29P - Tumour response to neoadjuvant chemotherapy in breast cancer: Routine pathologic markers improve the predictive power of a cell-loss metric based on release of thymidine kinase 1 into blood

Abstract

BackgroundEarly prediction of tumor response to therapy is essential for individualized treatment and sparing non-responders needless harm. A relationship has recently been found between pathologic response in breast cancer (BC) and a measure of cell loss based on serum levels of thymidine kinase 1 (sTK1), a macromolecule released when proliferating tumor cells are disrupted, and tumor volume. Objective: To establish whether the predictive power of this cell-loss metric can be further improved by baseline tumor characteristics. MethodsFifty-eight women with localized BC received neoadjuvant epirubicin/docetaxel in 6 cycles, supplemented with bevacizumab in cycle 3-6. The cell-loss metric, defined as the ratio between sTK1 (ng/ml) and tumor volume (cm3), was obtained prior to and 48h after cycle 2. The predictive value of this metric, and the improvement by adding routine baseline markers, was evaluated using pathologic response as endpoint (16 complete responses (pCR); 42 remaining tumors). ResultsCompared to baseline (median = 0.285 ng / ml), sTk1 increased 2-fold before cycle 2 and 3-fold 48h after cycle 2 while tumor volume (baseline 105 cm3) decreased by 70%. The cell loss method increased from 0.0032 units (IQR 0.0015-0.0099) to 0.0166 (0.0064-0.0423) and 0.0232 (0.0095-0.053), respectively, showing a strong association with pathological response (p = 0.002). Combination with histological markers, and especially the progesterone receptor, significantly improved the prediction of pCR, achieving positive and negative predictive values of 81% and 93%. ConclusionsThe usefulness of tumor markers in blood can be increased by combining them with other tumor properties. Thus, in neoadjuvant treatment of BC the cell-loss metric, which relates sTK1 to tumor volume, has been noticed as a predictor of pathologic response. Here we found that adding certain established tumor markers significantly improved the predictive power of the metric. Early prediction of tumor response makes the cell-loss metric potentially useful in personalized oncology and in the evaluation in new therapeutic modalities. Clinical trial identificationPROMIX: NCT00957125. Legal entity responsible for the studyThomas Hatschek, MD, PhD, Karolinska University Hospital. FundingHas not received any funding. DisclosureB. Tribukait: Shareholder / Stockholder / Stock options, minor stock owner: AroCell Ab.