Pathology Associates Of Lexington, P.A.
Pathology Associates Of Lexington, P.A.
Pathology Associates Of Lexington, P.A.
 Home | Pathology Group MembersOur Hospital  Search This Website:
        Melanoma, prognostic parameters & measures
Melanoma Pathology Parameters

For decades (at least since 1980), research scientists have hoped to come up with technology that will avoid tissue biopsies when "cancer" is suspected. Most technologies have ended up bringing MORE biopsies to pathologists (even though it avoids biopsies on a subset of lesions that previously would have been biopsied). MoleMap has the hope of dermatologist-directed comparative photography to help prove lack of significant interval change so as to avoid surgery. Dermatoscopy allows special magnified visualization by dermatologists of skin moles so as to discern benign 3 dimensional configurations and avoid some biopsies. In 2009, laser scanning screening is getting grant money as it stimulates the imagination of researchers & engineers hungry for research grants. In late 2011, MelApp came out for the iPhone for a nonmedical phone owner to (1) take a phone photo of his/her suspicious mole and (2) have the app calculate "riskiness" that it be melanoma (and even use the phone GPS to locate a nearby dermatologist). But, I found it not easy to take an optimal photo.

Primary tumor sites: We think of melanoma as a skin cancer; but, they can rarely originate ANYWHERE, including from node capsule nevi in various lymph node clusters (such as in the groin or axilla). ALSO, rarely, biphasic maligancies arising in non-skin locations can have a melanomatous-appearing component with, beginning in 2011, treatment implications [was melanomatous L12-8775; was not melanomatous L12-9826].

Pathology lab specimen processing: If your local pathologist has very carefully processed your skin specimen and carefully determined the parameters, you will be able to find out additional information on the internet or from medical contacts around the world. Careful processing includes a system of careful "grossing" of all skin specimens so that there is assured embedding of skin such that epithelial surface is at 90 degrees to the plane of microtome avoid incorrect Breslow thickness measurement due to tangential increasing of thickness. A FAXed, excellently informative pathology report contains essential information on which experts can give additional opinions and advice (there are usually only a few actual glass slides which could be shipped out long distances for other expert opinions). Especially important is the application of exact criteria to separate RGM & VGM in cases of thin level II, see below14. All of these parameters are considered in estimating a melanoma patient's prognosis if left untreated. If it is "bad", then various treatment measures are considered.There is developing genetic information6.

Pre-surgical predicting of likelihood of positive SLN: One parameter for assessing prognostic risk in a melanoma patient is "node status". And there are parameters which help to select those at risk for node positivity...tested by sentinel lymph node or other node biopsy or dissection if liklihood of positivity warrants it. A general online calculator based on common parameters is at MSKCC. But, if the calculator produces a low risk number, a search for presence or absence of other higher-risk histology features may result in a decision to check the nodes anyhow. Especially as to melanoma "growth phase", "vertical growth phase (some think radial growth phase...RGM... rarely if ever metastasizes17) melanoma" (VGM) has the ability to metastasize (13 % of VGMs have pos. SLN). So, RGM are extremely unlikely to metastasize. This 2006 study of 682 VGM SLN cases found 88 positive. Here are their SLN triage & prediction criteria: (1) if Breslow is 2mm or less & zero mitoses per square mm, 2.1% have pos. SLN & if any mitoses per square mm, 9.8% pos SLN (those negative for TILs have the highest rate in this group); and, if (2) Breslow is thicker than 2mm, then absence of significant TILs actually infiltrating through melanoma cell clusters has a 40% SLN positivity rate while presence of TILs only has a 20% positivity rate17. So any VGM [criteria] with no dermal-melanoma-cell mitoses and less than 2mm thick has only a 2% likelihood of a positive sentinel node17, 19! But, there is an on-line claulator20 for 10 year overall survival prognosis of thin (a mm or less in thickness) lesions.

Sentinel-node (SLN) status:

  • Positive vs. negative SLNs: in one study21 (a follow up period of up to just over 5-year follow up), 31% of SLN positive patients recurred & 19% of SLN positives died of melanoma, while SLN negatives were 6% & 4%, respectively.
  • SLN node-capsule nevus:
    • False positive SLN: On finding a positive SLN, the pathologist must be careful not to interpret small node capsule melanocytic nevi (amitotic and low Ki67) as metastatic melanoma. These nevi (many being minute) may be as common as 39% of SLN encounters when using a detection system of immunohistochemistry marking and automated digital image analysis 21. They are more common in SLNs than non-SLN and in patient's with congenital melanocytic nevi or melanoma associated nevi...their coarser but practical study (also with IHC) finding a rate of 11% in melanoma cases22.
    • Node-capsule primary melanoma: Likewise, it is possible that a melanoma initially diagnosed which originated as a pathological node as a capsular-nevus-derived primary [L08-3623] (if the previous history is totally negative and the patient is also currently negative for a primary melanoma of skin or mucosa).
  • Metastatic volume21: less than a half cubic millimeter vs. greater than half.

LONGTERM PROGNOSIS: In spite of particular case prognosis predictions, remember that melanoma is a malignancy infamous for NOT "playing by the rules". There are some amazing stories of long term survivors! I our one county and hospital during a 40 year career, the following came to my attention. (1) I personally knew a local professional man who was misdiagnosed about 1973 and only correctly diagnosed a year later when he had bulky axillary metastasis...he lived well until he developed a new primary melanoma & died in 2012 (39 years later). (2) Also in the early 1970s, a man in that same small town & friend of (1) had a bad melanoma diagnosis & lived maybe 25 years (both had gotten vaccine therapy at M. D. Anderson). (3) Another local man had a dark mole removed in 1974 & appeared for emergency laparotomy showing bulky stage IV melanoma in 1983. He then underwent a Christian "laying on of hands" some weeks later (the story). Some time further, he went on a macrobiotic diet; he died in early 1997. (4) In 2005, we had an 80 y/o F who presented with bone lesions suggesting myeloma who was later found to have had melanoma in 1960. (5) In 2015, a 70 y/o male had a benign FNA of a possible early pancreatic mass; hed'd had a flank melanoma removed in 1980 followed 5 years later with mets [S05-15202; least 35 years survival]. In 2016, we had a patient present with a cough & found to have a lung mass: melanoma (she'd had a melanoma of the face removed [where?] in 1978 & treated at Duke with some sort of immunotherapy [T16-63]).

Our group's pathology reports on the primary skin melanoma include the following numerous parameters (basic diagnostic criteria):

  1. tumor location: extremities B vs. trunk Intermediate vs. other.
    ( acral [volar & subungual], scalp, ears, and mucous membrane or immediately adjacent skin9)W.

  2. sex of patient (survival, men W vs. women B).

  3. Clark's melanoma histogenetic types: lentigo maligna B, SSMM, acral lentigenous, mucosal lentigenous, verrucous, desmoplastic, nodular W [microscopic diagnostic criteria & other pigmented lesions].

  4. Breslow thickness.

  5. Clark's level.

  6. mitotic rate.

  7. host inflammation.

  8. angiolymphatic invasion: detection of vascular invasion implies the need for chemo/immuno-therapy.

  9. tumor ulceration (not meaning traumatic excoriation [L05-517] or artifactual "ulceration"6)...defined as absence of epidermis over a major portion (probably 3MM or more [L-05-517]) of the melanoma & measurements should be made16. See below.

  10. surgical margin status.

  11. perineural space invasion: one may need to perform an IHC neural marker such as S100 in order to see small cutaneous twigs obscurred by perineural tumor. Positivity implies strong consideration of local radiation because of about 25% increased risk of local recurrence in such cases.

  12. lymp node status:we are now working up about 1-2 sentinel lymph node cases per week (our department is regularly involved in breast cancer sentinel node cases and a large experience of detailed and focused skill at lymph node dissection in search of metastatic malignancies). It is vital to not mistake a nodal nevus for metastatic melanoma (about IHC)...even taking care to discern a nodal nevus (criteria, above, this page) in a completion dissection in a case with positive sentinel node [L07-5254]!

A file on melanoma types and skin lesions which might clinically be melanoma

An overview page with link to melanoma on-line prognostic calculator

Below: B=better, W=worse.

Naked-eye (dermatologist) parameters:

  • is it located on an extremity B (arm or leg) or non-extremity W?

  • is it in a sun-exposed B or covered location W?

  • is it small B or large W?

  • is it nodular W or not completely nodular B?

  • is it pigmented B (melanotic) or not pigmented W (amelanotic)?

  • is it "tumor ulcerated" W (a scab or crust may cover the ulcer) or not ulcerated B? [don't misdiagnose traumatic or excoriative defects...they have hemorrhage and brightly eosinophilic fibrin exudation on the site surface...which are not really tumor-necrosis ulcerations6] True ulceration tends to be thick melanomas9.

  • is the patient young W or old B?

  • could it be a cutaneous metastasis from a previous melanoma (be sure to warn the pathologist if you previously had a melanoma)?

Predicting Metastatic Potential:

The various Clark levels, followed later by designations of radial growth or vertical growth melanoma, are a staging attempt to separate those which are highly unlikely (RGM & Clark levels I & II) to metastasize from the rest of melanomas which are more likely to metastasize. The Breslow thickness is thought to be a more reliable parameter, less subject to inter-observer variation. Warning13: Dr. Ackerman has collected a small group of cases of very thin melanomas which metastasized; we have seen one which was daignosed as MIS turn up later with brain mets [LMC-01-8112]...never guarantee a melanoma patient of a complete cure! Once in a while, a pathologist gets pulled into a case to the point of being moved to express advice & encouragement that the patient undergo a maximum attempt to cure, based on all pathological factors [L05-517].

Microscopic (pathologist) parameters:

  • Ulceration: true tumor ulceration seems related to "explosive" upward intra-epidermal melanoma cell scattering (whereas, the epidermis tends to bulge over non-ulcerated melanomas15). In thin melanomas (1MM or less thick), the level of invasion is more powerful prognostically than in thicker stage I-II where presence of ulceration indicates a 50% 10 year survival rate & without ulceration, a 78% survival rate.

  • what is the exact thickness (the Breslow thickness3 is measured in millimeters from the granular epidermal layer to the deepest point of tumor penetration)...thinner B or thicker W? A carefully interpreted marker stain such as MART-1 can help discern vague deeper cells as well as "early" progression of cells into the next deeper Clark level [S-01-9973; S-01-9340]. There is greater inter-observer agreement on thickness than on Clark level12.

Thickness (clinical stage I) 8-year survival rate10
<0.76 mm. 93.2%
0.76-1.69 mm. 85.6%
1.70-3.60 mm. 59.8%
>3.60 mm. 33.3%
  • less than 1.0 mm, lymph nodes usually negative1  

    • AJC TNM    pT1=  0.75 mm. or thinner (likely Clark I-II)

    • AJC TNM    pT2=  0.76-1.5 mm (likely Clark III or worse)

    • AJC TNM    pT3=  1.5-4.0 mm
      1.            pT3a= 1.5-3.0 mm
      2.            pT3b=  >3 but not more than 4 mm
    1. greater than 4.0 mm, high risk that nodes are positive1  
      • AJC TNM    pT4=   > 4 mm. and/or invades subcutaneous fat

  • what is the Clark level (I-V)...I B....V W?
Clark level3 Notes Growth phase
I in-situ, noninvasive intraepidermal only (lentigo maligna); MIS vs. early II can be problematic [S-03-16189] radial
II homogeneously widely spaced, small (<10 cells wide) nests in papillary dermis deeper & no dominant nodule that looks "different"; MART1 may help discern microinvasion [S-01-9340] radial (but there are some thin, Clark II verticals, see below14)
III (a nodular is no less than III);
usually heterogeneous "look" because of expanding tumor plaque/nodule filling and expanding papillary dermis; sometimes II-III problematic [S-01-13088; LMC-01-4912; S-04-9050; L-05-1936]; sometimes III-IV problematic [S-02-433]
IV active invasion of the reticular dermis15 (an IHC melanoma marker may help to discern [S-01-9973]) vertical
V invasion of subcutaneous fat vertical

is it radial growth phase (RGP/RGM) B or vertical growth phase (VGP/VGM) W?

Radial (tumorigenic) growth phase mel. (RGP/RGM) Vertical (tumor) growth phase mel. (VGP/VGM)
cells single or small groups
<10 (<159) cells wide
a dermal nest is larger than epidermal or DEJ nests9, 17
dermal clusters smaller than epidermal clusters14 one or more dermal clusters larger than the largest epidermal cluster14***, 17 or a dermal mitosis...either one17
epidermal component tends pagetoid or lentigenous at least one dermal nest looks different than parent dermal nests
dermal component profile horizontal at least one dermal nest >10 cells thick
(>25 cells thick nearly diagnostic VGM)
dermal and epidermal cytology similar9 dermal & epidermal cytology variation; especially if dermal atypia worse than epidermal9
dermal mitoses usually absent usually dermal mitoses (any dermal melanocytic mitosis makes it
VGM14, 16***, 17 )
even when heavy host response, it is usually not at tumor cells if host response, cells amongst and beneath tumor cells
Clark level I-II all VGP/VGMs are Clark level III-V

Presence of this type of cluster and/or dermal mitosis...especially presence of any dermal mitosis makes it VGP/VGM!!!

  • what is the mitotic rate10, 11 (number of mitotic figures per square millimeter of tumor tissue) it low B or high W? In Clark's study of 23 variables, 6 were independently prognostic; and the mitotic rate was the most powerful!

    • 8-year survival if zero mitoses/mm2: 95.1%

    • 8-year survival if 0.1-6.0 mitoses/mm2: 79.4%

    • 8-year survival if >6.0 mitoses/mm2: 38.2 %

  • What is the Schmoeckle & Braun-Falco prognostic index10 (PI) (product of Breslow thickness x mitotic rate)...low B or high W?
Breslow thickness PI survival rate
 1.5-3.99 mm. >19  [L-05-517] worse
1.5-3.99 mm. <19 better
  • Are there any dermal or subcutaneous metastatic cell clusters  within 5 cm. W (satellitosis)...or not B?

  • what is the host inflammatory cell response in VGM..."brisk" B vs. "non-brisk" W?

Host Vertical Growth (VGM) Tumor Infiltrating Lymphocyte (TILs should surround & disrupt VGM cells15) Response5


(1) if the lymphs are diffusely present throughout the substance of the VGM component, or
(2) peripherally layered across the entire underneath of the VGM component

NON-BRISK (1)one or more foci of diffuse intra-VGM, or
(2)focally peripheral underlayering
ABSENT no lymphocytic reaction at all, or none infiltrates VGM15

Plasma cells: does the host inflammation contain plasma (plasmas) cells or not (few) B? [yes (especially if clusters to sheetsW, 23 ...34 of 710 melanomas23 ) is more likely to have gone to lymph nodes] 1

3-parameter likelihood for positive lymph nodes1

upper extremity location + < 1mm + no plasmas 0 %
>1 mm. + upper extremity with plasmas or located elsewhere & no plasmas 20.4%
>1 mm. + elsewhere + plasmas 45.5%
  • is there any visible venous or lymphatic embolic lumenal tumor W or not B? ["yes" is awful2]

  • is there any visible perineural (neurotropism) space invasion W or not B? If so, may be (strongly recommend consideration [L07-132], NOTE HERE) a candidate for local control radiation.

  • are the surgical margins free of visible tumor B (did the surgery seem to get all of the tumor out?) or positive W for tumor?

  • NODES: do sentinel or other regional nodes contain metastatic cells or not B?

    1. S100 is the most sensitive  IHC marker7 (but it also marks nodal dendritic cells) while HMB-45 & MART-1 are more specific; HMB-45 can have some IHC positive mononuclear cells in which the nuclei are discordantly non-melanomatous [LMC-01-6340]. Watch out for nodal nevi, present even in cases with melanoma mets, in fibrous septae & capsule & with benign cytology & lacking nucleoli [L07-5254].
    2. if not, be sure to note whether the melanoma shows signs of regression6: papillary dermal deposition of loose, edematous collagen containing increased numbers of  melanophages & possibly ectatic vessels plus MAYBE some lichenoid vacuolar change in overlying basal keratinocytes8.
    3. it is the number of positive nodes (met = 0.2 mm or larger tumor cell cluster) and the percentage of the node occupied by the met (not the size of the met) that is better or worse15. Therefore, watch out for nodal nevi, present even in cases with melanoma mets, in fibrous septae & capsule & with benign cytology & lacking nucleoli [L07-5254]. Node capsule nevi have a very low Ki67 proliferation of around 1% & are HMB45 negative, while both melanoma, nevus cells, & nodal dendritic cells are S100 positive, and nodal nevi & melanoma are MART-1 positive18.
  • serum LDH repeatedly elevated (not just one time) is a stage independent prognostic factor for decreased survival15.
  • serum albumin at or less than 3.5-4.0 is an independent prognostic factor for decreased survival in stage IV cases15.

REFERENCES: [melanoma diagnostic criteria] [skin diseases index]

  1. The American J. of Dermatopathology, A. B. Ackerman, editor, 6(supplement):1-352, 1984 [EBS's office]
  2. A Trial in Philadelphia, a combined reprint,  A. B. Ackerman (compiler), Dermatopathology: Practical & Conceptual, 6(3):236-250, 2000; 7(1):21-27, 2001; 7(2):165-170, 2001. [EBS's office]
  3. Principles & Practice of Surg. Path. and Cytology, [3 vol. text] Silverberg, SC, 3rd Ed. 1997.
  4. Prognostic factors for Cutaneous Melanoma [editorial], Wick, Mark R., AJCP, 110:713-717, 1998.
  5. Prognostic Value of Tumor Infiltrating Lymphocytes in the Vertical Growth Phase of Primary Cutaneous Melanoma, Clemente, Mihm, et. al., Cancer 77:1303-10, 1996.
  6. Journal of Clinical Oncology, Mihm MF (et. al.) 19(16):3635-3646, 15 Aug. 2001 [EBS's office], Final Version of the American Joint Committee on Cancer Staging System for Cutaneous Melanoma.
  7. Archives of P. & L. M., 125:1295-1306, Mihm MF (et. al.), October 2001, Malignant Melanoma, an Update.
  8. Murphy GF, Dermatopathology, 1995 [textbook].
  9. Maize JC, et. al., Cutaneous Pathology, 1998 [textbook, WRA's office]
  10. Fitzpatrick TB (et. al.), Dermatology in General Medicine [text], vol. I, 4th Ed. (1993) Chapter 12 (Dr. Mihm, et. al.)
  11. Clark WH, J. Natl. Ca. Inst. 81:1893-1904, 1989.
  12. Fletcher A, et. al., J. of Pathology, 163:245-250, 1991.
  13. Ackerman AB, 9th Annual (Pittsburgh) Seminar in Pathology, 4 May 2002.
  14. Relevance of Vertical Growth Pattern in Thin Level II Cutaneous Superficial Spreading Melanomas, Lefevre M et. al., A. J. Surg. Path. 27(6):717-724, 2003 (an excellent case-controlled study in search of features predictive of relatively poor prognosis in thin melanomas). Abstract also in CAP Today October 2003, page 80.
  15. Compton CC, et. al., Protocol for the Examination of Specimens from Patients With Malignant Melanoma of the Skin, Arch. Path. & Lab. Med. 127(10):1253-1262, Oct. 2003.
  16. McKee, Calonje, & Granter, Pathology of The Skin...
  17. Kruper LL, Spitz FR, Czerniecki BJ, Fraker DL, Blackwood-Chirchir A, Ming ME, Elder DE, Elenitsas R, Guerry D, Gimotty PA, "Predicting sentinel node status in AJCC stage I/II primary cutaneous melanoma" Cancer 107(10):2436-45, 15 November 2006.
  18. Biddle DA, et. al., AJSP 27(5):676, 2003.
  19. Gimotty PA & Gerry D, "Prognostication in Thin Cutaneous Melanoma", Arch Path & Lab Med 134(12):1758-1763, Dec. 2010 (a special issue on Melanoma Update).
  20. Melanoma Molecular Map Project website HERE.
  21. Riber-Hansen R, et. al., Automated digital volume measurement of melanoma metastases in sentinel nodes predicts disease recurrence and survival, Histopathology, 59(3): 433–440, September 2011.
  22. Holt JB, et. al., "Nodal Melanocytic Nevi in Sentinel Lymph Nodes
    Correlation With Melanoma-Associated Cutaneous Nevi", AJCP 121:58-63, 2004.
  23. Medscape, about plasma cells & melanoma, June 2016, HERE, via Modern Pathology 29(4):347-358, 2016.

(posted 9 February 2002; latest addition 17 June 2016)

© Copyright 1999 - 2006, all rights reserved, Pathology Associates Of Lexington, P.A.