En route to a new in vivo diagnostic of malignant pigmented melanoma

Dear Editor, In the face of the continuous rise in the incidence of malignant melanoma, attention has turned increasingly to individual and public awareness for cancer prevention. One consequence is the increase of excisions of suspicious pigmented lesions. The rate of excisions per newly diagnosed melanoma reported in the literature varies at least between 1:4 and 1:179 depending on the study, the experience of dermatologists and tools such as dermoscopy (Carli et al., 2004; Schmitt et al., 2011). A significant reduction of excisions that turn out in the subsequent histopathological diagnosis to be benign nevi is highly desirable. This needs a safe diagnostic basis. Recently, we have shown that there is a distinct difference in the spectral shape of the melanin fluorescence from benign melanocytic nevi as compared to malignant melanoma. This fluorescence difference has been found to be a new, safe diagnostic tool for melanoma detection – so far studied with freshly excised samples of suspect pigmented lesions as well as with corresponding histopathologic specimens (Eichhorn et al., 2009; Leupold et al., 2011). In a continuation of this work, we have carried out the first in vivo fluorescence study with 87 volunteers with suspicious pigmented lesions. Until lately, in the fluorescence from skin tissue, contributions from melanin could hardly been detected because of the extremely low melanin fluorescence quantum yield. The so-called autofluorescence spectra of skin tissue are by far dominated by NAD(P)H, flavins and other endogenous fluorophores. Only with excitation in the spectral region around 800 nm, has NIR-fluorescence from melanin in human tissue been reported (Han et al., 2009). Eventually, access to the melanin fluorescence in the visible region could be reached by a special mode of excitation via a stepwise two-photon absorption process with 800-nm photons from a nanosecond laser (Scholz and Leupold, 2006). The possibility to absorb two 800-nm photons stepwise (consecutively) is unique for melanin. All other endogenous fluorophores can only absorb 800-nm photons simultaneously. Therefore, by using a comparably long pulse duration for the fluorescence exciting laser pulse (2.5 ns), all fluorophores except melanin are strongly discriminated, resulting in melanin-dominated fluorescence spectra from pigmented skin lesions. Only NAD(P)H and flavins (Huang et al., 2002) partially contribute to the nevus fluorescence spectra, whereas such contributions to the melanin-dominated fluorescence spectra of melanoma are almost negligible (Borisova et al., 2005; Eichhorn et al., 2009; Leupold et al., 2011). These 800-nm excited fluorescence spectra revealed a distinct red-shift of the fluorescence maximum as a result of malignant transformation from benign nevi to malignant melanoma. This red shift has been attributed to the higher pheomelanin/eumelanin ratio in melanosomes of melanoma as compared with those of nevi, which results in a faster relaxation of the intermediate melanin excited state in the process of the consecutive two photon absorption (for details see Leupold et al., 2011). This phenomenon of fluorescence red-shift is true for freshly excised skin tissue samples and the corresponding histopathologic specimens, as well as the intermediate state of samples in formaldehyde (Eichhorn et al., 2009; Leupold et al., 2011). But does this pertain also to pigmented lesions in vivo; is there also the said red-shift of fluorescence in melanoma? Although there have been hints in this direction, this preliminary finding is far from being trivial, and more extended in vivo investigations are imperative. Therefore, in this first in vivo study, the melanin-dominated fluorescence spectra of 87 pigmented lesions from volunteers of Caucasian skin type have been investigated. From an experienced clinical point of view and supported by commercially available equipment from DermoGenius (Munich, Germany) and SIAscope (Cambridge, UK), these lesions were all diagnosed as atypical (dysplastic) nevi and selected for excision. Therefore, the melanin-dominated fluorescence spectra could be measured also from all 87 freshly excised samples as well as from the respective histopathologic specimens. The main result is the extensive similarity of the in vivo fluorescence spectra with those measured immediately after excision. Representative examples of this result are given in Figure 1 for a benign nevus and in Figure 2 for a malignant melanoma. Melanin-dominated fluorescence spectrum of common nevi. (A) Characteristic example of fluorescence spectrum from a compound nevus in vivo. (B) Characteristic example of fluorescence spectrum from the same area of the compound nevus, measured immediately after excision. In both cases, excitation is realized by 2.5 ns-pulses via 810 nm two-photon absorption. The continuous line is a least-squares fit with a sixth-order polynomial of data points from 300 spectra repetitions. Melanin-dominated fluorescence spectrum of malignant melanoma. (A) Characteristic example of fluorescence spectrum from a superficially spreading melanoma in vivo. (B) Characteristic example of fluorescence spectrum from the same area of the SSM, measured immediately after excision. In both cases, excitation is realized by 2.5 ns-pulses via 810 nm two-photon absorption. The continuous line is a least-squares fit with a sixth-order polynomial of data points from 300 spectra repetitions. Based on this spectral correspondence of the in vivo and ex vivo spectra, the value of the stepwise two-photon excited melanin fluorescence as a unique diagnostic tool for the detection of melanoma, as already demonstrated for ex vivo pigmented lesions (Leupold et al., 2011), also holds true for the respective spectra measured directly on the patient (in vivo). This conclusion has been confirmed by the study with the 87 atypical nevi. Based on the melanin fluorescence diagnostics in vivo, five of these lesions were identified as melanoma and 70 lesions showed nevi spectra. The remaining 12 samples showed indications of beginning malignant melanocytic transformation. As such, we interpreted a dominance of spectral fluorescence shapes, which are intermediate between nevus (cf. Figure 1) and melanoma (cf. Figure 2). As described in Leupold et al. (2011), these intermediate shapes reflect partial quenching of autofluorescence [NAD(P)H- and flavin-dominated] and partial red shift of melanin fluorescence. The results of fluorescence diagnostics for the corresponding 87 ex vivo samples matched in each case the results of the in vivo diagnostics. The corresponding dermatohistologic findings agreed with the fluorescence diagnosis in all five cases of melanoma (four SSM and one MM situ) and in the 70 cases of benign samples, divided into 65 benign nevi (49 compound nevi, eight junctional nevi, four nevoid lentigo and four dermal nevi) and five cases of seborrhoic keratosis. In contrast, for the 12 samples of fluorescence-indicated beginning of malignant transformation the histologic findings did not reveal malignancy, although in some cases there were indications of atypia (nine compound nevi, one junctional nevi, one nevoid lentigo and one seborrhoic keratosis). Such differences have also been found in a few samples in a previous study with 167 ex vivo pigmented lesions (27 of 136 histopathologically characterized nevi) and were interpreted to reflect fluorescence indications of malignant changes in metabolism that have not yet manifested in obvious malignant morphologic changes (Leupold et al., 2011). The excision appears to be justified for the 12 cases of the fluorescence-indicated beginning of malignant transformation, but there remain 70 excised benign samples (80% of the altogether excised samples) that, according to the in vivo melanin fluorescence diagnosis, did not have to be removed. This number is enlarged from 80 to 94% if the decision is based exclusively on dermatohistology, the current gold standard of melanoma diagnostics. These measurements have been performed from case series in the following institutions: the Center for Dermatology, Berlin, Kurfürstendamm (January 2009 to June 2010; 69 cases) and the University Department of Dermatology in Bochum (February/March 2010; 18 cases). To make the in vivo fluorescence measurements, the LIMES 16-P device (LTB Lasertechnik Berlin GmbH, Berlin, Germany), as described in Leupold et al. (2011) and Eichhorn et al. (2009) was supplemented with a measuring head mounted on a flexible arm. This head realizes 810 nm/2.5 ns-laser pulse excitation of skin areas 50 μm in diameter and scans over the suspect skin regions. The latter is carried out by a stepped motor-controlled movement of the skin through the fixed optical axis for excitation and fluorescence detection. For each of the 87 in vivo and ex vivo samples, the fluorescence diagnosis was obtained on the basis of at least 100 spectra from a grid of measuring areas (50 μm in diameter each) over the pigmented lesion. The automated data analysis evaluated the quotient of the integrated fluorescence intensity in the two regions around 500 and 600 nm, each with a 35-nm bandwidth. If this quotient was in a certain critical range characteristic of melanoma, the corresponding point in the grid was marked by a red dot (for details and examples cf. Leupold et al., 2011). Histopathologic examination was performed by dermatopathologists from the Department of Dermatology, Vivantes Klinikum, Berlin-Neukölln, and the University Department of Dermatology in Bochum. In conclusion, it has been found that the melanin-dominated fluorescence spectra from pigmented skin lesions before (in vivo) and after excision (ex vivo) are virtually identical. Especially in vivo, malignant melanotic transformations are indicated by a distinct spectral red shift, qualifying the use of this fluorescence as a unique diagnostic tool for measurements directly on the patient. One can expect that this melanin fluorescence-based melanoma diagnostic will enable a significant reduction of excisions of suspicious pigmented lesions which would prove later to be benign nevi in the histopathologic diagnosis.