PRecision IMaging to Evaluate Kaposi Sarcoma (PRIME-KS)

ABSTRACT Monitoring treatment response for Kaposi sarcoma (KS) using present clinical judgement is inadequate. While objective monitoring of KS lesions can be performed, the measurements required are tedious and time- consuming, making them impractical for use in routine clinical care in busy low- and middle-income country

(LMIC) clinics. This leads clinicians in LMIC to rely on subjective monitoring strategies, which may unnecessarily prolong treatment and incur unnecessary chemotherapy toxicities. KS remains a major public health threat: it is the most common cancer and a leading cause of cancer-related death among men in Malawi, Mozambique,

Uganda, and Kenya.1-4 Current standard of care for monitoring KS treatment response is also imprecise. These methods are based on clinical assessment of lesion area and height and are subject to human variation. They perform particularly poorly in persons with dark skin, leading to racial inequities in adverse effects including

alopecia, cardiac toxicity, neuropathy, bone marrow suppression, financial toxicity, decreased quality of life, and premature treatment discontinuation, undermining effectiveness.7,8 To address these challenges, we propose to test a new technology — SkinScan3D (SS3D) — to elucidate its contribution in clinical practice, optimize

performance, and integrate into workflows in Africa where need is greatest. SS3D is a simple, low-cost, and user-friendly technology that combines liquid lens technology and artificial intelligence (AI), providing high- resolution 3D images/maps of KS skin lesions that incorporate lesion height and volume. We will formally refine

the technology to make it user-friendly and easy to integrate into workflows in Uganda and Kenya, determine its measurement improvements in real-world clinical practice, and assess its contribution to improved clinical outcomes. Our specific aims are: Aim 1: Refine SkinScan3D Technology and Usage Protocols for Routine

KS Imaging in East Africa. We will engage end-users (clinicians, nurses), and patients to refine SS3D through: a) focus group discussions; b) discrete choice experiments to identify and quantify preferences for re-design of hardware, software, and standard operating procedures; and c) iterative Human-Centered Design workshops.

Aim 2: Compare Reproducibility and Accuracy of KS Lesion Size Measurements Between SkinScan3D and Current Standard of Care Measurement. We will compare the concordance correlation coefficients (quantitative measure of reproducibility) and coefficients of determination (quantitative measure of accuracy)

between SkinScan3D and the manual method (ruler-based). Aim 3: Validate and Optimize SkinScan3D Clinical Workflow in Real-World Settings. We will roll out the refined SS3D package to a variety of practice locations in East Africa to determine whether the device will perform robustly and efficiently in different settings.

Specifically, we will measure: device usability with the System Usability Scale; device acceptability, appropriateness and feasibility with validated quantitative instruments; clinic workflows and time burden with time-and-motion studies; and implementation costs with activity-based micro-costing.