The indications for performance of a BSGI study include high-risk surveillance, alternative to MRI, palpable breast masses with a negative mammogram and ultrasound, a newly diagnosed breast cancer with occult foci, and in women with breast implants or following direct silicone injection to resolve a difficult question.

Initially studies were performed on a conventional gamma camera and hence there were issues related to optimal positioning of the breast and poor resolution; this technique did not detect small cancers less than 1 cm. The sensitivity was less than 50 %, making it a less attractive alternative [6, 9]. Over the last 20 years advancements in gamma ray detector technology (for example, the use of the semiconductor cadmium zinc telluride) and the use of dedicated dual head breast scanners have improved both energy and spatial resolution. This has enabled detection of tumors as small as 1–3 mm. Also, production of images that are oriented similar to standard mammograms has made it easier for radiologists to interpret these studies. These improvements have also resulted in decreased radiopharmaceuticals doses, making the test more acceptable [7, 16].

MBI uses the radiotracer technetium (^99mTc) sestamibi in doses of about 20 mCi (740 mBq) injected intravenously (IV) in one of the antecubital veins. Imaging starts immediately and continues up to the desired number of counts per image, approximately 100,000. Images are obtained with breast oriented in a manner similar to the standard mammogram; craniocaudal (CC) and mediolateral (MLO) images of both breasts. The image acquisition time is about 10 min for each view, to a total acquisition time of 40 min for a complete study [9, 17]. The breast compression is less than that in a mammogram (a pressure of 15 lbs/square in. as opposed to 45 lbs/square in. on a conventional mammogram).

In addition to its utility as an adjunct diagnostic tool, MBI is an attractive imaging test from the cost point of view because of the wide availability of the radiopharmaceutical and compact size of the imaging equipment. It is a useful problem-solving tool particularly in patients unsuited for MRI, because of metallic implants, renal dysfunction, or claustrophobia. BSGI has high sensitivity, specificity, and positive predictive value (PPV) compared to standard mammograms and ultrasound evaluation [7–9, 16–18]. Figure 1.2 shows an example. Weigert et al. [8] in a multicenter study determining the impact of molecular imaging concluded that statistically BSGI was more accurate (better sensitivity, specificity and PPV) than ultrasound, when findings were discordant from a standard mammogram. Lately, BSGI guided biopsy systems [6, 8] have become available which allows confirmation of pathology results and better correlation with imaging findings.

Poor visualization of chest wall and axilla are some of the drawbacks, which can be overcome with additional views. The inability to obtain all of the breast tissue in the field of view (FOV) to a large extent has been alleviated by offering nuclear medicine technologists training in breast positioning by mammographers [9]. In view of potential radiation risks the dose of the injected radiopharmaceutical has been steadily decreasing. Initial trials on BSGI used doses of 30 mCi (1110 mBq), currently this has dropped to 20 mCi and some centers use only 10–15 mCi of ^99mTc [7]. Trials at Mayo Clinic in Rochester, MN are experimenting with a dose as low as 4 mCi and by enhancing image quality with digital post-processing. At this dose the radiation dose to the breast is comparable to a standard two-view mammogram. The results from this study will determine if MBI has a role in screening, particularly for the intermediate risk category, where the benefit of MRI is not clearly defined. While one of the earlier limitations of BSGI was the poor sensitivity of molecular imaging in identifying sub-cm tumors, with recent advances, Hruska et al. [16] demonstrated sensitivities up to 80 % for tumors less than 1 cm.

Positron emission tomography (PET) imaging is a useful diagnostic test in many malignancies, but has not been accepted as standard of care in breast cancers [19]. Dedicated PEM scanners for breast have been developed, providing higher resolution than whole body PET scanners [20]. In PEM and PET imaging, the radiotracer fluoro-deoxyglucose (18-FDG) is used, providing a physiological measure of increased metabolic activity. While a promising tool it also demonstrates ‘hot spots’ at inflammatory and infective sites resulting in false positives.

Transillumination of the breast using light, referred to as diaphanography, was proposed in 1920s. Variants of this approach were investigated till the early 1990s. However, the approach was not recommended for breast cancer screening [24]. Better understanding of the contrast mechanisms, characterization of absorption and scattering properties of breast tissues at various wavelengths, and techniques for modeling optical photon transport through tissues have facilitated development of quantitative methods for diffuse optical spectroscopic imaging. Diffuse optical imaging using continuous wave, time domain, or frequency domain measurements at near-infrared (NIR) wavelengths can be used to provide noninvasive in vivo quantitation of attenuation and scattering properties of breast tissue. This can be used to determine total hemoglobin content, oxygen saturation (ratio of oxygenated hemoglobin to total hemoglobin), water, and lipid content. Extension of the approach to 3D imaging, similar to computed tomography (CT), has resulted in diffuse optical tomography (DOT) systems. Hand-held diffuse optical spectroscopy imaging systems have been developed and continue to be refined [25, 26]. Stand-alone DOT prototype systems for adjunctive use in diagnostic breast imaging have been developed by academic investigators [27, 28] and by commercial entities (SoftScan^®, Advanced Research Technologies, Inc., Montreal, Canada; CTLM^®, Imaging Diagnostic Systems, Inc., Fort Lauderdale, USA).

In a study of 90 subjects, DOT showed that the ratio of total hemoglobin in the abnormality to that in the normal contralateral breast was statistically different for malignant tumors [29]. However, the study noted that there may be a resolution threshold of approximately 6 mm. Development of multimodality systems combining NIRS with X-ray imaging has been reported [30–32]. In a study of 189 breasts from 125 subjects including 51 breasts with lesions, a statistically significant increase was observed for total hemoglobin in malignant tumors larger than 6 mm compared to fibroglandular tissue [33]. A hand-held probe combining ultrasound with NIR imaging has been developed, and in a study of 65 subjects with 81 lesions significantly higher concentration of total hemoglobin was observed in malignancies than benign lesions [34].

Development of NIR systems integrated with MRI [35, 36] that incorporates image information from MRI during NIRS reconstruction as well as clinical evaluation of such multimodality systems have shown that malignant tumor exhibit higher concentration of total hemoglobin and lower oxygen saturation. The use of exogenous contrast agents such as indocyanine green as well as tumor-targeted contrast agents that are under development can preferentially enhance lesions and could improve differential diagnosis. Monitoring of neoadjuvant chemotherapy response with a hand-held diffuse optical spectroscopic imaging system in a limited dataset showed that significant changes in oxygenated hemoglobin could be observed approximately 90 days after initiation of therapy [37]. In order to reduce re-excision rates following breast conserving surgery (BCS), NIR-based optical imaging systems to assess tumor margins are being investigated [38, 39]. In summary, the past two decades have witnessed substantial improvement in optical imaging techniques using NIR spectroscopy, and its transition to routine use for some clinical applications is highly probable in the near future.

In the management of early breast cancer, BCS is the standard surgical procedure, where excision of the least volume of breast tissue free of tumor cells at the margins is the goal. However, the current surgical literature [40] estimates positive margins at surgery in the range of 20–70 %. This necessitates revision excision. Currently, the quickest means to evaluate tumor margins is the ‘traditional frozen section’. This process is however laborious, time-consuming, does not include the entire tumor surface, and is limited by freezing artifacts of the surgical margins. Cauterization surgery also limits evaluation. Currently, there are many experimental, real-time, imaging options that are being evaluated. There is a need for such techniques to be cost-effective, reproducible, and dependable.

We performed an experimental trial [41–43] of excised lumpectomy specimens at the University of Massachusetts Medical School in collaboration with the Medical Physics department at University of Massachusetts, Lowell, MA. The investigators evaluated lumpectomy specimens from breast cancer patients with polarization techniques after staining the specimen with dilute methylene blue. Wide-field polarization for quick macroscopic survey and small FOV confocal microscopy for small FOV with high resolution was performed, images analyzed, and later correlated with findings at Hematoxylin and Eosin (H&E) stained pathology slides evaluated by a trained breast pathologist (Fig. 1.3). The difference in the reflectance and fluorescence polarization values for benign and cancerous tissue was exploited. In these studies, Patel et al. [41–43] observed good correlation between fluorescence polarization values and findings on H&E stained sections of benign and malignant breast tissue on histopathology. The reflectance polarization values did not correlate as well. While the researchers concede to slight misregistration between confocal microscopy images and H&E stained specimens, the ease of use has good future potential to evaluate ex vivo specimens. The instrument could also be used in vivo on patients on the operating table to discern any residual malignant tissue that merits excision. A clinical trial is underway to evaluate the utility of this imaging technique for intraoperative evaluation of margins during BCS.