The Evolution of Melanoma Diagnosis: 25 Years Beyond the ABCDs

Abstract

Early detection of malignant melanoma remains the key factor in lowering mortality from this cancer. Recognizing the importance of this issue 25 years ago, our group at New York University published in CA: A Cancer Journal for Clinicians the mnemonic “ABCD” to facilitate the early diagnosis of melanoma. Studies have demonstrated the usefulness of this paradigm in enhancing early melanoma diagnosis as a part of clinical examinations, mass screenings, and public education programs. Approaches to melanoma diagnosis have dynamically evolved during the ensuing quarter century. In the 1990s, dermoscopy enabled the recognition of new subsurface features to differentiate between malignant and benign pigmented lesions. During the last decade, new computer-based technologies have improved diagnostic sensitivity and specificity and may result in optimizing lesion selection for biopsy and pathology review. Despite all of the advances in melanoma diagnosis, timely recognition, detection, and rapid treatment of melanoma remain critical. Although pathologic examination remains the gold standard for diagnosis, this cancer has the potential to be diagnosed through noninvasive approaches because of its cutaneous location. From the development of the ABCDs through current attempts that use complex computer algorithms and genetic markers, a clinician's ability to detect melanoma in its earliest form has been augmented. However, a “good clinical eye” is still fundamental to selecting the lesions for evaluation among the sea of those that are prevalent. As current approaches are refined and new techniques are developed, the improved ability to diagnose this cancer will hopefully enhance reaching the goal of reducing melanoma mortality. CA Cancer J Clin 2010. © 2010 American Cancer Society, Inc.

Introduction

Melanoma is an increasingly important public health problem in the United States and worldwide. The incidence of melanoma has been increasing faster than that of any other cancer in the United States.1 Overall, melanoma incidence increased at an average of 4.6% annually from 1975 to 1985 and 2.7% annually from 1986 to 2007.2 Statistically significant increases are occurring for tumors of all histologic subtypes and thicknesses, including those greater than 4 mm. Invasive melanoma currently is the fifth most frequently diagnosed cancer in men and the sixth most frequently diagnosed cancer in women in the United States. In 2010, 68,130 newly diagnosed cases of invasive melanoma and 46,770 cases of in situ melanoma are expected.3 At current rates, the lifetime risk of an American developing invasive melanoma is approximately 1 in 58 (Fig. 1) overall and 1 in 39 for Caucasian men and 1 in 58 in Caucasian women. This contrasts dramatically with a lifetime risk of 1 in 1500 for Americans born in 1935.4 Approximately 8700 people are expected to die from melanoma in the United States during 2010, accounting for 65% of all skin cancer deaths.3, 5

Figure 1. Lifetime risk of an American developing invasive melanoma.

The importance of diagnosing melanoma early in its evolution cannot be overstated. Because prognosis in melanoma is directly proportionate to the depth of the neoplasm, detection of melanoma early in its evolution is of crucial importance in saving lives. Melanoma typically initially grows horizontally within the epidermis (melanoma in situ). In time, it then penetrates into the dermis (“invasive melanoma”).6 The vertical depth of the melanoma (measured downward from the top of the stratum granulosum [granular cell layer] of the epidermis) has been shown in multivariate analyses to be the factor that best correlates with prognosis (Fig. 2).6–8 Therefore, accurate diagnosis at an early stage, leading to earlier treatment, is crucial to successful management. Fortunately, there has been steady improvement in melanoma survival over decades with the 5-year relative survival for invasive melanoma rising from 82.6% for cases diagnosed from 1975 through in 1979, to 93.1% for cases diagnosed in 2002.2 Because the primary treatment for cutaneous melanoma, surgical excision, has not changed substantially during the past several decades, the improved 5-year survival rate can be primarily attributed to earlier detection. Earlier detection and therapy also lead to less morbidity and decreased cost of therapy. An estimated 90% of expenditures for melanoma therapy in the United States are related to advanced disease.9 Therefore, a significant savings in healthcare cost can be realized if melanoma can be detected in an earlier, more easily treatable phase.

Figure 2. Fifteen-year patient survival by melanoma thickness. Early detection and treatment significantly enhances survival.

The standard method to evaluate a skin growth to rule out melanoma is by biopsy followed by histopathological examination. The challenge lies in identifying the lesions that have the highest probability for being melanoma. Such lesions should be biopsied, and their histopathology appropriately evaluated at the earliest possible time in their development.

Basic Factors in Early Diagnosis

During the past 30 years, there has been significant evolution in the diagnosis of early melanoma (Table 1). Several factors have contributed to a marked improvement in detection of cutaneous melanomas at an early, curable stage. Before the 1980s, there had been little change in identifying melanoma, as the diagnosis was made by identifying clinically macroscopic features. Melanomas were often recognized only when they were large, ulcerated, and fungating (Fig. 3).10 By that time, prognosis was poor. Overall melanoma incidence and mortality continued to increase in the United States and elsewhere, making the early recognition of melanoma an important public health priority.11

Table 1. Recent Evolution of Primary Approaches to Melanoma Diagnosis

DECADE(S)	INPUT	EXAMPLES
Pre 1980s	Gross features	Bleeding
		Ulceration
1980s	Morphologic	ABCD(E)s
	clinical features	Mass screenings
	Screenings	Public and professional education
1990s	Subsurface features	Dermoscopy
2000+	Digital features	Computer analysis
	Subcellular features	In vivo diagnosis

Figure 3. Advanced melanomas demonstrate fungating, ulcerative, and nodular presentations. All of these patients died within 1 year from when these images were recorded.

In 1985, recognizing the need to educate physicians and the public to recognize melanoma in its early clinical presentation, members of our group at New York University devised the ABCD acronym (Asymmetry, Border irregularity, Color variegation, Diameter >6 mm).12, 13 The ABCD criteria were intended to be a simple tool that could be implemented in daily life, a mnemonic “as easy as ABC” to alert both laypersons and healthcare professionals to the clinical features of early melanoma. On the basis of our experience in evaluating patients at the New York University School of Medicine Melanoma Cooperative Group, we found that asymmetry, border irregularity, and color variegation were consistently associated with lesion diameter greater than 6 mm in early melanoma lesions. These observations led to the addition of D to the A, B, and C criteria. Recent studies have reconfirmed that although lesions smaller than 6 mm can occur in 25% of newly diagnosed lesions,14 the diameter guideline of larger than 6 mm remains a useful parameter for clinical diagnosis.15

The ABCDs were thus intended to help describe a subset of melanomas, namely early, thin tumors that might otherwise be confused with benign pigmented lesions (Fig. 4).10, 16 Therefore, both elevated and ulcerated lesions were excluded in the initial analysis because we sought to elucidate features of early melanoma. Pigmented skin lesions that were ulcerated without a history of antecedent trauma would already have been highly suspicious for advanced melanoma and would have required biopsy regardless of other features. Also, it should be emphasized that the criteria were developed to assist nondermatologists in differentiating common moles from cancer and were not meant to provide a comprehensive template for the recognition of all melanomas.

Figure 4. Comparison of early melanoma (top) exhibiting the ABCD features vs benign nevi (bottom). Photographs 1–3, and 6 were provided through the courtesy of Harold Rabinovitz, MD.

Since then, the ABCD criteria have been verified in multiple studies, which have documented the effectiveness and diagnostic accuracy of ABCD in clinical practice, and their efficacy has been confirmed with digital image analysis.17–21 The sensitivity and specificity of these criteria vary when used singly or in combination, and sensitivity declines as specificity increases. Sensitivity22 ranges from 57% to 90% and specificity18 from 59% to 90%. Barnhill et al investigated interobserver variability and reported moderate but statistically significant agreement in most clinical features, including irregular borders and multiple colors, among 4 physician evaluators,23 and inter-rater reliability and objectivity for these criteria has been demonstrated by others.24 The combination of reliable sensitivity and specificity in addition to adequate interobserver concordance in the application of the ABCD criteria support the ongoing utility of this screening instrument in clinical medicine.

These now well known parameters of Asymmetry, Border irregularity, Color variegation, and Diameter greater than 6 mm are used worldwide by groups such as the American Cancer Society, American Academy of Dermatology, and others and have been featured in the lay press to provide simple parameters for evaluation and identification of pigmented lesions that may need further examination. It should be noted that not all melanomas have all 4 ABCD features. It is the combination of features (eg, ABC, A + C, and the like) that render cutaneous lesions most suspicious for early melanoma.

Other melanoma early diagnosis paradigms have been developed to enhance early diagnosis. The most recognized of these is the Glasgow 7-point Checklist, which includes 3 major criteria (change in size, shape, color) and 4 minor criteria (sensory change; diameter of 7 mm or greater; and the presence of inflammation, crusting, or bleeding).25 The Glasgow checklist has been less widely adopted than the ABCD criteria possibly because of its greater complexity. The “ugly duckling” sign is another clinical approach because a pigmented lesion that “looks different from all of its neighbors” may be suspicious for melanoma.26 The ugly duckling sign has been shown to be sensitive for melanoma detection, even for nondermatologists.27

The importance of lesion evolution as a cardinal feature of cutaneous melanoma is also well supported.28 The need to recognize lesion change in our acronym was met by our enhancement of the ABCDs through the addition of “E” for “Evolving.”11, 29 This augmentation substantially enhanced the ability of physicians and laypersons to recognize melanomas at earlier stages. “E” for Evolving is especially important for the diagnosis of nodular melanomas, which frequently present as smaller lesions at more advanced stages (ie, thicker tumors) where early recognition is even more crucial. ABCDE is a simple, succinct, and memorable tool that has been demonstrated to effectively educate the public, the nondermatologist, and the dermatology medical community on the key features of melanoma, including lesion change.30

Skin Self-Examination and Office-Based and Mass Screening

In an attempt to detect melanoma earlier and to incorporate the ABCDs into national public and professional education campaigns, individual and mass evaluation programs were initiated in the mid 1980s (Fig. 5). Skin self-examination (SSE) was encouraged by many organizations for all individuals but especially for those at highest risk for melanoma. Instructing patients to perform regular SSEs is important for several reasons. Melanomas are commonly detected by patients, although it is far more common for dermatologists to detect second primary tumors.31, 32 Persons undergoing skin self-examination were found to be more aware of melanoma and to have the lesions they detected be thinner when biopsied versus persons who did not practice this approach.33 The main predictors of thorough skin self-examination performance are having been advised to do so by a physician, availability of a partner to help, and availability of a wall and hand-held mirror. The use of photographic examples of lesions with the ABCDE characteristics enhances the effectiveness of the procedure.34

Figure 5. (A-D) Skin self-examination steps are reprinted with permission of Rigel DS and Carucci JA. Malignant melanoma: prevention, early detection, and treatment in the 21st century. CA Cancer J Clin. 2000;50:215–236.

Office-based screening has also been shown to enhance earlier detection of melanoma.35 Total Skin Exam (TSE) is a noninvasive, quick, and sensitive (89% to 97%) screening procedure when performed by a physician who is qualified to identify skin cancers.36 A sensitivity of 93.3%, specificity of 97.8%, positive predictive value of 54%, and negative predictive value of 99.8% have been reported when TSE is conducted by dermatologists.37

Results from several studies comparing the thicknesses of initial versus subsequent primary melanomas, melanomas detected at first encounter versus at follow up office visits, or melanomas recognized by patients themselves versus by physicians have confirmed that regular screening is associated with diagnosis of thinner melanomas. In one surveillance program, the mean thickness of initially detected index lesions was 1.44 mm compared with 0.52 mm for surveillance melanomas.38 Additional studies have shown that primary melanomas are more likely to be smaller and thinner when diagnosed during routine surveillance of high-risk patients.39

In addition, nationwide mass screenings have been undertaken. The first Monday in May has been recognized as Melanoma Monday with associated public education events undertaken each year.40 Formal volunteer programs that have been sponsored by groups such as the American Academy of Dermatology, American Cancer Society, Skin Cancer Foundation, and others have screened more than 2 million individuals since 1985, and thousands of clinically presumptive melanomas have been detected.41 These programs have effectively been conducted worldwide.42

The specific recommended intervals for screening of persons at risk are varied. The American Cancer Society recommends a cancer-related check-up including skin examination at the time of periodic health examinations for those aged 20 years and older. The American Academy of Dermatology, Skin Cancer Foundation, and the 1992 National Institutes of Health Consensus Conference on Early Melanoma recommend annual screening for all patients. The National Cancer Institute encourages routine examination of the skin, with particular emphasis on high-risk groups. The US Preventive Services Task Force (USPSTF) believes that limited evidence prevents accurate estimation of the benefits of screening for skin cancer in the general primary-care population.43 The American Academy of Family Physicians and American College of Obstetrics and Gynecology44, 45 recommend screening only for high-risk populations with family or personal history of skin cancer, increased occupational or recreational exposure to sunlight, or clinical evidence of precursor lesions. The American College of Preventive Medicine recommends periodic total cutaneous examinations be performed on targeted populations at high risk for malignant melanoma. Skin-cancer screening can be conducted through self-examination or by a physician. However, the American College of Preventive Medicine recommends that practitioners who perform skin examinations undergo training to assure high-quality examinations and to reduce unnecessary biopsies.46

In a cost-effectiveness analysis for melanoma screening, Losina et al found that one-time melanoma screening of the US general population at age 50 years and screening of siblings of patients with melanoma every 2 years by dermatologists have cost-effectiveness ratios of $10,100/QALYs gained and $35,500/QALYs gained, respectively.47 These ratios are comparable to those for other types of cancer screening, including breast, cervical, and colorectal cancers, all of which are recommended by the USPSTF. However, no randomized prospective melanoma screening trial exists.48 For this reason, the USPSTF believes that the current evidence is insufficient to assess the balance of benefits and harms of screening for skin cancer by primary-care clinicians or by patient skin self-examination.49

In an attempt to better monitor change in specific lesions in higher risk patients, the use of baseline full-body imaging was developed.50 Approximately 25 to 40 segmental baseline images of the entire skin are taken using standard poses and including all existing nevi and “nevus-free” areas of skin. The images then serve as a baseline for comparison during follow up examinations to detect new or changing lesions. The photographs may be printed and kept in the patient's chart or electronically archived.

The effectiveness of total body imaging in early melanoma detection and reduction of biopsies in high-risk (personal or family history of melanoma, dysplastic or multiple nevi) patients has been demonstrated in multiple studies.51–53 This approach also enhances TSE by detecting new lesions and is effective in noting small changes in pigmented lesions over time (Fig. 6). However, based upon patient preferences, effectiveness may be somewhat compromised as some areas of skin may be covered by undergarments or hair and therefore missed in the images.54

Figure 6. Newly appearing melanoma (right) 14 months after baseline image (left). A comparison to the baseline image highlights change.

Currently there are several total-cutaneous imaging systems that are used to aid in localizing suspicious lesions with close-up clinical images. To overcome the expected variability in physician interpretation of clinical images, new software isolates and highlights new or changing lesions for the physician. Attempts are also being made to create automated systems for whole-cutaneous, 3-dimensional, skin, digital photography that can automatically detect new or changing lesions.55

Using Technology to Augment Early Melanoma Diagnosis

Dermoscopy

In the 1990s, light-based visual technologies were adopted to augment the early diagnosis of melanoma. It had been demonstrated that diagnostic accuracy could be improved through the use of surface microscopy, which allows an observer to examine pigmented skin lesions covered by a drop of oil and a glass slide through a stereo microscope, but this technique was time consuming and subjective.56 To obtain these benefits with an approach that was more easily used in the clinical setting, dermoscopy (also known as dermatoscopy or epiluminescence microscopy) was developed and is now a well established method that uses a hand-held lighted magnifier to analyze skin lesions by observing newly defined and descriptively named subsurface structures, eg, dots, streaks, veils, networks (Fig. 7). The initial instruments used an oil or alcohol interface to decrease light reflection, refraction, and diffraction. This made the epidermis essentially translucent and allowed in vivo visualization of subsurface anatomic structures of the epidermis and papillary dermis that are otherwise not discernible to the unaided eye. Dermoscopes usually facilitate a 10-fold magnification of the skin. Newer instruments make the process easier to use by using cross-polarizing light filters which eliminate the need for oil or alcohol. A systematic review of the diagnostic accuracy of dermoscopy in detecting melanoma reported that this technique improved the sensitivity and specificity of clinical diagnosis of melanoma from 71% to 90%.57 However, the effectiveness of this technique depends on experience. Piccolo et al58 reported that dermatologists with 5 years of experience using dermoscopy had a 92% sensitivity rate and a 99% specificity rate when they diagnosed melanoma from dermoscopic images, whereas inexperienced users had a 69% sensitivity rate and a 94% specificity rate.

Figure 7. View of a suspicious pigmented lesion through a polarized light handheld dermoscope.

The basic diagnostic strategy in dermoscopy typically involves a decision tree to guide a determination of whether particular lesions are melanocytic and should be biopsied. The presence of pigmented networks, globules, dots, or streaks favors a melanocytic lesion. The second step for melanocytic lesions is to classify them as benign, suspicious, or malignant on the basis of dermoscopic features by using the scoring systems or algorithms for dermoscopy. Results of the Consensus Net Meeting on Dermoscopy showed that 3 criteria were especially important in distinguishing a malignant from a benign pigmented lesion, asymmetry, atypical pigment network, and blue-white structures (a combination of the earlier categories of blue-white veil and regression structures; Fig. 8).

Figure 8. Dermoscopic images of (left to right) a banal nevus, a dysplastic nevus, and a melanoma (note blue-white veil). Photographs were provided through the courtesy of Alfred W. Kopf, MD.

Several different algorithms for dermoscopic diagnosis are used to differentiate melanoma from benign melanocytic lesions: pattern analysis; ABCD rule of dermoscopy; 7-point checklist; color, architecture, symmetry, and homogeneity (CASH); and Menzies method (Table 2).59–61

Table 2. Comparison of Sensitivity, Specificity, and Diagnostic Accuracy of Dermoscopy Algorithms

DERMOSCOPY DIAGNOSTIC ALGORITHM	SENSITIVITY	SPECIFICITY	DIAGNOSTIC ACCURACY
Pattern analysis59	85%	79%	71%
ABCD59	84%	75%	76%
7-Point checklist59	78%	65%	58%
CASH61	98%	68%	—
Menzies63	85%	85%	81%

Based on the results of the Consensus Net Meeting on Dermoscopy, sensitivity varied between 82.6% and 85.7%, and specificity varied between 70% and 83.4% for distinguishing melanomas from benign lesions.62 Although pattern analysis had the highest sensitivity, specificity, and diagnostic accuracy for melanoma detection when performed by dermatologists, the Menzies method proved better when used by nondermatology physicians.63 Clinicians may be discouraged from using dermoscopy because they feel it is too time consuming. However, Zalaudek et al determined that that use of dermoscopy increased the median skin examination time by only 72 seconds.64

Changes in melanocytic nevi over time can also be appreciated through side-by-side dermoscopic comparisons. Multiple studies have described monitoring nevi by comparison of serially taken dermoscopic images.65, 66 Whereas this approach appears sensitive for early melanoma detection in that a relatively higher percentage of melanomas were diagnosed as in situ, some melanomas might have been missed because they either presented as new lesions or arose from nevi that were not serially monitored.67

Dermoscopy can increase or decrease confidence that a melanocytic lesion is benign or malignant,68, 69 thereby improving the early detection of melanoma, while reducing the need for unnecessary biopsies.70, 71 Regular annual follow up monitoring is also needed to detect slow-growing melanomas in which subtle changes may become apparent only over time.72 Careful observation with particular attention to evolving changes in appearance should direct which lesions are biopsied or removed. The downside of dermoscopy is its potential failure to detect very early or “featureless” (amelanotic) melanomas.73

Digital Image Capture and Analysis

MoleMax (Derma Medical Systems, Vienna, Austria) is a computer-based polarized-light dermoscope.74 The polarized-light source is used with the hand-held video dermoscope for close-up imaging and does not require any oil immersion or contact fluids between the skin and the video head. The MoleMax software is conducive to follow-up examinations, as the transparent overlay feature performs a standardized comparison of images with previous data. Apart from live-video dermoscopy, MoleMax also allows total-body photography and creates a digital map of the skin of patients with high-risk factors and numerous pigmented lesions. These images can be used as a baseline for comparison when suspicious changes are found and for follow up melanoma screening visits.

Computer-Augmented Image Analysis

The idea of computer-assisted melanoma diagnosis was first introduced in the early 1990s.75, 76 There have been many approaches that have been proposed and developed since 2000 that have been promising. However, there is no method claiming to offer complete accuracy in melanoma diagnosis. Each approach has advantages and drawbacks for evaluating pigmented lesions (Table 3).

Table 3. Comparison of Emerging Technologies in Melanoma Diagnosis95

TECHNOLOGY	SENSITIVITY	SPECIFICITY	ADVANTAGES	DISADVANTAGES
This table originally appeared in the European Journal of Dermatology 18(6):617–31; Patel J K. et al., and is reproduced with permission.
MoleMax	N/A	N/A	Two camera system; no oil immersion required; transparent overlay for follow up; total body photography	No computer diagnostic analysis
MelaFind	95–100%	70–85%	Multispectral sequence of images created in <3 seconds; Handheld scanner	-
Spectrophotometric intracutaneous analysis	83–96%	80–87%	Diagnosis of lesions as small as 2 mm in diameter; observes skin structure, vascular composition and reticular pigment networks; handheld scanner	-
SolarScan	91%	68%	Empirical database for comparison; session, and image-level accuracy calibration; recorded on graphic map of body	Requires oil immersion
Confocal scanning laser microscopy (CSLM)	98%	98%	Histopathological evaluation at bedside with similar criteria; longer wavelengths can measure into papillary dermis; fiber-optic imaging allows for flexible handheld devices	Poor resolution of chromatin patterns, nuclear contours and nucleoli; can only assess to depth of 300 μm; melanomas without in situ component will likely escape detection
Optical coherence tomography (OCT)	N/A	N/A	High resolution cross-sectional images resembling histopathological section of skin; higher resolution than ultrasound and greater detection depth than CSLM	Photons are scattered more than once, which can lead to image artifacts; ointment or glycerol may be needed to reduce scattering and increase detection depth; observation of architectural changes and not single cells
Ultrasound technology	99%	99%	Cost effective; information about inflammatory processes of skin in relation to nerves and vessels	Tumor thickness may be overestimated because of underlying inflammatory infiltrate; melanoma metastasis cannot be separated from that of another tumor; images can be difficult to interpret
Tape stripping mRNA	69%	75%	Rapid and easy to perform; painless; practical for any skin surface; can retest same lesion	Results based upon small data set delay in getting test results need larger gene expression profiles for comparison
Electrical bioimpedance	92–100%	67–80%	Complete examination lasts 7 minutes	Electrical impedance properties of human skin vary significantly with the body location, age, gender, and season

Digital Dermoscopy

Computerized approaches have been used to augment the efficacy of dermoscopy in a variety of formats and have been shown to have better results than those of manual dermoscopy. In one study, 10 expert dermoscopists independently examined dermoscopic images of 99 small (<6 mm diameter) pigmented skin lesions (49 melanomas and 50 benign pigmented skin lesions) to determine whether biopsy should be performed. The identical set was also analyzed by a computer-augmented dermoscopy-based system, and sensitivity and specificity were compared. The dermoscopists recommended small melanomas for biopsy with a sensitivity of 71% and specificity of 49% with only fair interobserver agreement (kappa = 0.31 for diagnosis and 0.34 for biopsy). In comparison, the computer-based system achieved 98% sensitivity and 44% specificity.77 These computer systems obtain additional data beyond what can be seen by the naked eye. By using a variety of approaches, researchers have developed several systems that are or will soon be available for use in the clinical setting.

Multispectral Digital Dermoscopy and Image Analysis

The depth of penetration of light into the skin is directly related to wavelength. Information found at different depths is useful in differentiating benign pigmented skin lesions from malignant lesions.78 In multispectral digital dermoscopy, a sequence of images is obtained by using given bands of wavelengths (Fig. 9). Data obtained from differing depths are compared and analyzed by a computer that has a database of historical lesions. The lesion is then classified, and one of the classifications is whether biopsy is recommended (suspicious for melanoma). This technique offers the advantage of analyzing features indiscernible to the human eye, probing up to 2 mm below the surface of the skin.

Figure 9. Multispectral analysis of a pigmented lesion shows that longer wavelengths penetrate deeper so that diagnostic information can be obtained below the surface of the lesion.

This multispectral approach has been augmented through using artificial neural networks so that the diagnostic algorithm iteratively improves. This approach is objective, thereby limiting interphysician variability and could be used to support primary-care physicians in their identification of pigmented skin lesions that require further investigation.79

MelaFind (Electro-Optical Sciences, Irvington, NY) is a multispectral digital dermoscope with a specialized imaging probe and software to assist with differentiation between early melanoma and other skin lesions.80 A hand-held imaging device with an illuminator that shines light of 10 different, specific, wavelength bands ranging from 430 nm to 950 nm (including near infrared) controlled by narrow interference filters on a rotating wheel is used to collect data. An associated photon sensor is used to record the image. Proprietary processing software is used to extract specific features from the images.81, 82 The software determines the edge of the lesion and generates a 10 digital image sequence which is produced in less than 3 seconds (Fig. 10). The images are then analyzed for wavelet maxima, asymmetry, color variation, perimeter changes, and texture changes, and the output is a binary recommendation of whether or not to perform a biopsy. The MelaFind database of pigmented skin lesions includes in vivo MelaFind images and corresponding histological results of approximately 9000 biopsied lesions from approximately 7000 patients. Studies have demonstrated that MelaFind can achieve 95% to 100% sensitivity and 70% to 85% specificity.

Figure 10. Multispectral analysis of melanoma from shortest wavelength (most superficial)1 to longest wavelength (deepest)10 shows (A) a dermoscopic image and (B) a digitally detected margin. Images provided by MelaSciences.

SolarScan is an automated instrument (Polartechnics, Sydney, Australia) and has a 3-charge–coupled device (CCD) video camera for acquiring digital images of lesions. The video head is coupled to the skin with oil to eliminate surface reflections. The acquired image is compared with an empirical database of more than 1800 benign and malignant lesions. Changes in color, pattern, and size are recorded along with the position of each monitored lesion on a graphical map of the patient's body. Images of a lesion from different time points can be viewed simultaneously, and the corresponding analysis is displayed on 4 different graphs. Detection of 14 shades of dermoscopic colors, as well as the blue-white veil structure (which is one of the best features for invasive melanoma diagnosis), is achieved with a specificity for melanoma diagnosis of 97%.68 In a study of 2430 pigmented lesions, SolarScan was found to give a sensitivity of 91% and specificity of 68% for melanoma, which was comparable to that of expert clinicians.83

Spectrophotometric Intracutaneous Analysis (SIAscope; Astron Clinica, Cambridge, UK) is based on the principle that individual skin components vary in their optical properties. It analyzes the distribution of a lesion's position and quantity of chromophores (melanin, blood, and collagen) within the papillary dermis in the infrared region producing 8 narrow-band spectrally filtered images. This high resolution instrument visualizes the skin structure, vascular composition, and pigment networks with detail and clarity. A hand-held scanner that emits radiation ranging from 400 nm to 1000 nm is placed directly on the lesion surface and scans it with light. Computer images that show the location, quantity, and distribution of the skin chromophores within the epidermis and papillary dermis are generated. These features were found to be highly specific (87%) and sensitive (96%) for melanoma.84 In a study of 348 lesions examined with the SIAscope, the technique was noted to have a sensitivity of 83% and specificity of 80%.85 In a study comparing dermoscopy versus Siascopy, an analysis was performed on 65 patients with 83 lesions (12 melanomas) where the diagnosis of melanoma could not be ruled out on the basis of the clinical evaluation by a nondermatologist. Dermoscopy and SIAscopy had sensitivities of 92% versus 100% and specificities of of 81% versus 59%, respectively.86

Laser-Based Enhanced Diagnosis

Confocal Scanning Laser Microscopy

Confocal scanning laser microscopy (CSLM) (VivaScope 3000; Lucid, Rochester, NY) is a noninvasive technique that provides real-time in vivo imaging of skin lesions at variable depths in horizontal planes equivalent to the resolution of conventional microscopes (Fig. 11).87 The high resolution allows imaging of nuclear, cellular, and tissue architecture of the epidermis and underlying dermal structures without a biopsy. The confocal microscope uses a near-infrared laser of low power so that no tissue damage occurs.88 With this technique, a beam splitter separates the light mixture from the laser by allowing only laser light to pass through while reflecting fluorescent light onto a detection apparatus with a pinhole-sized spatial filter. The reflected light is then transformed into an electrical signal recorded by the computer. Images of horizontal sections are reconstructed into 3 dimensions by using multiple tomograms in the horizontal direction.89

Figure 11. Comparison of benign (A) and malignant (B) pigmented lesions using confocal microscopy. Note the similarity of the clinical appearance but the marked differences in confocal images. (A: Ring pattern with edge papillae at dermoepidermal junction. B: Pagetoid and dendritic cells infiltrating the epidermis). Photographs were provided through the courtesy of Harold Rabinovitz, MD.

CSLM relies on the interpretation of images of microanatomical structures, which resemble a histopathological evaluation with similar criteria.90 Longer wavelengths of light enable measurements of greater depth up to the papillary dermis.91 Distinct patterns and cytologic features of benign and malignant lesions can be identified, which correlate with the histological criteria for melanoma.92 Newer CSLM techniques use fiber-optic imaging instead of the pinhole-aperture detector. Therefore, CSLM allows more flexible hand-held devices for in vivo clinical use.92

Most of the diagnostic criteria for specific CLSM features are reliable and reproducible, indicating the definitions of morphological features.93 In a study of 125 pigmented lesions (37 melanomas) comparing CLSM and dermoscopy, dermoscopy had a sensitivity of 89.2% and specificity of 84.1% versus a sensitivity of 97.3% and specificity of 83.0% for CSLM.94 Sensitivity superior to the diagnostic accuracy achieved with dermoscopy was reached by this imaging modality.95 CLSM can also better detect amelanotic melanoma.96

Reflectance Confocal Microscopy

Reflectance confocal microscopy (RCM) allows a higher resolution analysis of dermoscopic structures than does CLSM but is more technically sensitive and expensive to use and is not as effective in analyzing deeper structures. The wavelength range of 550 nm to 1000 nm allows oblique imaging, which helps to distinguish between benign and cancer-prone skin lesions.97

The overall sensitivity and specificity of RCM has been found to be 90% and 86%, respectively.98 The presence or absence of monomorphic melanocytes as a single diagnostic criterion enhances diagnostic accuracy. RCM can assess microanatomical structures to a depth of only 300 μm or so.99 Thus, processes in the reticular dermis cannot be examined for the presence or absence of invasion. Although initial results are promising and encourage further exploration, more experience is needed to determine the role of RCM in the diagnosis of melanoma.

Optical Coherence Tomography

Optical coherence tomography (OCT) is a technique that enables an examination of the skin to a depth of about 1 mm. The light reflectivity of different tissue components (melanin and cell membranes) provides contrast in the images, and these findings correlate with pathology. Under OCT, melanomas demonstrate increased architectural disarray, less defined dermoepidermal borders, and vertically oriented icicle-shaped structures not seen in nevi.100 In addition, although the resolution of OCT is insufficient to reveal the morphology of single cells, lesion architecture can be evaluated and correlated with surface dermoscopic parameters. The utility of OCT for skin lesions has not been fully established because sensitivity and/or specificity studies for melanoma detection have not been reported. Because histopathologic structures may be less clearly observed in hyperkeratotic or raised lesions, OCT may be better suited for flat and nonscaling lesions.

Diagnostic Ultrasound

Ultrasound scanning is a safe noninvasive method that in some settings can be used to show subtle differences between nevi and melanoma. Transducers with higher frequency wavelengths are beneficial for diagnosing skin lesions because they allow better resolution of small lesions located near the skin surface. However, higher frequencies also lead to decreased depth of penetration by the ultrasound waves; thus, the choice of the probe frequency depends on the diameter and site of the lesion.101

High-resolution B-mode ultrasound has traditionally and primarily been used to assess the depth/thickness of melanoma tumors. Reflex transmission imaging (RTI) is a form of high-resolution ultrasound that can be used in combination with white-light digital photography for classification of pigmented lesions. Rallan et al used RTI to derive sonographic parameters to differentiate melanomas from benign pigmented lesions.102 Significant quantitative differences allow discrimination between melanomas, seborrheic keratoses, and nevi. With the use of a 20-MHz, ultrasound, B-scan imaging system interfaced to a computer, melanoma can be distinguished from basal cell carcinoma with 100% sensitivity and 79% specificity.103

Further refinements are needed to make this technique widely applicable. Despite various frequency transducers that can probe depths greater than 1.5 cm, melanoma metastasis often cannot be separated from that of another tumor. In addition, the consistency of ultrasound results depends heavily on the skill and experience of the examiner and the anatomic site of the lesion.

Using mRNA Patterns to Diagnose Melanoma

MRNA patterns that have been identified in melanomas can be tested in suspicious skin lesions. By using a noninvasive method called tape stripping, physicians can harvest and test cells in the upper epidermis. This technique has the advantage of being noninvasive, rapid, easy to perform, and painless. Adhesive tape is applied to a pigmented skin lesion and then briskly rubbed in a circular motion. The edge of the lesion is then outlined on the tape with a marker. Superficial cell layers of skin are stripped off by removing the tape, and mRNA is collected from these cells. The tape outside of the marked edge of the lesions is removed so that surrounding normal epidermis is not tested. Enough mRNA for gene-expression profiling is removed for analysis by ribonuclease-protection assay for the purpose of differentiating melanoma from benign lesions.104 In a study of 150 suspicious pigmented lesions that used tape-stripping to sample cells for staining with toluidine blue and cytological examination (without analysis of mRNA), sensitivity and specificity were 68.7% and 74.5%, respectively, demonstrating its potential as a helpful diagnostic tool for the early detection of melanoma.105

Wachsman et al tested suspicious pigmented lesions by tape stripping 4 times and then biopsying the lesions for pathological analysis. Normal uninvolved skin was also tape stripped to serve as the negative control. A 20-gene classifier that discriminated melanoma from atypical nevi was reported in the Wachsman study, and subsequent testing of this classifier found it to be 100% sensitive, 90.6% specific, and 92.4% diagnostically accurate for detection of both in situ and invasive melanoma.106, 107 Additional clinical trials are currently underway to refine these gene-expression profiles for identifying early melanomas.

Although still in early trials, tape stripping may eventually be most beneficial as a prescreen for patients with multiple suspicious lesions to identify which subset of lesions should be considered for biopsy confirmation. If the analysis is positive for a particular gene-expression profile associated with melanoma, the pigmented lesion should be biopsied.

Cellular Electrical Resistance (Bioimpedance)

Electrical resistance (bioimpedance) of lesions has been studied to assess differences between skin cancers and benign lesions.108 Bioimpedance levels are a function of cell shape and structure, cell membranes, and the amount of water present in the cells. Based on these features, electrical impedance of cancer and benign cells are different because cancer cells typically have a different shape, size, and orientation than benign cells do.109

Bioimpedance measurements of suspicious pigmented lesions are taken on both the center of the lesion and a noninvolved reference skin site. Lesional and reference skin are measured at 5 depth levels, approximately 0.1 mm to 2 mm into the tissue, and data are analyzed by a computer.110 The entire process takes approximately 7 minutes to complete.

Bioimpedance has a high sensitivity of almost 100% for in situ and thin melanomas and can differentiate melanoma from benign nevi with 92% to 100% sensitivity and 67% to 75% specificity.110 Spiked microinvasive electrodes may also be better for melanoma detection (92% sensitivity and 80% specificity) than the regularly used noninvasive probes.112 Because electrical impedance properties of human skin vary significantly by location, age, gender, and season, more studies are being undertaken to standardize results.113

Current Recommendations

Unlike other cancers that are internal, the cutaneous location of malignant melanoma allows diagnosis through noninvasive approaches. There have been significant advances in melanoma diagnostic techniques during the last 25 years. Beginning with the development of the ABCDs, through current attempts that use complex computer algorithms and genetic markers, all of these approaches have augmented the clinician's ability to detect melanoma in its earliest form. Currently, the major technologies readily available to the clinician are baseline image comparison and dermoscopy, and these can be effectively and inexpensively incorporated in the primary-care setting for evaluating a patient with pigmented lesions.114 Techniques such as confocal microscopy are available in institutional settings. Several of the computer-aided approaches are nearing US Food and Drug Administration (FDA) approval and, when available, will be straightforward enough to use to augment melanoma diagnosis in the primary-care setting.

However, the challenge of all of these techniques is that they still require a “good eye” to select the lesions for evaluation among the sea of lesions that are prevalent. As Dr. Neville Davis stated more than 40 years ago, “Melanoma writes its message in the skin with its own ink, and it is there for all of us to see. Some see but do not comprehend.”115 Primary-care physicians currently have a significantly lower sensitivity for melanoma detection (31%) than do skin-cancer specialists.116 Given that more than two-thirds of skin disease in the United States is treated by nondermatologists, better education must begin early in their training and must be provided to all healthcare professionals to improve diagnostic acumen and education of skin cancer. Only 25% of primary-care residents receive any training in this area, and performing just 4 total skin examinations per year significantly improves a resident's diagnostic skills.117

Future Outlook

Despite all the advances in melanoma diagnosis during the past 25 years, timely recognition, detection, and treatment of melanoma remain crucial. To be successful, we must “see” melanoma in its earliest form, and newer approaches beyond the ABCDs have the promise to make this a reality.

Well-designed studies are needed to more definitively understand the relative efficacies, advantages, and disadvantages of each of these techniques. As incidence continues to rise, however, one of our best weapons against this potentially deadly neoplasm will be multidisciplinary approach focused on early detection using the best skills of all involved. As current diagnostic approaches are refined and new techniques are developed, we will, hopefully, reach our goal to lower mortality from this cancer.