Introduction
Cutaneous melanoma is an increasingly significant public health concern worldwide. Although it accounts for only 10% of skin cancers, it is responsible for more than 75% of skin cancer-related mortality, and its incidence rate has been increasing for decades
[
1
]
. Early detection is essential to mitigate this burden
[
2
–
3
]
. Indeed, the five-year survival rate ranges from 91% to 95% when the Breslow thickness is less than one millimeter, and ranges from 63% to 79% when it reaches 2–4 mm, and even worse (less than 48%) above 4 mm. Due to a growing shortage of dermatologists in Western countries
[
4
]
, primary care is increasingly recognized as a suitable setting for skin cancer screening
[
1
,
5
,
6
]
. Dermoscopy has shown benefits in early melanoma detection while minimizing excisions of benign lesions
[
7
]
. Its effectiveness in the early detection of melanoma is supported by level A evidence-based recommendations
[
8
,
9
]
. The Primary Care Dermatology Society in the UK
[
10
]
and the Skin Cancer College Australasia in Australia and New Zealand
[
11
]
are successful examples of general practitioners (GPs) organizing themselves to participate in melanoma detection (with the notable help of dermoscopy) alongside dermatologists.
In France, this activity remained uncommon, as melanoma diagnosis is historically under the responsibility of dermatologists.
In 2021, the Association of French GPs Practicing Dermoscopy was created through the release of an open Facebook account
[
12
]
, modeled after that of the International Dermoscopy Society, “Dermatoscopy”
[
13
]
.
As of 10 November 2024, 2920 GPs have subscribed to the account.
This study aimed to evaluate the number and characteristics of melanomas detected by trained GPs three years after the foundation of the association.
Methods
Study Population and Authorizations
This study received approval from Centre Léon Berard institutional ethics committee (R201-004-511).
This retrospective study was conducted from 26 November 2024 to 26 December 2024.
A survey developed on Google Forms ([online computer software], Alphabet, 2023, Mountain View, USA) was distributed to all subscribers of the Association’s Facebook ([online computer software], Meta, 2024, Menlo Park, California, USA) account on 26 November 2024, with a reminder sent on 5 December.
Participants were asked to complete the survey and upload screenshots of anonymized histopathological reports of melanomas they had personally diagnosed; patients referred by other healthcare professionals for a suspicious lesion were excluded.
Responses were sent to a designated Google Drive Folder ([online computer software], Alphabet, 2023, Mountain View, USA).
The survey gathered the following information about GPs: sex, date of birth, start date of dermoscopy practice, frequency of dermoscopy use, medical practice setting (self-employed, salaried, et cetera), practice location (urban, rural, et cetera).
Participants who uploaded melanoma cases into the database with no survey completed were classified as “unknown” (n=9). GPs who completed the survey with no uploaded melanoma cases were interpreted as GPs who had not diagnosed any melanoma.
The study focused on the following histopathological characteristics: melanoma pathological subtype (superficial spreading melanoma (SSM), lentigo malignant melanoma (LMM), acro-lentiginous melanoma (ALM), nodular (NM), unclassifiable, metastatic) and Breslow’s thickness.
Cases of melanocytic intraepithelial neoplasia (MIN)/ intraepidermal atypical melanocytic proliferation of uncertain significance (IAMPUS) and melanocytic tumors of uncertain malignant potential (MELTUMP) cases were excluded to ensure better comparability with historical series in the literature.
Statistical Analysis
Qualitative variables are described as frequencies and percentages (n, %), and quantitative variables as medians and range (min–max).
The mean and median Breslow thickness were evaluated on invasive melanomas only, since Breslow’s index does not apply to in situ melanomas.
For comparisons with results from other cohorts
[
14
–
17
]
, Fisher’s exact test was used when the theoretical value of each category was less than 5, and a chi2 was used when this value was equal to or greater 5.
Results
As of 26 December 2024, the Facebook account had 2920 subscribers (
Figure 1
).
Figure 1
.
Flowchart of participants.
![]()
Among them, 69% (2015/2920) were female and 87% (2540/2920) were under under age 44 years.
A total of 2773 (95%) resided in France, including 44 from Overseas Departments and Territories, with additional followers from French-speaking countries like Belgium (52), Algeria (44), Switzerland (7), Morocco (7), and Tunisia (5).
Participants’ Characteristics
A total of 85 GPs participated in this study (
Table 1
), of the 628 who viewed the Facebook post (13.5%), representing 2.9% of the total number of followers (85/2920).
Table 1
Characteristics of Participants.
![]()
The median start of dermoscopy practice among participants was June 2023 (
Figure 2
).
Figure 2
.
Number of participants practicing dermoscopy over time among the responders to the survey
![]()
Melanomas Diagnosed
The mean Breslow thickness of melanomas detected in our study was 1.12 mm (median 0.5 mm) (
Table 2
).
Table 2
Histopathological Characteristics of Diagnosed Melanomas.
![]()
In 2000 (
Table 3
), the mean Breslow thickness was 1.9 mm for dermatologists (median 0.9 mm) and 2.9 mm for GPs (median 2.0 mm)
[
15
]
.
Table 3
Comparison of Melanoma Characteristics in this Study with Previously Published Studies.
![]()
In 2013, Grange et al. in France identified a mean Breslow thickness of 1.68 mm (median 0.85 mm) for melanomas detected by trained GPs
[
16
]
.
In 2020, Lorier et al., also in France, analyzed the melanomas detected during consultation by dermatologists in private practice, finding a median Breslow index of 0.4 mm
[
17
]
.
Our study identified:
a lower proportion of in situ melanomas than did Hay et al.
[
14
]
(39.5% vs 65%,
P
<0.001)
a higher proportion of in situ melanomas than did Grange et al.
[
16
]
(39.5% vs 28.2%,
P
<0.001)
A similar proportion of in situ melanomas than did Lorier et al.
[
17
]
(39.5% vs 43.8%,
P
=0.50)
Our study identified:
a similar rate of Breslow thickness >1 mm among the invasive melanomas compared with Hay et al.
[
14
]
(22.4% vs 27.8%,
P
= 0.39)
a lower rate of Breslow thickness >1 mm among the invasive melanomas than did Grange et al.
[
16
]
(22.4% vs 42.6%,
P
<0.001).
Our study identified a different distribution of histologic subtypes compared with Hay et al.
[
14
]
(
P
<0.001) and Richard et al.
[
15
]
(
P
<0.001), and one similar to that of Lorier et al.
[
17
]
(
P
=0.43)
Participants diagnosed an average of 3.2 melanomas each (median 2; min 0; max 36) (
Figure 3
).
Figure 3
.
Melanomas diagnosed per GP.
![]()
Each participant diagnosed a mean of 2.9 melanomas per year of dermoscopy practice (median 1.6; min 0; max 48).
Discussion
Early recognition of life-threatening skin tumors is a public health concern, but is nowadays made difficult by the growing shortage of dermatologists in many European countries. Moreover, recent studies have underlined that Breslow thickness at melanoma detection is higher in areas of medical deprivation
[
18
]
. In France, we count 2592 dermatologists for 52291 general practitioners
[
4
]
. Australia and, to a lesser extent the UK, have chosen to involve general practitioners in skin cancer detection. In Australia, the majority of melanomas are diagnosed by GPs. This practice is very new in France, and still developing, despite some limitations, such as GPs’ lack of confidence in their own diagnostic abilities, despite good results
[
19
]
. The Association of French GPs Practicing Dermoscopy, created in 2021, now includes 2920 members.
In this context, we conducted a survey to analyze the melanomas detected by Association members.
Dermoscopy-trained GPs from the Association of French GPs Practicing Dermoscopy were able to detect early melanomas, with a notable improvement in the mean and median Breslow thickness compared to previously reported values in 2000
[
15
]
and 2013
[
16
]
. The proportion of in situ melanomas increased (28.2% in 2013 vs 39.5% in present study), as did the proportion of invasive melanomas with Breslow index less than 1 mm (57.4% in 2013 vs 77.6% in the present study).
These findings suggest a significant improvement in melanoma detection, likely attributable to dermoscopy implementation in general practice.
Notably, Lorier et al.’s
[
17
]
evaluation of detected melanomas by dermatologists in France is consistent with our results. Using a similar study design, they reported a comparable median Breslow thickness and distribution of melanoma histological subtypes. This suggests that after appropriate training in dermoscopy, GPs can also achieve excellent performance in skin cancer screening and are qualified to perform such screening
[
19
]
. Developing this expertise may be crucial since a large proportion of patients consulted their GP in the year preceding a melanoma diagnosis
[
20
]
.
With a growing number of GPs participating in melanoma detection, including approximately 20% in rural areas (significant, given the shortage of dermatologists in these locations
[
21
]
), this could lead to a rapid increase in the number of melanomas detected by GPs, since the median number of melanomas diagnosed per year was 1.7.
Although this may seem low, melanoma remains relatively rare in general practice: with approximately 17,000 new cases per year and 52,291 GPs in France, the average GP encounters a new melanoma case only once every three years. Participants in our study detected melanomas at a rate three times higher than this national average.
Half of the participants used dermoscopy daily, but this practice remained relatively new for most French GPs, with a median dermoscopy experience of only 1.5 years.
Strengths and Limitations
To date, most publications on skin cancer screening and diagnosis by general practitioners have been limited to:
assessing the use of dermoscopy by general practitioners in various countries
[
22
,
23
,
24
,
25
,
26
,
27
]
evaluating general practitioners’ ability to detect suspicious skin lesions and refer them to dermatologists
[
28
,
29
,
30
]
Investigating the improvement in general practitioners’ skills after clinical and dermoscopy training of varying durations provided by dermatologists
[
31
,
32
,
33
,
34
,
35
,
36
]
determining the required training duration for general practitioners to gain proficiency in skin cancer detection
[
37
]
exploring the benefits of tele-expertise in improving diagnostic accuracy and reducing treatment delays for skin cancers
[
38
]
developing computer-assisted diagnostic systems and algorithms with artificial intelligence to aid in the dermoscopic image classification for general practitioners
[
39
]
Apart from the previously mentioned work from Australia, which aligns with a national health policy prioritizing melanoma due to its prevalence, no study has examined in detail the actual contribution of general practitioners to the complete and independent diagnosis of cutaneous melanomas. This research thus represents a pioneering study in Europe, highlighting the legitimacy and significant role of general practitioners in skin cancer screening.
The low participation rate, considering the number of the Facebook account followers, constitutes a limitation of this study. However, among the 2920 followers, not all are GPs (e.g., pediatricians, dermatologists, occupational doctors, physiotherapists, et cetera) or are replacement GPs or residents that may not have access to the final pathology reports of their cases. Additionally, some individuals (n=8) who were known to have diagnosed melanoma cases (previously published on the Facebook account, n=16) did not participate in this study.
Recruitment bias: Respondents may have a greater interest in skin cancer detection, potentially increasing the representation of dermoscopy practice among GPs. However, this does not question the capabilities or outcomes achieved by GPs practicing dermoscopy.
Most participants were young (mean age 38.0 years) and female (56.7% (42/73)), which is consistent with the demographics of the Facebook group (58% were under age 34, and 69% female), in contrast with the global GP population in France (mean age: 51 years old, female: 46%). Some GPs could have been worried about openly participating in a study related to a medical specialty outside their primary field.
Recall bias: Some survey questions relied on the GP’s ability to recall certain information, which could have led to slight inaccuracies in the reported start date of dermoscopy practice. This is a non-differential error, but it could result in an underestimation of the number of melanomas per participant, as some cases may have been forgotten or the pathology report may not have been forwarded to the general practitioner by the dermatologist who performed the excision.
Allocation bias: The study specified that only melanomas diagnosed directly by a GP should be included. Some GPs, especially those who accept referrals from other GPs, may have included these as their own diagnoses. However, we provided clear guidance to avoid this issue.
Future Directions
The Association of French General Practitioners Practicing Dermoscopy has recently been created, and these promising results need to be confirmed with an extended follow-up of five years.