Patent Application
20050288573
This invention relates generally to detecting melanoma using electron paramagnetic resonance.
Cutaneous malignant melanoma incidence continues to rise, with the American Cancer Society estimating 55,100 new cases of melanoma in 2004. Of these cases, about 7910 melanoma deaths will occur. The economic and emotional costs of this disease are large, with just the direct costs of treating newly diagnosed melanoma cases in 1997 alone being estimated at $546 Million. Furthermore, it is still unclear even exactly how sunlight exposure is linked to melanoma and these uncertainties have made it difficult to offer precise melanoma prevention recommendations. It seems certain, therefore, that the prevalence of melanoma, and the many costs associated with this disease, is high for many years to come.
One of the keys to improved melanoma outcomes is improved detection and diagnosis, so that suspect lesions are clinically evaluated better and earlier. However, despite continual advances, the overall diagnostic accuracy for melanoma has remained much lower that desirable for a disease with such high potential mortality rates, varying from 64% to 80%. While the index of suspicion for pigmented skin lesions remains high, lowering the numbers of false negatives (i.e. untreated melanomas) these will still occur, and the high numbers of false positives are associated with significant costs including unnecessary procedures, emotional distress and scarring etc. The sub-optimal accuracy of conventional diagnostic techniques has prompted the successful development of various imaging modalities for use in detecting and diagnosing melanoma, an area that has recently been reviewed very comprehensively.
These imaging techniques for detecting, diagnosing and staging melanoma are predominantly based upon optical techniques, as the melanin pigment in melanoma provides an inherent high contrast. However, a major limitation of optical techniques is that they cannot image deeply into the skin, as a consequence of both light scattering and absorbance, especially in highly pigmented lesions, and are limited to maximal depths of 0.2 to 0.3 mm. This is a major disadvantage, as tumor thickness is the single most powerful staging factor, and predictor of survival, with the important range of tumor thickness allowing this diagnostic prediction being from 0.5 to 10 mm. Therefore, this most important parameter is currently only available after excision. Although other imaging modalities such as computed tomography (CT) scanning, ultrasound, and magnetic resonance imaging (MRI) can image deeper melanomas, they are severely limited by lack of soft tissue contrast and/or melanoma specificity, in addition to a range of technique-specific technical difficulties. In addition to its optical properties, melanin contains a high concentration of stable free radicals formed during oxidation of its aromatic amino acid side-chains, so that Electron Paramagnetic Resonance (EPR) spectroscopy of these radicals is a useful technique for characterizing melanin.
The following publications provide additional information:
All publications listed above are incorporated by reference herein, as though individually incorporated by reference.
Embodiments, aspects, advantages, and features of the present invention will be set forth in part in the description that follows, and in part will become apparent to those skilled in the art by reference to the following description of the invention and referenced drawings or by practice of the invention. The aspects, advantages, and features of the invention are realized and attained by means of the instrumentalities, procedures, and combinations particularly pointed out in these embodiments and their equivalents.
FIGS. 4A-B show an EPR imaging of Xiphophorus melanoma.
FIGS. 6A-B show M. domestica (shaved) at months after a 2-stage DMBA application.
FIGS. 7A-C show H&E stained section of focal melanocytic hyperplasia and melanoma lesions in M. domestica treated with the 2-stage DMBA protocol.
FIGS. 9A-B illustrate a surface coil resonator.
FIGS. 10A-B show a whole body resonator.
FIGS. 14A-B illustrate in vivo EPR instrumentation for measurement of UV induced radicals.
In the following detailed description, reference is made to the accompanying drawings which form a part hereof, and in which is shown by way of illustration specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. Other embodiments may be utilized and structural, logical, and electrical changes may be made without departing from the spirit and scope of the present invention. The various embodiments are not necessarily mutually exclusive, as some embodiments may be combined with one or more other embodiments to form new embodiments. The following detailed description is, therefore, not to be taken in a limiting sense.
In an embodiment, electron paramagnetic resonance spectroscopy is used to provide a signal from melanin to image a melanoma, to detect metastatic melanoma in a sentinel lymph node, to measure light penetration in melanocytes in skin, or combinations thereof. In various embodiments, EPR techniques may be applied in vivo including in vivo techniques to detect primary cutaneous melanotic lesions.
Electron Paramagnetic Resonance is a spectroscopic technique that sensitively and specifically detects free radicals. EPR is analogous to nuclear magnetic resonance (NMR), using a magnetic field and radiofrequency source to sensitively and specifically detect free radicals. In an EPR procedure, a sample is placed in a magnetic field and the absorption of applied radio frequency (RF) energy by unpaired electrons of free radicals is detected. The EPR technique has recently been extended by instrument development to allow in vivo spectroscopy and even (in favorable cases) imaging of free radicals. These techniques have been used in vitro and in vivo, and in studies of free radical biochemistry in skin tumor promotion. Importantly as a major advantage, EPR directly and non-invasively detects, quantifies, and identifies reactive free radical intermediates, as compared to product studies where the reactive intermediates must be inferred and the sample processed. In an embodiment, EPR may be used to detect free radicals such as melanin radicals and ascorbyl radical both in vitro and in vivo.
Electron paramagnetic resonance spectroscopy unit 110 and analyzing unit 120 may be configured to provide an image of a melanoma at various depths from a skin surface. The image of a melanoma may include a depth of 0.5 mm to 0.10 mm from a skin surface. The image of a melanoma may include a depth of 0.10 mm or more from a skin surface. Electron paramagnetic resonance spectroscopy unit 110 and the analyzing unit 120 may be configured to determine the status of a sentinel lymph node with respect to metastatic melanoma. The determination of the status may be made in less than about 20 minutes. Apparatus 100 may be used to examine excised tissue or tissue in vivo.
Various embodiments may be used to image a melanoma using an EPR signal from melanin endogenous to the melanoma. In an embodiment, in vivo EPR techniques at 1 GHz frequencies and advanced surface and/or whole body coils may be used to provide an optimized combination of RF tissue penetration and sensitivity for studying free radicals in vivo. Such tissue penetration may be approximately 10 mm. Since melanin is already widely used optically as an endogenous contrast agent, an embodiment of the invention provides in vivo EPR techniques to detect and image even thick melanoma, using the inherent molecular contrast of melanin to allow rapid, noninvasive melanoma imaging, staging, and assessment of local lymph node involvement. In an embodiment, 1 GHz EPR spectroscopy with surface and whole body coils may be used to detect the melanin EPR signal in melanotic lesions. Both the premelanoma lesion focal melanocytic hyperplasia (FMH) and malignant melanoma may be observed, allowing for both early and later tumor detection.
According to an embodiment of the inventive subject matter hereof, EPR imaging may be used to image the depth of FMH and primary melanomas. Magnetic field gradients may be applied to allow spatial imaging of melanoma depth and margins, and correlate EPR images with histological measurement. According to an embodiment of the inventive subject matter hereof, EPR ‘virtual biopsy’ of lymph nodes or other tissue may be performed.
EPR imaging may allow 3-dimensional imaging of even thick melanomas, and can easily be linked to an optical imaging modality (e.g. dermoscopy or multispectral imaging) for coregistration and multimodal imaging. The molecularly based, high contrast imaging of melanoma for determination of thickness providing accurate initial staging of the disease can assist clinical choice of treatment options. Furthermore, it may allow accurate determination of melanoma morphology and extent in 3 dimensions, enabling resection margins to be accurately determined in a lesion-specific, rather than empirical manner. The detection of clusters of melanoma cells present in surrounding lymph nodes providing ‘virtual biopsy’ may also be very useful, as this is another powerful predictor. According to an example embodiment, tests of the methods hereof are tested on an appropriate animal model or using clinical human studies, where the utility and effectiveness of the methods are tested in human melanoma.
Tumor depth from 0.5 to 10 mm is the best staging and prognostic factor, while regional lymph node metastasis is also a major determinant of outcome, so determination of these factors is important. However, optical imaging techniques, though providing high contrast, are limited to a superficial depth range of 0.2 to 0.3 mm, and more penetrating imaging modalities, such as MRI, ultrasound, and near infrared (NIR) optical coherence tomography, have very low contrast properties for melanoma, as melanin is not well detected by these techniques. In an embodiment, EPR imaging may both detect and image melanoma at depths of 10 mm or greater, with high and molecularly specific contrast, due to its specific detection of melanin radicals.
Melanin comprises an average of about 1.5% wet weight of human melanoma tissues (with a 3:1 ratio of eumelanin to pheomelanin), and so represents a significant component of these tumors. Furthermore, melanin contains a large concentration of stabilized free radicals that are formed during its polymerization, so that EPR is an important technique for studying melanins. The EPR signal of melanin has been used to study photobiology in excised skin samples. EPR spectroscopy has also been used to study the melanin content of excised tissues. The potential for using the EPR signal of melanin to image melanoma was recognized some time ago. It is important to note that this was an extremely rudimentary experiment upon only the surface of an excised tumor, with somewhat questionable data. All these experiments were performed on isolated excised tissues, using ‘conventional’ X-band EPR techniques operating at approximately 9 GHz. However, the penetration of such microwave frequencies in ‘lossy’ aqueous samples is limited to approximately 0.1 mm, and this is why excised samples were used. Therefore, such instruments cannot either provide images of melanomas any deeper than approximately 0.1 mm, nor can it operate in situ. These and other technical details have prevented any useful development.
In the early 1990's several groups began using EPR spectrometers operating at lower frequencies to begin to study paramagnetic species in intact organisms, and this has developed into a very useful approach. Different groups have optimized instruments at different frequencies depending upon sample needs, but the L-band frequency range centered at approximately 1 GHz, developed by Walzcak and Swartz at Dartmouth and others has proven optimal for studies of tissues at depths up to about 10 mm. Several innovations, including advanced surface coil resonators, autotune-automatching microwave bridge-resonator combinations, and the development of magnetic field gradient coils and amplifiers for imaging, have made the instrumentation useful for in vivo and imaging studies. The improvements in sensitivity of these in vivo low frequency spectrometers have also extended the range of radical species that can be studied. In an embodiment, these developments in sensitivity have been sufficient to enable detection, and even imaging of the melanin in melanoma tumors in intact, non-excised samples.
Melanoma thickness is the single strongest parameter for staging and prognosis. Large studies have convincingly shown that tumor thickness is the single most powerful staging parameter and predictor of survival in localized melanoma disease. A continuous relationship between tumor thickness and survival has been shown, with the important range of thickness being 0.5 to 10 mm. This importance of depth is evident in a recent American Joint Commission on Cancer staging system. Conventional measurements of thickness prior to lesion excision are based upon determining lesion thickness from a biopsy, and so the entire tumor is not assessed for thickness. Although the lesion is palpated to choose a suitably thick biopsy site, it remains a point sample, and so may not accurately represent the global deepest penetration of a tumor. Likewise, after resection, the serial sectioning of the entire excised tumor and pathological examination of each section is currently the only way to ensure the maximum thickness in the entire tumor is measured. Yet, this can not usually be performed. In an embodiment, EPR imaging, as an imaging technique that globally and non-invasively may determine tumor thickness throughout the tumor, can offer significant advantages and may also be used to guide a point biopsy to the area of maximal tumor thickness. Thus, histopathology at the deepest and most invasive site may be determined. Additionally, in laterally invasive disease, a biopsy is guided to a similar site of maximal lateral invasion, or a resection margin determined directly according to the tumor, as opposed to margins that are more empirically derived.
Optical techniques are generally limited to superficial depth, while other techniques lack contrast. The majority of developing techniques for imaging and diagnosing melanoma have been based upon various optical techniques for both image acquisition and analysis, such as multi-wavelength analysis or artificial intelligence for diagnosis. However, a major limitation of these optical techniques is that they cannot image deeply into the skin, due to both light scattering and absorbance, especially in highly pigmented lesions. For example, dermoscopic, multispectral and confocal techniques are limited to maximum depths of 0.2 to 0.3 mm. Optical coherence tomography, using longer wavelength near IR light, has maximum depths of 1 mm but results in loss of melanin contrast.
This inability to detect and resolve non-superficial melanoma is a major disadvantage of optically-based techniques, as tumor thickness is the single most powerful predictor of survival in localized melanoma disease. The range of tumor thickness allowing this diagnostic prediction is in the range of the range 0.5 to 10 mm. Ultrasound techniques show some promise, although there is again a lack of contrast between melanoma and surrounding tissue, and under- and overestimates of thickness are common. This limits its primary usefulness to preoperative planning of diagnosed melanomas. MRI of melanoma lesions shows potential, with the development of high resolution surface coils, although its contrast specificity for melanoma remains low, and it is again primarily used for preoperative planning of pre-diagnosed melanomas. In an embodiment, EPR imaging, as a technique of melanoma imaging that has the ability to image deep into the skin to a depth of approximately 10 mm and that uses melanin as a high sensitivity endogenous contrast agent, may provide an advance in the field.
Existing developments in EPR technology may allow rapid clinical applications, at reasonable cost. Some considerations when developing new medical technologies are: a) the length of time to develop this nascent technology into a truly clinically-available application so that clinical trials can be performed, and b) whether this technology ultimately can be realized at a cost that can be justified by the benefits it brings. The NIH-funded in vivo EPR Center at Dartmouth has developed whole-body EPR magnets and instrumentation for clinical use and is currently performing clinical EPR trials such as trials of EPR oximetry. Overall, several of the key potential limitations to clinical and commercial development have been, or are being addressed.
Regional sentinel lymph node metastasis is an additional major determinant of outcome. The unfortunate propensity for melanoma to metastasize is primarily responsible for the mortality caused by this cancer. This metastasis occurs via the lymphatic drainage, and so local lymph nodes are usually primary initial sites for metastasis. Elective (total) lymph node dissection (ELND) in the local field is useful in staging, but does not significantly impact upon outcome in metastatic disease, and is associated with significant morbidity, except perhaps in selected subsets of patients. Recent efforts have focused upon minimally invasive identification of the first lymph node into which the melanoma drains—the ‘sentinel’ lymph node (SLN), followed by its excision and analysis. Although there has been some controversy, it has been broadly documented that this is a useful procedure, minimizing morbidity of ELND, while retaining the staging power of knowledge of metastasis and having been shown to especially be useful in melanomas with thickness greater than 1 mm. In thin melanomas (less than 1 mm) however, its overall cost-effectiveness is low, due to lower rates of metastasis in thin lesions. Overall, although controversial, SLN analysis shows significant promise. However, the intraoperative use of frozen section analysis of the SLN is not recommended because it is much less sensitive than conventional immunohistochemistry of fixed sections. Furthermore, the use of Positron Emission Tomography (PET) to detect and image metastasis only has a 14-22% detection rate. Even conventional post-operative sectioning has disadvantages, in that typically only a few sections are examined as total sectioning of every lymph node and pathological analysis can be inordinately expansive.
Because of these inherent disadvantages of current techniques, it can obviously be highly advantageous to have a sensitive technique to detect the presence of melanoma cells within the SLN at the time of operation, so as to guide surgical decisions at the time of surgery, as opposed to requiring subsequent procedures. Embodiments of EPR analysis of the SLN may allow a sensitive intraoperative determination of melanoma metastasis into the SLN, noninvasively so that subsequent pathology is not compromised. In an embodiment, evident lymph nodes may be analyzed in situ, without requiring dissection, in effect providing a ‘virtual biopsy’. In an embodiment, EPR detection of melanoma metastasis in lymph nodes may be performed such that the entire node is sampled at once, and then if desired, either imaged or excised.
In various embodiments, electron paramagnetic resonance spectroscopy may be used to detect metastatic melanoma in a sentinel lymph node. In addition to its visual properties, melanin present in melanocytes is paramagnetic, and so has properties that can be detected by various EPR embodiments. In an embodiment, these properties to very may be used sensitively to detect any melanoma metastasis in lymph nodes or other tissue, to allow accurate efficient and timely analysis of the metastatic status of this tissue. The sentinel lymph node (SLN) is the first node in the local lymphatic basin to which the tumor drains, and its status is thus uniquely and sensitively indicative of lymph node metastases. Accordingly SLN status is one of the most powerful predictors of the outcome of patients with stage 1 or II melanoma. The status of the SLN is also valuable in planning for melanoma patients, as both conventional and developing treatments are offered according to the cancer stage. SLN status is a major determinant of the cancer stage.
In an embodiment, EPR detection of melanoma metastasis in excised tissue may be performed. The EPR signal of melanin in any metastatic melanoma in the SNL, or indeed other tissue, can be detected on excised tissue. The status of an excised tissue may be accurately determined during the course of an operation, and the information given intraoperatively to the surgeon. The time taken would be less than about 20 minutes. The sensitivity may be very high, depending upon the microwave frequency band chosen for the EPR procedure. In an embodiment, specific properties of the EPR signal generated using EPR spectroscopy may be detected to unambiguously confirm that the signal is that of melanin. Such specific properties may include, but are not restricted to g value, linewidth and shape, power saturation properties, T1 and T2 of the EPR signal etc. The tissue is not physically changed by this process, so that it can be analyzed by conventional histopathology afterwards. Thus, the surgeon, knowing the status of the sentinel lymph node, could elect to not proceed with regional dissection if there is no overt disease, or, if the SNL is positive, continue with further dissection until the removed nodes are no longer positive. In an embodiment, an expert database of these properties may be constructed so that the EPR-based device automatically assesses the signal itself, without the need for an operator expert in EPR analysis, and gives the operator a readout with confidence intervals.
Depending on the microwave frequency used and the size of the tissue, it may be necessary to freeze the sample to prevent non-resonant absorption of microwave energy. Such knowledge is apparent to one skilled in the art. Thus a lymph node of mass 200 mg could be frozen and analyzed whole in an X-band spectrometer, and many other configurations are possible.
In an embodiment, EPR detection of melanoma metastasis in tissue in situ may be performed. Instead of excising the tissue, low-frequency EPR techniques may be used to assess the status of these tissues in situ. Although there may be slightly lower sensitivity than some detection protocols of excised tissues, this in situ procedure may be counterbalanced by the non-invasive nature of the procedure. The EPR detection scheme would be very similar, however, to a procedure excising the tissue. Other clinical implications and uses of the knowledge of lymph node status are also possible.
Various embodiments using EPR spectroscopy may be used for molecular assay of the UVA protective efficacy of sunscreens. Prevention by avoiding solar light and tanning booths, and use of sunscreen is central to public health programs. However, melanoma, unlike other skin cancers, is linked to longer wavelength UVA irradiation. The conventional SPF of a sunscreen relates to how well it protects against sun induced erythema, and therefore only measures UVB, not UVA. Yet, being able to measure a ‘UVA-SPF’ is very important in sunscreen design and testing. Some optical techniques have been described, but these have a number of disadvantages, both technically and in their conceptual development.
In an embodiment, electron paramagnetic resonance spectroscopy may be used to measure a UV induced radical signal from melanin to measure light penetration in melanocytes in skin. The effectiveness of a sunscreen may be determined to protect against melanoma based on the UV induced radical signal from melanin. The effectiveness of UVA sunscreens may be directly and accurately measured by its ability to reduce the UVA increase in melanin radical formation. A ‘melanoma protection factor’ or MPF mathematically calculated factor may be assigned to the sunscreens.
In addition to its optical properties, melanin contains a high concentration of stable free radicals formed by oxidation of its aromatic amino acid side-chains that may be detected by EPR techniques. Importantly, and central to its role in melanoma causation, it has also been shown in vitro that illumination of melanin by UV and blue light generates further meta-stable melanin radicals, causing oxygen consumption and also release of reactive oxidants such as superoxide (and hence hydrogen peroxide and potentially hydroxyl radical,) with essentially identical action spectra. Importantly, this suggests that these light-induced melanin radicals could be an appropriate surrogate measure for oxidant release into the melanocyte, and hence the potential of light to cause melanoma. The observation of oxidative DNA damage by UVA irradiation of melanin and precursors in vitro and in melanocytes bears witness to the importance of this process. These meta-stale light-induced melanin radicals have been observed by EPR in UV irradiated intact pigmented rabbit skin. This observation prompted a test as to whether an action spectrum of melanin radical formation could be generated in situ in the melanocytes of Xiphophorus skin, thereby gaining a deeper understanding of factors that result in the action spectrum of melanoma causation. As will be seen, the correlation between the two was in fact great, suggesting that oxidants formed in melanocytes by the action of UVA upon melanin playa central role in melanoma causation.
The measurement of UV induced radical signal of melanin to measure light penetration into the melanocytes in skin may be achieved in human skin using low frequency equipment and surface resonator. The received UV doses may be smaller than those required for other endpoints such as erythema, although extended illumination to test photostability is possible too. All skin types can be studied, and if required, the external loop can be centered on nevi (moles) to see how the sunscreen protects against UVA in these important sites too. Furthermore, the measurement is absolute, and hence free of potential bias that other visually scored tests, such as persistent pigment darkening (PPD), work upon. In addition to its use in humans, embodiments of the technique may also be applied to a wide range of animal models of melanoma, and the MPF of sunscreens calibrated against their ability to protect against melanoma, a major advance.
In an embodiment, EPR spectroscopy may be uniquely able to non-invasively determine and quantify these radical species in melanocytes in situ, in intact skin. In an embodiment, the response of non-excised, human skin may be determined, especially in melanocytic nevi prior to excision. Thus, direct comparisons between measurements made in models and humans may be possible. In an embodiment, electron paramagnetic resonance spectroscopy to measure a UV induced radical signal from melanin to measure light penetration in melanocytes in skin may be used to formulate a sunscreen having a high melanoma prevention factor.
Although the EPR apparatus is currently expensive, it is not exorbitant, and the manufacturers have experience in adapting the machinery to a specific protocol in EPR dosimetry of ionizing radiation. Volume savings from selling many of these e-scan machines have reduced the price considerably.
In various embodiments, electron paramagnetic resonance spectroscopy may be applied to animals and animal models to optimize in vivo EPR techniques to detect primary cutaneous melanotic lesions. Melanomas may be induced in the mammal Monodelphis domestica by topical 7,12 dimethyl(α)benzanthracene (DMBA) application, and 1 GHz EPR spectroscopy with surface and whole body coils may be used to detect the melanin EPR signal in melanotic lesions. Both the premelanoma lesion focal melanocytic hyperplasia (FMH) and malignant melanoma may be observed in this model, allowing determination of suitability for both early and later tumor detection. Using this melanoma model, magnetic field gradients may be applied to allow spatial imaging of melanoma depth and margins, and correlate EPR images with histological measurement. This use of EPR imaging may be conducted to image the depth of FMH and primary melanomas. In an embodiment, application of EPR to animals may be used to validate the EPR ‘virtual biopsy’ of lymph nodes. Known volumes of FMH lesions may be implanted post mortem at differing depths into tissues to establish the minimum detectable lesion sizes in deeper tissues.
The Monodelphis domestica melanoma model will allow testing of imaging and detection hypotheses. Although the Xiphophorus hybrid fish melanoma model possesses several advantages, and was used in initial method validation studies, the phylogenetic distance between this species and man is compounded by the very different skin structure, making this a suboptimal model for further development. A mammalian melanoma model that recapitulates many of the features of human skin and melanomas is required. The HGF/SF transgenic mouse model shows enormous promise in understanding melanoma causation, as it possesses human-like histopathology and benefits from murine models. However, HGF/SF mice do not express melanin as an albino FVB mouse background is used for their generation and propagation. This results in the formation of a melanotic melanomas (i.e. melanomas devoid of any melanin) that are not representative of human melanomas except in albinos. Their lack of melanin also precludes their imaging by EPR. Although future studies can benefit from establishing this transgenic model in a pigmented strain, the success of this approach is not guaranteed. Therefore, according to one example embodiment of the inventive subject matter hereof, it may be used and/or tested on a proven melanoma model in Monodelphis domestica, which possess several key advantages. Firstly, M. domestica is pigmented, and melanotic lesions in this model are darkly pigmented with melanin. Importantly, it is a model known to be susceptible to induction of the premelanoma lesion, focal melanocytic hyperplasia (FMH). FMH can be induced through treatment with either UVA, and/or UVB alone, but a much more rapid and efficient method is with topical 7,12 dimethyl(α)benzanthracene (DMBA) treatment. This produces multiple focal melanocytic hyperplasia lesions within 2 months, which progress to melanoma, and metastasize. Thus, this model will allow rapid production of both premelanoma FMH and malignant melanoma lesions, so that the application of EPR for the detection and imaging of these lesions can rapidly be achieved early in a scheduled research plan. M. domestica is a small laboratory animal (maximum grown weight 150-250 g) that is readily housed and bred and is used in a wide range of studies.
Melanin detection limits may be obtained with low frequency in vivo EPR techniques. In an embodiment, since melanoma tumors contain about 1.5% melanin, one use of the inventive subject matter hereof is to provide some estimate of the limits of detection of the in vivo low frequency EPR spectrometer, using purified eumelanin. Since many previous low frequency EPR spectrometers have been of relatively low overall sensitivity in comparison to the higher frequency conventional X-band machines operating at approximately 10 GHz, one goal is to make an improved sensitivity spectrometer and make certain that is has enough sensitivity. If only 30 mg of pure melanin can be detected, the method might be limited to detecting a minimum of about 2 grams of melanoma tissue, so it can be limited to quite large tumors. The eumelanin was from Sepia officinalis, a widely accepted standard for human eumelanin. Although this eumelanin does contain a certain amount of associated protein, this somewhat simplistic use of dry weight of this compound will actually underestimate the potential sensitivity.
In an embodiment, detection limits may be obtained in an animal model of melanoma. According to one example embodiment, the sensitivity limits were experimentally examined in an authentic melanoma. Although the sensitivity of this technique was obviously high for isolated melanin, and although one would not predict any major difference in the EPR signal or intensity of isolated, as opposed to in situ melanin, the sensitivity limits in an authentic melanoma may be experimentally examined. For this preliminary work, a well characterized fish model of melanoma in Xiphophorus hybrids was used, in which UV light induces locally aggressively infiltrating melanotic melanomas. Although there is obviously some phylogenetic distance between fish and humans, they are a useful model. Many features of human melanoma pathology are recapitulated in this model. Fish contain predominantly eumelanin and the relatively small size of melanomas apparent in these small fish is readily apparent by observing gross anatomy providing a further test of sensitivity.
The melanoma on the flank of the fish in
In an embodiment, EPR imaging in an animal model of melanoma may be conducted. Above, it is established that the detection sensitivity of the low frequency EPR system with the surface coil may be high and that the approach may well able to record a spectrum of the melanin in a melanoma in the Xiphophorus tumor model. A good signal:noise ratio was readily obtainable. It was also determined that it was possible to use the imaging capabilities of the Bruker E540 imaging spectrometer to image melanomas in the Xiphophorus model. It is important to note that these images were obtained using the volume resonator due to the gradient coils for the surface coil being unavailable at that time. This whole-body, volume resonator has a diameter of 3.5 cm and a length of 4.2 cm, and so the filling factor (essentially the ratio between sample volume and the active volume of the resonator) for this small sample was very poor. Despite this technical disadvantage, EPR imaging was able to successfully image melanoma in this model.
Shown in
The M. domestica model of melanoma possesses all the requirements and several advantages for study of an embodiment. Initial studies into the use of DMBA as an initiating agent used a 2-stage DMBA application protocol of 100 μl of 0.5% DMBA applied to the back, with a repeat treatment 3 days later. This resulted in extensive FMH lesions in every animal of approximately 20 to 50 lesions within 2 months that is greater and more rapid than the prevalence or multiplicity of FMH in UV treated animals.
Thus, according to one example embodiment, in vivo EPR techniques, according to example embodiments of the inventive subject matter, can detect and image even thick melanoma, using the inherent molecular contrast of melanin to allow rapid, noninvasive melanoma imaging, staging, and assessment of local lymph node involvement.
According to one example embodiment, in vivo EPR techniques are optimized to detect primary cutaneous melanotic lesions. In this embodiment, an experiment is conducted to induce premelanoma FMH and primary malignant melanoma lesions in M. domestica and use in vivo EPR with an advanced autotune/automatch surface coil resonator to detect the melanin radical signal in these lesions during their growth. The experiment may also be conducted to determine sensitivity thresholds as a function of systematic optimization of instrumental and experimental parameters.
Electron Paramagnetic Resonance is a spectroscopic technique that sensitively and specifically detects free radicals, such as melanin radicals and ascorbyl radical both in vitro and in vivo. The sample is placed in a magnetic field and the absorption of applied radio frequency (RF) energy by unpaired electrons of free radicals is detected. EPR directly and non-invasively detects, quantifies, and identifies reactive free radical intermediates, as compared to product studies where the reactive intermediates must be inferred and the sample processed. To assist in visualizing the unique configuration of in vivo EPR instruments,
An embodiment of a specific EPR methodology may be applied for in vivo studies of melanin radicals in FMH and melanoma induced in M. domestica. Shaved M. domestica is anesthetized using isoflurane, administered for induction in a suitable box, and for maintenance via a nose cone. To maintain rectal temperature at 33° C. the animal is placed on a thermostatically controlled heating pad for external-loop surface coil experiments. For whole body volume resonator use, the water cooling for the gradient coils is maintained at 33° C. A Bruker Elexsys spectrometer is operated at L-band frequencies, approximately 1.1 GHz, using a highly specialized and sensitive automatching/autotuning bridge. The static magnetic field may be provided using a 4″ air gap electromagnet, operating at approximately 40 mT (400 G) for resonance with a thermostatic hall effect probe controlling the field. Magnetic field modulation may be applied using calibrated field modulation coils that are specially designed to minimize microphonics. These are inbuilt for the whole body resonator, separate for the surface coil. Spectra are recorded using multiple-scan averaging techniques, as signal averaging is useful in in vivo EPR. Although numerically equivalent, a long-scan time single scan is susceptible to slight movement artifacts, whereas in multiple scans, any such movement is averaged to low levels. Two different resonator configurations are tested for their sensitivity and utility.
In an embodiment, an external loop surface coil resonator may be used.
In an embodiment, a whole body volume resonator may be used.
An embodiment of a methodology for optimization of melanin radical EPR signals in vivo may be performed. Due to the complex interactions of the RF and magnetic field modulation fields, microphonics can have complex effects on the noise floor, especially at the larger modulation amplitudes required to maximize sensitivity of detection of the somewhat broad (0.4 mT) linewidths of melanin. As a result, it is necessary to optimize the combination of incident microwave RF power, modulation amplitude, and modulation frequency to allow maximal RF and modulation amplitude, while minimizing microphonics. The automated multidimensional experiments possible with a Bruker Elexsys console allow a rapid ‘search’ of this multidimensional ‘noise floor space’ and greatly facilitate initial optimization. These multidimensional experiments may be automatically analyzed for signal height and signal:noise ratio. The experiments use the dimensions of incident microwave power (40, 20, 10, 5 and 2.5 mW) modulation amplitude (0.6, 0.5, 0.4, 0.3, 0.2 0.1 mT) and modulation frequency (100, 75, 50 and 25 kHz) requiring 120 different conditions to be initially tested. Since each scan is expected to take no more than 1 minute, though more like 20-40 seconds, this overall time (40-120 minutes) is within acceptable limits for an anesthetized animal preparation. Once a suite of 5-10 most favorable conditions is identified, these may be automatically programmed for rapid optimization in every subsequent experiment. However, where rigid quantitative comparison is required (for example comparing the EPR signal as a function of lesion volume or depth etc.), it is important to use identical instrument parameters. Thus, the signal from each of this ‘suite’ of 5-10 most favorable conditions is recorded and stored, allowing comparisons of the EPR signals between samples at each of these parameter combinations.
An example methodology for induction of FMH and melanoma in M. domestica is provided in the following. Treatments with 100 μl of 0.5% DMBA 3 days apart lead to the rapid development of FMH at high prevalence and multiplicity per animal, within 2 months. This protocol is used to develop FMH at high multiplicity. Male animals (2-3 months old) may be obtained from the Southwest Foundation for Biomedical Research, housed, and fed as previously detailed. They may be anesthetized with isoflurane and dorsal hair removed with clippers. Groups of animals may be treated with 2 treatments of 100 μl of 0.5% DMBA in ethanol, 3 days apart. These animals may be shaved weekly with a Remington microscreen shaver and FMH and cutaneous malignant melanoma prevalence and multiplicity noted. The back of each animal may be digitally photographed at 3 megapixel resolution and data archived for any retrospective analysis as may be required. These groups of animals may then have their lesions used for EPR studies as detailed subsequently. Experience shows, in some cases, that pigmented lesions that are not raised are usually found to be FHM upon histological examination, whereas those pigmented lesions that are raised (i.e. vertically invasive) are usually found to be melanoma on histology. Thus, all imaged lesions are so classified initially, but confirmed by histological analysis.
Experimental EPR detection and optimization of both FMH and primary malignant melanoma may be conducted. Two groups of animals may be treated with the DMBA protocol. The general in vivo EPR detection methodology with respect to previously described specific EPR methodology for in vivo studies of melanin radicals in FMH and melanoma induced in M. domestica may be used to detect the melanin EPR signal. Both the external loop surface coil and the whole body volume resonators may be used. The sample access size may preclude measurement of animals greater than approximately 6 months. So, the sample imaged may be the entire back of the animal after the gross dissection necessary before lesions are removed for histological and morphometric analysis. The methodology for optimization of melanin radical EPR signals in vivo discussed previously may be applied in each of these animals. Should it be apparent after at least half the group has been analyzed that use of the limited suite of 5-10 most favorable conditions covers all useable options, then the suite will be done. The animal will then be euthanized. The studied lesion(s) may be photographed, then carefully excised, and processed for morphometric and histological analysis. The optimized EPR signal intensity may then be correlated with lesion volume and depth.
Two groups of animals may be used. Group a may be categorized for EPR detection of early FMH lesions. A group of 10 animals may be used, as the high prevalence and multiplicity of FMH will allow study of at least one lesion in each animal. The animals may be studied at 2-3 months after DMBA treatment.
Group b may be categorized for EPR detection of malignant melanoma and late FMH lesions. A group of 20 animals may be used, as there is a lower prevalence of malignant melanoma, and not all animals are expected to progress to this end point. The animals may be continuously monitored and animals demonstrating primary malignant melanoma by gross examination and palpation (approximately 3-6 months after DMBA treatment) may be studied. At 9 months all non-primary malignant melanoma-bearing animals may be studied and euthanized with melanoma confirmed histologically.
Data analysis and correlation of EPR detection with lesion dimensions and histopathology may be conducted. Each EPR studied lesion may be excised, formalin fixed, grossly sectioned to a margin of 0.5 mm of normal tissue at the lateral margins of the lesion, and paraffin embedded using standard procedures. Transverse 10 micron thick sections may be made in triplicate every 0.5 mm across the diameter of the lesion. The area of melanotic tissue and the depth to which it extends may be determined using an Olympus BH2 upright microscope interfaced with the quantitative image analysis programs ImagePro Plus and Metamorph. The lesion volume may be then be determined by interpolating area between these transverse sections as a function of their section depth using the computer software. Maximal, median, and mean lesion depth may be determined. The optimized EPR signal intensity may then be correlated with lesion volume and depth to determine what the smallest detectable lesion is and also to determine how lesion depth affects signal intensity per unit lesion volume.
With respect to lesion multiplicity, it is anticipated that FMH is induced at high prevalence and multiplicity by the 2-stage DMBA protocol. Should lesion prevalence be unexpectedly high, so as to preclude the discrete lesions seen in the M. domestica model of melanoma discussed previously, group a will be repeated with half the DMBA concentration before progressing to group b. With respect to lesion sectioning, for very small diameter lesions (less than 2 mm in any dimension) the step of 0.5 mm between sections detailed above is too large for accurate volume determination, and steps of 0.25 mm may be used. Statistical analysis may be applied. Linear regression (Sigma Stat 3.0) analysis may be used to study EPR signal intensity as a function of lesion volume, lesion mean, median and maximal thickness, and the ratio of volume to thickness. As all these are ratio variables, linear regression is well justified. Although multiple regression analysis (by hierarchical stepwise regression) may also be performed, it is suspected that the use of relatively simple and robust analysis at this initial stage provides guides as to which variables are introduced in the hierarchical stepwise regression analysis. There is expected a high positive correlation coefficient between lesion volume and EPR signal intensity, simply as more melanin is present in a larger lesion. Also anticipated is a moderate correlation between signal intensity and ratio of depth to volume, as deeper melanin will have slightly less signal per amount than more superficial melanin, as there is some limitation of penetration of RF to deeper tissues.
At the conclusion of these experiments, the relevant EPR parameters may be optimized for in vivo measurements of melanin in melanotic lesions. Furthermore, these experiments may provide an understanding of the interplay between signal intensity, lesion volume, and lesion depth, so that one is able to accurately estimate the minimum detectable sizes of lesions at different depths, of particular relevance to the feasibility to EPR virtual biopsy of lymph nodes.
According to another example embodiment, EPR imaging may be used to image the depth of FMH and primary melanomas. In an embodiment, experiments may be performed to EPR image FMH and primary melanoma lesions, using both surface coil and whole body resonators, and to correlate the EPR images with histopathological assessment of lesion morphology and dimensions. The experiment may also be performed to critically evaluate the degree of correlation between EPR and morphometric analysis.
The field of EPR imaging, although well established by the early 1990's has undergone considerable recent progress for biomedical applications, as is exemplified by work of several groups worldwide. As with MRI, the spatial localization of spins is encoded by the application of magnetic field gradients, although the pulsed gradient sequences used in MRI are not generally applicable due to the much shorter relaxation times of electronic, as opposed to nuclear spins. Thus, images are obtained by obtaining projections over a swept range of magnetic field gradient angles and using filtered backprojection to produce the final image. The magnetic field gradients required are produced by anti-Helmholtz and/or figure of eight coils as described by Eaton (See, Eaton, G., Eaton, S. & Ohno, K. Epr Imaging and in Vivo Epr, CRC press, Boca Raton, 1991). High currents (2 to 20 amps) and continuous duty operation cycles usually necessitate air or water cooling. The maximal theoretical resolution is a function of EPR linewidth, attainable gradient strength, and the number of projections, although signal:noise ratios are also a factor.
An example of a specific EPR imaging methodology for M. domestica FMH and melanoma is provided in the following example. A Bruker E540 imaging EPR spectrometer possesses advanced software control of all image acquisition, processing and image analysis routines through proprietary XEpr and XView software. The calculated resolution as a function of linewidth, gradient strength, and number of projections is automatically calculated. The magnetic field gradient projections required for any given image are automatically determined and achieved through a dedicated hardware gradient controller and 3 individual current amplifiers for X, Y, and Z gradient coils capable of continuously supplying 20 amps at 35 volts in each axis. A whole body, volume imaging resonator contains water-cooled gradient coils that provide gradients in three orthogonal axes of up to 0.6 T m−1. The in-house water cooled gradient coils for an external loop surface coil resonator are capable of gradients in excess of 0.16 T m−1. An imaging protocol may use the optimized signal parameters from the methodology for optimization of melanin radical EPR signals in vivo previously discussed. After acquisition of an imaging dataset, the projections may be deconvoluted using a non-gradient EPR spectrum and filtered backprojection performed. Identical deconvolution parameters may be used to allow meaningful comparisons between gradient strengths.
An example experimental EPR imaging of both FMH and primary malignant melanoma may be performed. Two groups of animals may be treated with a DMBA protocol and the melanin EPR signal in the resultant melanotic lesions may be imaged. Both the external loop surface coil and the whole body volume resonators may be used to determine which provides optimal images. The surface coil experiments may be performed first, followed by the whole body resonator. Animals greater than age 6 months may not fit in the whole body resonator, and so the excised back of these animals may be imaged in this resonator after their in vivo study with the surface coil. The experiments may initially evaluate the use of 4 different gradient strengths, equivalent to 1, 2, 3 and 4 times the intrinsic linewidth per cm, (i.e. 0.4, 0.8. 1.2 and 1.6 mT cm−1) to determine the optimal resolution achievable within useful signal:noise. As an example, the EPR image in
The two groups of animals may be organized as group a and group b. Group a may be categorized as EPR imaging of early FMH lesions. A group of 20 animals may be used. In order to fully assess the correlation between EPR imaging and histological morphology, two sectioning protocols are necessary. The lesions are embedded with half of the lesions sectioned perpendicular to the uppermost surface of the lesion, i.e. perpendicular to skin surface, as shown in
Group b may be categorized as EPR imaging of malignant melanoma and late FMH lesions. A group of 20 animals may be used, as there is a lower prevalence of malignant melanoma, and not all animals are expected to progress to this end point. Depending upon final prevalence, perpendicular sectioning, as in
Data analysis and correlation of EPR imaging with lesion morphology and histopathology may then be conducted. As detailed above, all lesions that have been EPR imaged will have been marked with a surgical ink tattoo, and then photographed to allow co-registration of the EPR image with the histopathological and morphological analysis. The lesions may then be excised, formalin fixed, grossly sectioned to a margin of 0.5 mm of normal tissue at the lateral margins of the lesion, and paraffin embedded using standard procedures. The lesions may be physically sectioned in two different planes as detailed above with respect to FIGS. 11A-B.
The morphology of the lesions may then be determined using the same protocol and software as previously discussed with respect to data analysis and correlation of EPR detection with lesion dimensions and histopathology. Since the thickness of melanoma lesions is so central to the clinical outcomes observed, a critical evaluation of the cross sectional area and thickness of melanotic lesions in histological sections obtained perpendicular to the lesion, as in
With respect to numeric analysis of lesion morphometry, linear regression may be used to compare the numerical values of EPR image and histological lesion area and thickness (all ratio variables) in each pair of sections. A high positive correlation coefficient between the values obtained by EPR and histological analysis is expected. The extent of correlation between EPR and histology images may also be studied as a function of the applied magnetic field gradient strength in order to determine if correlation is maximized by any one particular gradient strength. It is anticipated that, as long as an adequate signal:noise ratio is obtained, the higher the gradient strength, the better the image resolution and the better the correlation between EPR and histological analysis. Multiple regression analysis may also be used to allow inclusion of the average signal:noise ratio of the EPR projections as a variable in the analysis of correlation and gradient strength, as it is likely that that higher gradients will result in lower signal:noise ratios. Multiple regression analysis may be realized by hierarchical stepwise regression and by Sigma Stat 3.0 software. EPR projections may be automatically calculable using the Xepr software.
With respect to objective analysis of morphology, infiltration, and pigmentation, the variables in this analysis are derived from scored scales and as such are ordinal variables, although the relatively large scale range and relatively linear assessment will allow analysis as interval variables. The correlation between EPR and histology image variables may again be analyzed as a function of applied magnetic field strength and the average signal:noise ratio of the EPR projections. Multiple regression analysis may be used including multiple regression analysis by hierarchical stepwise regression.
With respect to gradient strength optimization, should it be apparent in early images that image quality increases with magnetic field gradient strengths in the preliminary range of up to 0.16 mT cm−1, values of 0.2 and 0.24 mT cm−1 may be further added with analyze of the corresponding images. The preliminary data in the Xiphophorus model previously described with respect to EPR imaging in an animal model of melanoma indicates these vales represent useful maxima, beyond which lowered signal:noise ratios negate the improved spatial resolution. These findings may be confirmed experimentally in the Monodelphis model.
With respect to analysis of resonator suitability, after there is a good understanding of how the different experimental and image-derived variables are correlated, a sub-set of the most informative images may be analyzed as a function of the type of resonator used to determine which may have the most optimal features for clinical studies. Should there be no significant advantages in image quality, one may focus future developments upon the surface coil resonator, as this allows easy placement over the lesion of interest and also allows optical coregistration techniques via the free space in the middle of the loop.
At the conclusion of this series of experiments, there may be a rigorous and quantitative analysis of just how well EPR imaging may be able to determine the overall morphology of FMH and melanoma lesions, using histology as the ‘gold standard’. There may also be knowledge available as to how the magnetic field gradient strengths and resonator types affect the quality of the EPR images.
According to another example embodiment of the inventive subject matter, EPR ‘virtual biopsy’ of lymph nodes or other tissue may be validated. According to this embodiment, experiments may be provided to implant known small volumes of FMH lesions post mortem into tissue, and use a surface coil resonator to determine minimal detectable volumes of melanoma.
As all techniques have been previously described, these example sections are referred to for brevity. Unimaged excess FMH tissue may be harvested from sections of groups of animals for induction of FMH and melanoma in M. domestica to detect primary cutaneous melanotic lesions and from sections of groups of animals to image the depth of FMH and primary melanomas, as the high lesion multiplicity will mean not all lesions can be imaged. This tissue may be snap frozen and stored in liquid nitrogen. When needed, appropriate amounts of this tissue (˜0.5 mg, 1 mg, 2.5 mg, 5 mg, 7.5 mg, and 10 mg) after thawing may be accurately weighed. A lesion-free area of skin from the euthanized animals may be selected optically and the EPR spectra of this area may be determined with the surface loop resonator using the optimized parameter set obtained from optimizing in vivo EPR techniques to detect primary cutaneous melanotic lesions. A small flap may be raised and incision may be made at depths of either 2, 5, or 8 mm and accurately measured. The FMH samples may then be placed at the base of the incision and the flap temporarily closed with thin surgical tape. The EPR spectrum of the implanted tissue may then be obtained with the surface coil. The tape may be removed, the tissue removed, and the next largest tissue sample may be implanted. This overall experiment may be performed with 5 more replicates with analysis of the results of each. Should analysis indicate it, one may increase the number of replicates. The absolute signal intensity and the signal: noise ratio of each spectrum may be measured automatically by the Xepr software.
The signal intensity and signal: noise ratio (both ratio variables) may be analyzed as functions of FMH tissue mass and the depth to which it was implanted to determine the minimal detectable mass of FMH at differing depths. For example, a signal:noise ratio of greater than 2 may be used. This provides definitive quantitative measure of how useful this approach is to EPR virtual biopsy.
In an embodiment, EPR spectroscopy may be used in experiments on UVA induced melanin radicals. EPR may be applied to UVA induced melanin radicals in both mouse & fish skin to generate action spectra in models of melanoma causation. In an embodiment, experiments may provide EPR detection of UV-induced melanin radical species in ex vivo skin. Using Xe and Xe/Hg arc lamps with appropriate interference filters, action spectra for the generation of UV-induced melanin radical species in intact murine skin ex vivo may be generated utilizing EPR spectroscopy.
For comparison, the action spectrum for UV induced melanin radical formation is shown in
An embodiment includes in vivo, in situ detection of UV-induced radicals in skin. In vivo EPR technology has been developed by a number of groups over the last 15 years, with commercially available equipment available in the last 5 or so years. In an embodiment, through the use of a special microwave bridge operating at 1 GHz frequencies with an external loop EPR resonator, EPR detection may be performed in anesthetized, live animals, with UV irradiation via a fiber optic delivered to the skin inside the detection loop.
In various embodiments including various experiments, methods and apparatus use electron paramagnetic resonance spectroscopy to provide a signal from melanin to image a melanoma, to detect metastatic melanoma in a sentinel lymph node, to measure light penetration in melanocytes in skin, or combinations thereof. In various embodiments, apparatus may include instructions for computer control of the apparatus to perform embodiments of the functions described herein. Instructions for such automated control and analysis made reside in various forms of computer readable medium.
Although specific embodiments have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art that any arrangement that is calculated to achieve the same purpose may be substituted for the specific embodiments shown. It is to be understood that the above description is intended to be illustrative, and not restrictive, and that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Combinations of the above embodiments and other embodiments will be apparent to those of skill in the art upon studying the above description.
This application claims priority under 35 U.S.C. 119(e) from U.S. Provisional Application Ser. No. 60/576,403 filed 1 Jun. 2004, U.S. Provisional Application Ser. No. 60/609,599 filed 14 Sep. 2004, and U.S. Provisional Application Ser. No. 60/613,860 filed 28 Sep. 28, 2004, which applications are incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60576403 | Jun 2004 | US | |
| 60609599 | Sep 2004 | US | |
| 60613860 | Sep 2004 | US |
