Comparison of [68Ga]Ga-HER2 Affibody PET/CT and MRI for the early prediction of response to targeted neoadjuvant therapy in breast cancer.

OBJECTIVE

To evaluate early ΔSUVmax% for predicting pathological complete response (pCR) in HER2-positive breast cancer patients undergoing neoadjuvant therapy (NAT), to build a combined [68Ga]Ga-HER2 Affibody PET/CT model, and to compare its performance with magnetic resonance imaging (MRI).

METHODS

45 patients underwent both [68Ga]Ga-HER2 Affibody PET/CT and MRI at baseline and after two cycles of NAT. Univariate logistic regression identified parameters associated with pCR. After assessing multicollinearity, combined PET (ΔSUVmax%, TBR2) and MRI (ADC2,ΔADC%) models were constructed. The model performance was evaluated using leave-one-out cross-validation (LOOCV) method. Receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis were performed.

RESULTS

27 of the 45 enrolled patients achieved pCR (60%). After treatment, the pCR group showed significantly lower PET/CT parameters, a significantly higher ΔSUVmax%, and significantly higher MRI-derived ADC and ΔADC% values compared to the non-pCR group. The combined PET model showed a LOOCV AUC of 0.887 (95% CI: 0.781-0.969), significantly higher than that of the combined MRI model (0.642, 95% CI: 0.474-0.803, p = 0.015). Calibration was numerically better for the PET model (p = 0.404) than for the MRI model (p = 0.309), indicating a relative advantage for PET. Decision curve analysis indicated higher clinical net benefit for the PET model across threshold probabilities of 10%-90%.

CONCLUSION

ΔSUVmax% is a promising early predictor of pCR.The [68Ga]Ga-HER2 Affibody PET/CT combined model provides superior predictive value compared with MRI‑based model and could serve as a promising biomarker to guide personalized treatment decisions.