Accurate preoperative evaluation of breast cancer subtypes is critical, as distinct biological profiles guide clinical treatments and predict prognosis. Our present study illustrated the spatial heterogeneity observed in angiogenesis and protein in cells through APTw and IVIM parameters, enabling the classification of breast cancer subtypes. Our findings underscore the potential of APTw technology as a novel tool for distinguishing TNBC. The combination of APT and f values yielded an enhanced diagnostic accuracy (AUC = 0.950) in identifying TNBC and Luminal A type cancers.
Our results revealed that APT values were higher in TNBC than compared to HER2 type cancers, Luminal A and Luminal B type cancers. Liu et al. [19] investigated the four molecular subtypes and showed comparable APTw values without statistically significant difference (P = 0.073). This result may be because single slice 2D APT imaging was employed in this study. Such geometric localization may miss the plane with the largest tumor diameter. Our study employed 3D APTw to better cater to intratumoral heterogeneity. Kamitani et al. [20] used 3D APT in breast cancer and found that the APT value of TNBC was higher than Luminal type cancers and HER2 cancers, consistent with this study's findings. Patients with TNBC typically have a relatively poorer outcome than those with other breast cancer subtypes owing to inherently aggressive clinical behavior and a lack of recognized molecular targets for therapy[25, 26]. APTw extends the scope of molecular imaging technology to proteins. Chan et al.[27]suggested that APTw may have the potential to assess the aggressiveness of breast cancer. Therefore, it is reasonable that TN cancers showed a high APT value. In addition, large bare nuclei, nuclear atypia, ill‑defined borders, necrosis, and lymphocytic cell infiltrate were more common in TNBC as compared to non‑TNBC[28]. TNBC has limited treatment options and is usually treated with cytotoxic therapy with poor clinical efficacy[23]. Therefore, it is essential to distinguish TNBC from other subtypes of breast cancers and to find new therapeutic targets for TNBC.
Recent studies of IVIM in breast cancer demonstrated that D, D*, and f values significantly differ from molecular subtypes[29–31]. Our results showed that the f values were higher in TNBC compared to HER2-type cancers. This correlates with the research results of Vranic[32]; they thought TNBC exhibited increased f values compared with other subtypes, which suggests that this type of cancer undergoes increased tumor angiogenesis. This is the exact reason for the increase in APT values. However, some studies[15, 33] suggest that the f value in TNBC is lower; this may be associated with decreased perfusion in the center of tumors due to necrosis. In addition, our results showed that the f value was higher in Luminal A-type than in Luminal B-type cancers. It Indicates that the f value correlated positively with PR expression. Similarities in hormone receptor effects suggest a negative correlation, as with ER expression. Meng et al. [34] thought that f value showed a low correlation with ER expression. However, PR + expression tumors may increase tumor growth by angiogenesis by normalizing tumor vasculature, which would allow for proper blood flow. Studies have shown that progesterone may increase angiogenesis by regulating VEGF in breast cancer cells[35].
Interestingly, there was no correlation between tumor diameter and breast cancer subtypes in our study. Nakashima et al.[36] indicated larger volume increases in TNBC. Given the expected differences in breast cancer growth rates across subtypes, predicting growth rates at diagnosis could facilitate more tailored surgical plans based on patient-specific rates.
Ki-67 plays a role in overall and disease-free survival in breast cancer patients[37]. Predicting Ki-67 expression based on imaging is significant in clinical practice. Some studies [19, 20] have shown an association between APT and Ki-67 expression. This could be attributed to the presence of high concentrations of mobile proteins and peptides during active tumor cell proliferation, potentially leading to increased APT values. The microenvironment, including factors like micro-necrosis or fluid accumulation, might further contribute to heightened tumor APT values.
Histologic grade, despite tumor size and nodal status, remains prognostically relevant for overall survival. As an independent prognostic factor, breast cancer with a higher histological rate is less differentiated and has a poorer prognosis than breast cancer with a lower histological grade[38]. Most studies have demonstrated the association between APT value and grades [34, 39], our analysis thought there was correlation between APT value and grades, but some studies have shown that there is no correlation between them [20]. Several factors could explain this inconsistency, such as the histological type of breast cancer (tumor fever chymosin). Our study also found that APT values were relatively low in stage T1 breast cancer, which we considered to be related to the relatively slow proliferation of proteins in cells with low stage. More attention has been paid to patients with stage T1, and this finding could help in the diagnosis and treatment of patients with stage T1 breast cancer[40]. We expect more and more breast cancers are being detected in preinvasive or earlier stages.
The present study suggests that APTw offers advantageous over IVIM in differentiating between TNBC and Luminal A type cancers and prognosis evaluations. Combined IVIM and APTw led to improved diagnostic accuracy (AUC = 0.950). While previous studies [15, 33] demonstrated that IVIM diagnostic efficacy across breast cancer subtypes. Also, some previous [20, 34, 39] studies showed that APTw may be related to ER, PR, and HER2 expression. Our study combines the two imaging methods to reveal more possibilities for predicting breast cancer subtypes.
This study also has some limitations. (1) The number of patients was relatively small, especially in patients with TNBC. (2) We excluded the patients with small lesions because we thought small lesions impact the accuracy of APTw due to the influence of breast fat and water [41]. (3) ROI delineation was performed at the two-dimensional level on the largest lesion area, introducing potential bias. (4) The longer scanning time of APT and IVIM, especially the lengthening of the already long routine breast MRI examination time, could compromise patient comfort and trigger motion artifacts.
In conclusion, both APTw and IVIM parameters exhibited significant differences across TNBC and other subtypes of breast cancers. The APT and f values were independent risk factors for predicting the TNBC. A combination of APTw and IVIM imaging may be an effective noninvasive method for clinically assessing TNBC from breast cancer subtypes.