Tuesday, October 28, 2008

Lines.

Case 1, Woman 60's; 6mm diameter, mid back.



Case 2, Woman 70's, 12mm diameter, calf
More practice cases for Harald's algorithm. We look at lines today.

10 comments:

Dr Alan Cameron said...

No takers so far...

The first lesion shows one pattern lines reticular with eccentric hyperpigmentation. This is a VERY weak clue, and indeed this lesion is a Clark naevus.

The second lesion shows 2 colours (light and dark brown, more than one pattern, lines reticular and structureless (there are not enough circles (SW and E) to constitute a pattern, grey dots N are artefact).

Clues to melanoma are eccentric structureless (an instance where a feature may be both a pattern and a clue), lines reticular thick.
With the only colours being brown, we expect in situ melanoma, which is indeed what is seen on histology.

pbu7364 said...

Great cases and thanks for the Kittler site. What constitutes a weak clue to MM and what is the batting order or priority for what outranks what if there are conflicting clues....Paul B

pbu7364 said...

Another question...case 2..the stuctureless area is eccentric but does not go to the border..I assume it just has to be off centre and not involve the border??...Paul again.

Dr Cliff Rosendahl said...

Good questions Paul. There is no priority for clues as I understand it. Pattern is regarded as outweighing clues. For example if there is one pattern and one colour then that will probably outweigh a clue. For example if there are pseudopods but only one pattern it will probably be a Reed naevus.
Symmetry(one aspect of pattern) also tends to outweigh any clues to melanoma. In such a situation where there is conflict of "clues" each case is decided on its merits
If there is more than one pattern or more than one colour then only one clue to melanoma needs to be present to make biopsy necessary

Dr Cliff Rosendahl said...

"Structureless eccentric" means that there is a significant structureless area present and the lesion is not symmetrical. It does not have to extend to the edge ( different to the new Menzies criterion for light coloured melanoma of "peripheral structureless brown" which does extend to the edge.

Dr Alan Cameron said...

I wrote you a long answer this morning Paul but blogger has swallowed it. These are important questinos so I will have another go.

In general, pattern trumps clues. For example, if we imagine the second lesion without the structureless areas; it would be fairly symmetric but with thick reticular lines. Most often the pattern (lines reticular symmetric; naevus) would outweigh the clue (thick lines reticular, melanoma). From here it is a judgement call which diagnosis you should favour.
Eccentric hyperpigmentation is a very weak clue to melanoma; so much so that it really only indicates you should look closer for other clues in lesions pigmented this way. Eccentric structureless is much stronger, as is asymmetric combination of multile patterns.

Re clues, if you look in Scott's book he gives sensitivity and specificity for his clues, most of which parallel Harald's clues. So for example, grey dots are a relatively weak clue as they can be seen in pIEC, LPLK, and regressed naevi as well as melanoma. Likewise thick reticular lines is a weakins clue. PEripheral black dots/clods is a very strong clue.
When you have conflicting clues, Haralds is the only algorithm which makes explicit the needto use your judgement to weight the clues. I think acknowledging this judgement call is improtant in the respect that it makes you consider each of the differentials of any given lesion. This is more likely, in my opinion, to lead you to correct action.

Jean-Yves Gourhant said...

A small try for case1. Thank you Alan. Good practicing.
One pattern, lines reticular, 1 clr, thin and brown (I suppose) > Clark or solar lentigo.
My guess is Clark (dotted and comma vessels).

Jean-Yves Gourhant said...

And case 2 on sunday!

Jean-Yves Gourhant said...

Case 2.
One pattern, lines reticular > 1 color > then I would say black eccentric.
Dx between Clark and MM.
Clues for MM? Structureless eccentric, but it seems to be skin coloured. Some grey dots.
To sum up, not enough for MM, dx Clark.

Jean-Yves Gourhant said...

Ok, read the result.
Underestimated the eccentric structureless area, I judged it as "skin coloured".
Thick lines are not that easy to see.
I definitively see dots, which are, for some, in the lines.
Well not the dx, but a lesion I would have excised (one finds the consolation he can!)