Multispectral imaging system implementing concurrent NIR and color measurements. (A) Schematic representation of the setup. Measurements using two CCDs are under computer control. (B) Spectra of light transmittance in each imaging channel. Note the clear spectral separation in the detection channels, which was obtained by using a dichroic mirror and a set of bandpass filters. (C) Absorption spectra of two different NIR QDs. The black and red lines indicate the 760-nm-emitting QDs and the 800-nm-emitting QDs, respectively. (D) Photoluminescence intensity of QDs. The emission spectra of the QDs exactly match the spectra of the detection channels.
Optical resolution of the implemented imaging system. (A) An image of the U.S. Air Force 1951 standard resolution target. (B) Horizontal CTF. The transverse resolution of the system was measured to be by using the Rayleigh criterion (26.4% contrast). (C) Vertical CTF. The lateral resolution of the system was determined to be by using the Rayleigh criterion.
Imaging performance characterization with a phantom. All images were acquired from a phantom of three fluorescence tubes immersed in a diffusive fluid. The left tube contains 100 nM of 760-nm-emitting QDs whereas the right tube contains 100 nM of 800-nm-emitting QDs and the middle tube contains both, i.e., 100 nM of 760-nm-emitting QDs and 100 nM of 800-nm emitting QDs. [(A) and (E)]: Fluorescence images at the and the detection channels, respectively. [(B) and (F)]: Results of fluorescence normalization with the attenuation (intrinsic) image shown in (I). Note the normalized images correct for optical heterogeneity, such as a stronger illumination at the center, and better outline the extent of the tubes. [(C) and (G)]: Results of spectral unmixing for disentanglement of 760-nm-emitting QDs and 800-nm-emitting QDs. They show how the fluorescence image would have looked if it were only for 760-nm-emitting QDs (C) or only for 800-nm-emitting QDs (D) in a phantom. [(D) and (H)]: Results of combined processing of fluorescence normalization and spectral unmixing. Note that elimination of spatial heterogeneity and clear separation of overlapping fluorescence signals were successfully achieved.
Intracellular delivery of NIR QDs by electroporation. (A) Photograph of the homemade electroporation device. (B) Transmission electron micrograph of a lung cancer cell (NCI-H460) after electroporation by 18 pulses with a pulse width, a pulse interval, and a 1.3 kV/cm pulse amplitude. [(C)–(E)] Magnified views of the cell in (B). Note that successful subcellular localization of many QDs (arrows) is revealed. (F) Water-dispersed pure QDs.
In vitro long-term followup of QD retention in lung cancer cells (NCI-H460). After QD loading by electroporation, cells were cultured as a function of time. Left column: pseudocolored images of NIR fluorescence depicting QD retention and distribution. Middle column: counterstaining images with DAPI visualizing the nuclei. Right column: pseudocolored NIR fluorescence/DIC images. Note that QDs remained inside the cells for more than 1 month.
Long-term in vivo fluorescence imaging of tumor growth in the s.c. xenograft mouse tumor model. Approximately lung cancer cells (NCI-H460) were injected subcutaneously into an athymic immunodefieicnt nude mouse, and in vivo imaging was performed as a function of time by using the developed system. From the left, whole mouse images, attenuation (intrinsic) images at the excitation wavelength , original fluorescence images at the emission wavelength , and fluorescence normalized and spectrally resolved images are displayed. Corrected fluorescence images, formed according to Eqs. (5)–(7) described in the text, better delineate the tumor burden and margin because of a correction for the tissue’s optical attenuation and an elimination of the autofluorescence. Note that intracellular delivery of QDs by electroporation and use of the multispectral imaging system with a normalization and unmixing algorithm can be used to monitor tumor cell biology such as tumor development and progression in vivo.
Postresection inspection of the dissected tumor 35 days after s.c. injection of QD-labeled lung cancer cells (NCI-H460) by CLSM. (A) A NIR fluorescence image of tumor tissue. NIR QD retention inside the tumor cells is shown. Invisible NIR fluorescence is pseudocolored with red. (B) DAPI fluorescence image of tumor tissue depicting the nuclei. (C) An overlay of the NIR fluorescence and the DIC images of tumor tissue.
Intraoperative imaging of disseminated ovarian cancer with the i.p. xenograft mouse tumor model. QD-loaded human ovary cancer cells (SK-OV-3, cells) were transplanted i.p. into athymic immunodeficient mice to obtain diffused peritoneal carcinomatosis. 2 weeks after tumor cell inoculation, intraoperative imaging was performed with the developed imaging system. From the left, whole mouse images (A) and (B), color images, attenuation (intrinsic) images at the excitation wavelength , original fluorescence images at the emission wavelength , and fluorescence normalized and spectrally resolved images. Low-magnification images with skin open (C)–(F) show widespread peritoneal dissemination (solid arrows) like late-stage ovarian cancer. High-magnification images with abdominal open (G)–(J) show specific invasion of cancer cells (an arrowhead) into the peritoneum and fat. Note that the autofluorescences from the intestine (double arrows) and a false signal from lower-attenuating fat tissue (dotted arrows), which are seen in conventional fluorescence measurements (E) and (I) were significantly diminished in the corrected fluorescence images (F) and (J).
Postresection inspection of the peritoneal carcinomatosis 2 weeks after i.p. injection of QD-labeled ovary cancer cells (SK-OV-3). (A) A NIR fluorescence image. NIR QD retention inside tumor-cells is observed. Invisible NIR fluorescence is pseudocolored with red. (B) DAPI fluorescence image depicting the nuclei. (C) A NIR fluorescence/DIC image.
Imaging system specifications.
Article metrics loading...
Full text loading...