Two multi-gene tests designed to predict the risk of disease progression and response to chemotherapy in breast cancer produce broadly similar results for high- and low-risk patients, but do not always agree in their predictions for those at intermediate risk, a new analysis shows.
In recent years, several genomic tests have been developed to provide prognostic information for breast cancer. Dr Catherine Kelly from Mater Misericordiae University Hospital, Dublin, Ireland, and colleagues in the USA examined the agreement in prediction results between two multi-gene assays --the widely used Oncotype DX and a newer PAM50 breast cancer intrinsic classifier.
"At the present time we know there are at least four different breast cancer sub-types with clinical relevance --'luminal A', 'luminal B', 'basal-like' and 'Her2-enriched'. Several commercial assays are being introduced to assign clinical sub-type or predict risk of recurrence for breast cancer," Dr Kelly explained.
"An imminent clinical challenge for practicing physicians is to understand how these different assays relate to each other and what to expect when more than one assay is performed on the same cancer. This is the first study to compare two standardized, readily available but conceptually different prognostic risk predictors."
OncotypeDX is intended for use in patients with hormone receptor-positive, lymph node-negative breast cancer to identify women at low risk of breast cancer recurrence who safely can avoid chemotherapy. It measures the activity of 21 genes and generates a recurrence score (RS) which categorizes patients into low, intermediate or high risk groups depending on the risk of distant recurrence.
The PAM50 breast cancer intrinsic classifier measures the expression of 50 genes and stratifies breast cancers into five sub-types; luminal A, luminal B, basal-like, Her2-enriched and normal-like, so the target population is not just patients with ER-positive breast cancers. This test is currently undergoing clinical validation.
The researchers assessed the two tests on 119 breast cancer specimens from patients who the researchers classified as being at 'clinically intermediate' risk for recurrence based on several criteria: median tumor size 1.5cm, all estrogen receptor-positive, Her2 negative, lymph-node negative and most grade II.
"In other words, these patients would likely be a challenge in terms of whether they would benefit from chemotherapy or not," Dr Kelly said. "In these situations, multi-gene assays can provide additional independent prognostic information."
The results showed that all patients with high RS according to the OncotypeDX were classified as 'luminal B' or 'basal-like' by the PAM50 classifier, whereas the majority of low RS cases (83%) were "luminal A" type. Half (51%) of the Intermediate RS cancers were re-categorized as low risk 'luminal A' cancers by the PAM50. All 'luminal A' cancers are either Low (70%) or Intermediate (30%) risk by RS, whereas 'luminal B' cancers are comprised of a mixed risk group by RS including 33% that are high risk by OncotypeDX.
"These results indicate that there is reasonably good agreement between the two methods for high and low prognostic risk assignment," Dr Kelly said. "However, intermediate risk groups by one assay often include cases with discordant risk prediction by the other method. When discordant risk results are obtained it is currently unknown which assay will predict outcome more accurately."
"At the present time it is uncertain whether patients with an intermediate OncotypeDx RS benefit from chemotherapy," Dr Kelly added. "It is interesting that PAM50 re-categorized about half of the patients with intermediate RS to the low risk luminal A category which suggests that these patients may not benefit much from adjuvant chemotherapy due to their already very good prognosis and limited chemotherapy sensitivity."
However, she noted, PAM50 is currently undergoing validation. "We will have to wait for outcome data to see which test is most accurate at predicting outcome particularly."
Commenting on this study, which he was not involved in, Prof Roberto Labianca, from Ospedali Riuniti di Bergamo, Italy, noted that the data with PAM50 are very interesting, because they indicate that in the 'intermediate risk' group (as defined by the well-established method OncotypeDX) it could be possible to exclude additional patients from adjuvant chemotherapy.
"However, a validation of these findings in a larger population is mandatory before they can be translated into clinical practice."