A review of MRI (CT)/US fusion imaging in treatment of breast cancer

Even if magnetic resonance imaging (MRI)-detected lesions are found during contrast-enhanced MRI examinations before surgery, it is difficult to identify them at the same site via ultrasound examination [1]. The reliability of ultrasonography tends to depend on the operator. To solve the various problems of ultrasonography (and to ensure objectivity and reproducibility), the fusion of different (or similar) imaging modalities is effective. Fusion imaging technology has already been used in the biopsy of prostate cancer in the urological field [2‐6] and in the treatment of hepatocellular carcinoma in the liver field [7‐14]. We herein report on an ultrasonographic diagnostic technology using magnetic fields in the field of breast cancer.

Ultrasound, which is affordable and places little burden on patients, is the first choice for identifying MRI-detected lesions. However, considering that the patient position of an ultrasound is different from an MRI, it is sometimes difficult to accurately identify MRI-detected lesions using ultrasound. Figure 1 shows how the difference in patient position can change the spatial relationships between lesions. For this reason, the ultrasonographic detection of MRI-detected lesions depends greatly on the skill of the ultrasound technician. Detection rates vary between institutions [15], and ultrasound reportedly has a low identification rate for non-mass enhancement when comparing masses and foci [16].

MRI/ultrasound fusion technology is garnering attention as a way to overcome this limitation. The RVS^® system is an ultrasound fusion technology with magnetic position navigation developed in Japan by Fujifilm Healthcare. In this diagnostic imaging system (Fig. 2), spatial locational information created by a magnetic field generator is detected by a position sensor attached to an ultrasound probe, and the locational information is used to display MR/CT/ultrasound images together with the ultrasound images from the probe in real time [17‐31]. An auxiliary function that links contrast-enhanced MR images and ultrasound images makes it possible to identify MRI-detected lesions that may be difficult to observe with ultrasound alone. This system improves the identification of MRI (or CT)-detected lesions, especially in small tumors, non-mass enhancement, and sentinel lymph nodes [17‐32]. In this article, we report on a technique for preoperative diagnosis of breast cancer extent and identification of non-mass enhancement using CT/ultrasound fusion imaging at our institution. Since the fusion system uses magnetic fields during scanning, it is contraindicated for patients with pacemakers as it may cause malfunctions.

Prone contrast-enhanced MRI is the first step in determining breast cancer extent. If MRI detects additional lesions presumed to be intraductal extension, supine contrast-enhanced MRI is performed in the same position as that for the ultrasound examination. The results, taken together, are useful when conventional ultrasound alone is insufficient for diagnosis. We expect that performing two MRIs poses a tremendous challenge for some institutions, both from a medical standpoint and from the patient’s financial standpoint. In addition, this is a considerable challenge in promoting MRI/ultrasound fusion technology. However, contrast CT is performed in the same position as the ultrasound; thus, it is much easier to combine CT with ultrasound for fusion imaging than performing two MRIs (Fig. 3). At our facility, all breast cancer patients undergo prone-enhanced MRI to diagnose the spread of breast cancer before surgery. In addition, contrast-enhanced CT is also performed to stage breast cancer lesions. For this reason, our facility may have a better environment for fusing ultrasound and CT images than other facilities. Kousaka et al. reported that CT/ultrasound fusion is useful for identifying incidental findings [21]. Breast cancer lesions and spread presumed to be intraductal extension have been identified with prone-enhanced MRI (Fig. 4a). After confirming the lesion corresponding to the MR image in the contrast-enhanced CT image, CT and ultrasound fusion using RVS^® was performed (Fig. 4b). First, DICOM data from contrast CT (using GE Revolution^®, taking 1.25 mm slices) were fed into the ultrasound system. The scan was initiated at the nipple of the affected side. When the probe was moved to the region of the tumor, minor discrepancies may have occurred between the different imaging modalities. To obtain a more accurate fusion, it is important to align the positions of the fat around the lesion, the shape of the mammary gland and ribs, and the blood vessels [27, 28], and to visualize the same cross-section. The CT and ultrasound images were corrected again. In the surgically resected specimen, we pathologically confirmed the intraductal spreading from the tumor, and we believe it matched each image finding.

The technique described above for identifying the extent of intraductal spread from a tumor is a basic technique for RVS^®. In this section, we describe practical applied techniques to identify non-mass enhancement. In Fig. 5a, non-mass enhancement can be seen spreading regionally under MRI. Under ultrasound alone, identifying the lesion is difficult. RVS^® may be particularly effective in such cases. After confirming the area corresponding to non-mass enhancement visualized on the prone position MR image on the contrast-enhanced CT image, fusion of CT and ultrasound using RVS^® was performed (Fig. 5b). Our approach is not to immediately begin searching for the lesion but to begin in the anatomical structures surrounding the lesion and working inward to identify it. The site of non-mass enhancement was pathologically confirmed to be ductal carcinoma in situ in the surgically resected specimen.

In Japan, patients with hereditary breast and ovarian cancer (HBOC) syndrome having breast or ovarian cancer are eligible for surveillance with contrast-enhanced MRI under the National Health Insurance. Although MRI-guided biopsy for MRI-detectable lesions is covered by the insurance, only a limited number of institutions can provide such biopsies. Figure 6a shows the MRI-detected lesion revealed on the unaffected side (left breast) in a patient with HBOC (right breast cancer) syndrome. It was difficult to identify the MRI-detected lesion using ultrasonography alone. Therefore, after confirming the position of the MRI-detected lesion on the CT image, we fused the CT image with the ultrasound image using RVS^®. As we mentioned earlier, the key to fusion is to use the same sectional view from a CT scan and ultrasonography after registering the position using the shape of fat, a mammary gland, or a rib. Clear identification of the MRI-detected lesion was possible with fusion in this case (Fig. 6b). In this case, instead of tissue biopsy, we used the excised specimen in a risk-reducing mastectomy to search for the MRI-detected lesion. There were no malignant findings, but columnar cell hyperplasia without atypia was observed. As fusion imaging is useful for HBOC syndrome and other cases in identifying non-mass enhancement detected on MRI, its use is expected to increase.

In previous studies, the identification rate of MRI-detected lesions was 30–76% [17‐19, 22, 28, 30, 33, 34] using conventional ultrasound alone without fusion. In contrast, fusion of supine contrast-enhanced MR images and ultrasound images has been reported to improve identification rates to 78–100% [17‐19, 22, 28, 30, 31, 33, 34] (Table 1). In a previous study, a second-look ultrasound was performed on MRI-detected lesions via fusion technology using RVS^® or V Nav^® (GE Healthcare). After identifying the MRI-detected lesions, a pathological examination was performed under ultrasound guidance. The results showed that 44–77% and 23–56% of lesions were benign and malignant, respectively [18, 22, 28, 30, 31, 34].

At our facility, all patients with breast cancer are diagnosed using contrast-enhanced MRI to diagnose the extent of breast cancer before surgery. In addition, contrast-enhanced CT is also performed for staging all breast cancer lesions. If non-mass enhancement is observed on prone-enhanced MR images, an attempt is made to identify the lesion via ordinary second-look ultrasound without using RVS^®. If non-mass enhancement cannot be identified via second-look ultrasound, after confirming the region corresponding to the non-mass enhancement visualized on the contrast-enhanced MR image in the contrast-enhanced CT image, CT and ultrasound fusion using RVS^® is performed.

Kousaka et al. reported a 100% (11/11) rate of identification of the lesion site using fusion of CT and ultrasound images for lesions found incidentally on chest CT [21]. They also reported a breast cancer diagnosis in 36% (4/11) of pathological examinations. The fusion of supine contrast-enhanced MR images and ultrasound images is preferred, but the fusion of contrast-enhanced CT images and ultrasound images may also be clinically useful.

In this manner, a high rate of lesion identification can be achieved using the ultrasound fusion technique for MRI-detected lesions. In the future, MRI-guided biopsies may be indicated for cases that cannot be identified using the ultrasound fusion technique.

Fusion imaging allows clinicians to observe lesions that cannot be identified with ultrasound alone. This enables appropriate preoperative imaging, and increases the safety of the examinations and surgery. As fusion ensures reproducibility, we think that it is also useful for providing explanations to patients and as an educational tool. We firmly believe that this modality will gain widespread popularity in many clinical settings in the near future.