As described in the Introduction, the adenoma missed detection rate remains at approximately 22% (range, 15 to 32),2 and flat and depressed lesions are frequently overlooked. Moreover, the specificity of white light imaging for distinguishing neoplastic from nonneoplastic lesions is poor and depends on visible mucosal changes. The histological analysis of small and possibly less relevant lesions significantly increases workload and cost.17 Therefore, benign lesions are often removed, increasing both cost and risk.3

Imaging devices and fluorescent probes should be chosen according to the molecular imaging goal. For lesions to be detected at a high frequency, high sensitivity is mandatory. Therefore, wide-field endoscopes and high-sensitivity probes are needed. However, trials aimed at characterizing lesions might instead opt for high specificity by the resolving cellular details.18 In clinical practice, this technique may improve the detection rate of missed polyps, flat lesions, and residual neoplasia after endoscopic mucosal resection.8-10

Tumor-specific activation of molecular beacons is desired. In a study with Apc^min mice, a cathepsin-sensing near-infrared probe was used for fiberoptic endoscopic detection of colorectal adenomas,19 demonstrating that tumor-specific dequenching of smart probes can be used for endoscopic imaging.

Fluorescence probes for molecular imaging can be applied topically or systemically during GI endoscopy. For safety, topical administration of probes is preferred over intravenous injection. Topical administration is only possible when the target is directly accessible and the region of interest is limited. Injection into a vein is used for systemic administration when the probes cannot pass through the healthy epithelial barrier.

In vivo molecular imaging in the field of GI endoscopy has been used in various preclinical and clinical trials. For the diagnosis of high-grade dysplasia in Barrett esophagus, Li and Wang,20 one of the leading groups in the field of endoscopic molecular imaging, performed optical imaging after topical administration of fluorescence-labeled peptides. To uncover small peptides with a strong affinity for high-grade dysplasia in Barrett esophagus, a phage display technique was employed. For example, this group identified a seven-residue peptide, ASYNYDA, which specifically binds to high-grade dysplasia in Barrett esophagus using the M13 phage display. The fluorescein-labeled ASYNYDA peptide was topically administered over a region of intestinal metaplasia in the distal esophagus. After a 5-minute incubation period, the unbound peptides were rinsed off. Fluorescent images using a wide-field endoscope sensitive to fluorescence revealed peptide homing to a region of high-grade dysplasia that was histologically confirmed. In a recent study, Wang's group also identified peptides that bind to adenomas in the CPC;Apc mouse model using an in vivo phage display (Fig. 1).21,22 Fluorescence-labeled peptides bound to dysplastic epithelium in the CPC;Apc mice were observed using a wide-field endoscope. Using multispectral peptides, colorful images can be obtained that demonstrate a tumor's molecular profile.23,24

Very recently, Urano's group successfully developed a novel activatable fluorescence probe based on the concept of intramolecular spirocyclization.25 The novel spirocyclic rhodamine-based fluorescence probe for γ-glutamyltranspeptidase, called γ-glutamyl hydroxymethyl rhodamine green (gGlu-HMRG), has been successfully developed, and is upregulated in cancer cells (Fig. 2). In mouse models of peritoneal ovarian cancer, the highly fluorescent product HMRG was produced within 1 minute after topical spraying of gGlu-HMRG onto tissue surfaces with a high signal contrast between the tumor and the background. Use of the gGlu-HMRG probe can aid surgeons in detecting cancerous nodules by enabling accurate biopsy and tumor resection as well as delineating the tumor borders to ensure complete tumor removal.26

Although research is promising, several hurdles must still be addressed before optical molecular imaging is ready for clinical use. With regard to the optical molecular probes, the greatest problem to overcome is the clinical approval of fluorochromes and conjugates that have been used as optical molecular imaging probes in animal studies.

Despite the problems of integrating optical molecular imaging into colorectal cancer (CRC) clinical trials, several groups have advanced targeted optical molecular probe imaging strategies for the clinical translation of optical molecular imaging in the field of GI endoscopy in humans. Smart activatable probes are promising technologies for the optical molecular imaging of CRC. These probes are substrates of enzymes that are upregulated at the targeted mucosal region. Enzymatic cleavage of these probes leads to increased fluorescence intensity, which is inhibited by fluorescence energy transfer in the native state.27,28 Yoon et al.29,30 performed near-infrared fluorescence (NIRF) imaging using an matrix metalloproteinase-activatable probe to detect colon tumors at the early disease stage (Fig. 3). They observed diffuse inflammation within the tumors at the early stage of the azoxymethane/dextran sulfate sodium model. NIRF signal intensities were strongly correlated with each tumor stage from adenoma to adenocarcinoma. NIRF imaging also distinguished tumors from inflamed mucosa. This strategy allowed for the earlier assessment of the effects of therapy prior to the occurrence of morphological alterations in the underlying disease. NIRF imaging combined with a polymeric nanoparticle-based probe allows for enhanced tumor detection and estimates the malignant potential of the lesion. Thus, this approach could translate to improved accuracy of endoscopic diagnosis and lesion characterization without the need for invasive procedures such as biopsies.

CRC is a serious potential complication of inflammatory bowel disease (IBD), and the incidence of cancer in IBD is increasing in Korea.31 Regular surveillance colonoscopy is currently recommended for the early detection of dysplasia.32 Traditional surveillance recommendations include interval random biopsy throughout the colon because patients with IBD have a propensity for developing flat and subtle neoplasms that may evade detection and rapidly progress to advanced CRC.33,34 However, random biopsies are difficult to perform, and target biopsy is impossible because underlying inflammatory lesions are irregular. Previously, most dysplasia detected microscopically using random biopsies was thought to be invisible by white light colonoscopy.33,34 Recently, Gounaris et al.33 reported that cathepsin substrate-based probe imaging correctly identifies dysplastic foci within chronically inflamed colons. Therefore, they suggested that combined WLE and NIRF endoscopy has unique advantages that may increase sensitivity and specificity of surveillance colonoscopy in patients with colitis-associated cancer (CAC).33 In the future, molecular imaging might help in the early detection of CAC and dysplasia.35,36

Defects in intestinal barrier function are associated with IBD.37 Applying the concept that increased intestinal permeability is also a prognostic indicator of relapse, Kiesslich et al.37 reported that cell shedding and barrier loss detected by CLE predicts IBD relapse and has the potential to become a diagnostic tool for its management. In the future, molecular imaging targeting many inflammatory cells or mediators may be used to monitor disease severity and predict disease prognosis.

Molecular imaging can be applied to targeted therapies in oncology. In CRC, autonomous growth is considered a multi-step process that includes the upregulation of growth factors and their receptors such as epidermal growth factor receptor (EGFR). Cetuximab, a chimeric monoclonal antibody against the extracellular domain of EGFR, is a component of multimodal chemotherapy for metastatic CRC.38 Goetz et al.39 recently demonstrated that molecular imaging using fluorescence imaging could predict the cetuximab response. These results indicate that in the future, molecular imaging may be used for predicting patient responses to targeted therapy for CRC.

The development of apoptosis imaging agents is another translational research area in optical molecular imaging in the field of GI endoscopy. Molecular imaging of apoptosis is a useful tool for evaluating the response to chemotherapy and radiotherapy. During apoptosis, phosphatidylserine (PS) is exposed on the apoptotic cell surface. PS is a phospholipid that is normally expressed on the inner leaflet of the bilamellar cell membrane and is invariably translocated to the outer surface as an early event in apoptosis. Therefore, apoptosis imaging probes are based on PS recognition.

Using an M13 phage display, Thapa and coworkers40 identified a nine-residue phosphatidylserine-binding peptide (PSP), CLSYYPSYC. The fluorescein-labeled CLSYYPSYC peptide specifically binds to apoptotic cells, and optical imaging after systemic administration of this peptide to tumor-bearing nude mice (H460 cell xenograft model) treated with a single dose of an anticancer drug (camptothecin) revealed peptide homing to the tumor. This peptide-based specific probe has several advantages over proteins and antibodies, such as superior stability, nonimmunogenicity, low cost, easy labeling, and rapid clearance and tissue penetration. Moreover, the optical probe Cy5.5, which emits in the near-infrared range rather than the visible range, facilitates the detection of signals in vitro, ex vivo, and in vivo. Therefore, apoptotic imaging using PSP-Cy5.5 may be useful for clinical purposes.

One major challenge in radiobiology is predicting tumor radioresistance and determining normal tissue radiosensitivity for providing patients with personalized treatment. Optical imaging may more efficiently evaluate the radiation treatment response because it can be used to detect local signals from an irradiated lesion. Moreover, optical molecular imaging is more versatile than positron emission tomography molecular imaging because of its ability to intervene in the GI tract while the mucosal lesion is being identified. This is enabled because fluorescence imaging devices are coupled to the endoscopes and can provide real-time in vivo image information; thus, optical imaging can be more easily used for molecularly targeted procedure guidance.41 Optical imaging coupled with endoscopy is a good imaging modality choice for assessing the response to radiotherapy of intraluminal neoplasms, including rectal, cervical, bladder, or head and neck cancer. Using optical imaging coupled with endoscopy, the response to radiotherapy can be predicted promptly, facilitating more rapid therapeutic decisions. In this regard, the apoptosis optical imaging with fluorescence dye approach presented in the previously mentioned study may be applicable to functional endoscopy in the future.

Sumiyama et al.42 reported that they could examine enteric neuronal networks using endomicroscopy after colonic polyp removal. Although their observation suggests the future possibility of enteric neural network imaging using endoscopy, it is not clinically applicable at this time because current techniques, including CLE and AFI, have observation depth limitations.