Abstract P2-06-16: Addressing Thermal and Battery Efficiency in AI Enhanced Portable Ultrasound Screening Protocols for Breast Cancer
Bibliographic record
Abstract
Abstract Background: Breast cancer remains a significant cause of mortality in women, especially in rural and underserved communities where access to mammography is limited or nonexistent. The incidence of advanced-stage breast cancer is 66% higher in Hawaii than in the mainland US, and 5 – 9 times higher in the U.S. Affiliated Pacific Islands (USAPI). AI-enhanced point-of-care ultrasound (POCUS) breast imaging may be an effective method for detecting breast cancer while still in the early stages. The hypothesis is that AI detection and classification algorithms will increase the accuracy of POCUS such that it would approach that of mammography. Further, it may reduce the training levels needed to do the early detection POCUS scanning. One such candidate device is portable, wireless, and provides an SDK for inserting AI models before presenting the image to the user. We have been exploring a suitable protocol for using this POCUS device in remote conditions where infrastructure may be limited. Further, we asked if the portable battery-operated scanner could keep up with a scanning rate of 2 patients per hour for an indefinite amount of time. In this study, we seek to identify the performance parameters that need to be considered to use this and other POCUS systems in rural, remote, and underserved communities. Methods: The POCUS system consists of a portable handheld scanner (Clarius L7 HD3 model; Clarius Mobile Health Inc, Vancouver, Canada), a tablet computer (Samsung Galaxy Tab S9 Ultra; Samsung Electronics Co., Ltd, South Korea), and a laptop (Dell XPS 15 running Windows 11; Dell Inc., Texas, US). The scanner communicates with the tablet and laptop via an ad hoc WIFI connection. A breast phantom (Gphantom, EDM Medical Solutions, Florida, US) was scanned continuously for 10 minutes, followed by a 20-minute charging period, and repeated. Pretrained AI models were inserted into the image stream using the Cast API. Scanner temperature and battery levels were monitored to determine their time characteristics utilizing the Clarius app. Longer scan and charging periods were used as well to capture the full extent of heating, cooling, and charging cycles. Probe temperature and charging/discharging characteristics were fit to exponential functions.Results: The system was able to operate for 24 minutes continuously before hitting a thermal protection temperature of 48°C. The probe was able to fully recharge from 0 to 100% within 60 minutes. For the 2-patient-per-hour protocol, the probe was able to regain its starting charge and temperature for back to back phantom scans over 4 hours. Time constants were found to be 16.1 minutes for temperature increase and 49.6 minutes for battery discharge. The user found it difficult to hold the probe for temperatures above 42°C. A shorter time between scans may not be feasible but being tested.Conclusion: The findings show that the device can be used continuously for at least two patients per hour with 10 minutes of continuous scanning followed by 20 minutes of charging. However, the operating temperature of the probe is quite high and it may be difficult to handle. Healthcare workers can optimize usage protocol to maximize functional time without frequent recharging in remote areas. Future work will focus on refining AI models for real-time applications and exploring alternative devices with better thermal performance. Citation Format: Nusrat Zaman Zemi, Dustin Valdez, Arianna Bunnell, John Shepherd. Addressing Thermal and Battery Efficiency in AI Enhanced Portable Ultrasound Screening Protocols for Breast Cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P2-06-16.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".