Abstract PO1-07-05: Preliminary Evaluation in Breast Cancer of PointCheck: a Noninvasive AI-Based Neutropenia Monitoring Device

Abstract Background Febrile neutropenia (FN) is a common complication that can arise in breast cancer patients receiving myelosuppressive chemotherapy and increases the risk of fatal infections. Despite better preventive strategies there are between 140,000 to 200,000 FN-related hospitalizations yearly. Quick detection and awareness of grade III neutropenia (Absolute Neutrophil Count [ANC] < 1,000/µL) or greater can be critical to managing FN and treatment scheduling and dosing in patients with breast cancer. The current gold standard for neutropenia detection is a peripheral blood collection that requires patients to come into the clinic. PointCheck™ is a novel, noninvasive technology that can monitor for grade III neutropenia or greater and enable prompt detection in the home by acquiring microscopy videos of superficial capillaries through the nailfold skin and analyzing those videos with computer vision AI algorithms. To evaluate the usability, diagnostic performance, and preliminary clinical utility of PointCheck™ on FN detection we conducted a multi-center observational usability and diagnostic study. Methods The present study included an overall cohort of 175 diverse cancer patients, with 70 breast cancer patients. The primary endpoint was to achieve a score over 80.8 on a standardized System Usability Scale (SUS). Secondary endpoint was achievement of diagnostic performance with an area under the curve (AUC) above 0.80. We also performed an exploratory analysis on the device's clinical utility. Eligible participants were introduced to the device, watched a tutorial video and were given a user manual. They used the device autonomously under the supervision of the clinical team. Usability data was collected using the SUS, the scoring system that has a scale between 0-100. PointCheck™ measurements were analyzed using AI algorithms and compared to same-day Complete Blood Counts (CBCs) collected within 90 minutes of the measurement to assess accuracy in classifying patients as grade III neutropenic or greater (< 1,000/µL), or non-grade III neutropenic (≥1,000/µL). In the exploratory clinical utility analysis, clinicians provided narrative insights by completing a Likert survey assessing how PointCheck™ would support or impact their clinical decisions. Results 81.4% of breast cancer patients scored above 80.8 on the SUS scale across all sites, with a mean SUS score of 87.7 (SD=13.2). Furthermore, the AI-based PointCheck™ classifier accurately discriminated neutropenia patients with an AUC = 0.90 in the overall cohort. In the exploratory analysis, a majority of clinicians (65.2%) agreed that the device could have helped better evaluate their patients (Table 1). Whereas almost 3 out of 4 clinicians (72.4%), described that the use of PointCheck™ could have avoided preparing a chemotherapy infusion if the patient was not ready due to grade III neutropenia or greater. Conclusions The present study showed that there is a high perception of usability of PointCheck™, indicating an above average user experience and falling within the top 10% of systems. Furthermore, we found that PointCheck™ can accurately detect grade III neutropenia or greater in a large cohort of patients including breast cancer patients. Application of PointCheck™ to clinical practice as a monitoring system can contribute to the early detection of FN and antineoplastic therapy management. Table 1: Clinical Utility Survey Results Citation Format: Juan Garcia-Mosquera, Manuel Alva, Toshiaki Iwase, Ganimete Lamaj, Pablo Tolosa, Rodrigo Sánchez-Bayona, José Pelaez-Escobar, Alberto Pablo-Trinidad, Ryan Benasutti, Aurélien Bourquard, Ian Butterworth, Alvaro Sánchez-Ferro, Carlos Castro-González, José Manuel Pérez-García, Alejandro Martínez-Bueno, Eva Ciruelos, Azadeh Nasrazadani. Preliminary Evaluation in Breast Cancer of PointCheck: a Noninvasive AI-Based Neutropenia Monitoring Device [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-07-05.