Type of systemic therapy | Type of study | Timing of early 18F-FDG PET | pCR rate (%) in breast and lymph nodes | Best predictor of pCR | Confirmation in multivariate analysis | Study |
Epirubicin-cyclophosphamide (4 cycles and then docetaxel–trastuzumab) | Single center (n = 30) | After 2 cycles | 53 | Absolute residual SUVmax at PET 2 of ≤3 (breast or axilla) (AUC, 0.91) | Not done | Groheux et al. (46) (2013) |
Anthracycline (followed by taxane) or taxane + trastuzumab; 6–8 cycles | Single center (n = 14) | After 2 cycles | 29 | 50% change in SUVmax threshold with 100% NPV | Not done | Zucchini et al. (47) (2012) |
Trastuzumab, lapatinib, or their combination (later combined with paclitaxel) | Multicenter (n = 66) | After 2 and 6 wk of biologic therapy | 35 | Metabolic R (EORTC criteria) had doubling of pCR rate | No (hormone receptor status was strongest predictor) | Gebhart et al. (51) (2013) |
Docetaxel + trastuzumab | Single center (n = 57) | After 1 cycle | 44 | SUVmax at PET 2 of <2.1, with required SUVmax at PET 1 of ≥3.7 | Yes | Humbert et al. (49) (2014) |
Paclitaxel + carboplatin + trastuzumab | Single center (n = 25) | After 6–8 wk | 60* | 18F-FDG PET results not correlated with pCR | Not done | Koolen et al. (45) (2013) |
Trastuzumab + docetaxel, with or without carboplatin | Single center (n = 37) | After 1 cycle | 38 | ≥75% change in SUVmax; pCR accuracy, 76% | Yes | Humbert et al. (48) (2012) |
Trastuzumab + docetaxel in case of poor metabolic response, with or without bevacizumab | Multicenter (n = 142) | After 2 cycles | 54 in PET R; 24 in PET NR; 44 in PET NR receiving bevacizumab | ≥70% change in SUVmax | Not done | Coudert et al. (50) (2014) |
↵* Evaluated in breast only.
AUC = area under the curve; NPV = negative predictive value; R = responders; EORTC = European Organization for Research and Treatment of Cancer; NR = nonresponders.