Incorporating DecisionDx-Melanoma into Clinical Practice: Molecular Information for Personalized Management Decisions Eric D. Whitman, MD, FACS Medical Director, Atlantic Health System Cancer Care Director, Atlantic Melanoma Center Morristown and Summit, New Jersey
Castle Biosciences provides molecular diagnostic solutions for patients with underserved and dermatologic cancers Cutaneous Melanoma Development Programs Uveal Melanoma Informing clinical decisions Proven impact on management Standard of care decisions for 1 of 2 patients tested Active research and development programs in skin cancers ( e.g. cSCC) and other underserved 31-GEP test accurately identifies Gene 15-GEP test accurately identifies cancers risk of recurrence independent of Expression aggressive tumor biology traditional staging factors Profiling Development of biomarkers for and informs SLNB (GEP) treatment response prediction 12,000+ decisions Adds further prognostic accuracy melanoma test orders for UM patients in 2018 Sequencing capabilities to Next enhance precision medicine Generation NGS test identifies actionable NGS panel offers comprehensive Sequencing BRAF , NRAS and KIT mutations in genomic profiling (NGS) melanoma tumors Tests performed in CAP-accredited, CLIA-approved laboratory
Individual risk drives management decisions for melanoma patients – including informing SLNB Management of Stage I ‐ III melanoma patients What’s the risk? involves making decisions on: • Frequency of follow ‐ up • Frequency and modality of surveillance imaging • Referral to Surgical Oncology • Sentinel Lymph Node Biopsy Sentinel Lymph Node Biopsy • Referral to Medical Oncology • Adjuvant therapy NCCN guidelines recognize that a patient’s individual risk of recurrence should drive management decisions NCCN guidelines also recognize that a patient’s individual risk of SLN positivity drives SLN biopsy recommendations
Staging-only approach misses patients with aggressive tumor biology • While AJCC clinicopathologic factors are helpful clinically, the majority of deaths occur in early stage disease 1-5 Melanoma Deaths • Prognostic accuracy needs to be Stage at Diagnosis by Stage at Diagnosis improved as it has direct implications on patient follow up Stage I • Newer therapies and regional Stage I Stage II interventions have shown 34% 80% 26% effectiveness in the adjuvant setting Stage II 12% • Early intervention is consistently shown to be a significant (or the 8% most significant) predictor of Stage III Stage III 40% response 6 Excludes Stage IV Excludes Stage IV 1 AJCC v7 J Clin Oncol 2009 ; 2 SEER data release 2017; 3 Morton et al. N Engl J Med 2014; 4 Whiteman et al. J Invest Dermatol 2015; 5 Shaikh et al. J Natl Cancer Inst 2016; 6 Poklepovic and Carvajal. ONCOLOGY 2018
DecisionDx-Melanoma was developed to assess risk of recurrence independent from traditional clinico-pathologic factors using tumor biology Class 1: 1A Lowest risk Quantifies expression Low risk of melanoma recurrence within 5 of 31 genes from 1B Patients years primary tumor using RT- Low risk with PCR Stage I-III Applies a validated melanoma Class 2: 2A algorithm Increased risk High risk of melanoma Accurately classifies recurrence within 5 2B patients as low or high years Highest risk risk Gerami et al. Clin Cancer Res 2015; Gerami et al. JAAD 2015; Zager et al. BMC Cancer 2018; Gastman et al. JAAD 2019
DecisionDx-Melanoma test in clinical use: T1a / Stage IA melanoma patient case • 34 ‐ year ‐ old female with melanoma on arm • Breslow depth: 0.6mm, no ulceration (T1a) , mitotic rate <1/mm 2 • AJCC Stage IA • DecisionDx ‐ Melanoma ordered by Dermatology: • Class 2B result Clinical management of patient with Class 2B tumor • Referred to Medical Oncologist for high ‐ intensity surveillance • CT scan ‐ clear • CT scan six months later ‐ biopsy proven oligomet to the lung, BRAF negative • Radiotherapy to lung metastasis • Started on combination ipilimumab/nivolumab, and doing well (clear scans) after 36 months Case contributed by Medical Oncologist, Topeka, KS
DecisionDx-Melanoma validation, performance and clinical impact studies include >2,900 patients across 19 peer-reviewed publications Validation & Validation & Prospective Studies Clinical Impact Studies Performance Studies Performance Studies • Vetto et al. Future Oncol 2019 - • Keller et al. Cancer Med 2019 - • Mirsky et al. J Drugs Dermatol 2018 - NEW NEW NEW • Dillon et al. SKIN 2018 • Gastman et al. JAAD 2019 - NEW • Podlipnik et al. JEADV 2019 - • Schuitevoerder et al. J Drugs NEW • Gastman et al. Head & Neck Dermatol 2018 2019 - NEW • Greenhaw et al. Dermatol Surg • Hyams et al. WCD 2018 2018 • Zager et al. BMC Cancer • Svoboda et al. J Drugs Dermatol 2018 • Hsueh et al. J Hematol Oncol 2018 2017 • Gerami et al. CCR 2015 • Farberg et al. J Drugs Dermatol 2017 • Renzetti et al. SSO 2017 • Gerami et al. JAAD 2015 • Berger et al. CMRO 2016 • Cook et al. Diagn Pathol 2018 Newly validated clinical utility Prospective studies Management decisions impacted to inform SLNB discussions confirm validity for 1 in 2 patients tested
Combined validation cohort demonstrates accurate, independent prognostic value of DecisionDx-Melanoma test to identify patients’ risk for metastasis • DecisionDx ‐ Melanoma test was clinically validated to predict risk of recurrence in three published studies: • Gerami et al. CCR 2015 • Gerami et al. JAAD 2015 • Zager et al. BMC Cancer 2018 • Gastman et al. JAAD 2019 publication combines unique patients from these previous validation studies to enable analysis of clinically relevant subgroups: • SLNB negative • Stage I ‐ IIA • Thin tumors ( ≤ 1 mm) Gastman et al. JAAD 2019
Cumulative validation cohort for DecisionDx-Melanoma includes 690 evaluable patients that met inclusion criteria • Attribute Summary IRB approved protocol to collect Stage I ‐ III melanoma with recurrence/metastasis or Age (years) more than 5 years of f/u without recurrence Median (range) 59 (18 – 94) Breslow depth (mm) • Eighteen U.S. dermatologic and surgical Median (range) 1.3 (0.1-29) centers Ulceration present 190 (32%) • Mitotic rate ≥ 1/mm 2 352 (74%) Diagnosed between 1998 and 2014 • Enrollment began in 2011 with data Node status positive 200 (29%) updates in Oct 2017 Stage I 333 (48%) • 782 patients were enrolled in the studies II 150 (22%) • 690 met the inclusion criteria and were III 207* (30%) evaluated GEP Class 2 298 (43%) • 483 of those were Stage I/II * Seven cases had Stage III, SLN negative disease All non-metastatic cases with ≥ 5 years follow up • 57% of patients were low risk (Class 1); 43% were high risk (Class 2) Gastman et al. JAAD 2019
Clinical validity and prognostic value of DecisionDx-Melanoma have been demonstrated in 690 Stage I-III patients RFS DMFS MSS % Distant Metastasis Free % Recurrence Free % Survival p<0.0001 p<0.0001 p<0.0001 Time (Years) Time (Years) Time (Years) Class 1A Class 1B Class 2A Class 2B n=80 n=312 n=84 n=214 Median time to distant recurrence for patients with Class 2 result: • Stage I ‐ II patients: 1.6 years (70 Class 2 patients recurred) • Stage III patients: 1.0 year (92 Class 2 patients recurred) Gastman et al. JAAD 2019
DecisionDx-Melanoma test further informs risk obtained by AJCC 8 th edition staging Stage NCCN III I II Risk Category 100 99.6% >99% 94.8% Low Risk 98% ≈ IA Stage I ‐ IIA ≈ IA ≈ IIA Melanoma-Specific Survival (%) 89.5% 90% 90 ≈ IIIA 84.7% ≈ IIIB 80 High Risk 77% Stage IIB ‐ III AJCC MSS 70 Class 1A Class 2B 61.2% ≈ IIIC+ 60 N=690 Prado et al. SKIN J Cutan Med 2018:suppl 2
Meta-analysis objective and methodology • Objective: To evaluate the cumulative prognostic effect of the DecisionDx ‐ Melanoma test across multiple cohorts, with a focus on risk difference between Class 1A and 2B Multivariate analysis Meta ‐ analysis DecisionDx ‐ Cohort 1: Greenhaw et al. 2018 DDx ‐ Melanoma Comparisons across Melanoma effect result Cohort 2: Hsueh et al. 2017 and within studies (Class 1A vs. Class 2B) (Class 1A, 1B, 2A, 2B) and tests for RFS Cohort 3: Gastman et al. 2019 Breslow thickness heterogeneity DMFS Ulceration Cohort 4: Novel cohort Node status Age 1,479 Stage I ‐ III patients Greenhaw et al. Presented at AAD, Late ‐ Breaking Abstract 2019
Meta-analysis across four independent studies: DecisionDx-Melanoma class and correlated survival outcomes (n=1,479) Class 1A Class 1B Class 2A Class 2B GEP Result 5-year RFS Event Rate GEP Result 5-year DMFS Event Rate 1A (n=790) 91.4% 6.7% 1A (n=597) 94.1% 5.5% 1B (n=169) 85.1% 14.2% 1B (n=148) 88.1% 12.2% 35.8% 24.1% 2A (n=159) 64.0% 2A (n=133) 75.9% 2B (n=361) 43.6% 50.1% 2B (n=345) 55.5% 38.8% Greenhaw et al. Presented at AAD, Late ‐ Breaking Abstract 2019
Recommend
More recommend
