier regarding RT benefit.

These results give the possibilities in the future to omit radiation for many of the patients with a low POLAR score. The test must be further validated and streamlined before going into clinical practice but we have got a possibility here to select patients for radiotherapy and also to safely omit part of the patients for breast radiotherapy.