The value of diffusion-weighted imaging in assessing the ADC changes of tissues adjacent to breast carcinoma

Conservative surgery has become a well-established alternative to mastectomy in the treatment of breast cancer. However, in case of larger lesions or small-size breasts, the removal of adequate volumes of breast tissue to achieve tumor-free margins and reduce the risk of local relapse may compromise the cosmetic outcome, causing unpleasant results [1]. An issue of critical importance is thus to know well the transition from the tumor tissue to normal tissue in breast carcinoma in order to helpfully decide the surgical scope. Currently, the evaluation of tumor scope in clinical works relies on pathologic examination of margins free of gross tumor tissues. However, previous studies reported that in some patients with histologically negative margins, a relatively high recurrence rate was still observed [2]. It was thought that genetic and molecular alteration precedes phenotypic changes, therefore histologic assessment alone may be inadequate to detect the presence of transformed cells in surgical margins [3]. In another word, histologic assessment alone may be insufficient for the detection of transition from the tumor tissue to normal tissue. A recent study [4] discovered that there existed the geographic zones of the normal tissue adjacent to invasive cancers in which methylation changes could be identified and the authors summarized that multiple gene promoters could be used as a surrogate biomarker to define the molecular margin of lumpectomy in the future. Studies of other malignant lesions [3, 5‐9], including head and neck cancer, pancreatic cancer, lung cancer, colorectal carcinoma, rectal cancer and hepatocellular carcinoma, also demonstrated that there existed a molecular border of tumor tissues with various molecular methods. However, all of these studies were performed on the biopsy tissues or resected tissue specimens and no preoperative assessment of peri-tumor was reported. There is, therefore, a need for a non-invasive technique to detect the change of peri-tumor tissue for the purpose of providing more information about surgical scope.

Magnetic resonance imaging (MRI) has a high resolution and can provide much more detailed images than mammography and ultrasonography (US), which makes it a widely-used tool for the diagnosis of breast lesions. However, the conventional breast MRI (plain MRI) is still a morphological diagnostic technique which only provides general anatomical information such as signal, shape, size and location. One of the latest advancements in MRI technology is the application of diffusion-weighted imaging (DWI). The principle underlying DWI is that the thermal motion of water molecules in extracellular fluid enables the acquisition of images that reflect both histological structure and cellularity [10] and therefore it can detect the changes of tissue structure at molecular level. It also enables the quantitative evaluation of apparent diffusion coefficient (ADC), which may be useful for distinguishing malignant from benign tissues and monitoring therapeutic outcome [11, 12]. Compared with benign lesions, diffusion of malignant tumors with high cellular tissue decreased and the ADC value in malignant tumors is lower than that of benign lesions [13‐15]. Several latest studies have shown that ADC has a potential for clinical appreciation in differentiating benign and malignant lesions with good specificity [16‐19].

On the other side, DWI, based on its imaging mechanism, could detect the changes of ADC in different tissues. Measurement of the ADC provides a quantitative estimate of the restrictive nature of the motion of water molecules within tissue for each voxel in a diffusion-weighted image. This study was thus designed to compare the ADC values between malignant and benign lesions in breast and to study the change of ADC values in peri-tumor tissues through echo planar diffusion weighted imaging (EPI-DWI). Our main aim was to study the DWI changes in tissues adjacent to breast carcinoma, which would be helpful for the clinical surgeon to decide the scope and pattern of operation.

This study was approved by the ethics committee of Xi'an Jiaotong University and all patients gave written informed consent before beginning the study.

Despite the improvement in the detection of breast cancer with the widespread application of mammography and ultrasound, breast lesions still remain difficult to diagnose and characterize, especially in dense fibroglandular breasts. The main advantage of MRI in the breast is that they can improve the detection and characterization of multiple and/or small lesions even in the dense fibroglandular breasts. However, the low specificity of MRI remains a problem [20].

In recent years, the DWI has been extensively applied in evaluating cerebral tumors and the correlation between the ADC value and the cellular density has been verified. Briefly, the higher the cellular density is, the lower the ADC value will be in DWI, and vice versa [21]. For malignant tumors, they have a relatively high cellular density and therefore will produce a low ADC value in DWI, while for benign lesion, its density is generally low and thus will produce a high ADC value in DWI. Application of DWI in the diagnosis of breast lesions has been reported recently [14, 16, 22‐24]. These studies showed that in malignant breast tumors, the ADC was significantly lower than that in benign tumors. These authors concluded that ADC might help to differentiate benign and malignant lesions with good specificity, and may increase the overall specificity of breast MRI, which is consistent with our results.

In ROC curves, the specificity of the ADC is dependent on the threshold value that determines the differentiation between benign and malignant tumors. In our study, we obtained two threshold values: 1.24 ± 0.25 × 10^-3mm²/s (b = 500) with 93% sensitivity and 100% specificity; and 1.20 ± 0.25 × 10^-3mm²/s (b = 1000) with 96% sensitivity and 97% specificity. The threshold of ADC value in our study was close to the data reported by Luo JD [16] (1.22 × 10^-3mm²/s) and Rubesova E [14] (1.13 ± 0.10 × 10^-3mm²/s). The reason of difference was that in our study we chose b = 500 and b = 1000 while in Luo JD's study, b = 0 and b = 800 and in Rubesova E's study, three-dimensional fast low-angle shot (3D-FLASH) with contrast injection was applied. An exception in our study is that one breast malignant phyllode had a high ADC value (b = 500, 1.72 ± 0.14 × 10^-3mm²/s; b = 1000, 1.73 ± 0.14 × 10^-3mm²/s), which was similar to that in Woodhams R's report (1.67 ± 0.59 × 10^-3mm²/s) [12]. The reason of a high ADC in tumor was that in this case the lesion was mainly liposarcoma histologically and the liposarcoma, similar to adipose tissue, has a relatively high ADC value.

Another important aim in our study is try to study the DWI changes in tissues adjacent to breast carcinoma. It is well known that over the past 30 years in the field of breast carcinoma surgery, the extent of surgery has been progressively reduced, which leads to less disfigurement and a significant improvement in life quality of patients [25]. From an oncological perspective, outcome of conservative breast surgery was found to be equally effective when compared with mastectomy [26]. However, tumor recurrence after breast-conserving surgery still remains a problem [27]. Many studies [28‐31] reported that obtaining negative surgical margins in conservative breast surgery influenced the incidence of local disease recurrence and probably overall survival. Preoperative evaluation of the change in peri-tumor tissue will plays an important role for the success of conservative operation.

Nowadays in clinical works, the preoperative estimation of the excision scope of breast tumor is generally made by surgeon according to his own experience, based on the results of mammography, ultrasound and conventional MRI. However, because the capacity for malignant growth is acquired by the stepwise accumulation of defects in specific genes regulating cell growth and tissue homeostasis [32], the genetic and molecular alteration prior to phenotypic changes in peri-tumor tissues usually could not be reflected by conventional mammography, ultrasound and MRI examinations. On the other hand, although intraoperative pathological diagnosis was regarded as an important standard second only to routine paraffin section for the evaluation of negative or positive margin of tumor, it can only be performed approximately 30 minutes post excision of the lesion and the decision must be made intraoperatively [33‐37]. Routine paraffin section diagnosis is the "gold standard", however, the data will be decided after operation a couple of days. Therefore, it is essential to explore new techniques that should be noninvasive and could detect the change of peri-tumor tissue before operation. For these reasons, DWI, due to its distinct characteristics mentioned above, might provide some information on the microstructure change of breast tissues and probably becomes a potentially valuable method for evaluating the change of peri-tumor tissue in breast carcinoma.

Besides, since this work was basically a methodological study and our aim was to study the change of ADC values change in peri-tumor tissue by the novel MRI technique, all the lesions recruited in our study were from mastectomy and relatively large in size. Therefore, further researches are needed to assess whether the results and threshold in the present study could be also applied to the smaller lesions.

In conclusion, our study showed that the ADC value was a sensitive and specific parameter that can help to differentiate benign and malignant breast lesions with a threshold value of 1.24 ± 0.25 ± 10^-3mm²/s (b = 500) and 1.20 ± 0.25 ± 10^-3mm²/s (b = 1000). Our results also showed that the ADC values increased gradually from the tumor to peri-tumor to normal tissues and there existed a region with a thickness of about 5 mm surrounding the border of tumor (by routine MRI), which has an abnormal ADC value, while beyond this region, the ADC value of tissue return to a normal range gradually.

Our next stage studies will include: 1) Measure ADC value of each ROI layer with a thinner thickness to detect the gradual change from carcinoma to normal tissue; 2) Compare the examination of DWI with that of pathological histology about the change of peri-tumor tissue in breast carcinoma and study the correlation of diagnosis between these two methods.