The 72-gene classifier (72-GC) was developed for recurrence-risk prediction for patients with estrogen receptor–positive and node-negative breast cancer. 72-GC could differentiate the high-risk from the low-risk patients with a high statistical significance, and is considered to be applicable to formalin-fixed, paraffin-embedded (FFPE) tumor tissues because the results of 72-GC on fresh-frozen tissues and FFPE tissues showed a high concordance.Background:
The 95-gene classifier (95-GC) can classify patients with estrogen receptor (ER)–positive and node-negative breast cancer into those with low and high risk of relapse with an accuracy similar to that of 21-GC (Oncotype DX). Because 95-GC uses RNA from fresh-frozen (FF) tumor tissues, we herein attempted to develop a gene classifier that is applicable to RNA from formalin-fixed paraffin-embedded (FFPE) tumor tissues.Patients and Methods:
25 paired FF and FFPE tumor tissues were subjected to DNA microarray for gene-expression analysis. Of the 95 probes included in the 95-GC, 72 were selected for construction of the gene classifier for FFPE tumor tissues, because the gene expression detected by these 72 probes was well preserved in the FFPE tumor tissues.Results:
The 72-GC was constructed with these 72 probes for the training set comprising 549 FF tumor tissues and validated with 434 FF tumor tissues (relapse-free survival at 10 years was 91% for the low-risk and 74% for the high-risk group (P = 3.74 × 10−7). The predictive capability of 72-GC for prognosis was found to be comparable to that of 95-GC. The 25 paired FF and FFPE tumor tissues from each of 25 patients were classified into the same risk group by 72-GC for 23 patients (92% concordance). 72-GC using the FFPE tumor tissues showed that the prognosis for the low-risk group was significantly (P = .007) better than for the high-risk group.Conclusion:
72-GC is comparable to 95-GC in terms of accuracy of prognosis prediction, and may be effective for FFPE tumor tissues.