Automated Placental ROI Selection for Quantitative Ultrasound Analysis Using Boundary-Aware Thresholding
Bibliographic record
Abstract
Quantitative ultrasound (QUS) analysis of placental tissue holds significant promise for evaluating pregnancy health and detecting placental pathologies. However, widespread clinical adoption of QUS is limited by the need for manual selection of a proper region of interest (ROI) for the calculation, which increases the analysis time and introduces operator variability. This study presents an automated ROI determination method that uses deep learning segmentation directly on minimally processed raw radio frequency (RF) data (along with corresponding B-mode data for comparison) to replace manual ROI selection in placental QUS analysis while also taking advantage of the model's internal confidence thresholds to reduce inclusion of nonplacental tissue near ambiguous tissue boundaries. Using a diverse, multinational dataset of 1084 annotated ultrasound images, the proposed automated ROI model achieved a mean dice similarity coefficient (DSC) above 0.8 for confidence thresholds between 50% and 75%. At thresholds above 75%, precision improved, but both DSC and recall performance declined, and the amount of placental tissue identified for an accurate QUS calculation fell below the minimum required ROI size. Bland-Altman analysis showed tight agreement between the QUS values derived from ground-truth ROIs and automated ROIs up to a 75% confidence threshold, but reduced agreement at higher thresholds consistent with reducing the placental pixel area for below the minimum. These findings demonstrate both the feasibility of automated ROI determination and the need for careful consideration of the confidence thresholds when assessing model performance.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".