For thousands of years, clinical decision-making during surgery has been based on human vision. Discolorations and anatomical appearances often indicate abnormalities that can be correlated to various pathologies: inflammation is red, hypoxia is blue, and jaundice appears yellow. Tactile information also offers valuable information during surgery, but it comes with limited sensitivity and specificity.
For all the versatility of human vision, however, the surgeon's eye cannot penetrate far below the tissue surface. Discolorations can be misleading due to their varying biological origins, and scant visual information exists to aid in detecting specific cells, remnants, or molecules. Interventional detection of diseases, and more particularly of cancer, still relies on crude visual markers typically present only in progressed disease. Practitioners must often rely on a postoperative histopathological analysis to confirm surgical success or failure, risking an adverse prognosis or additional procedures for the patient.
A number of noninvasive radiologic imaging technologies—including magnetic resonance imaging (MRI), x-ray computed tomography, and positron emission tomography—have been employed in diagnosis and tumor localization, to decide whether surgical resection is necessary, and to plan operations. However, during an actual procedure, the resolution, sensitivity, and specificity offered by these modalities are scarcely adequate to provide surgical vision or tactile sense for guidance. Moreover, intraoperative imaging based on low-field MRI or x-ray imaging involves bulky and expensive machinery (e.g., a movable apparatus called a C-arm) that imposes complex requirements and carries side effects, such as magnetic field compatibility and exposure to ionizing radiation. These features are not appropriate for wide dissemination in the operating room, which is why many endoscopic procedures rely on the harvesting of multiple random biopsies, whose invasive destructiveness is justified by the better account they give of the disease.
We have recently considered fluorescence molecular imaging for enhancing surgical vision and interventional procedures. The development of custom fluorescence markers has seen significant growth in the past decade, offering many alternative designs for biomarkers that can target cancer cells. To make clinical use of this potential, however, two major hurdles must be overcome. One is the development of an imaging method that can accurately and reproducibly capture the fluorescence from tissues, providing true biodistribution in real time while remaining insensitive to tissue discoloration. The other is the translation of potent chemical agents for staining human cancer cells from in-vitro to in-vivo use.
The challenges posed by these two obstacles are not negligible. We have observed that images obtained photographically, even with a sensitive camera and bandpass filter, do not always reflect true fluorochrome biodistribution. Instead, the detected intensity depends to a great extent on three independent parameters: fluorochrome concentration, optical absorption and scattering within tissues, and the depth of the fluorescence activity.1, 2 Tissue auto-fluorescence may further degrade the image by reducing contrast. The effects of these parameters vary, depending on the wavelength employed and the tissue type, and an accurate camera system needs to account for them. While auto-fluorescence can be subtracted from the signal via linear spectral unmixing, the nonlinear effects resulting from depth and the optical properties of tissues are more difficult to decompose. For example, a relatively opaque or dark lesion may mask some of the fluorescence passing through it, possibly leading to a false negative, while a light-colored lesion may result in a false positive. Meanwhile, the clinical translation of fluorescent agents presents a different set of challenges related to toxicity and the efficacy of using a new molecule in humans. For this reason, in-vivo agents have until recently been limited mostly to fluorescein, indocyanine green, and other dyes lacking specificity to cancer cells.
A paper we recently published in Nature Medicine3 demonstrated perhaps the first clinical translation of a targeted dye. The system it describes offered real-time fluorescence imaging and the ability to correct for variations in light absorption. Our clinical study was conducted on ovarian cancer patients, using an agent targeted to folate receptor alpha (FRα) and labeled with fluorescein isothiocyanate. Absent in healthy cells, FRα is overexpressed by more than 90% of epithelial ovarian cancers. The contrast agent was injected intravenously into patients previously diagnosed with ovarian cancer and scheduled for surgery. Expression of FRα was confirmed on biopsy. As shown in Figure 1, the resulting capacity to visualize cancer lesions went well beyond the ability of the human eye. Surgeons could identify five times as many cancer lesions under fluorescence guidance as they could via optical observation.
(a) Quantitative fluorescence camera placed above a patient prepared for surgery. (b) Color image from an ovarian cancer patient. (c) Fluorescence molecular image (in green) superimposed on (b). (Photos from van Dam et al.3
Fluorescence-enhanced surgical vision continues to evolve toward comprehensive systems that will offer true quantification of the amount of fluorescent agent present while being insensitive to illumination and tissue parameters. Though developed originally for surgical intervention, the technology has lately been adapted to endoscopes and surgical microscopes, promising to change the paradigm of diagnostic endoscopy. A recent focus on fluorescent agents operating in the near-infrared may permit imaging deeper in tissues. These and other developments are shaping and directing the future of surgical procedures by delivering improved sensitivity and accuracy in tumor delineation (i.e., circumferential resection margin), locoregional involvement, and lymph node interrogation. Using accurate fluorescence imaging systems and validated fluorescent agents, the results are expected to be comparable from lesion to lesion, patient to patient, and overall in multi-center clinical trials, resulting, we hope, in clinical approval of fluorescence molecular imaging.
The authors acknowledge financial support from the Bundesministerium für Bildung und Forschung.
Vasilis Ntziachristos, Jürgen Glatz
Vasilis Ntziachristos, PhD, is a professor and the director of the Institute for Biological and Medical Imaging, Helmholtz Center Munich and Technical University of Munich.
1. V. Ntziachristos, Going deeper than microscopy: the optical imaging frontier in biology, Nat. Methods 7
, no. 8, pp. 603-614, 2010. doi:10.1038/nmeth.1483
2. V. Ntziachristos, J. Ripoll, L. Wang, R. Weissleder, Looking and listening to light: the evolution of photonic imaging, Nat. Biotech. 23
, no. 3, pp. 313-320, 2005. doi:10.1038/nbt1074
3. G. van Dam, G. Themelis, L. M. A. Crane, N. J. Harlaar, R. G. Pleijhuis, W. Kelder, A. Sarantopoulos, J. Bart, P. S. Low, V. Ntziachristos, Intraoperative tumor-specific fluorescent imaging in ovarian cancer by folate receptor-α targeting: first in-human results, Nat. Med.
17, pp. 1315-1319, 2011. doi:10.1038/nm.2472