Home > Skin cancer - Referral Guidelines for...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Clinical Governance Research and Development Unit (CGRDU), Department of Health Sciences, University of Leicester. Referral Guidelines for Suspected Cancer in Adults and Children [Internet]. London: Royal College of General Practitioners (UK); 2005 Jun. (NICE Clinical Guidelines, No. 27.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

16Skin cancer

General recomendations


A patient presenting with skin lesions suggestive of skin cancer or in whom a biopsy has been confirmed should be referred to a team specialising in skin cancer. [D]


All primary healthcare professionals should be aware of the 7-point weighted checklist (see recommendation 1.10.8) for assessment of pigmented skin lesions. [C]


All primary healthcare professionals who perform minor surgery should have received appropriate accredited training in relevant aspects of skin surgery including cryotherapy, curettage, and incisional and excisional biopsy techniques, and should undertake appropriate continuing professional development. [D]


Patients with persistent or slowly evolving unresponsive skin conditions in which the diagnosis is uncertain and cancer is a possibility should be referred to a dermatologist. [D]


All excised skin specimens should be sent for pathological examination. [C(DS)]


On making a referral of a patient in whom an excised lesion has been diagnosed as malignant, a copy of the pathology report should be sent with the referral correspondence, as there may be details (such as tumour thickness, excision margin) that will specifically influence future management. [D]

Specific recommendations



Change is a key element in diagnosing malignant melanoma. For low-suspicion lesions, careful monitoring for change should be undertaken using the 7-point checklist (see recommendation 1.10.8) for 8 weeks. Measurement should be made with photographs and a marker scale and/or ruler. [D]


All primary healthcare professionals should use the weighted 7-point checklist in the assessment of pigmented lesions to determine referral:

Major features of the lesions: [C]

  • change in size
  • irregular shape
  • irregular colour.

Minor features of the lesions:

  • largest diameter 7 mm or more
  • inflammation
  • oozing
  • change in sensation.

Suspicion is greater for lesions scoring 3 points or more (based on major features scoring 2 points each and minor features scoring 1 point each). However, if there are strong concerns about cancer, any one feature is adequate to prompt urgent referral.


In patients with a lesion suspected to be melanoma (see recommendation 1.10.8), an urgent referral to a dermatologist or other suitable specialist with experience of melanoma diagnosis should be made, and excision in primary care should be avoided. [C]

Squamous cell carincomas


Squamous cell carcinomas present as keratinizing or crusted tumours that may ulcerate. Non-healing lesions larger than 1 cm with significant induration on palpation, commonly on face, scalp or back of hand with a documented expansion over 8 weeks, may be squamous cell carcinomas and an urgent referral should be made. [C]


Squamous cell carcinomas are common in patients on immunosuppressive treatment, but may be atypical and aggressive. In patients who have had an organ transplant who develop new or growing cutaneous lesions, an urgent referral should be made. [C]


In any patient with histological diagnosis of a squamous cell carcinoma made in primary care, an urgent referral should be made. [C]

Basil cell carcinomas


Basal cell carcinomas are slow growing, usually without significant expansion over 2 months, and usually occur on the face. Where there is a suspicion that the patient has a basal cell carcinoma, a non-urgent referral should be made. [C]



All pigmented lesions that are not viewed as suspicious of melanoma but are excised should have a lateral excision margin of 2 mm of clinically normal skin and cut to include subcutaneous fat in depth. [B(DS)]



There were 6,062 recorded cases of malignant melanoma in England and Wales in 2001. Of these 3,424 were in females and 2,638 in males. The incidence of melanoma increases with age in both males and females rising steadily in both sexes from age 15 years onwards.

Figure 19. Newly diagnosed cases of skin melanoma in 2001 in England and Wales. (77).

Figure 19Newly diagnosed cases of skin melanoma in 2001 in England and Wales. (77)


The age specific mortality rates for melanoma of the skin are similar for both men and women, and numbers mirror the increase in incidence with increase in age. There were a total of 1,480 deaths from malignant melanoma in England and Wales in 2002, of which 784 were males and 696 females (see Figure 20).

Figure 20. Mortality figures from skin melanoma for 2002 in England and Wales. (78).

Figure 20

Mortality figures from skin melanoma for 2002 in England and Wales. (78).

Non-Melanoma (basal and squamous cell carcinoma)


Non-melanoma skin cancers are the most common cancer occurring in the UK (http://www.cancerresearchuk.org). There were an estimated 59,000 cases diagnosed in 1999 across the UK, but the true figure may be higher because of under-reporting of cases. Basal cell carcinoma (BCC) is now thought to be the most common of all human malignancies closely followed by squamous cell carcinoma (SCC) (307).

Incidence rates increase with age in males and females, and in 1994 incidence reached 718 in males and 407 in females per 100,000 population in those aged 85 years and over.

Figure 21. Incidence rates per 100 000 population for non-melanoma skin cancer, England and Wales, 1994(17).

Figure 21Incidence rates per 100 000 population for non-melanoma skin cancer, England and Wales, 1994(17)


The provisional mortality rates for 1999 show that mortality from non-melanoma skin carcinoma remains low in those less than 50 years of age. Mortality in males increases steadily from age 60 years onwards peaking at 18/100,000 population in those aged 85 years or over. Female mortality rates begin to rise steeply from age 70 years and peak at 7/100,000 population in the same age group. In 2002, there were 259 deaths among men from non-melanoma skin cancers and 187 in women.

Figure 22. Mortality per 100 000 population from non-melanoma skin, England and Wales, 1999 (provisional)(17).

Figure 22Mortality per 100 000 population from non-melanoma skin, England and Wales, 1999 (provisional)(17)

16.1. Symptoms and Signs

16.1.1. Key Clinical Questions

How common are skin cancers in certain population groups? Which symptoms, signs and other features raise a suspicion of skin cancers (melanoma and basal cell carcinoma [BCC], squamous cell carcinoma [SCC]), and those that make cancer less likely as a diagnosis? Does family history discriminate patients who should be referred?

16.1.2. Evidence Question

In people attending primary care services with dermatological problems, which symptoms and signs and other features including family history when compared with the ‘gold standard’ are predictive of a diagnosis of cancer; and which symptoms and signs are not?

16.1.3. Evidence Statements

The incidence of melanoma is rare below aged 15 years, and increases with age. The incidence has been increasing over the last 20 years. (III)

The incidence of non-melanoma skin cancers increases with age, and has been increasing over the past 20 years. (III)


Features of skin lesions associated with melanoma include: increase in size, change in shape, change in colour, size >5–7 mm, irregular outline, ulceration, inflammation and bleeding. (III)

The evidence about the use of checklists such as the ABCD or seven-point checklists is limited. (III)

There is insufficient evidence about the effectiveness of educational interventions to improve the ability of general practitioners to identify melanoma. (III)

Several factors are associated with melanoma, but the level of risk conferred by these factors is insufficient to discriminate between those who should or should not be suspected of having melanomas. (III)

Non melanoma (basal or squamous cell carcinomas)

Squamous cell carcinomas present as keratinizing or crusted tumours that may ulcerate. (III)

Basal cell carcinomas may be nodular, cystic or ulcerated. (III)

The risk of squamous cell carcinomas is increased in people who are immunosuppressed. (III)

Several other factors increase the risk of squamous cell carcinoma, but insufficiently to distinguish those patients who should be suspected of having these tumours. (III)

Several factors, including immunosuppression, have been associated with basal cell carcinoma. (III)


(SIGN, Cutaneous melanoma: A National Clinical Guideline, 2003)(308)

The SIGN guidelines were developed following a detailed literature review and included the following recommendations:

  • Clinicians should be familiar with the seven point or the ABCDE checklist for assessing lesions. (D)
  • Clinicians using hand held dermatoscopy should be appropriately trained. (D) Health professionals should be encouraged to examine patients’ skin during other clinical examinations. (D)
  • Patients with suspicious pigmented lesions should be seen at a specialist clinic in a time commensurate with the level of concern indicated by the general practitioner referral letter. (recommended best practice)
  • Emphasis should be given to the recognition of early melanoma by both patients and health professionals. (recommended best practice)
  • Targeted education can enhance professionals’ ability to diagnose melanoma. (recommended best practice)
  • Healthcare professionals and members of the public should be aware of the risk factors for melanoma. (B).
  • Individuals identified as being at higher risk should be:
    • advised about appropriate methods of sun protection
    • educated about the diagnostic features of cutaneous melanoma
    • encouraged to perform self-examination of the skin. (C)
  • Brochures and leaflets should be used to deliver preventive information on melanoma to the general public. (D)
  • Leaflets and brochures used in melanoma prevention work should be non-alarmist. (recommended best practice)
  • If computer-based learning programmes are used they should be interactive in nature. (recommended best practice)
(Australian Cancer Network, 1999)(309)

These guidelines for melanoma were based on a systematic review of evidence that was considered by a multidisciplinary panel. The recommendations relating to clinical diagnosis were:

  • Good lighting and magnification is recommended when lesions are examined.
  • All clinicians should be trained in the recognition of early melanoma.
  • A good clinical history of the change in the lesion (if any), a past history of skin lesions, and a family history of melanoma should be obtained.
  • A family history is defined as melanoma in a direct-line family member – grandparent, parent, sibling or child of the patient.
  • Lesions which are suspicious or cannot be diagnosed after a period of observation should be biopsied, or the patient referred for a specialist opinion.
  • High risk individuals should be advised of the specific changes which suggest melanoma and encouraged to perform self-examination.
(Roberts et al, 2002)(310)

These guidelines for melanoma were produced jointly by the British Association of Dermatologists and the Melanoma Study Group. The seven-point checklist was recommended for both patient and general practitioner education. Lesions with any of the three major features (change in shape, irregular shape, irregular colour) or three of the minor features (largest diameter 7mm or more, inflammation, oozing, change in sensation) are suspicious of melanoma, and should ideally be seen by specialists (that is, clinicians routinely treating large numbers of patients with pigmented lesions). Specific recommendations were:

  • Patients with lesions suspicious of melanoma should be referred urgently to a dermatologist or surgeon/plastic surgeon with an interest in pigmented lesions.
  • These specialists should ensure that a system is in place to enable patients with suspicious lesions to be seen within two weeks of receipt of the referral letter.
  • All patients who have had lesions removed by their general practitioner that are subsequently reported as melanoma should be referred immediately to specialists.

(Grade C, level III)

Squamous Cell Carcinoma
(Motley et al, 2002)(311)

These British Association of Dermatologists/British Association of Plastic Surgeons guidelines addressed squamous cell carcinoma. Squamous cell carcinoma was defined as a malignant skin tumour of keratinizing cells of the epidermis or its appendages, which is locally invasive and has the potential to metastasize. The guidelines state it usually presents as an indurated nodular keratinizing or crusted tumour that may ulcerate, or may present as an ulcer without evidence of keratinization.

Other forms of squamous cell carcinoma include (a) actinic and radiation keratoses, which are scaly erythematous papules or plaques on sun damaged or irradiated skin that may develop into invasive squamous cell carcinoma; (b) pre-invasive carcinoma (carcinoma in situ): (i) Bowen’s disease, which is crusted, keratotic or a velvety erythrematius plaque; (ii) erythroplasia of Queyrat, which appears on the glans penis as a red, velvety patch; (iii) erythroplakia and malignant leukoplakia, on mucous membranes other than the glans penis; (c) verrucous carcinoma, a warty tumour that occurs most often on the hands, feet, anogenital area and oral cavity; (d) keratoacanthoma.

Basal Cell Carcinoma
(Telfer et al, 1999)(312)

These guidelines were produced on behalf of the British Association of Dermatologists, and dealt with basal cell carcinoma. Basal cell carcinoma was defined as a slow-growing, locally invasive malignant epidermal skin tumour, which occurs most commonly in caucasians. Metastasis is extremely rare, and morbidity is related to local tissue destruction, particularly on the head and neck. The clinical appearances are diverse, and include nodular, cystic, ulcerated (‘rodent ulcer’), superficial, morphoeic (sclerosing), keratotic and pigmented variants.

Primary studies

(Elwood et al, 1998)(313)

This article was a report from a larger case control study of risk factors. Information on all histologically confirmed cases of newly diagnosed cutaneous malignant melanoma was obtained from treatment centres and cancer registries in four provinces of Canada. Identified patients were interviewed about initial presentation and symptoms. In total, 801 patients aged 20 to 79 years were approached for interview, and 665 (83%) were successfully interviewed. Of these, 14 had acral lentiginous melanoma and were excluded from the report, leaving 651 patients who had cutaneous melanomas.

Population matched controls were selected for all cases. The interviewer was unaware of the case or control status of the interviewee. Pathological slides were obtained and reviewed by one of two pathologists, although slides could not be obtained for 121 patients (20%). The original pathology report was used for these patients.

Of the 651 patients, 60% were female. The mean age of the sample was 49.7 years (range 21 to 79 years). 415 patients (64%) had superficial spreading melanoma, 128 (20%) had nodular melanoma, 52 (8%) had unclassified or borderline melanoma and 56 (9%) had lentigo maligna melanoma. Most patients (65%) reported one or more of a set of four symptoms related to an existing mole or pigmented spot:

Each of the 651 patients presenting with melanoma were asked to describe, without prompting, the first indications of their disease, the results are shown in Table 13.

Table 13. First symptoms of melanoma(313).

Table 13

First symptoms of melanoma(313).

(Brady et al, 2000)(314)

This case series included 471 newly diagnosed patients with cutaneous melanoma presenting to a US specialist cancer centre between July 1995 and May 1998. Patients with an unknown primary site, noncutaneous melanoma, distant metastases or recurrent disease were excluded. All patients were asked to complete a questionnaire at their first visit to the cancer centre. Information regarding the Breslow thickness of the melanoma was available in 454 patients.

There were approximately equal numbers of males and females, but females were younger than males (51 years vs. 55 years, P<0.01). Most of the patients were Caucasian (N=456 of 471 patients; 97%) and most patients presented with American Joint Committee on Cancer Stage I and Stage II disease.

Most patients presented with melanoma > 0.75mm in Breslow thickness (62%; N=283 patients). The remaining patients (38%) had thin melanomas (≥0.75mm; N=122 patients) or in situ disease (N=49 patients). The majority of patients detected their own melanomas (N=270; 57%). Patterns of detection were influenced by patient gender. Females were more likely to self-detect than males (69% vs. 47%; P<0.0001). Physicians detected the melanoma in 16% of patients (N=74), followed by spouse in 11% (N=51). Physicians were three times more likely to detect thin lesions (≤0.75 mm) compared with nonphysician detectors (95% confidence interval [95% CI] 2.1, 6.5; P=0.0001). Physician detection occurred in only four of 84 males under age 50 years compared with 43 of 166 males age ≥ 50 years (P<0.0001). Patients who reported a family history of melanoma had a 2.7 fold increased likelihood of presenting with a thin lesion (95% CI, 1.6, 4.7; P=0.003). Family history information was available for 451 patients. Of these, 84 patients (19%) reported a family history of melanoma, and 366 patients (81%) reported no first or second degree relative with the disease.

Despite a trend towards thinner melanomas in females, the difference in the median Breslow thickness between females and males was not significantly different (1.10 mm vs. 1.13 mm; P=0.07). There was no significant association between tumour thickness and age, gender or lesion visibility.

(Schwartz et al, 2002)(315)

In this US case series, 1515 consecutive patients presenting to a US cancer centre between January 1998 and December 1999 with in situ or invasive cutaneous melanomas were questioned about their signs and symptoms. All histology slides were reviewed by a skin pathologist to confirm the diagnosis of primary cutaneous melanoma.

The mean age at diagnosis of the first primary melanoma was 52.6 years. The majority of patients (72%) were between the ages of 21 and 65, 26% being older than 65 years, and only 2% younger than 21 years. Females (48.9 years) were younger than males (56.1 years) at diagnosis of their first primaries (P<0.001). Physician detected lesions were thinner (0.40mm) than either self-detected (1.17 mm; P<0.001) or spouse-detected (1.00 mm; P<0.001) lesions. In males the Breslow depth of self-detected lesions (1.42 mm) was greater than that of the lesions detected by either the spouse (1.04 mm; P<0.005) or physician (0.42 mm; P<0.001). In females, the mean Breslow depth of self-detected lesions (0.98 mm) was greater than physician detected lesions (0.35 mm; P<0.001) but was not significantly different from spouse-detected lesions (0.72 mm; P=0.2).

The most common changes noted by patients were the colour, size, and/or shape/elevation of a lesion. Less common changes included ulceration, bleeding, tenderness, and itching. Mean Breslow depths associated with a change in colour (1.15 mm), size (1.33 mm), shape/elevation (1.47 mm) and itching (1.70 mm) were less than mean Breslow depths associated with ulceration (2.69 mm), bleeding (2.63 mm) and tenderness (2.44 mm; all P<0.005).

(Sober et al, 198)(316)

The study included a total of 598 patients attending two US hospitals. A questionnaire was administered by a trained interviewer to evaluate the frequency with which signs and symptoms were associated with melanoma. All patients in this sample had clinical stage 1 cutaneous melanoma. They were seen either with the primary tumour intact or within 30 days of its removal. The frequency of each sign and symptom was cross tabulated with four thickness ranges: <0.85 mm, 0.85 mm, 0.85 to 1.69 mm, 1.70 to 3.64 mm, and ≥ 3.65 mm.

For thin lesions (<0.85 mm) increase in size was noted in more than half and was the most frequent sign or symptom present for ‘thin’ tumours. This was closely followed by colour change, which was present in half. Bleeding, ulceration and tenderness were infrequently seen (present in five to 13%). Conversely, increase in height was the most frequent feature noted with the thickest tumours (≥ 3.65 mm), observed by more than 80% of patients. Bleeding and ulceration were reported in more than half. There was a direct relationship between increase in height and increasing tumour thickness. Itching of the lesion occurred in 20–46% of patients.

(Wick et al, 1980)(317)

The clinical characteristics of the primary tumour in 786 US patients with histologically confirmed superficial spreading melanoma were investigated in this case series from five US hospitals.

The most useful features for early diagnosis were change in size and change in colour, present in 71% and 55% respectively of patients with level II lesions. Increase in height of lesion correlated with more advanced disease. Ulceration and bleeding were predominantly found in advanced primary lesions and were judged of limited use in early recognition. The data revealed that primary lesions were of substantial size and generally much larger than acquired naevi (<7mm) from which they must be differentiated. The results suggested that site was not a major determinant for the presentation of early lesions. There was however a higher proportion of level II lesions (42%) on the head and neck. Conversely, a higher percentage of deeper lesions were encountered on the foot. Characteristic features of early (II, III) lesions associated with tumour growth were colouration and size. The features characteristic of advanced lesions were tenderness, ulceration and bleeding. Elevation became common at level III and above.

(Cassileth, 1987)(318)

In this case series, a retrospective analysis of the charts of 568 patients treated between 1972 and 1981 for superficial spreading melanomas was undertaken. The sample was composed of patients who had attended a single specialist US centre, and only data for patients over 17 years of age and with no prior primary melanomas were included. Information was recorded routinely for all patients by clinic nurses using a structured interview guide during the patient’s first clinic visit. Patients were asked about the presence of each of seven symptoms (size, elevation, colour, bleeding, ulceration, itching and tenderness) plus other features. Information was recorded about the type, number and duration of individual symptoms noticed by the patient; catalyst symptoms or the particular event that preceded the patient’s request for medical attention; and location, thickness and level of the melanoma.

Forty-eight percent of patients who met the eligibility criteria were men. Forty-six percent of patients reported the simultaneous occurrence of more than one catalyst symptom; 35% reported experiencing one catalyst symptom only; and 19% claimed that they had noticed no changes in existing lesions. The most common catalyst symptom pattern, a combination of size, elevation and colour was reported by 60 patients, who were diagnosed an average of 11.2 months after observing this combination. The mean tumour thickness at diagnosis for this group of patients was 1.26 mm (±1.8 mm). The second most common catalyst symptom, bleeding, was reported by 49 patients, who were diagnosed after an average of 2.3 months. A total of 75 different catalyst symptoms or symptom combinations were described.

Patients who sought medical attention in response to bleeding alone (N=49) had thicker lesions (mean 1.77 mm) than did the 45 patients who sought medical attention in response to changes in both size and colour (mean 0.54 mm). A total of 109 patients, 19% of the sample, could not identify any change in an existing lesion. The average lesion thickness for these 109 patients was 0.93 mm (±1.4 mm) compared with the average lesion thickness of 1.37 (±1.8 mm) for all other patients (P<0.01).

Symptom/Sign Checklists for Melanoma

Two checklists have been developed as diagnostic aids to assist identification of melanoma lesions from the presenting features indicative of malignancy.

The seven point checklist(310) recommended criteria used to evaluate lesions suggestive of melanoma. The checklist specifies major and minor features, they are; major features: change in size, irregular shape and irregular colour; minor features: largest diameter 7mm or more, inflammation, oozing and change in sensation. The checklist was developed mainly for use by primary care physicians to facilitate referral decisions. In a revision of the list, ‘irregular shape’ was replaced by ‘change in shape’. The major criteria are therefore change in size, shape and colour. The minor criteria are inflammation, crusting or bleeding, sensory change and a diameter greater than or equal to 7mm.

Table 14ABCD(E) System for clinical diagnosis of malignant melanoma(319)

BBorder irregularity
CColour variegation
DDiameter of 6mm or more

often excluded

The ABCD method(319) developed in the USA is another tool for diagnosing malignant melanoma. The system was expanded into ABCDE list to include an ‘E’ which has been used to indicate either ‘elevation’ or ‘evolutionary change’. Many benign naevi however, may be elevated. Asymmetry refers to one half of the lesion not matching the other. Border irregularity describes edges that are ragged, notched or blurred. Colour irregularity involves pigmentation that is not uniform; shades of tan, brown and black are present with dashes of red, white or blue; and a diameter of at least 6 mm is also classified as important.

(Whited, 1998)(319)

This was a systematic review of the accuracy of skin examination for melanoma using the ABCD(E) and revised seven point checklists. A literature search was performed using MEDLINE for the years 1966 through 1996 to identify relevant retrospective and prospective studies.

Evidence contained in the articles were evaluated and included if they had been given a quality rating of C or above. Twelve studies were included in total. Two studies reported information about the sensitivity for the ABCD checklist, in one it was 92%; (CI 95%, 82%–96%), and in the other 100% (95% CI 54%–100%); one study reported specificity to be 98% (95% CI, 95%–99%). The revised seven point checklist has been reported to have a sensitivity of 79% (95% CI, 70%–85%) to 100% (95% CI 94%–100%) and specificity of 30% (95% CI, 21%–39%) to 37% (95% CI, 21%–39%). Physicians’ global assessments for detecting the presence or absence of melanoma were estimated to have a specificity of 96% to 99%, while sensitivity ranges widely from 50% to 97%. Non-dermatologists’ examinations were less sensitive than those performed by dermatologists.

(Osborne, 1999)(320)

The aim of the study was to investigate possible predictor variables for false negative gradings using the seven point checklist in a population (107) of patients with histologically confirmed malignant melanomas presenting in Leicestershire between 1982 and 1996. The case notes of the included patients were examined retrospectively. False negatives were defined as those patients in whom another diagnosis was made or in whom there was evidence in the case notes that the diagnosis was thought not to be malignant melanoma. Demographic data were recorded together with clinical diagnosis, clinical features of each lesion according to the revised seven point checklist, and site of the lesion.

No clinical diagnosis had been given in the records for 43 of the 778 lesions, 599 were suspected of being melanoma, and 136 had not been suspected on clinical grounds. The clinical false negative diagnosis rate was 18.5% and the diagnostic sensitivity 81.5%. There were 476 females and 257 males, giving a ratio of 65% females. Sex had no effect on false negative rate; the proportion of females in the diagnosed group being 66% and the non-diagnosed group 60% (=0.20). The false negative rate varied markedly with site and was lowest for the trunk and leg (12 and 13%), but was 21% for the arm. More rarely occurring sites gave higher false negative rates from 31% to 42%. Comparing the false negative rate on the trunk (the lowest rate) with the other sites, the odds ratio for the face was 3.4 (P=0.0007), head and neck 5.1 (P<0.0001), arm 2.0 (P=0.02), leg 1.0 (P=0.6), sole 3.4 (P=0.06) and subungual 5.5 (P=0.007).

The false negative clinical diagnosis rate varied markedly with the presence of features of the seven point checklist (P<0.00001). It was lower if major features were present (8–18%), and greater if the minor features were present (13–35%). Major features associated with a particularly low rate were irregular shape and irregular pigmentation, 8 and 10%, respectively. Clinical features of lesions associated with a higher false negative rate were lack of irregular pigmentation and shape, altered sensation, the presence of inflammation and size < 7mm.

The multivariate logistic regression of all parameters showed that the relationships of false negative rate and melanoma site, irregular pigmentation, irregular shape, sensation, inflammation and diameter >6 mm were significant and independent. For the individual sites, results of univariate and multivariate analysis were similar, although the adjusted odds ratio and its significance, for the face compared with the trunk increased markedly on multivariate analysis. The results suggested that the face is a particularly difficult site. All of the clinical features except surface oozing/crusting/bleeding retained significance on multiple regression.

Risk Factors

(SIGN, 2003)(308)

The SIGN guidelines involved a systematic literature search that included assessment of risk factors. The findings were presented in a table, reproduced here. In the table, odds ratios are given, based on the findings of one or more primary studies, odds ratios being the odds in favour of exposure to a risk factor in people with melanoma to the odds in favour of exposure to the same risk factor among people who have not developed melanoma. The SIGN guideline observed that the odds ratios for someone who has skin that does not tan easily (1.98) is modest in comparison with the ten fold or greater risk of developing lung cancer in someone who smokes cigarettes compared to a person who has never smoked.

SIGN recommended that:

Genetic testing in familial or sporadic melanoma is not appropriate in a routine clinical setting and should only be undertaken in the context of appropriate research studies (D).

The SIGN guidelines cited a consensus document, which estimated that one to two percent of melanomas were attributable to the inheritance of melanoma susceptibility genes.

‘Members of such families are at significantly increased risk of developing melanomas. Many more melanoma patients have only one relative who also has melanoma. An intensive search for putative melanoma susceptibility genes has identified mutations in the CDKN2A gene in 20–30% of melanoma prone families in Scotland, reflecting rates reported in other parts of the world. Current expert consensus recommends that genetic testing in familial or sporadic melanoma is not appropriate in a routine clinical setting and should only be undertaken in the context of appropriate research studies and when appropriate counselling services are available.’(308)

Table 15Established Risk Factors for cutaneous melanomas(308)

Risk FactorORInformation
11–50 common moles >2mm1.7 – 1.9The risk of melanoma rises with the number of common moles.
51 – 100 common moles >2mm3.2 – 3.7
>100 common moles >2mm7.6 – 7.7
Family history of melanoma1.8Melanoma in a first degree family member (parent, sibling or child of the patient.
Previous history of melanomaStandardised incidence ratio range 4.5 – 25.6
Presence of 1–4 atypical moles1.6 – 7.3Atypical moles: ill-defined or irregular border; irregular pigmentation; diameter >5mm; erythaema (blanchable in lesion or edge); accentuated skin markings.
Red or light-coloured hair1.4 – 3.5
Presence of actinic lentigines1.9 – 3.5Actinic lentigines: flat, brown skin lesions associated with acute and chronic sun exposure. No direct malignant potential.
Giant congenital melanocytic naevi ≥20cm in diameterRelative risk range 239–1, 224 for extracutaneous as well as cutaneous melanoma.
Unusually high sun exposure2.6
Reported growth of a mole2.3
Skin that does not tan easily1.98
Light coloured eyes1.55 – 1.60
Light coloured skin1.40 – 1.42
AffluenceRelative risk approximately 3.0 for people residing in areas defined as Carstairs deprivation category 1 (least deprived) compared to Carstairs category 7 (most deprived).
Non-melanoma skin cancers

Basal cell carcinoma

Secondary studies
(Wong et al, 1989)(321)

This authoritative review concluded that exposure to ultraviolet radiation was the main causative factor in the pathogenesis of basal cell carcinoma. However, the precise relationship between risk of basal cell carcinoma and the amount, timing and pattern of exposure to ultraviolet radiation were unclear. The magnitude of the risk associated with increased exposure seemed to be insufficient to explain why particular people get these tumours whereas others did not. Several studies showed an association between cumulative ultraviolet exposure and risk of basal cell carcinoma, although the magnitude of risk conferred was small, with odds ratios in the region of 1.0 to 1.5. Other studies failed to find a significant association between estimated cumulative sun exposure in adulthood and the presence of basal cell carcinoma.

Skin type 1 (always burns, never tans), red or blonde hair and blue or green eyes have been shown to be risk factors for the development of basal cell carcinoma with an estimated odds ratio of 1.6. Development of basal cell carcinoma was reported to be more frequent after freckling in childhood and also after frequent or severe sunburn in childhood. This was in contrast to a story of sunburn as an adult, which does not seem to be associated with the development of basal cell carcinoma. Recreational sun exposure in childhood was identified as an important risk factor.

A positive family history of skin cancer seemed to be a predictor of development of basal cell carcinoma with an odds ratio estimated at 2.2. Several genetic conditions associated with the risk of developing basal cell carcinoma were albinism, xeroderma pigmentosa, and Bazex’s syndrome. Gorlin’s syndrome (the naevoid basal cell carcinoma syndrome) is a rare autosomal dominant condition in which patients develop multiple basal cell carcinomas and have other abnormalities including spine and rib anomalies, cataracts, and pitting of the palms and soles of the feet. Patients on immunosuppressive treatment also had an increased risk of basal cell carcinoma. The risk of developing a squamous cell carcinoma was increased slightly after a basal cell carcinoma, with a 6% risk at three years.

(Telfer et al, 1999)(312)

The British Association of Dermatologists guidelines stated that the most significant aetiological factor was chronic exposure to ultraviolet light, and consequently the head and neck were the most frequently affected sites. It mainly affects Caucasians, and increasing age, male sex, and a tendency to freckles were also identified as known risk factors.

Squamous cell carcinoma

Secondary studies
(Hawrot et al, 2003)(322)

This authoritative review described several risk factors. Patients with psoriasis treated with psoralens and ultraviolet A light (PUVA) were show to have an increased risk of squamous cell carcinoma that was associated with the number of treatments and the intensity of therapy. Long term follow-up studies of patients who underwent treatment with high doses of PUVA showed a relative risk of four to six compared with individuals not exposed to such treatments. PUVA effects appeared to be dose related and although lesions may occur as early as five years after therapy, the strongest correlation was seen in the second decade after therapy completion.

The incidence rate of cutaneous squamous cell carcinomas was increased in organ transplant recipients. Patients with transplants were at a three to four fold increased risk of systemic and cutaneous. An increased incidence rate of squamous cell carcinomas after transplantation was associated with time after transplantation, decreasing latitude and older age as well as childhood, duration of immunosuppression, intensity of immunosuppression, and history of skin cancer before transplantation.

In some studies the relative risk of squamous cell carcinomas was found to be approximately three times higher in people born in geographic areas receiving high amounts of ultraviolet radiation than in residents who moved to such areas only in adulthood; two to five times higher in those with very light skin colour, hazel or blue eyes and blonde or red hair; five times higher in individuals with exclusively outdoor occupations and three to eight times higher in people with severe versus no solar elastosis, freckling and facial telangiectasias. Although fair skinned caucasians, especially men in their 60s and 70s are at highest risk for cutaneous squamous cell carcinomas, other racial and ethnic types with intermediate skin types may be susceptible given predisposing environmental conditions.

Human papilloma virus (HPV) types 5, 16 and 18 were positively associated with squamous cell carcinomas. HPV types 16 and 18 were shown to produce cell line immortalization and tumour development in situ. The authors also considered that arsenic induces tumour formation. Therefore, metal ore workers and those with substantial exposure to insecticide were at risk. The carcinogenic effects of arsenic seemed to be dose dependent and may indicate internal malignant disease, especially if the skin tumour was in a non-exposed area.

Organ transplant recipients were at significantly increased risk for cutaneous squamous cell carcinomas and other cutaneous lesions. Persons infected with HIV showed a slightly higher incidence rate of cutaneous squamous cell carcinomas at relatively earlier ages than nonimmunosuppressed individuals although this finding was not confirmed. Individuals with chronically injured or inflamed skin with longstanding ulcers, sinus tracts, osteomyelitis, radiation dermatitis or burn scars were also at increased risk.

(Motley et al, 2002)(311)

Squamous cell carcinoma was usually related to chronic ultraviolet light exposure and was therefore especially common in sun damaged fair skinned individuals, in albinos and in those with xeroderma pigmentosum. It may develop de novo, as a result of previous exposure to ionising radiation or arsenic, within chronic wounds, scars, burns, ulcers or sinus tracts and from pre-existing lesions such as Bowen’s disease (intraepidermal squamous cell carcinoma). Individuals with impaired immune function, for example those receiving immunosuppressive drugs following allogeneic organ transplantation or those with lymphoma or leukaemia, showed increased risk of this tumour; some squamous cell carcinomas are associated with human papillomavirus infection. There was good evidence linking squamous cell carcinomas with chronic actinic damage and to support the use of sun avoidance, protective clothing and effective sunblocks in the prevention of actinic keratoses and squamous cell carcinomas; this was particularly important for patients receiving long-term immunosuppressive medication.

16.1. Investigations

16.1. Key Clinical Question

Should any investigations be undertaken in primary care, before referral?

16.2. Evidence Question

In people attending primary care services with dermatological symptoms, which investigations when compared with the “gold standard” are predictive of a diagnosis of cancer, and which are not?

16.3. Evidence Statements

Biopsy with histological examination is the standard investigation in people presenting with skins lesions that may be cancer. (III)

In comparison with specialists, a greater proportion of lesions excised by general practitioners are incompletely excised, and general practitioners are less able to predict malignancy. (III)

The investigation considered in this paper was excision biopsy, a procedure often undertaken in primary care.


(Department of Health, 2000)(2)

The Department of Health guidelines stated: ‘It is not recommended that patients with suspected melanoma are biopsied in a general practice setting. Patients should be referred with the lesion intact to the local specialist.

(SIGN, 2003)(308)

The SIGN guideline included the following recommendations:

GPs should refer urgently all patients in whom melanoma is a strong possibility rather than carry out a biopsy in primary care. (recommended best practice).

The local availability of fast-track services for patients in whom melanoma is suspected should be advertised widely to general practitioners. (recommended best practice).

A suspected melanoma should be excised with a 2mm margin and a cuff of fat. (D)

If complete excision cannot be performed as a primary procedure a full thickness incisional or punch biopsy of the most suspicious area is advised. (D)

A superficial shave biopsy is inappropriate for suspicious pigmented lesions. (C)

Primary studies

Nine articles were identified that reported studies of aspects of excision biopsy by general practitioners. Many of these were prompted by the general practitioner contract introduced in 1990, which included financial incentives for general practitioners to undertake minor surgical procedures. The ‘gold standard’ in all the studies was histological diagnosis.

(Bricknell, 1993)(323)

This study reviewed histopathology reports at one UK hospital with an aim of to examine the difference between skin biopsies of pigmented skin lesions taken by general practitioners and those taken by hospital specialists. It included 1205 biopsies involving 1000 patients, 15 of those patients had melanomas. General practitioners had undertaken 55% of the biopsies on the 1000 identified patients.

Features recorded on pathology forms included size increase (general practitioner 15.0%, specialists 25.1%), bleeding 13.6% vs. 6.6%, colour change 4.8% vs. 11.7% (all P<0.001). Hospital specialists excised significantly more lesions that had increased in size (P < 0.001) or changed in colour (P < 0.001). General practitioners excised more lesions that had bled (P < 0.001). Hospital specialists excised more of the 15 melanomas diagnosed (80%) (P < 0.05), and general practitioners excised more squamous papillomas (P < 0.01).

Of the melanomas excised, 40% were not suspected by the clinician. Although the study found that general practitioners were able to detect the majority of suspicious lesions, it concluded that all specimen’s should be submitted for histopathological diagnosis due to the uncertainty of clinical diagnosis. Additionally the paper commented that in order for general practitioners to carry out this minor operation, training is required in technical and diagnostic aspects of skin biopsy.

(Cox 1992)(324)

In this study, the findings of skin biopsies by general practitioners and examined at one UK hospital were reported. Of the total of 1017 biopsies, 56 (5.5%) were for malignant lesions. Of 21 basal cell carcinomas, nine had been considered by the general practitioner to be malignant. Six of the 21 had been inadequately excised. None of the four melanomas had been suspected, although they had been adequately excised. Additionally 21 squamous cell carcinomas were excised. Excision was adequate in eight, and the diagnosis had been suspected in only one.

(Khorshid, 1998)(325)

This study involved a survey of 819 pathology reports and interviews of 55 UK general practitioners who had submitted samples for analysis. 819 melanoma biopsies were identified, of which 59 were excised by the general practitioner. Various specialists excised the remaining melanomas.

15% of general practitioner excisions compared to 36% of non-general practitioner excisions were complete and adequate (P<0.001). General practitioners made an accurate clinical diagnosis in only 17% of cases.

(Herd, 1992)(326)

This UK retrospective case-control study included 42 biopsies performed by general practitioners which were found to be melanoma, compared to 84 randomly selected biopsies carried out in hospitals.

The Breslow thickness of lesions was not significantly different. Ten of the general practitioner excisions were incomplete compared with only three incomplete in the hospital sample (P<0.001).

Only six (15%) of the 40 general practitioner request forms mentioned the possibility of melanoma. Six had been excised for cosmetic reasons alone. The other reasons were change in size (N=25), and patient worry about malignancy (N=16).

(Hillan, 1991)(327)

This study reviewed 149 specimens referred by UK general practitioners to one hospital laboratory. The specimens included one melanoma, and two basal cell carcinomas. No squamous cell carcinomas were identified. 10% of the general practitioner specimens and 11% of a comparison group of specimens referred from the hospital were inadequately excised.

(Lowy, 1997)(328)

This study reviewed pathology specimens before and after the introduction of a policy of referring all removed tissue in the UK in order to examine whether histological examination of all tissue removed by general practitioners in minor surgery increases the rate of detection of clinically important skin lesions. A random sample of specimens sent by 257 general practices referring to 19 pathology laboratories was undertaken.

During the intervention period 5723 specimens were sent, compared with 4430 during the control period. The referral rate increased by an estimated 1.34 specimens per 1000 patient years (95% confidence interval 0.93 to 1.76, P < 0.0001). During the control period general practitioners sent 204 specimens (188 non-melanoma and 16 melanoma), compared with 188 specimens sent during the intervention period (173 non-melanomas and 15 melanomas).

This multi-centre study that showed no increase in detection of malignancy through a policy of referring all general practitioner skin excisions for histological examination.

(McWilliam, 1991)(329)

This study reported a retrospective analysis of histology records at one UK hospital, and included 292 skin biopsy specimens by general practitioners and 324 by general and plastic surgeons. General practitioner cases included six (2%) basal cell carcinomas, five (1%) squamous cell carcinomas, and one (0.3%) melanoma. 36% of all general practitioner’s samples compared with 16% of surgeons’ samples were incompletely excised. Agreement between clinical and pathological diagnosis in malignant cases was 29% for general practitioners and 90% for surgeons.

(O’Cathain, 1992)(330)

This study reported a UK prospective comparison of patients undergoing minor surgery in general practice and at one hospital. A total of 161 patients were compared, 67 of those in general practice and 94 in hospital. 9.8% of general practitioner cases and 1.2% of hospital cases were malignancies diagnosed as benign. 4.9% of general practitioner cases compared to 0% of hospital cases had not been adequately excised.

(Williams, 1991)(331)

This retrospective review of pathology records in one UK hospital evaluated 571 skin biopsy specimens from general practitioners. 26 (4.6%) biopsies were malignant (14 basal cell carcinomas, eight squamous cell carcinomas, four melanomas). The study did not assess completeness of excision

The articles generally indicated that a greater proportion of skin lesions performed by general practitioners had been inadequately excised in comparison with specialists, and that general practitioners were less able than specialists to predict malignancy among the lesions they excise. A small proportion of lesions excised by general practitioners turned out to be malignant (basal cell carcinomas, squamous cell carcinomas and melanomas). Mandatory referral of excised specimens for pathological examination has been recommended in joint guidelines of the Royal Colleges of General Practitioners and the General Medical Services Committee with the support of the Royal College of Surgeons of England, the Royal College of Surgeons of Edinburgh, and the Joint Committee of Postgraduate Training for General Practice and other organisations,(332) but the value of this policy was been questioned by Lowy et al (1997).(328) The guidelines stated: ‘Occasionally malignant lesions will be encountered which were not diagnosed clinically. All lesions removed by minor surgery should be sent for histological examination. There should be a written procedure in the practice which ensures that the pathology report is seen either by the general practitioner operator, or by the patient’s own general practitioner if different, and that any necessary action is initialled in writing.’

The joint guidelines also promoted training for general practitioners, audit of minor surgery, adequate staffing, premises, equipment, sterilization, and attention to the hepatitis status of the general practitioner and staff. General practitioners who complied with these guidelines would be eligible for inclusion in the minor surgery list.

16.3. Delay and Diagnostic Difficulties

16.3.1. Key Clinical Questions

In people attending primary care services with skin cancer symptoms, which psychosocial and socio-demographic factors are associated with delayed presentation? Which factors influence delay by patient and which delay by provider?

What diagnostic difficulties do primary care practitioners themselves report in determining whether a patient who presents with dermatological symptoms/signs may or may not need urgent referral with suspected cancer?

16.3.2. Evidence Questions

In people attending primary care services with skin cancer symptoms, which psychosocial and socio-demographic factors are associated with delayed presentation? Which factors influence delay by patient and which delay by provider?

What diagnostic difficulties do primary care practitioners themselves report in determining whether a patient who presents with dermatological symptoms/signs may or may not need urgent referral with suspected cancer?

16.3.3. Evidence Statements

Delay in the detection of melanoma may be associated with patients’ lack of awareness of the significance of signs of the condition (III).

Other factors may also influence patient delay in melanoma, including the site of the lesion, educational level, and anxiety. (III)

In countries that do not have primary care gatekeeping, delay is less when specialists make the diagnosis. (III)

Misdiagnosis on inspection of skin lesions by clinicians can lead to delay in diagnosis in some patients. (III)



A summary of the available evidence on diagnostic delays for skin cancers is presented below. Although a large number of studies explored delays in diagnosis of melanomas, no relevant studies were identified that addressed diagnostic delays for both basal cell carcinomas and squamous cell carcinomas.

Therefore, all of the studies included in this review were surveys of patients with cutaneous melanoma. Most explored not only the relationship between patient/doctor characteristics and delays in diagnosis, but also the relationship of these characteristics and tumour thickness. It is of note that in the case of cutaneous melanomas, longer delays in diagnosis did not necessarily mean a worse prognosis. The relationship between diagnostic delay and tumour thickness was far from linear, probably indicating that melanoma thickness was not only a product of the delay in diagnosis but also of the biological aggressiveness of the tumour.

Approximately 70% of cutaneous melanomas were detected by the patients themselves. Longer medical delays and greater tumour thickness at diagnosis appeared to occur more frequently in men, the elderly, poorly educated individuals, residents of rural areas, and people with little knowledge about melanocytic tumours. Common reasons given by patients for failing to seek medical advice were absence of systemic signs, absence of awareness of the urgency, fear and anxiety.

Malignant melanoma was misdiagnosed in approximately one in every six affected patients. Dermatologists appeared to have shorter diagnostic delays than general practitioners. Tumour thickness tended to be lower in melanomas diagnosed by dermatologists, and also in melanomas diagnosed coincidentally by any physician. The fact that most of the studies described took place in non-UK health care systems, where access to specialists was not dependent upon a previous referral by the patient’s general practitioner, limited the extrapolation of these findings to the NHS.


Secondary studies
(Silfen et al, 2002)(333)

In this authoritative review, the authors investigated the role of the physician and the patients in diagnostic delay of melanoma.

  • Physicians
    Tumour characteristics had an important effect, a shorter medical delay occurring for nodular and lentigo melanoma than for acrolentiginous melanoma. Longer diagnostic delays were also associated with tumours deriving from nevi compared with de novo melanomas.
  • Patients
    In one case-control study, monthly skin self-examination was associated with a 63% reduction in mortality from melanoma.
Primary studies
(Betti et al, 2003)(334)

The aim of the study was to investigate factors related to early detection of melanoma and factors associated with delay. Consecutive patients referred to an Italian hospital with cutaneous melanoma between September 1994 and December 2000 were interviewed by a trained dermatologist. The questionnaire included demographic, tumour and behavioural data. All patients had histologically confirmed melanoma. Patients who were not able to respond accurately to the questionnaire were not included in the study.

216 out of the 270 patients approached were enrolled in the study. Mean patient delay was 6.11 months (range ± 9.75 months), and mean medical delay was 1.53 months (range ± 5.34 months). There were no differences among causes of patient delay and mean age, anatomic site of lesions, level of education, knowledge of the problem, civil status or pigmentation. 51% of the patients delayed the consultation of a physician because of anxiety, fear, or lack of no time or being too busy. They tended to have a longer patient delay and a higher Breslow thickness (0.99 ± 1.41) (P < 0.001).

22 cases (10.19%) were observed in which the practitioner or the specialist delayed diagnosis or treatment. No correlation between physician delay and anatomic location of the lesion was observed. Pigmentation of the lesion significantly delayed the time of diagnosis by the physician (4 ± 9 months vs. 1.34 ± 5 months for the pigmented melanomas) (P < 0.04).

(Brochez et al, 2001)(335)

The aim of this study was to describe the diagnostic pathway for cutaneous melanoma in a Belgian community, to quantify both patient and physician delay and to define factors related to it. Patients were recruited both from a university hospital setting and from practices (population based melanoma register). All patients with a diagnosis of cutaneous melanoma between January 1995 and December 1999 were included, and asked to complete in a questionnaire about delay in diagnosis.

131 questionnaires were completed. The time from the first noticing a new or changing lesion to consultation with a physician (patient delay) was a mean of 169 days (median, 61 days). Worried patients tended to have a longer patient delay, although the difference did not reach statistical significance. There was no difference in patient delay for lesions difficult to self-examine compared with lesions more easily self-examined such as head and neck, chest, abdomen, arms, extensor side of the legs. Colour change and itch were associated with longer patient delay (median 64 days vs. 24 days if no colour change, P < 0.05; and 137.5 days vs. 29 days if no itch, P < 0.01). Patient delay was not influenced by age, gender or socio-economic factors.

General practitioners and dermatologists were the physicians most frequently involved in the first medical encounter about a lesion (55 and 33% of all cases, respectively). Of the physicians who first observed the lesion, 34 of the 43 dermatologists suspected the lesion immediately, compared with 38 of 72 general practitioners (x² = 7.95, P = 0.005). There were significant differences in the time to excision if the physician took immediate action, referred the patient or took no immediate action.

(Oliveria et al, 1999)(336)

The purpose of this large population-based case control study was to examine the relationship between patients’ knowledge and awareness of the signs and symptoms of melanoma and delay in seeking medical attention for suspicious lesions. The study included 650 Caucasian residents of Connecticut 18 years of age or older with cutaneous melanoma newly diagnosed between 1987 to 1989, who were part of a population-based control study. Patients who had their melanoma identified by a physician during a visit for an unrelated condition were excluded (N= 395). Cases were identified from pathology reports and hospital tumour registry logs (N= 255, participation rate = 75%). Personal interviews were conducted to obtain information on patient’s knowledge of melanoma signs and symptoms, skin awareness, and delay in seeking medical attention.

The mean delay time for patients seeking medical attention was two months with a range from 0.5 to 22 months. Overall, the results revealed an inverse relationship between both knowledge and awareness and delay in seeking medical attention for melanoma. The odds ratios for knowledge of melanoma characteristics and delay ranged from 0.42 to 0.81 after controlling for age, gender, prior history of cancer, and skin self-examination. Patients who were aware of skin changes and or abnormalities had a reduced likelihood of delay in melanoma diagnosis after adjusting for age, gender, prior history of cancer, and skin self-examination practices (OR = 0.30, 95% CI = 0.12 –0.71). The findings suggested that knowledge of two or more signs or symptoms of melanoma reduces the likelihood of a delayed diagnosis (OR = 0.34, 95% CI = 0.13–0.88).

Skin awareness was associated with a reduced thickness (OR = 0.50, 95% CI = 0.28–0.89). Increased knowledge of melanoma signs and symptoms also decreased the likelihood of being diagnosed with a thick tumour (≥ 0.75 mm). The odds ratio ranged from 0.69 to 0.95 for the knowledge variables (except for larger diameter and abnormal shape, odds ratios = 1.17 and 1.14 respectively).

(Carli et al, 2003)(337)

The aim of this study was to investigate patterns of detection and variables associated with early diagnosis of melanoma in a population at intermediate melanoma risk. The study included 816 patients with cutaneous melanoma diagnosed in 2001, in 11 Italian clinical centres.

The patients, all caucasians with newly diagnosed lesions, were included in the study at the first visit after surgery, when the diagnosis of melanoma was histologically confirmed.

Each patient received a questionnaire about first identification of the lesion, the interval before diagnosis by a dermatologist or another specialist (patient’s delay), and the interval before the lesion was removed (physician’s delay). Patients were also asked about their knowledge of the criteria for early diagnosis of melanoma, their skin self-examination habits, and periodic medical consultation aimed to screen for melanoma. The main outcome measure was the relationship between patterns of detection and patients’ and physicians’ delays with melanoma thickness.

The mean (± SD) age of patients was 53.8 (± 14.8) years for men and 49.6 (± 15.7) years for women.

Patterns of melanoma detection

Most patients self-detected melanoma. Their spouse detected 12.5% of the lesions, while physicians first detected 38.7% of the lesions. The percentage of melanomas detected by a spouse differed according to sex (18.5% in male patients vs. 6.4% in female patients; x² test, P = .000). More than half of the subjects (68.9%) waited no more than three months before obtaining a diagnosis. The main reasons for longer waiting were the feeling that it was not important (56%), fear about a possible diagnosis of cancer (10.1%), lack of time (7.3%), and the mistaken opinion that to remove a naevus is dangerous (5.6%). Fifty-two patients (21%) reported waiting more than three months because another physician, seldom the family physician, did not think it was really a lesion suggestive of being a melanoma.

Effects on mean thickness

A lower mean thickness was significantly associated with female sex, high educational level, and the habit of performing skin self-examination. Age older than 60 years was associated with a higher mean thickness, compared with age younger than 40 years. Paradoxically, a lower mean thickness was found in those patients who waited more than one month before surgery once a definite diagnosis of a lesion suggestive of a melanoma was established (adjusted mean thickness, 0.74 vs. 0.89 mm).

Association with diagnosis of thin lesions

A statistically significant association with early diagnosis was found for female sex (odds ratio [OR] for a lesion >1mm in thickness, 0.70; 95% confidence interval [CI], 0.50–0.97), higher educational level (OR, 0.44; 95% CI, 0.24–0.79), and the habit of performing skin self-examination (OR, 0.65; 95% CI, 0.45–0.93). The association with age was of borderline statistical significance.

(Montella et al, 2002)(338)

The study’s aims were to test the relationship between tumour thickness and social and clinical variables (including diagnosis/treatment delay), and the relationship between delay and clinical variables. The authors undertook a retrospective study of 530 consecutive patients who underwent surgery for histologically confirmed melanoma between January 1996 and December 2000 at a single Italian hospital. Patients with an unknown primary site and metastatic tumour were excluded.

Data obtained at interview included: age, education, occupational status, diagnosis mode (symptomatic, asymptomatic, or incidental), visibility of tumour, and first symptom. Medical records were also inspected to extract information on some patient characteristics.

The most frequently reported symptoms were a lesion with increasing size (50.8%), bleeding (17.8%), colour change (15.2%), and itching (12.0%).

Breslow thickness

A larger proportion of females (72.1%) compared with males (64.4%) had a Breslow tumour thickness < 1.5 mm (OR = 1.8, 95% CI = 1.2–2.8, P = 0.005). A significant risk of having a Breslow tumour thickness ≥ 1.5 mm was noted in patients who had a low level of education (OR 3.0, 95% CI 1.9–5.0, p = 0.0001) or who were unemployed (OR = 1.7, 95% CI = 1.1–2.8, P = 0.001). A significant risk of Breslow tumour thickness ≥ 1.5 mm was reported for patients who were examined by a physician other than a dermatologist (OR = 1.8, 95% CI = 1.2–2.8).

Patient delay

A greater than three month delay was observed for anatomic locations visible to patients (OR = 1.7, 95% CI = 1.1–2.6, P = 0.02). Anatomic site of the primary lesion was also related to patient delay: patients who had the primary lesion on an extremity were more likely to delay > three months (OR = 1.6, 95% CI = 1.1–2.5, P = 0.02), especially females (OR = 2.2, 95% CI = 1.3–3.7, no P value given).

Medical delay

A significant association was observed between medical delay and the physician who made the diagnosis: a delay > three months carried a higher risk (OR = 2.0, 95% CI = 1.2–3.4, P = 0.01) in patients examined by a dermatologist. A medical delay of one to three months for patients with a primary lesion on an extremity was associated with an increased risk of melanoma (OR = 1.8, 95% CI = 1.0–2.9, P = 0.03).

None of the other variables studied (gender, age at diagnosis, education, and occupational status) were significantly associated with either patient or medical delay.

(Blum et al, 1999)(339)

This study included all patients (N= 429) with histologically confirmed cutaneous melanoma who had undergone surgical treatment at a Swiss hospital between 1993 and 1996. Patients were interviewed using a standardised questionnaire, the information obtained being merged with the data on tumour characteristics and case history contained in the medical records. Delay in melanoma diagnosis was defined as the time period between a patient’s first observation of a suspicious skin lesion and definite tumour treatment.

429 patients were interviewed (184 men, 245 women), median age 52 years. The melanoma was detected in 67% of women and 45% of men by the patients themselves (inter-gender comparison: P < 0.0001). The tumour was detected in about 50% of the remaining patients by a physician. Earlier diagnosis and treatment of melanoma were not significantly related to prognostic tumour parameters such as Breslow thickness or Clark’s level of invasion. Women were significantly more aware than men of the possible benefit of early treatment (P= 0.004). However, increased melanoma awareness was not associated with an earlier visit to a physician. Patients who detected the lesions themselves sought medical attention later than patients in whom attention had been called to their skin changes by other persons (median 122 vs. 59 days), and therefore were treated significantly later (P < 0.01). A misdiagnosis by the first physician visited was reported by 18% of patients, and 60% of these physicians were dermatologists. Misdiagnosis increased the period of time between first observation and treatment (median 122 vs. 31 days, P < 0.0001) as well as between the first visit to a doctor and treatment (median 61 vs. 28 days, P < 0.0001). When more than one physician omitted the diagnosis of melanoma (in 8% of all patients), there was a significant additional delay in treatment (median 303 vs. 89 days, P < 0.001).

Multiple regression analysis revealed the following factors to be significantly related to delay in melanoma diagnosis: denial of melanoma diagnosis by the first physician visited (P < 0.001, regression coefficient = 0.192), invasive melanoma of the head and neck (P < 0.05, regression coefficient = 0.134), self detection of melanoma vs. detection by other persons (P < 0.05, regression coefficient = 0.129), and patient’s knowledge about the induction of skin cancer by sun exposure (P < 0.05, regression coefficient = − 0.107). No correlation was found between delay in diagnosis/treatment and gender, age, Breslow tumour thickness, Clark’s level of invasion and histological type of melanoma.

(Richard et al, 2000a)(340)

This paper evaluated the role of a patient in contributing to delay in diagnosis of skin cancer. Consecutive patients referred for cutaneous melanoma to 18 French dermatological departments of the public hospital system participated in the study conducted between 1995 and 1996. Inclusion criteria were: at least 12 years of age, histological confirmation of diagnosis of melanoma, and interview within 12 weeks after melanoma resection. Patients were included only when the report forms were completed, when a histological slide was available, and when two experts confirmed the diagnosis. A total of 645 were entered by the centres, but only the 590 fulfilled all these criteria and included in the analysis.

All patients were examined and interviewed by a specially trained dermatologist in each centre. The questionnaire addressed patients’ characteristics such as age, sex, residence, social level, and education level, amongst others.

42.4% of the sample were males and 57.6% females. Tumour thickness in coincidentally diagnosed melanoma was significantly lower than in self-diagnosed melanoma (median 0.93 mm vs. 1.30 mm, P < 0.001). Median tumour thickness was significantly lower when the lesion was first detected by the patient than when it was detected by the family (1.22 mm vs. 1.40 mm, P < 0.001, Kruskal-Wallis test).

Reasons for delay according to the patient Patients delayed presentation to a physician beyond two months in 48.1% of cases. The reasons given were: innocent appearance of the lesion together with the absence of systemic signs in 39.3%, absence of awareness about the urgency in 34.8%, occupational reasons in 20.4%, familial reasons in 16.9%, fear of diagnosis in 9.4%, passivity until family urged consultation in 5.5%, negligence in 4.5%, and absence of pain in 1.0%.

Comparison of the self-detected and the coincidentally diagnosed melanoma Melanomas were more often self-detected by women than by men: 74.1% vs. 66.8%, respectively (x² test, P = 0.053). The patients with a self-detected melanoma had a significantly higher educational level than the patient with a coincidentally diagnosed melanoma (53.1% vs. 65.7%, x² test, P = 0.03). The patients with a coincidentally diagnosed melanoma lived more frequently in the countryside than the patient with a self-detected melanoma (29.6% vs. 20.3%, x² test, P = 0.02). Previous history of melanoma was more frequent in the patients with a coincidentally diagnosed melanoma than in the patients with a self-detected melanoma (27.9% vs. 16.5%, P < 0.001). The degree of awareness about skin, sun, and cancer was higher in patients who later detected their melanoma themselves than in those whose tumour was coincidentally detected.

Univariate analysis showed that people older than 65 years sought medical attention more quickly than people younger than 50 years (P = 0.003), but they tended to develop thicker tumours (P = 0.51). Gender did not influence significantly any component of the delays, although Breslow thickness was higher in men than women (P < 0.001). Delays did not differ in patients with high and low level of education, although those with low education level had thicker tumours (P < 0.001). There was no difference in the socioeconomic profile of the patients in regard to delays or Breslow thickness. Delays or tumour thickness were not influenced by marital status. People living in the countryside, although seeking medical attention more rapidly (P = 0.003), developed thicker tumours (P = 0.045). Awareness and information about melanoma did not have any significant impact on patient delay. Tumour thickness was significantly thinner when the patient had already heard about melanoma and was previously aware of the early signs of melanoma.

In a multivariate analysis, none of the candidate variables related to patient delay significantly predicted independently patient delay in multivariate analysis. In a stepwise multiple linear regression using all variables influencing tumour thickness, three variables were predictive of a high Breslow: ulceration, the fact that the patient said that raising was the reason for consultation, and nodular histological type.

(Richard et al, 2000b)(341)

The purpose of the study was to assess all doctor-related components in the delay before melanoma resection. Consecutive patients referred for cutaneous melanoma to 18 French dermatological departments of the public hospital system participated in the study conducted between 1995 and 1996.

Inclusion criteria were: at least 12 years of age, histological confirmation of diagnosis of melanoma, and interview within 12 weeks after melanoma resection. Patients were accepted only when the report forms were completed, when a histological slide was available, and when two experts confirmed the diagnosis. A total of 645 were entered by the centres, but only the 590 who fulfilled all these criteria were included in the analysis.

All patients were examined and interviewed by a specially trained dermatologist in each centre. The questionnaire investigated patient characteristics and habits, tumour clinical features, circumstances of melanoma detection, causes of delay in diagnosis, and doctors attitudes before removal. Physician delay was defined as the interval between the date the lesion was first examined by a physician and the date when a physician first proposed resection.

The median delay before the doctor proposed tumour resection was 0 (mean 103, range 0–5,783) days. For comparison, the median delay under patient responsibility was 912 (mean 3,829, range 0–5,261) days.

The first advice from the first doctor was considered to be appropriate in 85.8% of cases.

The delay to propose resection was much longer when the attitude of the first physician was inappropriate than when removal was proposed at the first visit (median 109 days vs. 0 days, P< 0.001). Although there was a higher tumour thickness when the attitude was inappropriate (median 1.40 vs. 1.15 mm, mean 3.15 vs. 2.00 mm), the difference was not significant (P = 0.99).

Tumour thickness was significantly lower when first seen by a dermatologist than by another physician (median 0.94 mm vs. 1.50 mm, mean 1.88 mm vs. 2.82 mm, respectively; P< 0.001). The delay to propose removal was significantly shorter when the first physician was a dermatologist than when he or she was a general practitioner or another specialist (median 0 vs. 25 days, mean 60 vs. 153 days, respectively; P < 0.001).

In self-detected tumours, doctors proposed removal significantly later for acrolentiginous melanoma, amelanotic melanomas, and melanomas of the hand and foot than for other tumours.

In a stepwise multiple linear regression, the most predictive factors influencing physician delay were histoclinical type and the ability of the first physician seen to recognise melanoma. The shorter delays were observed with lentigo melanoma and melanomas first seen by dermatologists. In a stepwise logistic regression, the factor most predictive of a long physician delay (> 30 days) remained the specialty of the first physician (other physicians vs. dermatologists; coefficient 2.27, SE 0.32, OR 9.7, 95% CI 5.16–18.2, P< 0.001).

(Schmid-Wendtner et al, 2002)(342)

The aim of the study was to investigate the extent and consequence of patient and professional delay in diagnosis and treatment of cutaneous melanoma. Between 1999 and 2001, 233 patients with histologically confirmed primary cutaneous melanoma diagnosed and treated at a German university hospital, were within three months of diagnosis. The interview investigated melanoma-associated symptoms, the site and features of the cutaneous melanoma, time intervals, and reasons for delay in diagnosis.

Patients with knowledge about melanoma presented with a median tumour thickness of 0.7 mm, whereas patients without knowledge had a median tumour thickness of 2.1 mm (P < 0.0001). Knowledge about melanoma was associated with the educational status of patients. More than 90% of patients with a high or medium educational status had knowledge about melanoma, and less than 10% had no knowledge about melanoma (P < 0.001). In contrast, only 71% of patients with low educational status were knowledgeable about melanoma.

Medical attention was sought within 1 month of noticing the appearance of a new lesion or the onset of changes in a pre-existing lesion by 15.5% of patients. Longer periods of patient delay were not associated with greater tumour thickness. The majority of patients asked about the reasons for delay had initially thought that the pigmented lesion was benign or not important (63.5%). A smaller group of patients did not delay the consultation of a physician (12.0%), 9.9% of patients were afraid of the physician’s diagnosis, 8.1% of patients could not detect the lesions themselves because of its anatomical site, and 6.9% mentioned that they were too busy to consult a physician. In 3% of patients the reasons for delay remained unclear.

(Cassileth et al, 1988)(343)

In this study, consecutive patients with cutaneous malignant melanoma referred to two US hospital-based melanoma clinics by community physicians between 1984 and 1986 participated in the study. Patients were white and over the age of 18.

The authors conducted interviews with all the patients (N = 275) and also the physicians (N = 437) whom they had consulted regarding their suspicious lesions before their eventual referral to a melanoma centre. Histology data were obtained for all patients.

A mean of six months elapsed (median one month) between the time that patients first noticed a new mark or a change in an existing lesion and the time that they became suspicious about it. The particular characteristics of lesions noted by patients did not influence length of time to suspicion. A mean of 2.6 additional months elapsed following suspicion until patients sought medical attention. The median delay during this period was one month. No lesion signs or characteristics were related to how quickly patients sought medical attention. The most common reason given by patients to explain this delay was that the lesion “did not represent an urgent problem”.

For the entire subject population, the mean time from the initial physician visit to the diagnosis of malignant melanoma was 3.9 months. Time from initial physician visit to diagnosis was shorter only for lesions with pigmentation (P = 0.002). No other lesion characteristic was associated with length of delay from initial visit to diagnosis.

Physicians alerted primarily by the lesion’s pigmentation and/or by its diameter or border, recalled having assessed the lesion clinically as a melanoma in 74% of patients. There was a significant relationship between correct identification of melanoma and physicians’ specialty (chi square, P < 0.05). Surgeons and dermatologists were more likely than other physicians to have identified the lesion correctly. The relationship between self-rated knowledge and correct identification of melanoma did not achieve statistical significance.

Physicians’ actions in response to this initial evaluation were associated with type of specialty practice (chi square, P < 0.001). Internists were most likely to make an immediate referral to a melanoma clinic, and surgeons were least likely to do so. Lesion characteristics were not associated with melanoma referral. Half of physicians interviewed reported that they did not examine the patient’s entire cutaneous surface. 52% of patients were seen by more than one physician prior to melanoma clinic referral. Patients who saw more than one physician were diagnosed as having melanoma a mean of 6.8 months after becoming suspicious about their lesions, compared to 4.1 months for patients who saw only one physician prior to melanoma clinic referral (Mann-Whitney U test, P= 0.006). Further, the interval from the initial physician appointment to diagnosis was greater for patients seen by more than one physician (5.8 months) than for patients seen by only one physician (1.8 months; P < 0.0001 by Mann-Whitney U test).

Of all the demographic variables analysed (sex, occupation, education, marital status, health insurance, and age), only sex was significantly associated with delay. Men waited an average of 1.9 months and women an average of 3.3 months before seeing a physician after becoming suspicious about their lesions (P < 0.005 with the Mann-Whitney U test). Neither patients’ self-rated awareness of body changes nor their scores on the preoccupation with appearance test were associated with any component of delay, with tumour thickness, or with level of invasion.

(Rampen et al, 1989)(344)

The aim of the study was to relate possible delay factors to the most important prognostic features at the time of diagnosis (the clinical stage of the disease for all patients, and the maximal tumour thickness). The study comprised consecutive patients (N = 284) with cutaneous melanoma presenting with primaries or metastases to 12 Dutch hospitals. Patients with non-invasive (Clark level 1) melanoma, patients who refused taking part in the study, and patients who were mentally unsuitable for the enquiry were excluded (N = 16).

All patients were interviewed shortly after diagnosis using a detailed questionnaire about the patient’s history, tumour characteristics, treatment particulars, and pathology.

The interval between the onset of signs and the first visit to a doctor tended to increase with age (P = 0.055). Females presented with less advanced disease than males, particularly in stage I disease (P = 0.004). Visibility of the primary lesion had no impact on the stage of the disease. The average interval between the appearance of the first signs and doctor’s consultation was similar in males and females. For both sexes, the interval was considerably longer for the easily visible melanomas than for the more hidden ones (P < 0.001, adjusted for sex). If patients suspected they had cancer, this tended to have a favourable impact on the stage of the disease (for the microstage P = 0.079, for the clinical stage P = 0.049). There was no evidence that patients in the higher socio-economic class have a better knowledge of the malignant nature of their disease (P = 0.076). Even if patients were aware of the possible malignant character of the growth, they often displayed a delay of more than one month before they consulted a doctor (54% of cases, N = 63). The reasons given for this delay were a feeling that the situation was not pressing in 41, lack of time in 24, fear of cancer in 15, aversion of going to the doctor in ten, and miscellaneous reasons in nine patients (many patients gave more than one reason).

Patients who had waited until their symptoms became severe enough to seek medical care by themselves, had a more advanced clinical stage of the disease than those who had been persuaded by someone else to go to the doctor, or than those whose melanoma had been discovered by chance (P = 0.018). Patients who presented their melanoma secondary to another reason for visiting the doctor had a more favourable clinical stage and the primary melanomas were considerably thinner (P < 0.001).

When doctors found a primary melanoma by chance, the microstage appeared to be much more favourable than when patients themselves had noticed a suspicious lesion (P < 0.001). Patients with amelanotic melanomas had more unfavourable microstages than those with melanotic primaries (P < 0.006). Melanoma suspicion was highest for melanotic and lowest for amelanotic tumours (P = 0.049).

Basal and squamous cell carcinomas

No suitable studies were identified


One systematic review of studies comparing the diagnostic performance of general practitioners and dermatologists in diagnosing melanoma was identified, and two additional primary studies that had not been included in the review were also assessed. No studies addressed the difficulties in diagnosis of basal cell carcinomas or squamous cell carcinomas.

Diagnostic difficulties

Secondary studies
(Chen, 2001)(345)

This systematic review was undertaken in order to compare the diagnostic accuracy and biopsy or referral accuracy of dermatologists and primary care physicians. Studies that presented sufficient data to determine the sensitivity and specificity of dermatologists’ or primary care physicians ability to correctly diagnose lesions suggestive of melanoma and to perform biopsies on or refer patients with such lesions. Studies published between January 1966 and October 1999 in MEDLINE, EMBASE and CancerLit databases were retrieved.

Two reviewers independently abstracted data on the sensitivity and specificity of the dermatologists and primary care physicians for diagnostic and biopsy or referral accuracy. Strict criteria for inclusion were applied to ensure results were comparable across studies. Thirty-two studies met the inclusion criteria. Ten of these were prospective, of which nine provided data for diagnostic accuracy and only one reported data for biopsy or referral accuracy. The nine prospective studies provided data from 583 dermatologists and 2314 primary care physicians for a comparison of melanoma diagnostic accuracy. Five of the nine prospective studies (plus an additional one) provided data for the biopsy or referral accuracy analysis. These studies included data on 106 dermatologists and 886 primary care physicians. Two of the studies used histopathologic analysis as the gold standard to define whether the biopsy or referral decision was correct, whereas another four used an expert panel of physicians.

The study designs differed considerably and were either prospective assessments or retrospective histopathologic reviews. Some studies permitted the physicians to give a list of differential diagnoses. Other researchers accepted only one diagnosis or only considered the first diagnosis if a list was given. For diagnostic accuracy, sensitivity for dermatologists was 0.81 to 1.00 for diagnostic accuracy (calculated from six studies) and 0.42 to 1.00 (from nine studies) for primary care physicians. None of the studies reported specificity for dermatologists. One study reported specificity for primary care physicians (0.98). For biopsy or referral accuracy, sensitivity ranged from 0.82 to 1.00 (from five studies) for dermatologists and 0.70 to 0.88 (from six studies) for primary care physicians. The range of specificity was 0.70 to 0.89 (from three studies) for dermatologists and 0.70 to 0.87 (from four studies) for primary care physicians.

Most of the studies included in the review evaluated only diagnostic accuracy and not biopsy or referral and did not report either sensitivity or specificity, and did not have an adequate sample size or describe the lesions shown to subjects.

Primary studies
(Brochez, 2001)(346)

This study was not included in the review (Chen, 2001(345)). It aimed to compare the diagnostic abilities of general practitioners and dermatologists in Belgium concerning pigmented skin lesions in general and melanoma in particular. The study design was a ‘before and after’ evaluation of a health education programme for general practitioners. A test set of 13 pigmented skin lesions on 35 mm colour slides as presented to 160 participating general practitioners and 60 dermatologists during a monthly educational course.

An invitation was addressed to 67 educational groups (representing 1956 general practitioners). Eight groups (160 general practitioners) accepted the invitation. The 160 general practitioners participating in the study represented 8% of all general practitioners in East-Flanders and 1% nationwide. Sixty dermatologists attending a monthly educational course were given the same test. The participating dermatologists represented about 7% of all those in the country.

The frequency of melanomas encountered was one in seven years for the general practitioners and one in eight months for dermatologists. Consultations for advice about pigmented lesions were encountered once in 30 days by general practitioners and once per day by dermatologists.

Sensitivity of general practitioners before the course in diagnosing melanoma from the slides was 72%, and 84% afterwards (dermatologists 91%). Specificity among general practitioners was 71% before and 70% after, and 95% among dermatologists. The positive predictive value (PPV) of general practitioners before was 61%, and 63% after (dermatologists 92%). The negative predictive value was 80% before and 87% after among general practitioners (dermatologists 95%).

(Girgis et al 1996)(347)

Questionnaires were sent to 141 randomly selected family physicians in one region in Australia to investigate their beliefs and practices in relation to skin cancer prevention, early detection and management. A total of 97 (69%) responded. Compared with family physicians throughout Australia, the survey had significantly fewer family physicians aged less than 30 years, and a significantly higher proportion aged 40 to 49 years.

Ninety-one percent of family physicians (N=86) indicated that they thought skin examinations were very or extremely worthwhile in the early detection of melanoma and other skin cancers. The three issues in which they felt most confident were performing a surgical excision (72%), diagnosing a basal cell carcinoma (71%), and advising patients on signs of skin cancer (69%). A total of 65% (53) of family physicians considered that they currently detected 90 to 100% of their patients with melanoma. Family physicians indicated that the factors most likely to encourage them to offer screening were patients being more informed about its benefits (82%; N=79), patients initiating the procedure (64%; N=62), having instructions about the signs to look for (61%; N=59), having long consultation times and a reduced patient workload (59%; N=57), and having consistent information about who needs screening and how often (57%; N=55). The factors that were most likely to discourage family physicians from screening their patients included lack of time (32%; N=31), forgetting (26%; N=25), lack of financial incentive (20%; N=19), not being familiar with the patients’ screening history (14%; N=14) and inability to convince patients who refuse (13%; N=13).

Copyright © 2005, National Collaborating Centre for Primary Care.
Cover of Referral Guidelines for Suspected Cancer in Adults and Children
Referral Guidelines for Suspected Cancer in Adults and Children [Internet].
NICE Clinical Guidelines, No. 27.
Clinical Governance Research and Development Unit (CGRDU), Department of Health Sciences, University of Leicester.

NICE (National Institute for Health and Care Excellence)

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...