If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Melanoma Institute Australia, The University of Sydney, Sydney, NSW, AustraliaFaculty of Medicine and Health, The University of Sydney, Sydney, NSW, AustraliaDepartment of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Sydney, NSW, AustraliaNSW Health Pathology, Sydney, NSW, Australia
Melanoma Institute Australia, The University of Sydney, Sydney, NSW, AustraliaFaculty of Medicine and Health, The University of Sydney, Sydney, NSW, AustraliaCharles Perkins Centre, The University of Sydney, Sydney, NSW, Australia
Melanoma Institute Australia, The University of Sydney, Sydney, NSW, AustraliaFaculty of Medicine and Health, The University of Sydney, Sydney, NSW, AustraliaDepartment of Melanoma and Surgical Oncology, Royal Prince Alfred Hospital, Sydney, NSW, Australia
Melanoma Institute Australia, The University of Sydney, Sydney, NSW, AustraliaFaculty of Medicine and Health, The University of Sydney, Sydney, NSW, AustraliaDepartment of Melanoma and Surgical Oncology, Royal Prince Alfred Hospital, Sydney, NSW, Australia
Melanoma Institute Australia, The University of Sydney, Sydney, NSW, AustraliaFaculty of Medicine and Health, The University of Sydney, Sydney, NSW, AustraliaDepartment of Melanoma and Surgical Oncology, Royal Prince Alfred Hospital, Sydney, NSW, Australia
Melanoma Institute Australia, The University of Sydney, Sydney, NSW, AustraliaFaculty of Medicine and Health, The University of Sydney, Sydney, NSW, AustraliaDepartment of Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital, Sydney, NSW, AustraliaNSW Health Pathology, Sydney, NSW, AustraliaCharles Perkins Centre, The University of Sydney, Sydney, NSW, Australia
Desmoplastic melanoma (DM) is an uncommon subtype of melanoma with distinct clinicopathological features. It is classified into pure desmoplastic melanoma (PDM) when the proportion of desmoplastic melanoma is ≥90% of the dermally-invasive component, and mixed desmoplastic melanoma (MDM) when the proportion of desmoplastic melanoma is <90%. Studies have reported a lower sentinel lymph node biopsy (SLNB)-positivity rate in PDM compared to MDM and non-DM. As a result, some have recommended not performing SLNB in PDM patients. When PDM is identified in a partial biopsy of a melanoma, there is a risk that sampling bias may under-recognise MDM, but to the best of our knowledge this has not been previously assessed or quantified.
The aim of this study was to assess the concordance of the proportion of desmoplastic melanoma in an initial partial biopsy of PDM with the proportion in the entire tumour following complete excision, in patients with cutaneous melanoma. A secondary aim was to determine how frequently this potentially resulted in a patient not receiving a SLNB.
Seventy-eight cases of cutaneous melanoma were identified from the Melanoma Institute Australia (MIA) database and 23 cases from the Memorial Sloan Kettering Cancer Centre (MSKCC), where an initial biopsy contained PDM and a subsequent wide excision had residual invasive melanoma. Clinicopathological features were analysed in all patients, including whether a SLNB was performed, the results of SLNB, and any subsequent recurrence.
Ninety percent (91/101) of cases were still classified as PDM in the complete wide excision specimen while 10% (10/101) of cases were reclassified as MDM, which was a significant change in classification of final desmoplastic melanoma subtype (p<0.001). The proportion of desmoplastic melanoma was also significantly different between the initial and excisional biopsies (p=0.004). Forty-eight (48/101) patients had a SLNB, of which two (4.5%) were positive for metastatic melanoma; both cases were PDM in the excision specimen. Of the 10 cases demonstrating MDM in the excision specimen, the initial biopsy was a punch biopsy in six cases, shave biopsy in two cases and subcutaneous tissue was sampled in two patients (one punch biopsy, one incisional biopsy). Four of these 10 patients underwent SLNB which was negative in all cases. Twenty-two patients developed recurrence in the follow-up period (median 30 months, range 1–192 months), three with MDM in their excision specimen. One patient did not have a SLNB and developed regional lymph node recurrence.
In this study there was a 10% risk that the percentage of desmoplastic melanoma in an initial biopsy of PDM was not representative of the entire lesion, resulting in reclassification as MDM in the excision specimen. If a SLNB is not performed in such cases, a positive SLNB may be missed (one patient in our study) which could impact treatment options for the patient. We recommend caution in not offering a SLNB in the setting of an initial biopsy of PDM if the biopsy is small compared with the overall lesion. If a SLNB is not procured at the time of wide excision in such cases, the SLNs should still be mapped by lymphoscintigraphy to facilitate careful follow up and to enable earlier detection and treatment of nodal disease.
as a variant of spindle cell melanoma with prominent dense collagen amongst the lesional cells. It represents less than 4% of all primary cutaneous melanomas
and shows distinct clinical and pathological features when compared to more common variants. Clinically, DM is more likely to occur in older people and in males. It is strongly associated with chronic sun exposure, most commonly occurring in the head and neck region,
but also rarely occurring at sun-protected locations such as mucosal or acral sites. Pathologically, DM is characterised by spindle cells which are separated by thickened, hyalinised fibrous (desmoplastic) stroma in the invasive foci
(Fig. 1, Fig. 2) and may or may not have an associated overlying in situ component in the epidermis, which when present is most commonly of lentigo maligna subtype.
Due to the often subtle features of the lesion, DM can be confused with low grade lesions such as dermal scar, dermatofibroma or benign nerve sheath tumours and if DM displays high-grade features there is a risk of misdiagnosis as high-grade spindle cell tumours such as sarcomas.
Fig. 1(A–D) Pure desmoplastic melanoma from back of 57-year-old female (A, low power H&E). The invasive melanoma consists of a predominant desmoplastic component (B, medium power; C, high power; H&E) comprising paucicellular ovoid to spindle cells dividing thickened collagen with scattered lymphoid aggregates (∗). A minor non-desmoplastic component is present in the superficial aspect of the specimen (D, medium power H&E) comprising atypical naevoid cells (white arrow), with adjacent desmoplastic component (black arrow). Overall the non-desmoplastic component was 5% of the invasive melanoma and a pure desmoplastic melanoma was favoured.
Fig. 2(A–F) Mixed desmoplastic melanoma from scapula of 50-year-old male with predominant desmoplastic component (75% of lesion, white arrows) and minor non-desmoplastic component (25% of lesion, black arrows) (A, low power H&E). The desmoplastic component comprises paucicellular ovoid to spindle cells dividing thickened collagen with scattered lymphoid aggregates (∗) (B, medium power; C, higher power; H&E). Transition from desmoplastic to non-desmoplastic component (D, medium power H&E). Non-desmoplastic component comprising hyper cellular atypical ovoid to spindled cells with scattered mitoses (E, medium power H&E). Transition from desmoplastic to non-desmoplastic component (F, medium power SOX10).
Busam et al. noted in 2004 that there was considerable heterogeneity within the literature regarding the diagnostic criteria used for defining DM, which likely contributed to reported differences in the prognostic significance of this diagnosis.
To further investigate this, Busam et al. subdivided DM into pure or combined (later termed mixed) subtypes. Pure desmoplastic melanoma (PDM) was defined as when the invasive melanoma showed overwhelming (≥90%) stromal fibrosis with spindled melanocytic cells separated by dense collagen with an overall low cellularity. MDM was defined as a melanoma in which the desmoplastic component occupied less than 90% of the invasive melanoma.
Following this report, a number of studies have proven the importance of this distinction in DM by finding that PDM is less likely to metastasise to local regional lymph nodes.
The distinction between MDM and PDM also has been shown to be of importance because patients with MDM had inferior survival data compared to those with PDM.
Sentinel lymph node biopsy (SLNB) in patients with otherwise clinically-localised primary cutaneous melanoma is a highly effective prognostic and staging tool
However, the role of SLNB in the management of patients with DM has not been clear-cut. Numerous studies have suggested that due to the lower rates of positive SLNs in patients with PDM compared with MDM and non-DM melanoma, SLNB in these patients may not be indicated.
When performing wide local excision of a cutaneous melanoma, the procedure can interrupt lymphatic drainage pathways and therefore reduce the accuracy of SLN identification if it is performed subsequently. Several studies have demonstrated that SLNB can be performed after definitive wide local excision but the results may be inaccurate.
Therefore, if partial biopsy of a primary cutaneous melanoma contains PDM and SLNB is not performed, because of the perceived lower risk of metastasis compared to a MDM or non-DM melanoma with similar prognostic features, there is the possibility that the initial biopsy's proportion of DM may not be representative of the entire lesion. In this scenario, the wide local excision may contain a non-DM component in the invasive melanoma and the lesion would be classified as MDM overall. The consequence of this would be that the patient missed the opportunity of having the appropriate SLNB, and any subsequent attempt to locate the SLN may be inaccurate.
This study assessed the concordance of the proportion of desmoplastic melanoma in the initial biopsy with that in the entire lesion, following wide excision, in patients with primary cutaneous melanoma. We also determined how frequently this may have resulted in a SLNB not being performed and how frequently this could have had clinical implications.
Methods
The study was conducted with Human Research Ethics Committee approval (The Sydney Local Health District Human Research Ethics Committee, Protocol No. X17-0312 and 2019/ETH07604, and No. X15-0454 and 2019/ETH06874). Between 1994 and 2022, patients who were diagnosed with a primary cutaneous melanoma, who had both a record of an initial biopsy reported as PDM, and a subsequent wider excision performed within 3 months, were identified from the archives of Melanoma Institute Australia (MIA) and the Department of Pathology of Memorial Sloan-Kettering Cancer Centre (MSKCC). PDM was defined as desmoplastic melanoma (elongated or spindle cells with separate thickened/hyalinised stroma with an overall paucicellular appearance) comprising ≥90% of the dermally invasive melanoma. MDM was defined as invasive melanoma where the DM component accounted for <90% but >5% of the invasive melanoma and non-DM if there was <5% desmoplastic component in the invasive melanoma. A careful examination of the histopathology report of all biopsies was performed to confirm the diagnosis. Any melanomas where a proportion of desmoplastic melanoma was not present were excluded. Any case where there was no residual invasive melanoma in the wider excision specimen was also excluded.
In all cases, patient details including age and sex were recorded. Clinical details collected were date of biopsy, type of biopsy, site of primary, whether SLNB was performed, if there had been a recurrence, and where it occurred, and date of last follow-up or death. Histopathological features collected were Breslow thickness for each biopsy; mitoses per mm2; presence or absence of ulceration; and Clark level across the lesion as a whole; result of SLN (positive or negative); percentage of the invasive melanoma accounted for by a desmoplastic component; and presence or absence of neurotropism.
In patients where there was a discordance between the two biopsy specimens (initial partial biopsy, and the second, wider excision specimen) and the overall lesion was defined as a MDM compared with a PDM in the initial biopsy, the histopathology slides were retrieved to confirm the proportion of desmoplastic melanoma in each biopsy by two histopathologists experienced in the assessment of melanocytic lesions (RVR, RAS, KJB). The overall classification of PDM, MDM and non-DM was calculated by taking into account the entire tumour over the two biopsies. Only unequivocal cases of PDM were classified as such on the initial biopsy; if there was any doubt whether the biopsy of the lesion met the histological criteria for PDM, it was classified as MDM.
Statistical analyses comparing histological variables were performed using the unpaired non-parametric Mann–Whitney test for continuous variables and a Chi squared test for non-continuous variables. An exact binomial test was used to assess the concordance of DM subtype between the initial and excisional biopsies and a paired Mann–Whitney test was performed to assess the difference in the proportion of desmoplastic melanoma between the initial and excisional biopsies. A p value of <0.05 was considered statistically significant.
Results
Clinicopathological features
Eighty-one cases of cutaneous melanoma were retrieved from the MIA database and 23 from MSKCC archives which fulfilled the inclusion requirements. All had an initial biopsy reported as PDM with a subsequent wider excision containing dermally invasive melanoma, with histopathology reports of both biopsies being available for assessment.
Of the 104 cases with an initial biopsy showing PDM, 13 cases were found to be MDM in the subsequent excision. However, following review of histopathological sections, three of these cases were reclassified as MDM on their initial biopsy and were excluded from further analysis.
The clinical features of the remaining patients (n=101) are outlined in Table 1. In summary, of the 101 patients included, 73 (72%) of the patients were male and 28 (28%) were female. The ages of the patients ranged from 31 to 94 years (mean 67 years, median 68 years). Fifty-eight (57%) melanomas were in the head and neck region, 24 (24%) cases involved the trunk, 15 (15%) cases were from the upper limb and four (4%) cases were located on the lower limb.
Table 1Summary of clinicopathological features of 101 cutaneous melanomas with an initial biopsy of PDM; following subsequent wider excision, 91 cases remained classified as PDM and 10 cases were reclassified as MDM
MIA n (%)
MSKCC n (%)
PDM n (%)
MDM n (%)
Total n (%)
Total patients
78 (77)
23 (23)
91 (90)
10 (10)
101
Males
59 (76)
14 (61)
66 (73)
7 (70)
73 (72)
Females
19 (24)
9 (39)
25 (27)
3 (30)
28 (28)
Age, years
Mean
67
63
64
69
67
Median
69
65
67
70
68
Range
31–94
31–86
31–93
33–94
31–94
Anatomical site
Head and neck
47 (60)
11 (48)
52 (57)
6 (60)
58 (57)
Trunk
18 (23)
6 (26)
22 (24)
2 (20)
24 (24)
Arm
11 (14)
4 (17)
13 (14)
2 (20)
15 (15)
Leg
2 (3)
2 (9)
4 (4)
0 (0)
4 (4)
Initial biopsy type
Punch
30 (38)
16 (70)
40 (44)
6 (60)
46 (46)
Shave
5 (6)
6 (26)
9 (10)
2 (20)
11 (11)
Excision
26 (33)
0 (0)
26 (29)
0 (0)
26 (26)
Incision
9 (12)
0 (0)
7 (8)
2 (20)
9 (9)
Unknown
8 (10)
1 (4)
9 (10)
0 (0)
9 (9)
MDM, mixed desmoplastic melanoma; MIA, Melanoma Institute Australia; MSKCC, Memorial Sloan Kettering Cancer Centre; PDM, pure desmoplastic melanoma.
For the initial biopsy of the melanomas, 46 (46%) were punch biopsies, 11 (11%) were shave biopsies, 26 (26%) were excision biopsies, nine (9%) were incisional biopsies and in the remaining nine cases the biopsy type could not be ascertained from the reviewed pathology reports.
The histopathological features of all cases are shown in Table 2, where the cases are grouped according to their final designation as either PDM or MDM after assessment of both biopsies. Ninety-one (90% of cases) were PDM and 10 (10% of cases) were MDM. The Breslow thickness ranged from 0.5 to 23 mm (99/101 cases had a thickness ≥1.0 mm) and neurotropism was absent in 35 cases, present in 61 cases and not reported in 5 cases. Mitotic rate ranged from 0–57 per mm2 (median=1) and ulceration was present in 18 cases and absent in 83 cases. There was no statistically significant difference between the PDM and MDM group for these prognostic histological features.
Table 2Summary of pathological features and outcomes of 101 cutaneous melanomas with an initial biopsy of PDM; following subsequent wider excision, 91 cases remained classified as PDM and 10 cases were reclassified as MDM
Overall n (%)
PDM n (%)
MDM n (%)
p value
Cases
101 (100)
91 (90)
10 (10)
Breslow thickness, mm
0.72
Range
0.5–23
0.5–23
0.8–11
Mean
6.4
6.5
5.2
Median
5.2
5.2
5.2
Neurotropism
0.8
Present
61 (60)
55 (60)
6 (60)
Absent
35 (35)
31 (34)
4 (40)
Not reported
5 (5)
5 (5)
0 (0)
Mitoses,/mm2
0.45
Range
0–57
0–57
0–25
Mean
4
3
4.2
Median
1
1
2
Ulceration
0.85
Present
18 (18)
16 (18)
2 (20)
Absent
83 (82)
75 (82)
8 (80)
Sentinel lymph node biopsy
Performed
48 (48)
44 (48)
4 (40)
Not performed
50 (50)
44 (48)
6 (60)
Unknown if performed
3 (3)
3 (3)
0 (0)
Sentinel lymph node biopsy result
Positive
2 (4)
2 (5)
0 (0)
Negative
46 (96)
42 (95)
4 (100)
Total patients recurred
22 (22)
19 (21)
3 (30)
Site of first recurrence
Local
5 (23)
4 (21)
1 (33)
Locoregional lymph nodes
5 (23)
3 (16)
2 (67)
Distant
12 (55)
12 (63)
0 (0)
Time to first recurrence, months
Range
3–40
4–40
3–36
Mean
20
19
23
Median
18
18
30
MDM, mixed desmoplastic melanoma; PDM, pure desmoplastic melanoma.
Forty-eight (48%) patients had a SLNB procedure at the time of wide local excision, of which two (4%) were positive for metastatic melanoma. Both positive SLNBs were in patients with PDM. During the follow-up period (median 30 months, range 1–192 months), 22 patients developed a recurrence of melanoma of which 19 were PDM (21%) and three were MDM (30%). Of the recurrent PDM cases, 12 of 19 (63%) recurred distantly, four (21%) recurred locally and three (16%) recurred in locoregional lymph nodes. Of the three MDM cases which recurred, one (33%) recurred locally and two (67%) recurred at locoregional lymph nodes; there were no distant recurrences. The mean time to recurrence was 20 months (median 18 months).
Proportions of desmoplastic melanoma
The proportions of desmoplastic melanoma in both the initial biopsy and the excision biopsy are shown in Fig. 3. There was a 90% (91/101 cases) concordance of designation of PDM on an initial partial biopsy with the final DM subtype (PDM/MDM) on the subsequent complete excision specimen which represents a statistically significant difference between biopsies (Exact binomial test, p<0.001). There was also a significant difference between the proportion (percentage) of desmoplastic melanoma in the initial biopsy when compared to the excision biopsy (paired Mann–Whitney test, p=0.004). When assessing this difference by initial biopsy type, the difference was only significant for the punch biopsy group (n=46, paired Mann–Whitney test, p=0.039) and not for excisional biopsy group (n=26, paired Mann–Whitney test, p=0.35). Differences were also not significant for the incisional (n=9) or shave biopsy (n=11) subgroups, albeit only a small sample size was available for each of these groups.
Fig. 3Proportion of desmoplastic melanoma in initial and excision biopsies in 101 patients with an initial biopsy of pure desmoplastic melanoma. (A) Stratified by initial biopsy type; (B) stratified by desmoplastic melanoma component in excision biopsy.
Fig. 4, Fig. 5 show images for two cases (Patients 10 and 4, respectively) which were reclassified from PDM to MDM.
Fig. 4(A–F) Melanoma excised from cheek of 59-year-old female. (A–C) Initial shave biopsy showed a PDM with greater than 95% desmoplastic melanoma component within the dermally invasive melanoma (A, low power H&E; B, low power S100 stain). The invasive melanoma is paucicellular with the malignant spindle cells dissecting the dermal collagen (C, high power H&E). (D–F) Wide excision showed an invasive melanoma (D, low power H&E) with paucicellular desmoplastic melanoma component similar to previous biopsy (E, high power H&E) and also more cellular non-desmoplastic melanoma component with fascicles of more cytological atypical and mitotically active spindle cells. Overall the lesion was classified as a mixed desmoplastic melanoma with 75% of the lesion desmoplastic melanoma (Patient 10, Table 3).
Fig. 5(A–F) Melanoma excised from nose of 80-year-old male. (A–C) Initial punch biopsy showed an invasive melanoma with 90% desmoplastic melanoma component within the dermally invasive melanoma. The invasive desmoplastic melanoma consisted of atypical spindle cells dividing thickened collagen bundles (A, low power; B, medium power; C, high power; H&E). (D–F) Wide excision showed an invasive melanoma (C, low power H&E) with desmoplastic melanoma component centrally (white arrow) (E, high power H&E) and a cellular non-desmoplastic melanoma component towards the periphery of the lesion (black arrow) (F, high power H&E). Overall the lesion was classified as a mixed desmoplastic melanoma with 70% of the lesion a desmoplastic melanoma component (Patient 4, Table 3).
The histological features and recurrence data for the 10 cases reclassified from PDM to MDM after excision are shown in Table 3. These 10 cases represent 10% of all of the cases with an initial biopsy showing PDM (n=101). In these patients, four had SLNB performed and all were negative for metastatic melanoma. Three of the 10 patients in this group have experienced a subsequent recurrence, two in regional lymph nodes and one locally. Of the two patients with a locoregional lymph node recurrence, one had a previous negative SLNB. Six of the 10 cases had an initial punch biopsy, two had shave biopsies and two had incisional biopsies. Of these discordant cases, the percentage of the lesion sampled in the initial biopsy ranged from 5 to 50% (mean 18.5%, median 15%). Only two of these cases sampled subcutaneous tissue in the initial biopsy.
Table 3Ten cases with initial biopsy of PDM and overall classified as MDM following subsequent wide excision
Case
First biopsy
Second biopsy
Change in desmoplasia
Breslow thickness (mm)
SNLB performed
SNLB status
Recurrence
Site of first recurrence
Type
Estimated % of entire lesion sampled
% Desmoplasia
Type
% Desmoplasia
1
Punch biopsy
50
100
WLE
50
50
7.0
Yes
Negative
No
–
2
Punch biopsy
25
95
WLE
60
35
2.9
Yes
Negative
Yes
Locoregional lymph node
3
Shave biopsy
20
95
WLE
70
25
2.7
Yes
Negative
No
–
4
Punch biopsy
5
90
WLE
70
20
11.0
No
–
No
–
5
Incisional biopsy
30
100
WLE
60
40
0.8
No
–
Yes
Local
6
Punch biopsy
10
100
WLE
75
25
5.4
No
–
No
–
7
Incisional biopsy
20
100
WLE
80
20
7.5
No
–
Yes
Locoregional lymph node
8
Punch biopsy
5
90
WLE
85
5
7.5
No
–
No
–
9
Punch biopsy
10
90
WLE
85
5
5.0
Yes
Negative
No
–
10
Shave biopsy
10
90
WLE
75
15
3.1
No
–
No
–
MDM, mixed desmoplastic melanoma; PDM, pure desmoplastic melanoma; SLNB, sentinel lymph node biopsy; WLE, wide local excision.
Clinically it most frequently involves the head and neck areas of the elderly and histopathologically is characterised by spindle cells which are dispersed in a prominent fibrous stroma which is frequently hyalinised or slightly myxoid. DM can be misdiagnosed clinically and histologically as a scar, dermatofibroma, neurofibroma or miscellaneous low grade spindle cell lesion, and is one of the leading causes of malpractice claims associated with the misdiagnosis of melanoma.
This was most likely related to the heterogenous range of tumours which were classified as DM, whereby even within a single institution the interpretation of what is required for a designation as DM was variable.
To further understand this entity, Busam et al. classified DM further as either a PDM ‘if the overwhelming majority (≥90%) was associated with prominent stromal fibrosis’, or as a MDM if the desmoplastic component was <90% of the invasive melanoma. This distinction was proved to correlate with clinical outcomes, with MDM showing worse disease-free survival and melanoma-specific survival when compared to PDM.
SLNB is a widely used prognostic and staging tool in patients with clinically-localised melanomas >1.0 mm thick (and sometimes in those with thinner tumours) and is required as part of the current (8th edition) AJCC staging system for cutaneous melanoma.
Due to conflicting reports on the positivity rates in DM, the appropriateness of SLNB in patients with this variant of melanoma has been questioned. In a recent systematic review, the SLNB positivity rate in all reported desmoplastic melanomas in the literature was 6.5%.
As far as we are aware, our study is the second largest to assess SLNB positivity in PDM and showed a positivity rate of 4.5% (2/44) in PDM. This compared to 16.6% (13/78) for MDM (results from a search of the MIA database for the same period as the study, unpublished data). These results are reasonably similar to the results in a recent systematic review.
This has led to most current clinical management guidelines not being definitive or not addressing SLNB recommendations specifically in patients with PDM and MDM.
Cancer Council Australia Melanoma Guidelines Working Party. Clinical practice guidelines for the diagnosis and management of melanoma. Sydney: Melanoma Institute Australia. https://wiki.cancer.org.au/australia/Guidelines:Melanoma.
Sentinel lymph node biopsy and management of regional lymph nodes in melanoma: American society of clinical oncology and society of surgical oncology clinical practice guideline update.
Due to the lack of current clear guidelines for management of patients with DM, we recognise that in the situation in which an initial biopsy of a melanoma is reported to be a PDM, some clinicians, following consultation with their patients, will decide not to proceed with SLNB at the time of wide excision, due to a perceived low risk of metastasis to regional lymph nodes. In such cases, if the SLN is not mapped at the time of wide excision, and the excision is subsequently proven to contain sufficient non-DM melanoma to designate the case overall as a MDM, then the opportunity to perform a SLNB may have been lost. This could impact the ability to gain important prognostic information and, if a positive SLNB is not procured, the opportunity to offer adjuvant systemic treatment options, including immune checkpoint or targeted therapies. Therefore, it is of clinical importance to understand the risk of the proportion of desmoplastic melanoma in an initial biopsy not being representative of the entire lesion. To our knowledge, ours is the first study to address this question. We found that in 10% (10/101) of cases where there was an initial diagnosis of PDM in the initial biopsy, the overall lesion was interpreted as a MDM following wide excision. The risk of a possible missed positive SLNB, due to over-representation of PDM in the initial biopsy in our study, was 1% (1/101) of cases with an initial biopsy of PDM in our study. In the discordant group of MDM cases, six of the 10 cases had initial punch biopsies, two were shave biopsies and only two sampled the subcutaneous tissue (one incisional biopsy and one punch biopsy). Therefore, caution should be used by the clinician if the decision is made not to pursue SLNB after an initial biopsy of PDM when the initial biopsy contains only a small amount of tumour such as with a small punch biopsy and/or when only the dermis is sampled, as the risk of a non-representative sample is higher in such biopsies because of limited sampling.
It is important for clinicians to understand the small risk of a non-representative proportion of desmoplastic melanoma component in the initial biopsy when they have discussions with their patients regarding the appropriateness of SLNB in PDM. If the decision was made not to proceed with SLNB in such cases, mapping with lymphoscintigraphy could be considered to enable close ultrasound monitoring of SLN if the lesion proved to be a MDM on wide excision.
We recognise that ours is the first study to evaluate the clinical question of the representativeness of PDM in partial biopsies and we only assessed 101 cases with an initial biopsy of PDM. Further studies need to be performed to confirm our findings. However, of equal if not greater importance, is the need for greater reproducibility of the criteria used by pathologists to diagnose desmoplastic melanoma and distinguish it from non-desmoplastic and spindle cell melanomas. As treatment decisions depend on the accurate percentage of any desmoplastic melanoma component, there need to be strict guidelines about its defining features. We recommend that, to qualify as desmoplasia, the melanoma should be low in cellularity with prominent fibrous stroma. Anecdotally, the most frequent cause of over-estimation of desmoplasia in our experience is a spindle cell melanoma with areas of low cellularity in keeping with desmoplastic melanoma but elsewhere the lesional cells become more cohesive with minimal stroma, greater atypia and more mitoses, which should be interpreted as a non-DM focus. This was the reason for the redesignation of the three cases from PDM to MDM on review of the initial biopsy in our study. As the designation of a PDM may result in a change in management by not offering SLNB, if there is any ambiguity or uncertainty regarding the diagnosis of PDM in an initial biopsy, we recommend that the pathologist designate the lesion as MDM and that only unequivocal PDMs be reported as such.
This study, to our knowledge the first of its kind, provides important practical information to clinicians and those drafting guidelines and considering the appropriateness of SLNB in patients with pure desmoplastic melanomas.
Conflicts of interest and sources of funding
No author specifically received support for the work reported in this manuscript. RVR is supported by a Clinical Researcher Scholarship from Sydney Research. RAS and JFT were recipients of an Australian National Health and Medical Research Council (NHMRC) program grant (APP1093017). RAS is supported by NHMRC Practitioner Fellowship. Research reported in this publication that was conducted at Memorial Sloan Kettering Cancer Center was supported in part by the Cancer Center Support Grant of the National Institutes of Health/National Cancer Institute under award number P30CA008748. JFT has received honoraria for advisory board participation from BMS Australia, MSD Australia, GSK and Provectus Inc, and travel and conference support from GSK, Provectus Inc and Novartis. RAS has received fees for professional services from F. Hoffmann-La Roche Ltd, Evaxion, Provectus Biopharmaceuticals Australia, Qbiotics, Novartis, Merck Sharp & Dohme, NeraCare, Amgen, Bristol-Myers Squibb, Myriad Genetics, GlaxoSmithKline. RPMS has received honoraria for advisory board participation from MSD, Novartis and Qbiotics and speaking honoraria from BMS and Novartis.
Acknowledgements
Support from The Cameron Family as well as from colleagues at Melanoma Institute Australia and the Royal Prince Alfred Hospital is gratefully acknowledged.
References
Conley J.
Lattes R.
Orr W.
Desmoplastic malignant melanoma (a rare variant of spindle cell melanoma).
Cancer Council Australia Melanoma Guidelines Working Party. Clinical practice guidelines for the diagnosis and management of melanoma. Sydney: Melanoma Institute Australia. https://wiki.cancer.org.au/australia/Guidelines:Melanoma.
Sentinel lymph node biopsy and management of regional lymph nodes in melanoma: American society of clinical oncology and society of surgical oncology clinical practice guideline update.
00689-4&id=gr1.jpg){kind=link}
00689-4&id=gr2.jpg){kind=link}
00689-4&id=gr3.jpg){kind=link}
00689-4&id=gr4.jpg){kind=link}
00689-4&id=gr5.jpg){kind=link}