Cost-Benefit Analysis of the Pigmented Lesion Assay When Introduced Into the Visual Assessment / Histopathology Pathway for Lesions Clinically Suspicious for Melanoma

The current care pathway for evaluation of pigmented lesions is visual assessment, followed by surgical biopsy and histopathologic assessment (VAH). 1 However, the significant overlap in clinical and histopathologic features that exists between benign and malignant nevi make the classification of pigmented lesions a for clinicians and 3 The VAH relies on image and which is subjective in nature. 1, 3, 4 This, paired with concern regarding potential consequences of missing a melanoma, has resulted in a VAH pathway with a relatively low sensitivity (65% for up to stage pT1a – 84% for all ABSTRACT Objective: To evaluate the potential savings to health plans when the Pigmented Lesion Assay (PLA) is incorporated into the assessment of pigmented lesions clinically suspicious for melanoma. Methods: A Return on Investment (ROI) model was developed from a US payor perspective to determine the per member per month (PMPM) net savings impact of incorporating PLA into the visual assessment/histopathology (VAH) pathway. Using 2019 claims data for patients with lesions suspicious for melanoma (N=239,854), use of PLA in year 1 was modeled and followed through subsequent years. Costs were assessed through the pathway of initial visual assessment, surgical procedure(s), histopathology, and subsequent management. Results: The ROI model predicted annual net savings of $0.54 PMPM for commercial health plans over a three-year period with incorporation of PLA into the VAH pathway. In this analysis, 95.7% of surgically assessed lesions clinically suspicious for melanoma were diagnosed as benign, with 30.4% of patients with benign lesions undergoing a more advanced procedure (e.g., excision), either initially or following a biopsy. Melanoma diagnosis rates associated with biopsy only, excision only, and biopsy followed by excision procedures in the VAH pathway were 0.9%, 0.1%, and 17.9%, respectively. Conclusion: Incorporation of the PLA into the VAH pathway for assessing suspicious pigmented lesions results in savings for commercial health insurance plans. Use of the PLA improves patient care by using genomic assessments to minimize avoidable surgical procedures on benign lesions, enrich the population of melanomas diagnosed, and decrease downstream costs of late-stage melanoma diagnoses.

The current care pathway for evaluation of pigmented lesions is visual assessment, followed by surgical biopsy and histopathologic assessment (VAH). 1 However, the significant overlap in clinical and histopathologic features that exists between benign and malignant nevi make the classification of pigmented lesions a challenge for even highly experienced clinicians and pathologists. 2,3 The VAH pathway relies on image and pattern recognition/interpretation, which is subjective in nature. 1,3,4 This, paired with concern regarding potential consequences of missing a melanoma, has resulted in a VAH pathway with a relatively low sensitivity (65% for up to stage pT1a -84% for all INTRODUCTION melanomas) and a negative predictive value (NPV) of 83%. 3,[5][6][7] These findings demonstrate the need for objective, costeffective technologies to help improve the assessment, classification, and management of pigmented lesions.
The PLA is a noninvasive test that objectively measures genomic markers associated with melanoma within skin tissue samples collected via adhesive patches. [8][9][10] The PLA is used to identify high-risk lesions (severely dysplastic nevi as well as in situ and invasive melanomas) 6,[11][12][13] and guide management to either a) biopsy and histopathologic evaluation or b) clinical surveillance. 6,14,15 Use of the PLA is supported by the National Comprehensive Cancer Network (NCCN) Guidelines, which state that pre-diagnostic noninvasive patch testing may be helpful to guide biopsy decisions (category 2A recommendation). 16 The PLA has a 91% sensitivity for melanoma and an NPV of over 99%; a lesion clinically suspicious for melanoma that tests negative by PLA has a less than 1% probability of being diagnosed histopathologically as a melanoma. 8,17 Previous studies have demonstrated that utilization of the PLA increases accuracy of pigmented lesion classification as either benign or malignant and thus dramatically reduces surgical procedures currently used to rule out melanoma. 14 The current study aims to build on the clinical value of the PLA by investigating the economic value of the PLA in the assessment of suspicious pigmented lesions.
Return-on-investment (ROI) modeling is a method used by actuaries to prioritize risks and determine the most cost-effective option for managing those risks. ROI models developed for healthcare payors estimate the potential incremental financial "investment" (costs) relative to the "return" (cost-offsets/savings) for specific cohorts of health plan members. ROI models incorporate net-cost, member churn, and medical-cost offset scenarios, leveraging proxy client data from claims databases with sensitivity testing of various population and assumption scenarios. In this study, an ROI model was developed to evaluate potential savings for commercial health insurance plans when the PLA is introduced into the VAH pathway to guide management of pigmented skin lesions clinically suspicious for melanoma.
The ROI model was developed from a health plan perspective and designed to assess the per member per month (PMPM) net cost/savings impact of incorporating PLA into the VAH pathway. The model was built from claims data, financial factor inputs, costs associated with the VAH pathway, PLA investment costs, assumptions related to lesion progression in non-detected melanomas, and the number of target members for PLA in a plan. The inputs are modifiable, allowing the user to customize the PMPM net cost/savings impact calculation to their plan's data and extrapolate aggregate year-over-year cost/savings to the plan.

Claims Data and Member Cohorts
Data were sourced from proprietary databases that contain claims data for approximately 13.8 million commercially insured members across the United States who had a plan covering both medical and pharmacy benefits between January 1, 2019 and December 31, 2019. 18  for different levels of costs associated with suspicious skin lesions of various diagnoses/management and estimate cost to the plan, four cohorts of skin lesions clinically suspicious for melanoma were classified based on diagnosis codes and/or procedure codes (Table 1 and Appendices 1-3).

Cost Estimation Methodology
The health plan cost for diagnosis and management of melanoma through the VAH pathway was determined and then compared to a pathway incorporating PLA to help guide patient management decisions. Using trended 2019 claims data, a population using PLA in year 1 was modeled and followed through subsequent years. 18 Assessments were made through the pathway of initial diagnosis of lesions suspicious for melanoma and procedures for biopsy, pathology, and subsequent treatments. Several differences were noted and modeled: 1) The use of the PLA compared to standard biopsy/pathology; 2) The shift in cost to the plan when providers use the PLA to guide patient management decisions based on the genomic results of the PLA instead of performing surgical procedures on pathologically benign lesions; 3) The shift in cost from genomic assessment and treatment of future melanoma, through objective criteria, compared to late-stage cancer treatment in future years.
Future costs 1 to the plan were estimated by using the historic costs for these pathways and trending them forward by historical medical cost trend (4%). Member churn rate (15%), the percentage of members in a plan who will leave the plan each year, and plan cost share percentage (84%) were applied to the model to determine the true cost to the plan. 19 From this, the model aggregated future years and discounted the cost by prevailing interest rates estimated from the 10-year U.S. Treasury T-Bill rate (1%) to arrive at the cost in terms of dollars today. 20 Inputs for cost trend, churn rate, and cost share were based on claims dated between January 1, 2017 and December 31, 2019. 19 Estimating Return-on-Investment Genomic Assessment Savings is a measurement of the cost savings realized due to genomic assessment of pigmented lesions suspicious for melanoma through incorporation of the PLA into the VAH pathway. The PLA identifies lesions at high risk for melanoma based on the presence of genomic markers associated with melanoma. 8,11,13 With histopathology, cellular atypia is detectable, but PLA detects genomic atypia, which cannot be ascertained visually. 11,13 The additional nonadvanced melanoma (managed by biopsy/excision) and advanced melanoma (metastatic and/or managed with advanced treatment) cases included in the calculation are assumed to be melanomas that would not otherwise be detected by the VAH pathway during a particular office visit and were derived using biostatistics from sensitivity analyses comparing VAH and PLA pathways. 3,8,21 The nonadvanced/advanced melanoma costs per patient represent the costs that would be incurred by the plan based on the likelihood that the melanomas missed by the VAH Figure 1. ROI Model Calculations i Excludes melanomas detected via VAH pathway alone. Non-advanced melanoma: managed by biopsy/excision Advanced melanoma: metastatic and/or managed with advanced treatment pathway during an office visit would be identified during the subsequent visit and how quickly the melanomas missed on the initial visit were likely to progress (Table  2). 22,23 Financial factor inputs of cost trend, cost share, and interest rate were also used to calculate nonmelanoma/melanoma costs ( Table 2). 19 Procedure Avoidance Savings is a measurement of the cost savings realized due to not performing biopsies or follow-up treatments on non-malignant lesions as determined by the PLA. The number of patients with benign lesions that underwent biopsy with or without follow-up procedures and associated costs were based on 2019 claims data. A benign lesion follow-up procedure rate of 14.9% was used based on the claims data ( Table 2). 18 PLA Costs is a measurement of investment required for incorporating the PLA into the VAH pathway. The number of patients with lesions that underwent biopsy was based on 2019 claims data. PLA Allowed Cost inputs include the CMS allowable cost per PLA ($760) 24 as well as the average cost of an office visit to evaluate a suspected pigmented lesion ($190) and the average plan cost share (84%) based on 2019 claims data ( Table 2). 18 Figure 1B shows the calculation for plan net PMPM cost savings, calculated as the aggregate Final Savings divided by the Total Health Plan Member Months. Final Savings was calculated as Genomic Assessment Savings + Procedure Avoidance Savings -PLA Costs ( Figure 1A). Total Health Plan Member Months is a direct measurement from experience data, which was derived from the 2019 claims data for the current study (Table 2). 18 The demographics of the study population used for the ROI analysis and for the larger National Claims Database are summarized in Table 3. The study population subset included health plan members with a lesion clinically suspicious for melanoma (N=239,854). Claims associated with other dermatologic conditions (basal cell carcinoma, squamous cell carcinoma, actinic keratoses, warts, etc; see Table 1) were excluded from this analysis. The study population was 57.7% female; 95.7% of members were ≥18 years old. The proportion of lesions in Cohorts 1, 2a, 2b, 3, and 4 were 16.5%, 68.0%, 11.9%, 3.1%, and 0.5%, respectively (Table 4).

Return-on-Investment With PLA
The ROI model predicted the annual net cost/savings for commercial plans incorporating PLA into the VAH pathway. When using default values (Table 2), the ROI model predicted an annual net savings of $0.54 PMPM over a three-year period (aggregate savings of $5.66 million for a plan of 1 million members). Net savings were driven primarily by the increased sensitivity of genomic assessment compared to visual assessment (Genomic Assessment Savings).
Scenario and threshold testing was done to determine the sensitivity of the net PMPM savings at three years to various assumptions used within the Genomic Assessment Savings calculations of the ROI model ( Table 2). The default value for additional melanomas identified by PLA was 26%, which was driven by sensitivity analyses comparing VAH and PLA pathways 3,8,21 ;

Table 2. Variables and Default Values
Variable Name

Cost Trend
The annual trend applied to Medical and Pharmacy costs year-on-year changes 19,20 Commercial: 4%

Interest Rate
Interest rate used for discounting, based on the 10-year U.S. Treasury T-Bill rate 20

Churn Rate
The percentage of members in a plan who will leave the plan each year 19,20 15%

Cost Share
The percentage of plan's cost share 19,20 Commercial: 84%

PLA Cost
Average allowed cost of PLA treatment 24 $760

Suspicious Lesion Office Visit Cost
Cost of an office visit, based on the percentage of plan's cost share, to assess a lesion suspicious for melanoma. 18,20 Commercial: $190

Benign Lesion with No Follow-up Procedure Cost
Direct cost (Medical + Rx) for benign lesion biopsy without follow-up procedure (subset of Cohort 2a) 18 Represents total cost of receiving a biopsy.

Benign Lesion with Advanced Procedure Only with No Followup Procedure Cost
Direct cost (Medical + Rx) for advanced procedure on a benign lesion, without follow-up procedures (subset of Cohort 2a) 18 Represents total cost of receiving an advanced procedure instead of a biopsy.

Benign Lesion with Follow-up Procedure Cost
Direct cost (Medical + Rx) for suspicious/benign lesion biopsy with follow-up procedure(s) (Cohort 2b) 18 Represents total cost of receiving a biopsy and follow-up procedures.

Benign Lesion Follow-up Procedure Rate
The number of patients with benign lesions that underwent biopsy with follow-up procedures (Cohort 2b) divided by the number of patients with benign lesions that underwent biopsy with or without follow-up procedures (subset of Cohort 2a + Cohort 2b) 18 Commercial: 14.9%

Advanced Procedure Without Biopsy Rate
The percentage of cases in which the physician performed a more advanced surgical procedure without an initial biopsy (subset of Cohort 2a

Patient Outcomes
Surgical procedures and diagnostic outcomes were assessed in the study population. Of patients who received a diagnosis of "Neoplasm of uncertain behavior of skin" (D48.5) or "Neoplasm of unspecified behavior of bone, soft tissue, and skin" (D49.2), 83.5% received at least one surgical procedure to rule out melanoma ( Table 4, Cohorts 2-4). Of these surgically assessed lesions, 4.3% were diagnosed as melanoma (number needed to biopsy: 23), and 95.7% were diagnosed as benign. In 15.5% of benign cases, the patient underwent a more advanced procedure (e.g., excision) instead of a biopsy for diagnosis. Another 14.9% of patients received additional advanced   surgical procedure(s) following a benign diagnosis, demonstrating a total advanced procedure rate of 30.4% for benign lesions.
Since the PLA is intended to be used for clinically suspicious pigmented lesions that are not overtly benign or malignant, 8 data from Cohorts 2 and 3 of the study population (Table 1) were used to determine the potential impact in both cost and quality of care that the PLA may provide. Each of these cohorts used one of three procedure categories to reach a diagnosis of benign (Cohort 2) or non-advanced melanoma (Cohort 3): A) biopsy only (N= 135,473); B) excision only (N= 28,086); and C) biopsy followed by excision (N= 34,818). Diagnosis rates of each procedural category are summarized in Table 5.
In category A, when a biopsy was the only procedure performed, 99.1% of patients received a diagnosis of benign (i.e., D22.XX), and 0.9% of patients received a diagnosis of nonadvanced melanoma (i.e., C43.XX, D03.XX). In category B, when an excision was the only procedure performed, 99.9% of patients received a diagnosis of benign, and 0.1% of patients received a diagnosis of nonadvanced melanoma (i.e., C43.XX, D03.XX). In category C, when a biopsy was performed followed by an excision, 82.1% of patients received a diagnosis of benign (i.e., D22), and 17.9% of patients received a diagnosis of nonadvanced melanoma (i.e., C43.XX, D03.XX). Categories B and C represent situations when the treating clinician decides to excise the lesion. When combining instances where a lesion is excised, 90.0% of patients receive a diagnosis of benign (i.e., D22), and 10.0% of patients receive a diagnosis of nonadvanced melanoma (i.e., C43.XX, D03.XX).  5 In agreement with these findings, the ROI model in this study indicates that incorporating the PLA into the VAH pathway can provide cost savings to commercial health insurance plans. This study also indicates where the PLA can improve outcomes of patients with lesions suspicious for melanoma.
The ROI model showed that adoption of PLA may result in savings of $0.54 PMPM over a three-year period. These cost savings were realized due to the use of genomic assessments to guide biopsy decisions for pigmented lesions clinically suspicious for melanoma. The performance metrics of the PLA (91% sensitivity, 99% NPV) improve upon those of the VAH pathway (65%-84% sensitivity, 83% NPV), 3, 5-8 resulting in more accurate classification of pigmented lesions based on genomics. Genomic aberrations precede visual changes in melanoma; 25 thus genomic assessments with the PLA are able to identify high-risk lesions for melanoma that may not otherwise be evaluated by the VAH pathway. 6,11,13 Using the PLA's genomic assessments to guide biopsy decisions could reduce overall treatment costs associated with VAH while improving patient outcomes.
In addition to cost savings, our study uncovered opportunities to improve upon the quality of care offered via the VAH pathway. Traditionally, clinicians have had to rely on image and pattern recognition to classify pigmented lesions as benign or malignant, which is a subjective and challenging exercise. [1][2][3][4] With the goal of never missing a melanoma, clinicians have exercised an abundance of caution when assessing pigmented lesions. For example, 83.5% of clinically suspicious pigmented lesions were surgically biopsied in this study, with 95.7% of those diagnosed as benign. Furthermore, the three procedural assessment pathways employed (biopsy only, excision only, and biopsy followed by excision) all yielded low rates of melanoma diagnoses (0.9%, 0.1%, and 17.9%). Though the approaches and outcomes above are understandable given the tools historically available, these data demonstrate the need for additional, reliable information that can more accurately classify pigmented lesions for enhanced decision making and improved outcomes.
NCCN Guidelines indicate that prediagnostic noninvasive patch testing may be helpful to guide biopsy decisions. 16 The PLA aligns with this guidance by noninvasively providing objective genomic information that indicates if a suspicious pigmented lesion is at high risk for melanoma. 8,10,11,13 Improved outcomes for patients occur when they receive appropriate care for their condition, receive more effective care, and/or avoid unnecessary procedures/therapy or testing. 26 With the PLA, approximately 15%  of  lesions  test  positive  and  are  recommended  for  biopsy  and  histopathologic evaluation, and the remaining 85% test negative and are recommended for clinical surveillance, thus avoiding a surgical procedure. 27 A separate analysis of PLA claims from 2020 independently validated these outcomes:

DISCUSSION
~20% of claims received a surgical procedure (i.e., presumed positive PLA result) and ~80% of claims did not receive a surgical procedure (i.e., presumed negative PLA result). 28 A recent long-term follow-up study demonstrated that clinical surveillance of PLA-negative lesions is appropriate; in this study, <1% of 1,535 PLA-negative lesions exhibited clinical changes leading to melanoma diagnoses (in situ or pT1a) during 24-36 months of follow-up. 17 Taken together, these data show that the PLA can improve patient outcomes by 1) identifying the small subset of clinically suspicious lesions at high risk for melanoma, and 2) avoiding surgical procedures of benign lesions.
Incorporation of the PLA into the VAH pathway enhances decision making for lesions clinically suspicious for melanoma, improves the quality of patient care, and drives lower overall costs to commercial health insurance plans.