Density and volume of cardiac calcifications detected on planning CT predicts cardiotoxicity after hypo-fractionated whole breast Radiotherapy.

Abstract

Cardiac calcifications (CAC) are emerging as predictors of cardiac toxicity after breast cancer Radiotherapy. Main purposes of this study were:1) to test the association between CAC scores and cardiac events in a cohort treated with moderate hypo-fractionation; b) to assess interaction between CAC and dosimetry/clinical predictors. Data of 1172 consecutive patients treated at our hospital with 3DCRT whole breast irradiation (40 Gy/15fr) were available. Heart was automatically segmented and the mean heart dose (MHD) was assessed. The Agatson score (AS), the CAC_volume and the max HU value in the heart (Max_HU) were assessed using an in-house script. Their association with cardiac events was tested by logistic regression, including the combined impact of MHD and relevant clinical parameters (including chemotherapy, age and smoking). CAC_volumes/Max_HU values were compared against the values obtained from our clinical planning system for 75 patients with calcifications. Twenty-nine patients experienced cardiac events (median follow-up:6.5y). AS/CAC_volume/Max_HU were highly significant predictors (p < 0.005). MHD (mean ± SD: 0.8 ± 0.1/2.4 ± 0.7 Gy for right/left) was predictive only if encoded according to the optimal cut-off based on Youden index (MHD > 1 Gy). The resulting multivariate model combined MHD > 1 Gy, age and CAC scores, with similar performances for the three different scores (AUC = 0.77/0.755, p < 0.0001, R² = 0.99), with the three different scores always being the major predictor. TPS gave consistent values for Max_HU while CAC_volume showed larger differences, although highly correlated. CAC are the most important predictors of cardiotoxicity. Max_HU is promising for fast assessment of patients at risk. The greatest clinical benefit of further heart sparing is expected in left breast cancer patients with moderate-severe CAC scores (Max_HU > 250).