Hematoporphyrin-mediated fluorescence reflectance imaging: application to early tumor detection in vivo in small animals
- First Online:
- Cite this article as:
- Autiero, M., Cozzolino, R., Laccetti, P. et al. Lasers Med Sci (2009) 24: 284. doi:10.1007/s10103-007-0523-1
- 54 Views
The in vivo early detection of subcutaneous human tumors implanted in small animals was studied by laser-induced fluorescence reflectance imaging (FRI), with a hematoporphyrin (HP) compound as an exogenous optical contrast agent. Tumor detection was shown to be possible just 3 days after the inoculation of tumor cells, when tumors were neither visible nor palpable. However, this detection capability is limited to a temporal window of approximately 100 h from HP administration and to a low optical contrast of the tumor (<2).
It is widely reported in the literature [1–18] that hematoporphyrin (HP) compounds accumulate preferentially in tumor tissues, although their uptake mechanisms in tumors are not completely understood.
Recently, our group [19–23] has shown that laser-induced fluorescence of HP dichlorohydrate, accumulated in subcutaneous solid tumors grown in mice, allows in vivo tumor detection. HP present in tumors and optically excited in the green (at 532 nm), emits long-wavelength light (emission peaks at 630 nm and 690 nm wavelengths) that can be captured by a charge-coupled device (CCD) camera as externally diffused light seen through a long-pass (cut on) filter. The HP targeting is tumor unspecific, simple to perform, since it needs only a systemic HP injection; in addition, HP, in the doses we used, is resistant to photobleaching and is chemically stable, non-toxic and does not show photosensitization effects in mice.
In the above-mentioned studies [19–23], the tumors were in late evolution stage, so that they were visible to the naked eye. On the other hand, since early diagnosis has become more and more a basic medical requirement, in this work we investigated the in vivo early detection of a high malignancy tumor by exploiting the performances of the ORCA camera and HP-selective uptake in tumor tissues.
Materials and methods
The experimental apparatus for measuring the fluorescence reflectance (FRI) has been described elsewhere .
In the present set-up the optical system provided a field of view of 10.2 × 7.8 cm2 so that the whole mouse body could be imaged.
We used 6-week-old Crl:CD-1 athymic nude mice (Charles River Laboratories, Calco, Italy). The tumor cell line (ARO) and its culture procedures have been described elsewhere .
We imaged ten anesthetized mice implanted with ARO tumor cells. Each tumor-bearing mouse was imaged prior to the HP injection so that we could check the absence of an endogenous fluorescence signal from the implanted tumor cells.
The fluorescence was measured on the third day after cell injection, 6 h after HP injection, and repeated daily on the successive four days. On the tenth day after cell inoculation (i.e., on the seventh day after the first HP administration), a second HP injection (with the same dose as the first one) was done, and a second set of FRI measurements were taken, identical to the first one.
The fluorescent marker, HP dichlorohydrate (Vit-porphyrin II, Teofarma, Italy), was injected intramuscularly into the right posterior leg of the mouse, on the opposite side with respect to the tumor. HP dichlorohydrate was injected into the mice as a water solution at a concentration of 4 mg HP/ml (dose of 28 mg HP/kg mouse body weight).
The measurement protocol included the acquisition of a white-light image and a fluorescence image. We obtained both these images by putting the cut-on filter (λ > 600 nm) in front of the CCD camera.
The first image showed the position and dimensions of the structures we were interested in as they visually appeared in the animal in the measurement conditions. The second image provided the map of the fluorescence emission and was registered with an exposure time ranging from 5 s to 10 s and a camera gain between 50 and 120.
Before and after each set of measurements, the image of the laser beam section on the mouse holder was acquired to evaluate its long- and short-term stability and its spatial uniformity.
Furthermore, from each image we subtracted the image of the CCD camera background noise.
The analysis of the laser beam images showed that the laser emission fluctuations were of the order of 5% and that the spatial uniformity of the laser beam cross-section was of the order of 10%.
Note that, owing to the high sensitivity of the CCD camera, the faint red fluorescence of the black cardboard (where the mouse was put) is visible outside the geometrical shadow produced by the excitation laser beam.
On the first day of measurement, the fluorescence images were already showing a large fluorescent area related to the injection site of HP and two bright spots located approximately in the region where the tumor cells had been implanted. Successively, the brightness of the tumor spots increased, and the fluorescence of the injection region spread to more and more large body areas. This gave rise to a background fluorescence that worsened the spot contrast. The second hematoporphyrin injection appeared not effective to increase the contrast once again, since the background fluorescence of newly injected HP added to the first one.
We estimated the contrast of fluorescent spots with respect to the background by considering the intensity profile along a straight line drawn across the maximum intensity pixel in the bright spot. We checked that the direction of this line did not strongly affect the calculated contrast and that the line lay in the same position with respect to the spot on the images of the same mouse on the different days. A typical line profile across the tumor region exhibited a sharp peak of fluorescence, corresponding to the tumor region, superimposed on a smooth background fluorescence profile (Figs. 1 and 2, third and fourth column). Sometimes, a steep drop in fluorescence appears, corresponding to the boundary of the mouse’s body. The optical contrast was calculated as the ratio Imax/I0, where Imax and I0 represent the maximum gray level in the spot and the background gray level at the maximum position. We estimated I0 by fitting the data points on both sides of the peak that are representative of the underlying smooth fluorescence profile.
Note that on the eight day (200 h) after the first HP administration the mouse was injected with a second dose of HP. From these plots we can see that, for both spots, the fluorescence contrast has a value of approximately 1.3, for the images taken after the first administration of HP, and decreases to 1.1 after the second administration.
The averaged contrast of all the bright regions on the mice studied showed a similar trend with respect to that we observed in the mouse described above. Indeed, the first five measurements taken after the first HP injection showed a higher contrast (mean value 1.30) than the successive four measurements taken after the second administration of the fluorescent marker (mean value 1.17). These mean values were statistically different, as was shown by a t-test (P < 0.05).
FRI is well known, and it was extensively investigated and used in the past. Nowadays, it has been ousted by more sophisticated techniques, but we think that it could be reconsidered in connection with the technological advances in the acquisition systems.
In this work, by using a low noise, high sensitivity and gain CCD camera, we performed HP-FRI measurements in vivo on mice bearing small tumors on their backs, subcutaneously implanted. The purpose of the work was to test the capability of these measurements for the early detection of the tumor and in the study of its successive growth.
The possibility of recognizing small tumors is related both to the sensitivity and the dynamic range of the imaging device and to the spatial resolution of the measurements. The detection of a tumorous area by a possible small increase in fluorescence with respect to the healthy tissue is ensured by the high sensitivity and dynamic range of the imaging device. As far as space resolution is concerned, although the device’s space resolution is of the order of 100 μm, the measurement resolution, strongly affected by the scattering of the fluorescence radiation in tissues, rises to approximately 700 μm (unpublished data, 2005). Nevertheless, we investigated very early stage tumors whose dimensions were greater than this last resolution value (i.e., of the order of a few millimeters at least). This tumor dimension was determined by the relatively high number of injected tumor cells (106) and by the relatively long time from their injection that we have to wait in order to be sure of real tumor implantation.
We showed that HP, at least for the studied cases, is a fluorescent marker suitable for detecting the early stages of tumor development.
Already on the third day after tumor cell inoculation, when the tumor was not distinguishable with the naked eye, or palpable, the region injected with the tumor cells exhibited a fluorescence 30% higher than that of the surrounding healthy tissues. This optical contrast between the tumor and the background remained constant, within the error bars, during the successive 96 h after the first HP injection, then it vanished.
To follow further tumor growth by FRI, we injected the mice’s legs with a second dose of HP, with the same modalities as the first one. In the successive 4 days we observed that the constant fluorescence enhancement of the tumor region was lowered to 17%.
After the first HP injection they show a steep increase, starting from a low fluorescence level, that can be considered approximately equal to the endogenous fluorescence level. After the second HP injection, the fluorescence intensity starts from a higher level and increases more slowly. The first difference can be imputed to the presence in the tissues of residual HP from the first injection. As far as the second aspect is concerned, we can guess that the repeated light dose administered to the same mouse produced photosensitization of the tumor cells with consequent cellular necrosis and a reduced HP uptake capability. In addition, if we compare quantitatively the time behavior of I0 and Imax, we can see that after the first HP injection, the factor by which Imax increased was much greater than that of I0. On the other hand, after the second HP injection, the difference between the two increase factors is very slight. This difference can explain the time dependence of the contrast as reported in Fig. 3.
The FRI system is simple enough to be integrated with a pixel radionuclide imaging system [24, 25], since the optical and radionuclide modality supply complementary information on the fluorescence-labeled and radionuclide-labeled biomedical targets, as our preliminary tests seem to indicate.
We used the HP-FRI technique to assess its capability to detect tumors in their early stages. The HP-FRI system, by exploiting the HP selective tumor uptake and the high sensitivity, spatial resolution and photometric dynamic range of the CCD camera, allows one to detect superficial tumors (both experimental and spontaneous) at a very early stage, i.e., when they are not visible to the naked eye. The limit of the technique is the low value of the optical fluorescence contrast (<2) that, nevertheless, allows clear tumor recognition. The detection is limited to a temporal window of approximately 100 h after HP injection and prevents the monitoring of the tumor growth for long intervals. However, these drawbacks do not prevent its use in the localization, evolution, dissemination and monitoring of surface tumors on small animals. Long-term monitoring could be obtained by administering the successive doses of HP at times much greater than the clearance time of this compound from tissues.
The authors wish to thank Mr Salvatore Sequino for the animal care.