Mid-treatment radiologic response as an early predictor of pathologic response in TNBC in the era of pembrolizumab

Abstract Background Breast MRI is used to monitor neoadjuvant chemotherapy (NAC) response in triple-negative breast cancer (TNBC), yet its predictive value for pathologic complete response (pCR) in the era of carboplatin and pembrolizumab remains incompletely defined. Methods We analyzed 517 TNBC patients who underwent serial MRI (baseline, mid-treatment, post-NAC) across three regimens: non-carboplatin (n = 225), carboplatin (n = 148), and pembrolizumab (n = 144). Associations between radiologic complete response (rCR) and pCR were evaluated. Results Mid-treatment rCR rates were highest with pembrolizumab (24.3%), followed by carboplatin (12.8%) and non-carboplatin (11.6%). Pembrolizumab independently increased mid-treatment rCR (adjusted OR 3.33; 95% CI 1.74–6.34; p < 0.001). Mid-treatment rCR was a strong predictor of pCR across all regimens (adjusted OR 26.33; 95% CI 9.26–74.81; p < 0.001), with a stronger association than post-NAC rCR (adjusted OR 8.85; 95% CI 5.52–14.18; p < 0.001). Conclusions Within each NAC regimen, mid-treatment rCR was a strong predictor of pCR, and pembrolizumab was associated with a higher likelihood of early radiologic response. Given the retrospective, single-center design spanning an extended period, these findings should be interpreted as hypothesis-generating and provide a rationale for prospective, response-adapted trials evaluating mid-treatment MRI in TNBC.