Monday, October 27, 2008




These cases are posted for those who wish to practice the "Diagnosis Accurate" Kittlerian algorithm. That is the long algorithm on the poster that is available from me at
Please do your analysis stepwise following either the "one pattern" or "more than one pattern" arm of the algorithm as appropriate. I will post the answers in a few days. If you want them earlier send me an email.
Alan and I will respond to any questions you ask to the best of our ability. Harald may also respond when possible.


Harald Kittler said...

Both lesions have the same pattern, i.e clods, and nearly the same color, i.e melanin predominates (brown and gray). The differentials for the first case are BCC, melanoma, and SEBK (because there is asymmetry of color, and cong. nevus, combined congenital nevus, or Spitz for the second (because there is symmetry of color)

the differential is solved by clues:
1. serpentine vessels make this lesion most probably a BCC
2. No more clues. You cannot solve the differential. No speculation!

There is a contradiction of criteria in case 2, i.e. benign pattern but also a clue to melanoma (gray clods). Usually (not always) pattern wins! It depends on your individual threshold and on the circumstances whether you excise lesions with a benign pattern and a single clue to melanoma.

Jean-Yves Gourhant said...

Harald, how can you say there is a symmetry in colour for case 2?
It's difficult to imagine. What does it mean?

Harald Kittler said...

The clods all have the same color in case 2 but not in case 1. The clods are equal in size and shape in case 2 but not in case 1. I acknowledge that it is difficult to make this judgement with certainty. But let us assume you cannot decide. Then you have to add all differential diagnoses of all possible branches:
congenital nevus, combined nevus, BCC, melanoma, sebk. And now you do the same thing again, you solve the differentials by clues! Case 1 is obvious because the best clue=serpentine vessels=BCC. There is no clue in case 2 except that the clods are equal in size and shape, which favors a nevus!

Dr Cliff Rosendahl said...

This is my assessment of case 1
More than one pattern (Clods,dots,?lines)
Melanin dominant
> 1 colour (brown,blue)
Leads to differential BCC,Melanoma,seb K

Clues - serpentine vessels (BCC)
Lines radial segmental (north and SE) These lines are connected to a common base and this is another clue to BCC

It is a BCC

Dr Cliff Rosendahl said...

Harald, I would say there is "asymmetry" in case 1 (rather than asymmetry of colour). Is that what you meant? The colour is pretty symmetrical throughout but the pattern in the lower half of this lesion is not the same as in the top half.

Jean-Yves Gourhant said...

Thank you Harald.
I will have a look at Cliff's poster this night to make a better interpretation of your post.
Another question, not related to these 2 cases. I have difficulties in differentiating big clods and structureless areas (especially when they are black or grey). I will give you some examples later, time to find them back. One is from Derm 101, the other frome the blog.

Dr Cliff Rosendahl said...

This is my analysis of case 2 following the algorithm
1 pattern (clods)
1 colour (brown)
Differential is naevus only. Nothing else! Could be Congenital (Including Unna or Meischer) or Spitz

This was reported:-
Sections show a benign intradermal naevus with heavily pigmented
nests in the upper dermis indicative of past reactivation. Excision
appears complete.
Why did I cut it out? Because I didn't trust the system!
We live and learn...

Dr Cliff Rosendahl said...

Harald mentioned grey clods in case 2. On my monitor I can't see that(except where hairs overlie the clods) but if you saw grey clods with the dermatoscope that would certainly be a clue to melanoma! Even so as Harald suggests pattern should win (over clues) and it should turn out to be a naevus.