RESUMO
AIM: The aim of this paper was to prospectively evaluate FDG PET/CT in the assessment of metabolic response to neoadjuvant chemotherapy and correlation with tumor cellularity in locally advanced breast cancer. METHODS: Images were acquired with a PET/CT scanner in 50 patients at baseline and after completion of treatment, just before surgery. All findings were confirmed by histopathological analysis. PET/CT quantification (SUVmax) at baseline and after finishing neoadjuvant chemotherapy (4 cycles of epirubicin + cyclophosphamide +/- taxanes) were compared using RECIST criteria and Miller & Payne (M&P) scale. RESULTS: Baseline mean tumor size was 4.4±1.6 cm. Thirty eight patients were considered responders and 12 nonresponders. According to M&P scale, 10 patients had good prognosis (grades 4-5) and 40 patients had bad prognosis (grades 1-3). All patients with grade 5 M&P had no significant postchemotherapy FDG uptake. Patients with bad prognosis had lower SUVmax variation (∆SUVmax) than patients with good prognosis (60.7% vs. 80.5%, P=0.0016). ∆SUVmax was lower in nonresponders than in partial responders according to RECIST criteria (38.9% vs. 67.6%, p<0.001), and was also lower in partial responders than complete responders (67.6% vs. 85.4%, P=0.005). A cut-off ∆SUVmax value of 52% differentiates responders from nonresponders with a sensitivity of 86% and a specificity of 90%. Probability densities of the ∆SUVmax (%) for stable disease (<45), partial (>45 to <82) and complete response (>82) showed an overall accuracy of 78% (Weighted Kappa=0.74). CONCLUSION: PET/CT is useful to monitor response to neoadjuvant chemotherapy in locally advanced breast cancer. ∆SUVmax on PET/CT correlates with tumor cellularity after completion of neoadjuvant chemotherapy.