Objective: In recent literature, increasing number of studies have analyzed associations between the MRI findings and histopathological features of breast cancer. Ki-67 index is an important prognostic factor in breast cancer which provides information about the mitotic activity and the growth rate of the tumor. The proliferation index markers like Ki-67 and PHH3 (Phosphohistone H3) play an important role in decision making in the treatment regimes. PHH3 is a relatively new entity which is specific only to the proliferation giving it a unique characteristic because it is not observed at the interphase. PHH3 provides a more sensitive and accurate mitotic index with less interobserver variability when compared with conventional H&E staining, thus emphasizing its potentially increased value in practice. In this study, our aim is to correlate both proliferation index markers (Ki-67 and PHH3) with breast MRI features in cases of invasive breast carcinoma.

Material and Methods: In this IRB approved study, out of 126 consecutive patients' retrospective data with written consent, 29 patients (mean±std age 48.86±13.10 and range [28-82]) with full apparent diffusion coefficient (ADC), Ki67 and PHH3 data were selected. In all of these surgical specimens, immunohistochemically administered Ki-67 and PHH3 and immunopositive cells in the digital pathology were counted. MRIs were performed in a 1.5T MR unit (MagnetomEspree with Syngo MR B15 software; Siemens, Erlangen, Germany). Images were evaluated by two radiologists in consensus. Mean ADC values were obtained from radiologist drawn ROIs. The results were analyzed with principle component analysis (PCA) and pairwise covariance analysis.

Results: Mean ADC value was (mean±std) 0.92±0.18 with range [0.67-1.3] (10-3mm2/s), mean Ki67 was 18.62±13.32 with range [5-62] ('%' for Ki-67) and PHH3 27.17±30.81 with a range [1-111] ('10HPF' for PHH3). When PCA was conducted in (ADC, Ki67, PHH3) space, after normalizing each of the variables 45.223% of total variance was explained mainly by the new variable obtained as a linear combination of the normalized values -0.0866*ADC - 0.9879* Ki67 - 0.7910*PHH3, 37.282% by -0.9879*ADC + 0.1545Ki67 -0.0101*PHH3 and 17.494% by 0.1283*ADC + 0.7806*Ki67 - 0.6117*PHH3 respectively. The analysis demonstrates that while Ki67 and PHH3 values explain much of the variance (45.223% + 17.494%) and thereby are highly correlated as expected, ADC on its own determines a significant portion(37.282%) of the data as well. Moreover, pairwise covariance analysis resulted in ADC, Ki67 and PHH variance values of 2.170, 1.528 and 2.197 respectively and, covariances of ADC|Ki67 (-0.1169), ADC|PHH3 (0.2048) and Ki67|PHH (1.31).

Conclusion: Ki-67 and PHH3 both show good correlation in the assessment of proliferative status of breast cancer. However, breast MRI ADC values cannot be reliably used for the prediction of Ki-67 and PHH3 status of the invasive breast tumors. More studies with larger series are needed to assess the potential of breast MRI to serve as a clinical and/or prognostic factor.