Challenges in Melanoma Diagnosis and Management

Challenges in Melanoma Diagnosis and Management Winter Clinical Dermatology Conference - Hawaii Darrell S. Rigel, MD MS Clinical Professor of Dermatology New York University Medical Center DISCLOSURE OF RELEVANT RELATIONSHIPS WITH INDUSTRY Darrell S. Rigel, MD Challenges in Melanoma Diagnosis and Management Castle A, H, I DS Rigel 2010 Other Cancers 1,584,500 Cancer USA - 2017 Skin Cancer 2,900,000 Skin Cancer USA - 2017 Melanoma 87,110 More Skin Cancers than all other cancers combined NMSC (BCC SCC) 3,000,000 Melanoma - USA Melanoma - USA 1/50 Lifetime Risk Invasive MM 1/150 Lifetime Risk Invasive MM 1/74 1/65 1/250 1/100 1/1500 1/600 1/1500 1/600 1/250 1/150 1930 1950 1980 2000 Kopf et al, J Derm Surg Oncol, 1982 1930 1950 1980 1985 1993 2000 2004 2017 Rigel et al, NYU Melanoma Cooperative Group, 2017 1

Melanoma - USA Lifetime Risk 1/65 Invasive MM 1/74 1/100 1/150 1/250 1/1500 1/600 1/40 Melanoma US 1/50 Projected 2017 Invasive = 87,100 In-situ = 63,410 1930 1950 1980 1985 1993 2000 2004 2017 2020 Rigel et al, NYU Melanoma Cooperative Group, 2017 Melanoma USA 2017 Melanoma incidence rising worldwide Lifetime Risk Invasive and Insitu MM 1/25 150,510 total cases 1930 1950 1980 1985 1993 2000 2017 Erdmann et al, Int J Cancer, 2013 Rigel et al, NYU Melanoma Cooperative Group, 2017 US Cancer Statistics, 2017 US Melanoma % Annual Incidence Change by gender 6% 5% 5.6% 5.5% 4% 3% 2% 3.2% 2.4% 1.7% 1% 0.4% 1975-85 1985-2005 2005-13 Male Female Siegel et al, Ca J Clincians, 2017 Siegel et al, Ca J Clincians, 2017 2

2016 2017 2016 2017 11% 18% Siegel et al, Ca J Clincians, 2017 Siegel et al, Ca J Clincians, 2017 US Lifetime Risk for Invasive Melanoma Non-Hispanic Whites US MM Incidence by Age 4% 3% 2% 1% 3.5% 1 in 28 2.4% 1 in 44 Male Female Siegel et al, Ca J Clincians, 2017 MM Anatomic Distribution Age 15-39 4 MM Anatomic Distribution Age 15-39 Males Females 3.5 3.6 49% 19% 15% 38% 22% 3 2.5 2 2.7 2.1 3.1 Males Females 17% Trunk Arms Legs Face H N 8% 32% Trunk Arms Legs Face H N 1.5 1 0.5 1.1 0.9 0.9 0.8 Weir et al, J AmerAcad Dermatol. 2011 0 Trunk Arms Legs Face H N Weir et al, J AmerAcad Dermatol. 2011 3

MM Anatomic Distribution Age 15-39 Melanoma incidence was higher among females (ageadjusted incidence rate = 9.74) vs. males (age-adjusted incidence rate = 5.77) and increased with age Increases occurred over time for Females but not Males These data suggest areas for etiologic research around geneenvironment interactions and the need for targeted cancer control activities specific to adolescents and young adult populations MM US 5 Year Survival by Year of Diagnosis 10 93% 88% 75% 82% 5 25% 1975-77 1987-89 2005-13 % Surviving Weir et al, J AmerAcad Dermatol. 2011 Siegel et al, CA J Clinicians, 2017 % Surviving 100 90 80 70 60 50 40 30 20 10 0 Melanoma US 10 Year Survival 92 Localized Regional Distant 54 12 Stage Pollack et al, J AmerAcad Dermatol, 2011 10 75% 5 25% MM US by Stage at Diagnosis 2004-2010 84% 9% 4% Localized Nodal Distant Siegel et al, CA J Clinicians, 2017 MM US 5 Year Survival by Year of Diagnosis Localized Disease % Surviving 10 75% 5 25% 96% 98% 99% 1992-95 1996-99 2004-10 Siegel et al, CA J Clinicians, 2017 US Annual Deaths from Melanoma 9940 10130 9710 9730 10000 9480 9180 8650 8700 8790 9000 8450 7910 8110 8000 7000 6000 5000 4000 3000 2000 1000 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 ACS 4

Skin Cancer Deaths US - 2017 For Melanoma: Melanoma 9,730 Other 3,860 Do factors relating to Melanoma Incidence = those related to Melanoma Mortality? Over 1 American dies of Melanoma every hour For Melanoma: Melanoma Pancreatic Cancer Low Death % High Death % Incidence = Mortality Early detection effective Incidence = Mortality Early detection rare Incidence = Mortality Melanoma Highest Incidence HI WY AZ FL NM AR CO Highest Mortality OK VT UT DE OR NC NV Pancreatic Cancer Highest Incidence CT LA ME WY WA NV WI Highest Mortality MA WI NV NJ LA ME CT Melanoma Prevention Primary Secondary 5

UV and Melanoma Comparison of Regional and State Differences in Melanoma Rates in the United States: 2003 vs 2013 MM Incidence is the measure of UV exposure behavior and MM development MM mortality, is a measure of access to care and early detection Therefore, when evaluating melanoma prevention, we need to look at both types of prevention to be effective Mounessa et al, JAMA Dermatology, 2016 Comparison of Regional and State Differences in Melanoma Rates in the United States: 2003 vs 2013 The Northeast, specifically New England, is the only US geographic region in which most states experienced a reduction in both death and incidence rates. Strong skin cancer prevention programs likely played a role in this region s success such as an initiative which funded sunscreen dispensers in public and recreational areas throughout Boston and other New England cities. Conclusion: Such programs may enhance public awareness about skin cancer and may suppress the continual rise in melanoma. Mounessa et al, JAMA Dermatology, 2016 Cost-Effectiveness of Skin Surveillance Through a Specialized Clinic for Patients at High Risk of Melanoma Decision-analytic model to compare the costs and benefits of specialized surveillance vs. standard care over a 10-year period Specialized surveillance through the High Risk Clinic was both less expensive and more effective than standard care. Mean saving was A$6,828 (95% CI, $5,564 to $8,092) per pt Mean quality-adjusted life-year gain was 0.31 (95% CI, 0.27 to 0.35) Main drivers of the differences were detection of melanoma at an earlier stage resulting in less extensive treatment and a lower annual mean excision vs. standard care Specialized surveillance was a cost-effective strategy for the management of individuals at high risk of melanoma Watts et al, J Clin Oncol, 2016 Indoor Tanning in Diverse US Youth Indoor Tanning in Diverse US Youth Data from 2015 Youth Risk Behavior Survey Non-heterosexual youths are at increased risk for skin cancer Non-heterosexual youths (White, Hispanics and Blacks)_ have an increased usage of tanning beds Non-heterosexual black youths have a higher tanning usage than white females Blashill et al, JAMA Dermatol. 2016 Blashill et al, JAMA Dermatol. 2016 6

Indoor Tanning in Diverse US Youth Data from 2015 Youth Risk Behavior Survey Non-heterosexual youths are at increased risk for skin cancer Non-heterosexual youths (White, Hispanics and Blacks)_ have an increased usage of tanning beds Non-heterosexual black youths have a higher tanning usage than white females Need to incorporate sexual orientation and ethnicity when developing skin cacner prevention programs for youth Blashill et al, JAMA Dermatol. 2016 Early MMs Increasing Faster Than Advanced MMs Cromwell et al, Sem Oncol, 2010 Population-Based 20-Year Survival Among People Diagnosed With Thin Melanomas 26,736 people diagnosed with thin melanomas 20-year survival was 96% Most influential determinants of prognosis were thickness 0.75 mm (HR 4.33 vs. tumors < 0.25 mm) and patient age at diagnosis >65 years (HR 2.8) vs. age <25 years. Acral lentiginous and nodular tumors, male gender, tumor site on the scalp or neck, or tumor invasion of the entire papillary dermis each independently increased the risk of dying from thin invasive melanoma Clinical vigilance is warranted for patients with thin melanomas Green et al, J Clin Oncol, 2014 Prediction of Survival in Patients With Thin Melanoma: Results From a Multi-Institution Study 2,243 patients with thin MMs (<1mm) were retrieved from databases at 6 centers with 124 mons follow up Worst prognosis categories were age >60 years, Breslow thickness >0.75 mm, Mitotic Rate 1, ulceration, presence of LVI, and regression 5. Breslow thickness >0.75 mm, Mitotic Rate 1, ulceration, and LVI were significantly associated with SLN+. Mortality associated with thin melanomas Pts need to be followed closely Maurichi et al, J Clin Oncol, 2014 More people die from thin melanomas than thick melanomas 4,218 Australians who died from melanoma between 1990 and 2009, thin melanomas (<1mm) accounted for 23% of melanoma deaths overall More people died from thin melanomas (296 deaths, 23%) than from thick melanomas more than 4 mm in thickness (186 deaths, 14%) or from metastatic presentations (207 deaths, 16%). More people with thin melanomas die than with thick melanomas because there are so many more thin lesions Whiteman and Olsen, WCCS 2014 Melanoma US by Disease Stage 86% 11% 3% Localized Regional Distant Pollack et al, J AmerAcad Dermatol, 2011 7

Recurrence Probability by Stage Invasive MM US Cases by Thickness SEER 1992-2003 16% 72% 8% 4% Turner et al, J Clin Oncol, 2011 <1mm 1-1.99mm 2-3.99mm 4+mm Landow et al, SID poster, 2016 Invasive MM US Deaths by Thickness SEER 1992-2003 29% 27% 17% 27% <1mm 1-1.99mm 2-3.99mm 4+mm Landow et al, SID poster, 2016 Detection of Occult Invasion in Melanoma In Situ Unequivocal MMIS without associated nevi or regression was identified using a consecutive sample of 33 cases 3 sequential slides were stained with H&E and melan-a. Melan-A stained slides showing definitive invasion were double-stained with Sry-related HMg-Box gene 10 (SOX10) to confirm the melanocytic nature of the cells Occult invasive melanoma was detected in 11 of 33 consecutive cases (33%) of previously diagnosed MMIS 6 of 11 melanomas (55%) were diagnosable only by immunohistochemistry History and physical examination including regional lymph nodes, education, and surveillance recommendations should be based on a very low, but not zero, risk of metastasis for MMIS Bax et al. JAMA Dermatol, 2016 Pathology Review of Thin MM and MMIS Impact on Treatment Decisions Overall pathologic discordance rate in diagnosis 4% (15/420 pts) Overall change in tumor staging rate 24% (97/405 pts) Changes in surgical excision margins in 12% of pts (52/420 pts) Decision about performing a sentinel lymph node biopsy in 16% of pts (67/420 pts) Review of thin MM or MMIS by an expert dermatopathologist results in frequent, clinically meaningful alterations in diagnosis, staging, prognosis, and surgical treatment Santillan et al, J Clin Oncol, 2010 DS Rigel 2010 Impact of Genetics on Melanoma Diagnosis and Prognosis 8

What does the mypath test do? Premise: Benign and malignant melanocytic lesions behave differently (invasion, metastasis, immune function etc.) and this is associated with expression of different (and/or different amounts of) RNAs. Identified a panel of 23 genes that are differentially expressed in benign and malignant melanocytic lesions and has developed a mypath test that purifies RNA from the tissue; quantifies how much of each of the 23 RNAs is expressed; and applies a mathematical algorithm to objectively determine if the lesion is benign or malignant based on the expression pattern.. Prognostication in Melanoma: Holy grail in melanoma biology Problems with Histologic Variables and Prognosis AJCC Recommendations Heavily Weighted Toward Use of Histology To Prognosticate in Melanoma Inability to apply to partial biopsies in many cases. AJCC recommendations based on excisions Stages are confusing; difficult to understand for both patients and physicians Many criteria are subjective, ie, ulceration, regression, Clark s levels, vertical growth phase, Breslow s thickness, vascularity, presence of TIL Some criteria cannot be quantified, ie, ulceration: reported as present or absent Questions about mitoses Ignores host immunity and other host factors A more objective, more accurate method is sorely needed Currently, most significant prognostic techniques for melanoma are AJCC criteria and Sentinel Lymph Node status What if we could non invasively identify patients who will have aggressive disease? 9

Genetic Testing to Assist in Prognosis of Melanoma FDA approved test developed by Castle Biosciences, Friendswood, Texas Uses formalin fixed, paraffin embedded tissue Quantifies expression of 31 genes from primary tumor to develop a Gene Expression Profile (GEP) Applies a validation algorithm to classify patients as Class 1 (low) vs Class 2 (high) risk of developing metastatic disease within 5 years Could be early reflex test for melanoma to suggest prognosis if becomes practical and cost effective GEP Uses formalin-fixed, paraffin-embedded tissue Quantifies expression of 31 genes from primary tumor Applies a validation algorithm Classifies patients as low vs. high risk Class 1 test result: Low Risk of metastasis within 5 years Class 2 test result: High risk of metastasis within 5 years GEP-31 Kaplan Meir Survival Curves GEP-31 Kaplan Meir Survival Curves Uveal Melanoma Cutaneous Melanoma 31 GEP Melanoma Analysis Summary This analysis shows that both SLNB positive status and DecisionDx- Melanoma Class 2 are important predictors of DMFS and OS. SLNB identified ~3 of patients who died, but 7 of patients who died were SLNB negative. Performing the 31 GEP-Melanoma assay in the SLNB negative cohort identified over 8 of those SLNB negative patients who developed distant metastasis and died. 10 75% 5 25% SLNB SLNB+ All Deaths n=62 Class 2 31 GEP Melanoma Class 2 identified 8 of SLNB deaths Wide versus narrow excision margins for high-risk, primary cutaneous melanomas: randomized trial survival analysis Previously published randomized trial (same researchers) of narrow (1 cm) versus wide (3 cm) excision margins in pts with thick cutaneous MMs showed narrow margins were associated with an increased frequency of locoregional relapse Current guidelines advise a 2 cm margin for MMs >2 mm in thickness Multicenter trial at 59 hospitals 900 pts with one primary localized MM greater than 2 mm in Breslow thickness on the trunk or limbs (excluding palms or soles) were randomly assigned (1:1) to receive surgery with either a 1 cm or 3 cm excision margin following an initial surgery Hayes et al, Lancet Oncol, 2016 10

Management of MM by US Dermatologists Email survey of US Dermatologists (n=510, 8% response rate) performed in August 2015 Asked questions on how they evaluated and managed MM 6 4 2 MM Survey AAD Data 15% 16% Practice Setting Validation 13% 1 45% 48% 27% Academic/University Dermatology Group Multispecialty Solo 26% 6 4 2 Experience Validation: MM Survey AAD Data 27% 35% 52% Years in Practice (YIP) 49% 21% 16% 2 15% 1 5% 13% 1 Geograpic Validation: MM Survey AAD Data 13% 13% 8% 1 11% 14% 8% 8% Zip Code First Digit 1 1 7% 6% 6% 6% 5% 5% 19% 18% 0 10 YIP 11 39 YIP 31+ YIP 0 1 2 3 4 5 6 7 8 9 Melanoma Surgical Margins Surgical margins for melanoma in-situ 1072 patients with 1120 MMIS was studied. All lesions were excised by Mohs micrographic surgery with frozen-section examination of the margin. The minimal surgical margin was 6 mm, and the total margin was calculated by adding an additional 3 mm for each subsequent stage required. 86% of MMIS were successfully excised with a 6-mm margin; 9 mm removed 98.9% of MMIS. The superiority of 9-mm to 6-mm margins was significant (P <.001). 5-mm margins for MM in-situ may be inadequate Kunishige et al, J Am Acad Dermatol, 2012 11

8 MMIS Margin Used 8 Melanoma thickness: <1mm Margin Used 6 62% 6 61% 4 36% 4 34% 2 2 2% <5mm* 6 10mm 1.1 1.9cm >2cm *consistent with guidelines 2% 3% <5mm 6 10mm* 1.1 1.9cm >2cm *consistent with guidelines 6 4 Melanoma thickness: >1mm Margin Used 54% 10 8 6 Treat vs. Refer out 11% 22% 67% 89% 78% 2 14% 32% 4 2 33% 0.1 <5mm 6 10mm 1.1 1.9cm* >2cm *consistent with guidelines MMIS Melanoma <1mm Melanoma >1mm Treat Refer Out Accuracy of initial shave biopsy as a predictor of final stage in pts with thin melanomas 81 thin melanomas ( 1mm) diagnosed via shave biopsy with positive deep and/or lateral margins Four cases (5%) had a change in stage or T level. In the majority of cases (95%) the thickness from shave biopsy is predictive of the final stage Kahana et al, J Am Acad Dermatol, 2013 Misdiagnosis of Melanoma by Type of Biopsy Negative impact of punch biopsy on accurate dx Ng, J. C. et al. Arch Dermatol 2010;146:234-239. 12

4 Preferred Biopsy Method Biopsy Preference Practice Setting 3 31% 36% 6 Shave Narrow Excision 2 4 41% 42% 1 3% Wide Local Excision Narrow Local Excision Shave Biopsy 13% Saucerization/ Scoop 11% 7% Punch Biopsy Other 2 23% Academic/Derm Group P<0.001 29% Multispecialty/Solo 6 Follow Up Dx <5yrs 8 Follow Up Dx >5yrs 4 49% 6 63% 4 2 25% 24% 2 3 q3month q6month q12month other 2% 2% 5% q3month q6month q12month other 10 75% 95% 87% Years in Practice (YIP) Treat vs. Refer out 86% 74% <10 YIP >10 YIP P<0.001 10 75% 82% 9 Practice Setting Treat vs. Refer out Academic P<0.05 All other practice types 7 5 5 51% 49% 25% 26% 5% 13% 14% MMIS treat MMIS refer out MM<1mm treat MM<1mm refer out 25% 3 18% 1 MMIS treat MMIS refer out MM<1mm treat MM<1mm refer out 13

Years in Practice (YIP) MMIS: Treat vs. Refer out Practice Setting MMIS: Treat vs. Refer out <10 YIP >10 YIP Academic All other practice types MMIS treat MMIS refer out MMIS treat MMIS refer out MMIS treat MMIS refer out MMIS treat MMIS refer out 5% 13% 18% 1 95% 87% 82% 9 P<0.001 P<0.05 Years in Practice (YIP) MM<1mm: Treat vs. Refer out Practice Setting MM<1mm: Treat vs. Refer out <10 YIP MM<1mm treat MM<1mm refer out >10 YIP MM<1mm treat MM<1mm refer out MM<1mm treat Academic MM<1mm refer out All other practice types MM<1mm treat MM<1mm refer out 14% 26% 3 49% 51% 86% 74% 7 P<0.001 P<0.05 Management of MM by US Dermatologists Email survey of US Dermatologists (n=510, 8% response rate) performed in August 2015 Asked questions on how they evaluated and managed MM Guidelines only partially followed Large differences in approaches Differences in approaches by experience Educational opportunity exists Maybe guidelines may be need to be reviewed/revised Challenges and Controversies in Treating Advanced Disease 14

Targeting Approaches to Systemic MM BRAF inhibitors interrupts the B Raf/MEK step on the ac va on pathway if the B Raf has the V600E mutation MEK inhibitors inhibits the mitogen activated protein kinase enzymes MEK1 and/or MEK2 PD 1 blockers programmed death 1 (PD 1) receptor is a negative regulator of T cell effector mechanisms that limits immune responses against cancer CTLA 4 antibodies CTLA 4 inhibits T cell responses Targeting Approaches to Systemic MM BRAF inhibitors interrupts the B Raf/MEK step on the ac va on pathway if the B Raf has the V600E mutation MEK inhibitors inhibits the mitogen activated protein kinase enzymes MEK1 and/or MEK2 PD 1 blockers programmed death 1 (PD 1) receptor is a negative regulator of T cell effector mechanisms that limits immune responses against cancer CTLA 4 antibodies CTLA 4 inhibits T cell responses Safety Profile of Nivolumab Monotherapy: Pooled Analysis of Patients With Advanced MM 576 patients, 71% experienced any-grade treatment-related AEs (most commonly fatigue [25%], pruritus [17%], diarrhea [13%], and rash [13%]) 1 experienced grade 3 to 4 treatment-related AEs AEs (occurring in 49% of patients) were most frequently skin related Treatment-related AEs with nivolumab monotherapy were primarily low grade, and most resolved with established safety guidelines Weber et al, J Clin Oncol, 2016 Programmed Death-Ligand 1 (PD-L1) and MM Response to the Anti Programmed Death 1 Antibody Pembrolizumab in MM 451 MM patients received pembrolizumab 344 (76%) had PD-L1 positive tumors Higher response rate and longer PFS (HR 0.76) and OS (HR 0.76) was observed (P <.001). PD-L1 expression in pretreatment tumor biopsy samples was correlated with response rate, PFS, and OS However, patients with PD-L1 negative tumors may also achieve durable responses Daud et al, J Clin Oncol, 2016 Pembrolizumab Better Than Ipilimumab in MM Regardless of PD-L1 Expression 834 patients, of which 8 were PD-L1-positive and 18% were PD-L1-negative Pembrolizumab Better Than Ipilimumab Regardless of PD- L1 Expression PD-L1 expression was associated with improved outcomes with pembrolizumab vs ipilimumab Daud et al, J Clin Oncol, 2016 Psychoeducational Intervention to Reduce Fear of Cancer Recurrence in Pts with a Prior MM Pts with a MM hx commonly report a fear of cancer recurrence (FCR) Participants were randomly assigned to intervention (n = 80) or usual care (n = 84) using a newly developed psychoeducational resource with 3 telephone-based psychotherapeutic sessions over a 1 month period timed in accordance with dermatologic appointments. At 6 months, the intervention group reported significantly lower FCR severity, trigger, and distress scores than the control group in the baseline-adjusted models This evidence-based psychoeducational intervention was effective in reducing FCR and stress and increasing melanoma-related knowledge in people concerned about recurrence or new melanoma Dieng et al, J Clin Oncol, 2016 15