Evidence review of the prognostic 40-gene expression profile test for cutaneous squamous cell carcinoma Methods Disclosures References › This study was sponsored by Castle Biosciences, Inc. › All authors are employees and shareholders of Castle Biosciences, Inc. 1. Jambusaria-Pahlajani, et al. Arch Derm 2010 2. Farberg, et al. Dermatol Clinics 2017 3. Colomer et al. Clin Transl Oncol 2018 4. Plasseraud et al. J Oncol. 2016 5. Teplitz, et al JDD 2019 6. Litchman, et al CMRO 2020 7. Farberg, et al. in preparation 8. Wysong et al., JAAD 2021 9. Ibrahim, et al Future Oncology 2021 10. Borman et al. Diagn Pathol under review 11. Farberg et al. CMRO 2020 12. Aaron et al. JDD 2021 Presented at Winter Clinical Dermatology Conference; January 14-19, 2022. Matthew S. Goldberg1, Sarah J. Kurley1, Jennifer J. Siegel1, Alison L. Fitzgerald1, Robert W. Cook1 1Castle Biosciences, Inc. For more information: mgoldberg@castlebiosciences.com Objective › To describe the performance of the 40-GEP test as a method of accurate assessment of a patient’s risk for metastasis after diagnosis of cSCC with one or more risk factors. ›The data across the three critical pillars of molecular testing demonstrate the robustness, accuracy and utility of the 40-GEP test. ›Clinical utility data illustrates that physicians understand 40-GEP test results and how to appropriately integrate these results into their clinical considerations for treatment of cSCC patients with one or more risk factors, ideally leading to a more personalized treatment pathway. ›Clinical validity data supports the use of the test as an adjunct to current risk assessment to better evaluate a patient’s metastatic risk. ›Analytical validity data exhibited robust technical reliability of the 40-GEP on clinical samples along with high concordance rates across multiple performance experiments. Conclusions › Previously published and unpublished clinical utility data addressing the impact of the 40-GEP on patient management plans were summarized.5-7 › Previously published clinical validation data from independent cases with verified clinicopathologic information and known outcomes were assessed by Kaplan-Meier survival analysis and Cox regression analysis.8,9 › Analytical validation of the performance of the 40-GEP test included precision experiments to assess inter-assay and intra- assay reliability and the assessment of its technical success rate.10 Synopsis › High-risk cutaneous squamous cell carcinoma (cSCC) is a subset of cSCC commonly requiring a more aggressive treatment regimen, due to an increased probability of recurrence, nodal/distant metastasis, or disease specific death. › Guidelines and staging criteria for cSCC are overall vague, creating a burden for clinicians when establishing an appropriate treatment plan. 1 › In many cancer types, molecular prognostics have had a significant and appropriate impact in patient care.2,3,4 › The purpose for the development of the prognostic 40-GEP test was to identify high-risk cSCC early in the disease state, such that its result could complement current risk assessment methods for development of more personalized management plans to reduce the risk of poor outcomes for cSCC patients. Results › Three separate clinical impact surveys were distributed to dermatologic clinicians (n=598).5-7 Table 1 presents the results of physician responses to “no 40-GEP” and post-40-GEP test results regarding management changes. Responses for all surveys demonstrated that the prognostic information garnered through the 40-GEP could aid in cSCC patient management in a risk appropriate manner. A representative patient vignette highlights that 97.5% of clinicians would change patient management intensity recommendations based on the 40-GEP with reduced intensity for Class 1 and increased intensity for Class 2B post-test management decisions (Figure 1). › Figure 2 demonstrates the ability of the 40-GEP test to classify patients based on risk of metastasis.8,9 › Regardless of the specific risk factor or clinicopathologic risk assessment method included in the multivariable regression analysis, the 40-GEP demonstrated independent and statistically significant prognostic value with hazard ratios (HR) for Class 2A and 2B similar to or beyond that of clinicopathologic factor-based systems. (Table 2). › Reliability of the 40-GEP test for class call assignments was verified by inter- and intra-assay concordance of 93% (n=27/29) and 98% (n=45/46), respectively.10 Over the duration of one year, 98% of all clinically tested samples with sufficient tumor content gave actionable Class call outcomes, highlighting the low multi-gene failure rate of the test. (Table 3).10 Figure 2. The 40-GEP accurately classifies patients by metastatic risk Figure 1. Example of 40-GEP impact on overall management strategy intensity6 Multivariate Cox Regression Risk Factor n Hazard Ratio p value 40-GEP Result Class 1 212 1.00 --- Class 2A 185 2.33 0.013 Class 2B 23 6.86 <0.001 Clinicopathologic Risk Factors Poor Differentiation 58 2.29 0.011 Perineural Invasion 53 1.22 ns Deep Invasion 72 2.05 0.039 Tumor Diameter N/A 1.07 ns 40-GEP Result Class 1 212 1.00 --- Class 2A 185 2.92 <0.001 Class 2B 23 9.50 <0.001 NCCN Risk Group High 255 1.00 --- Very High 165 1.99 0.009 Table 2. The 40-GEP provides independent prognostic value to existing risk assessment methods Cases were comprehensively staged based on medical records, pathology reports, and definitive surgical reports. All cases were high-risk by NCCN guidelines for localized cSCC or met Mohs micrographic surgery appropriate use criteria. Scan here for more info Analytic Validity10Clinical Validity9 Clinical Utility5-7 Multivariate Cox Regression Risk Factor n Hazard Ratio p value 40-GEP Result Class 1 212 1.00 --- Class 2A 185 2.97 <0.001 Class 2B 23 11.4 <0.001 AJCC8 T Stage T1/T2 340 1.00 --- T3/T4 80 2.69 <0.001 40-GEP Result Class 1 212 1.00 --- Class 2A 185 2.98 <0.001 Class 2B 23 9.42 <0.001 BWH T Stage T1/T2a 364 1.00 --- T2b/T3 56 2.38 0.002 Intra-assay concordance 98% Inter-assay concordance 93% Sample longevity and stability 96% Overall technical success 98% Table 3. The 40-GEP shows robust repeatability and reproducibility10 Clinical Impact Studies of 40-GEP5-7 Clinicians Patients Specific clinical recommendation changed with 40-GEP Overall change in management plan recommended with 40-GEP 34 real-world test users 6 real-world cases F/U, SLNB, baseline nodal imaging, adjuvant radiation, adjuvant chemotherapy, surveillance imaging Integration of the 40-GEP Class call significantly impacted recommended patient management plans in a risk- appropriate manner while staying within guidelines. 162 dermatologists* 2 patient vignettes F/U, SLNB, nodal imaging, adjuvant radiation, adjuvant chemotherapy 402 dermatologists 3 patient vignettes F/U, SLNB referral, radiation, chemotherapy, immunotherapy F/U = follow up schedule; SLNB = sentinel lymph node biology; *Majority dermatologists with 8.6% dermatology NP/PA, 1.2% dermatopathologist, 1.9% other Table 1. 40-GEP testing consistently changes management recommendations across 3 clinical impact studies 67-year-old male 1.2 cm scalp lesion with poor differentiation, 1.2 mm depth (2 high-risk clinicopathologic factors) NCCN High-Risk cSCC ・ AJCC T1 ・ BWH T2a (Adapted from Litchman et al.) Publications on proposed incorporation of 40-GEP testing: › Cross-specialty expert panel reports decision-making points where 40-GEP testing could inform clinical management12 › Proposed risk-aligned incorporation of 40-GEP testing into management strategies within NCCN guidelines11 The vast majority of clinicians would change management intensity for this patient based on the 40-GEP Class result Reduced clinical management intensity Increased clinical management intensity 40-GEP impact by clinician Class 2A 97.5% 59.9% Class 2B 93.2% Class 1 59.9% 45.1% Overall Change Overall Change Neutral Increase Neutral Decrease (Adapted from Ibrahim et al.) https://castlebiosciences.com/research-development/publications/