Correlations between diffusion-weighted imaging and breast cancer biomarkers

Objective We evaluated whether the apparent diffusion coefficient (ADC) provided by diffusion-weighted imaging (DWI) varies according to biological features in breast cancer. Methods DWI was performed in 190 patients undergoing dynamic contrast-enhanced magnetic resonance imaging (MRI) for local staging. For each of the 192 index cancers we studied the correlation between ADC and classical histopathological and immunohistochemical breast tumour features (size, histological type, grade, oestrogen receptor [ER] and Ki-67 expression, HER2 status). ADC was compared with immunohistochemical surrogates of the intrinsic subtypes (Luminal A; Luminal B; HER2-enriched; triplenegative). Correlations were analysed using the Mann- Whitney U and Kruskal-Wallis H tests. Results A weak, statistically significant correlation was observed between ADC values and the percentage of ERpositive cells (-0.168, P00.020). Median ADC values were significantly higher in ER-negative than in ER-positive tumours (1.110 vs 1.050×10 ^-3 mm ²/s, P00.015). HER2- enriched tumours had the highest median ADC value (1.190×10 ^-3mm ²/s, range 0.950-2.090).Multiple comparisons showed that this value was significantly higher than that of Luminal A (1.025×10 ^-3 mm ²/s [0.700-1.340], P00.004) and Luminal B/HER2-negative (1.060×10 ^-3 mm ²/s [0.470-2.420], P00.008) tumours. A trend towards statistical significance (P0 0.018) was seen with Luminal B/HER2-positive tumours. Conclusions ADC values vary significantly according to biological tumour features, suggesting that cancer heterogeneity influences imaging parameters. Key Points ̇ DWI may identify biological heterogeneity of breast neoplasms. ̇ ADC values vary significantly according to biological features of breast cancer. ̇ Compared with other types, HER2-enriched tumours show highest median ADC value. ̇ Knowledge of biological heterogeneity of breast neoplasm may improve imaging interpretation.