Kluger and Thomas 2008, The Dragon With Atypical Mole Syndrome

  • “(1) avoid getting the tattoo at too young an age when all the nevi have not appeared yet; (2) choose a small design; (3) choose an area devoid of nevi, if possible; and (4) choose light colors rather than dark ones”

Monfrecola et al 2023, Could wide tattoo delay the early diagnosis of cutaneous melanoma?

  • “Tattoo extension, especially in the back, leads to an increasing risk for diagnosis delay, as reported by our patient: this remarks the difficulty to detect growing lesions due to the size of tattoo as well as the colour.”
  • “Lesions identification could be hampered by the presence of tattoo, which makes more difficult the detection and the surveillance of pigmented lesions.”

Kluger 2023, Tattoos, tattooists, moles and melanomas

  • “only 20% of the cutaneous melanoma on tattoos develop on a preexisting naevus”
  • “there is no ground to suggest that any given tattooed individual undergoes ‘tightened’ surveillance if he is not at higher risk of melanoma or if the tattoo is not covering numerous moles […] Education of the general population, more practical advice regarding self-inspection, can be efficacious for early detection and has been the main goal of the yearly Euromelanoma campaigns.”
  • “Regarding tattooists, they are usually aware that they should not tattoo over a mole as a rule, even though there are always mistakes.”
  • “most of the professional tattooists are inspecting the skin of their customers before tattooing. Half of them have already spotted what they considered as a ‘suspicious’ lesion on one of their customers last year. Almost 70% of them recommended seeking advice from their healthcare provider.”

It’s interesting to me that there are reasons why tattoos might make it harder to spot melanomas (design interfering with skin checks) but also easier (closer inspection of skin by self or artist).

A rough calculation, even playing devil’s advocate and ignoring that positive effect:

  • For the average adult human, the skin has a surface area of 1.5–2.0 square metres (Wiki).
  • That makes 17,500 square centimetres.
  • Consider a typical tattoo, let’s say it’s 5 cm x 5 cm. That makes it 25 square centimetres.
  • This would cover 0.14% of the body.
  • Assume that the tattoo is placed on a portion of the skin with an exactly average probability of getting melanoma. (This is obviously wrong—e.g. I’d guess tattoos are more likely to be on visible areas of the skin, which are also areas exposed to the sun and thus more likely to get melanoma. Bear with me!)
  • A random Australian website tells me that 1 in 22 Australian women are diagnosed with melanoma before age 85. That makes it 4.5%.
  • So, for each of our theoretical tattoos, there would be a 0.14 % * 4.5 % = < 0.01% chance of having a melanoma on the area of the tattoo.
  • So even if a tattoo makes it more difficult to spot a melanoma by some percentage, the actual lifetime risk of facing the situation “has melanoma on tattoo and failed to detect it because of the tattoo” would also be < 0.01%.

Regarding the suggestion of Kluger and Thomas (2008) to wait to get tattoos until all the nevi have appeared:

  • Nevi numbers peak at age 30 and then gradually decrease, in a study of “Caucasian female subjects (2,786)” from the UK (Bataille et al 2007)
  • “Total nevus counts tend to increase until the 4th decade of life before decreasing in the elderly. The prevailing explanation for this age‐related effect has been that people develop new nevi in early life and lose nevi in later life. In fact, this nevus volatility has been observed in paediatric cohorts—over a follow‐up period of 3–7 years, up to 75% of patients had new nevi, and 28% had nevi that disappeared. In later life, there could be a shift, whereby disappearing nevi exceed new nevi. However, in adult cohorts, data on change in nevus counts with age has been inconclusive. An account for the conflicting data could be that the association between nevus counts and age is an artefact of cross‐sectional studies, and differences in ultraviolet radiation (UV) exposure across birth cohorts may explain the disparity in nevus counts by age.” (Reiter et al 2022)
  • “In terms of nevus counts, our results suggest that patients continue to acquire new nevi and most patients experience an overall increase in nevus counts over time.” (ibid) (and eyeballing their results, it seems like the new nevi stop appearing around age 30 or 40)

And lastly, my paper award of the year goes to: Rahmat et al 2023, Skin cancer incidence in transgender individuals receiving gender-affirming hormone treatment: a nationwide cohort study in the Netherlands

  • “The cohort consisted of 2,436 trans women and 1,444 trans men. The median age at the start of GAHT was 31 years (IQR 24–42) in trans women and 24 years (IQR 20–32) in trans men. The median follow-up time for trans women was 8 years (IQR 3–18) with a total follow-up time of 29,152 years and 4 years (IQR 2–12) with a total follow-up time of 12,469 years for trans men.”
  • “Eight trans women were diagnosed with melanoma (SIR 1.80 [95% CI 0.83–3.41] vs. all men; SIR 1.40 [0.65–2.65] vs. all women), and seven developed squamous cell carcinoma (SIR 0.78 [0.34–1.55] vs. all men; SIR 1.15 [0.50–2.27] vs. all women). Two trans men developed melanoma (SIR 1.05 [0.18–3.47] vs. all men; SIR 0.77 [0.14-2.70] vs. all women).”
  • “GAHT did not appear to affect skin cancer incidence in this large cohort of transgender individuals.”
  • “Melanoma incidence among trans women has previously been assessed in a study from the United States on overall cancer incidence in 4,889 transgender individuals from the Kaiser Permanente database linked with the Surveillance Epidemiology and End Results database and was found not to be different compared with the cisgender population. However, information on hormone use was not provided in this study. More recently, Singer et al. examined the relation between gender identity and self-reported lifetime prevalence of skin cancer in a population, also in the United States, that included cisgender, transgender, and gender nonconforming individuals, also without taking GAHT into account. In accordance with our analysis, no significantly different odds ratios of skin cancer history among trans men or trans women, compared with cisgender men, were reported.”

Update 20241116
I’ve been thinking more about the increased risk of skin cancer due to a family history of skin cancer. Specifically, two of my second-degree relatives (an auntie and an uncle) died at a tragically young age due to melanoma.

Wu et al 2019, Melanoma risk assessment based on relatives’ age at diagnosis

  • “FDR” = first-degree relative; “SDR” = second-degree relative; “HR” = hazard ratio
  • “Overall, FDR of cases had a 92% increased risk of melanoma compared to FDR of their respective controls (HR 1.92, CI 1.79–2.07).”
  • “Familial risk of melanoma among SDR was typically lower (most HRs between 1.4 and 2.1) than that for FDR (Online Resource 3). Among SDR, higher risk for melanoma was observed for SDR ages 29 or younger and SDR between ages 30 and 39 related to cases diagnosed between ages 30 and 39 (HR 2.46, CI 1.18–5.14; HR 3.18, CI 1.77–5.74, respectively).”
  • eyeballing Fig. 1, the effect of FDR and SDR look very similar! though this may be in part due to the way that the graph is designed, as the HR scale goes up to 8 and most of the actual HRs are below 3ish so probably the two groups look more similar than they actually are.
  • in any case, looking at supplemental table 3, for cases aged 40-49, the HR of SDRs ranges from 1.13 to 1.68, and only one is statistically significant. compare this to the HR of FDRs in supplemental table 2 which are much higher, between 1.55 and 4.89 and statistically highly significant with only one non-significant value.

Following my existing maths above:

  • A random Australian website tells me that 1 in 22 Australian women are diagnosed with melanoma before age 85. That makes it 4.5%.
  • If I’m interpreting Wu et al correctly, and allowing an oversimplification in not stratifying the calculation by my own age, a HR of ~1.3 due to SDR would mean that my baseline probability of lifetime melanoma is 5.85%.

I can’t find much quantitative information in the literature about the risk from having multiple relatives with melanoma.

Clinical Applications of Melanoma Genetics

  • “As the number of cases of melanoma in one family increases, so does the likelihood of a hereditary melanoma predisposition, such as CDKN2A. However, the baseline threshold for number of cases in the family differs based on geographic region. This is due to multiple factors that also affect sporadic melanoma incidence, including ethnicity and sun exposure [2, 29, 30]. In regions or ethnicities with a relatively high incidence of melanoma the likelihood of detecting a CDKN2A mutation is low unless multiple melanomas, or other CDKN2A-related features, are present within a family [1, 31]. Within a North American cohort of families with hereditary melanoma, approximately 30 % of families with three cases of melanoma had a CDKN2A mutation. A higher CDKN2A mutation frequency of approximately 80 % was reported for families with six or more cases of melanoma [1].”

So it seems like we’d be justified in believing that having 2 SDRs with melanoma increases your risk more than 1 SDR does. But I don’t have enough information about the specific level of this risk to do anything more than make a guess.

Also I’m not satisfied with my above rough analysis that uses the area of skin to calculate risk on a particular body part. But I found this study:

Conte et al 2021, Population-Based Study Detailing Cutaneous Melanoma Incidence and Mortality Trends in Canada

  • see Figure 1B, which shows the risk by body part. For female arms, this is 29.2% (as a percentage of all melanomas on any body part). This study is about Canada which is a huge limitation (Canada = snow and moose; Australia = the surface of the sun).

All this together (you’d need to change these numbers to do a similar analysis for yourself):

  • Baseline lifetime melanoma rate: 4.5%
  • Hazard risk due to one second-degree relative = ~1.3
  • Hazard risk due to an additional second-degree relative beyond the first = ??? let’s say ~1.2
  • Proportion of melanomas on arms = 29.2% (this includes two arms)

Thus, my very rough chance of getting a melanoma on my left arm is:

4.5% * 1.3 * 1.2 * 0.292 * 0.5  
= ~ 1.0%

So the chance of getting a melanoma on a certain part of the arm would be proportionally less. e.g. if a new tattoo covers, say, 30% of the area of my arm, we’d be looking at 0.3% of getting a melanoma in the area of the tattoo over my life. This is much larger than my more rough analysis (the original “< 0.01%” that I arrived at, above). But in absolute terms, it is still a very small risk.

So even if a tattoo makes it more difficult to spot a melanoma by some percentage, the actual lifetime risk of facing the situation “has melanoma on tattoo and failed to detect it because of the tattoo” would also be ~0.3%. (It’s not “dying of melanoma” or even “failing to detect that melanoma”, it’s “getting the melanoma in that spot at all”.)

To give this number some context: 0.3% is about 1 in 333 people. This is less likely than a random natural birth resulting in twins (1 in 250) (source: this random magazine article). It’s about as likely as your car being stolen in a one-year period (if you live in the United States and own a car). It’s less likely than being named Emma (if you are a girl). It’s actually less likely than dying in a car crash. It’s about as likely an encounter on Pokémon: Let’s Go! resulting in a shiny Pokémon—after the player has already maxed out their shiny rate. And—this one’s my favourite—it’s about half as likely as a women’s soccer player in college going on to play in the NWSL! (source)

I’m not sure if I’d feel justified in changing life plans based on any of the above probabilities. e.g. beyond the usual and obvious precautions, I don’t take active steps to drive less because of the risk of dying in a car crash. I wouldn’t bother preparing two cots in the nursery because of a slight probability of having twins. And I wouldn’t make any firm life plans based on the assumption that I would make it into the NWSL!