Organizations: Macao Polytechnic University, Faculty of Applied Sciences, Macau, China · The Netherlands Cancer Institute, Department of Radiology, Amsterdam, 1066 CX, The Netherlands · Radboud University Medical Center, Department of Medical Imaging, Nijmegen, 6525 GA, The Netherlands · Shenzhen University, Medical Ultrasound Image Computing (MUSIC) Lab, Shenzhen, 518060, China · Chinese Academy of Sciences, Hong Kong Institute of Science & Innovation, Centre for Artificial Intelligence and Robotics, Hong Kong, China · University Medical Center Utrecht, Imaging Division, Utrecht, 3584 CX, The Netherlands · Westlake University, School of Medicine, Affiliated Hangzhou First People’s Hospital, Department of Ultrasound, Hangzhou, 310006, China · Xidian University, State Key Laboratory of ISN, Xi’an, 710071, China · Shenzhen University, School of Artificial Intelligence, Shenzhen, 518060, China · Nanjing Medical University, School of Biomedical Engineering and Informatics, Nanjing, 211166, China · Zhejiang Cancer Hospital, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Center of Interventional Oncology, Hangzhou, 310022, China · Zhejiang Cancer Hospital, Hangzhou Institute of Medicine (HIM), Chinese Academy of Sciences, Department of Diagnostic Ultrasound Imaging & Interventional Therapy, Hangzhou, 310022, China
Medical ultrasound (US) enables real-time, non-invasive diagnosis, yet building US-based AI systems is limited by data scarcity, costly expert annotation, and limited cross-center generalizability. Privacy constraints and instance-level reporting hinder the collection of diverse abnormal cases and frame-level labels, while repeated retraining for new centers increases computational cost. We introduce EvoDiff, a generative active learning framework that integrates diffusion modelling with information-theoretic objectives, evolutionary search, and curriculum-based sampling to select informative samples across heterogeneous US datasets. EvoDiff formulates sample selection as a generate-to-retrieve process, in which diffusion-based optimization in feature space guided by information gain produces synthetic probes, and only the real unlabeled samples retrieved by these probes are annotated and used for downstream training. Across six benchmarks covering fifteen datasets and 16,696 images, we evaluate EvoDiff on convolutional network and transformer-based foundation model. EvoDiff approaches fully supervised performance with substantially fewer labels, outperforms active learning baselines, and improves foundation model generalization under cross-domain external validation on unseen cohorts. A reader study with six radiologists indicates the potential to support clinical annotation workflows, with a model-based estimated 24% to 84% reduction in annotation workload at matched downstream performance.
Figures & tables
Figure 1: Overview of the EvoDiff pipeline. a) Organ-conditioned coarse pretraining of the diffusion model and the guidance classifier on multi-organ, organ-identified but pathology-unlabeled ultrasound images, followed by pathology-level fine-tuning on the limited labeled subset. A flame symbol indicates that the parameters of a network are updated in that stage, and a snowflake symbol indicates that the network is frozen. b) Information-gain (IG)-guided sampling combined with the Local Optimization Evolutionary Algorithm (LOEA), which performs gradient-free evolutionary refinement along the frozen diffusion trajectory and uses entropy-based IG signals to steer latents toward informative regions. The triangular maps show the predictive-entropy landscape of the guidance classifier before and after LOEA refinement, coloured from dark blue for low entropy to yellow for high entropy. As the guidance strength increases, generation moves from certain samples, drawn as green disks, to uncertain samples, drawn as orange disks, as indicated by the green-to-orange arrow. The generated probes and the unlabeled candidate pool then undergo feature extraction and cosine-similarity ranking, and the expert oracle, shown in the red dashed box, annotates the top α % most informative real samples. The generated probes are never annotated and are never used as downstream training samples. c) Newly annotated data are merged with the existing labeled data to fine-tune the downstream models, a ResNet-50 and the ultrasound foundation model (USFM), and each subsequent active learning (AL) iteration increases the uncertainty of the training set. After all AL iterations, the downstream models are validated on 15 datasets, 8 internal and 7 external, spanning 6 diagnostic tasks.
Figure 2: Performance comparison of six active learning methods across annotation rounds on eight ultrasound image datasets. Results are reported as the area under the receiver operating characteristic curve (AUC, %) of the downstream ResNet-50 classifier. Round 0 represents the baseline performance of the initialized ResNet-50, and rounds 1 to 6 correspond to cumulative annotation budgets of 5%, 10%, 20%, 30%, 40%, and 50% of the unlabeled candidate pool. a) Appendicitis diagnosis on Appendix. b) Breast tumor diagnosis on BUSI. c) Breast tumor diagnosis on BUS-BRA. d) Breast tumor diagnosis on UDIAT. e) Cardiovascular event prediction on CUBS. f) Fatty liver diagnosis on Fatty-Liver. g) Ovarian tumor assessment on MMOTU. h) Thyroid nodule diagnosis on TN3K. In every panel, the blue dashed line denotes Random, the orange dashed line denotes VAAL, the green dashed line denotes Core-Set, the red dotted line denotes Max-Entropy, the purple dash-dotted line denotes BALD, and the brown dashed line denotes EvoDiff, each with circular markers at the evaluated rounds. The grey horizontal dashed line marks the fully supervised upper bound trained on the complete labeled set. The shaded band around each curve, drawn in the same colour as the curve, indicates the s.d. (standard deviation) across the five cross-validation folds.
Figure 3: Validation of USFM performance comparing random sampling with EvoDiff IG-guided curriculum learning across multiple ultrasound datasets. Each row corresponds to one internal dataset. a) Appendix. b) BUSI. c) BUS-BRA. d) UDIAT. e) Fatty-Liver. f) MMOTU. g) TN3K. Within each row, the leftmost bar chart reports the balanced accuracy (bAccuracy) of the ultrasound foundation model (USFM) fine-tuned on samples selected by random sampling at cumulative annotation budgets of 10%, 20%, and 30%, drawn as light-to-dark red bars, and by EvoDiff at the same three budgets, drawn as light-to-dark blue bars, with the value printed above each bar. The six confusion matrices to the right of the bar chart, titled from random 10% to EvoDiff 30%, report the corresponding predictions on the fixed internal test set, with rows giving the true label and columns the predicted label for the two task-specific classes, normal versus abnormal or benign versus malignant. Matrix cells are shaded in blue, with darker shading indicating a larger number of samples, and the sample count is printed in each cell. IG, information gain.
Figure 4: Foundation model performance in cross-domain validation and annotation efficiency analysis. a) External validation pipeline. The frozen EvoDiff model, marked by a snowflake symbol, selects informative samples from the internal source datasets (BUS-BRA, BUSI, UDIAT, and TN3K) for expert annotation, the ultrasound foundation model (USFM) is fine-tuned on the annotated samples, and the fine-tuned model is evaluated on four previously unseen external datasets (ALN-Ultra, BUS_UC, QAMEBI, and TN5000). b) Balanced accuracy of USFM on each external target dataset at annotation ratios of 10%, 20%, and 30%, arranged from the top row to the bottom row. Bars are grouped in pairs by source dataset, where the lighter bar of each pair gives the result without EvoDiff, that is with randomly sampled annotation, and the darker bar gives the result with EvoDiff-selected annotation. Orange bars denote BUS-BRA as the source dataset, green bars denote BUSI, blue bars denote UDIAT, and red bars denote TN3K, with the value printed above each bar. c) Model-based estimates of annotation cost across query rounds for the six organ tasks, computed using the mean unit annotation time within each experience stratum of the six-reader panel. The teal dashed line with circular markers shows the estimated annotation time for random sampling in the senior stratum, the dark blue dashed line shows the estimate for EvoDiff selection in the senior stratum, the orange dashed line shows the estimate for random sampling in the junior stratum, and the red dotted line shows the estimate for EvoDiff selection in the junior stratum. Error bars indicate the between-reader s.d. (standard deviation) within each stratum. The pale blue band and the pale orange band between the paired curves mark the estimated reduction in annotation workload with EvoDiff for the senior and junior strata, respectively.
Figure 5: EvoDiff-generated synthetic probes across varying guidance strengths. Each row shows synthetic ultrasound images generated for one organ and class, and the five columns correspond to guidance strengths (GS) of 1, 3, 5, 7, and 9. In the row labels, N denotes normal, A denotes abnormal, B denotes benign, M denotes malignant, and CVE(+) and CVE(-) denote cases with and without a follow-up cardiovascular event. Progressive intensification of the information-gain (IG) guidance across active learning (AL) iterations creates a curriculum learning framework in which sample difficulty systematically increases, moving generation from typical appearances toward diagnostically ambiguous appearances.
Figure 6: Systematic uncertainty reduction in downstream model predictions across multi-organ ultrasound test sets. Uncertainty maps are shown for six anatomical sites (appendix, breast, carotid, liver, ovary, and thyroid) with two representative cases per organ, where the suffix 0 or 1 in each row label gives the binary ground-truth label, benign 0 and malignant 1, or normal 0 and abnormal 1. Within each half of the figure, the first column shows the original test image, the second column shows the uncertainty map of the initialized model, and the following three columns show the maps after active learning (AL) rounds 1, 3, and 5, generated with guidance strengths of 1, 5, and 9, respectively. Each map is coloured from dark purple for an uncertainty of 0, through blue and green, to yellow for an uncertainty of 1, as indicated by the colour bars, and the shrinking bright regions from left to right illustrate decreasing predictive uncertainty.
Figure 7: Technical pipeline and implementation details of the EvoDiff framework. a) Pre-training phase. Organ-conditioned coarse pretraining of the diffusion model and the guidance classifier uses the 80% pathology-unlabeled pool, followed by pathology-level fine-tuning on the 10% pathologically labeled subset, with the remaining 10% reserved as the test set for final validation. A flame symbol indicates that a network is updated in that step, and a snowflake symbol indicates that it is frozen. b) Sampling phase. Information gain-guided sampling with the frozen diffusion model generates synthetic sample sets under guidance strengths (GS) from 1 to 9, forming the certain-to-uncertain curriculum sets depicted by the green-to-orange cylinders and arrow. c) Selecting phase. Features of the synthetic probes and of the unlabeled candidate pool are extracted by a ResNet-50 pretrained on natural images, cosine similarity between the two feature sets is computed within each organ-level category, and the top α % most similar real images, outlined in red in the ranking columns, are selected for expert annotation. d) Active learning (AL) iteration phase. The newly annotated samples are merged into the training set to fine-tune the downstream models, a ResNet-50 and the ultrasound foundation model (USFM), which are then validated on the frozen test set, and successive cycles progressively increase the curriculum difficulty, as shown by the green-to-orange arrow.
Organ
Dataset
Challenge
Sample Distribution
Country
Collection Period
Ultrasound Equipment
Patient Count
Age Range (Mean±SD) a
Dataset Type b
Normal
Benign
Abnormal
Total
Appendix
Appendix [ 43 ]
Appendicitis diagnosis
174
–
300
474
Germany
2016–2021
Toshiba Xario, Aplio XG machines
579
0–18
Internal
PrivAppendix
Appendicitis diagnosis
21
–
24
45
China
–
–
–
–
External
Breast
BUSI [ 3 ]
Breast tumor recognition
133
437
210
780
Egypt
2018
GE Logiq E9
–
25–75
Internal
BUS-BRA [ 17 ]
Breast tumor diagnosis
–
1,268
607
1,875
Brazil
–
GE Logiq 5 & 7, Aplio 300, U-Systems
1,642
16–89
Internal
UDIAT [ 66 ]
Breast tumor diagnosis
–
109
54
163
Spain
2012
Siemens C512
163
–
Internal
Table 1: Overview of 6 clinical applications and dataset distributions across multi-organ US datasets. Fifteen datasets comprising 16,696 images across six anatomical regions were used, with 10,182 images for model training and internal validation, and 6,514 images from ALN-Ultra, BUS_UC, QAMEBI, TN5000, and the three private cohorts for external validation.
Cumulative Annotation Budget
Organ
Dataset
Method
5%
10%
20%
30%
40%
50%
100%
FSL (upper bound)
-
-
-
-
-
-
0.6404± 0.0104
Random
0.4783± 0.0583
0.4870± 0.1163
0.4913± 0.0668
0.5391± 0.0422
0.6000± 0.0379
0.6087± 0.0550
-
VAAL [ 54 ]
0.4635± 0.0967
0.4891± 0.0763
0.4936± 0.0869
0.5279± 0.0448
0.5435± 0.0967
0.5832± 0.1383
-
Core-Set [ 53 ]
0.4845± 0.0692
0.5096± 0.0149
0.4858± 0.0918
0.5392± 0.0573
0.5502± 0.0270
0.5745± 0.0854
-
Max-Entropy [ 65 ]
0.5016± 0.0708
0.5277± 0.0645
0.5556± 0.0794
0.5804± 0.0940
0.5932± 0.0809
0.6217± 0.0927
-
Table 2: Comparison of classification accuracy between EvoDiff and other AL methods on the test sets. Values are reported as mean ± s.d. across the five fold-specific models obtained by five-fold cross-validation. One ResNet-50 was trained independently for each dataset, with no parameter sharing across datasets or organs. Strategy-versus-Random statistical comparisons, including the two-sided paired t -tests and the Benjamini–Hochberg multiplicity correction applied across the five strategy-versus-Random comparisons, are reported separately in the Supplementary Information. ( Bold represents the best result, Underline represents the second best result).
IG Guidance
LOEA
CL
Annotation Ratio=10%
Annotation Ratio=30%
Annotation Ratio=50%
Upper Bound
ACC
F1-score
AUC
ACC
F1-score
AUC
ACC
F1-score
AUC
ACC
F1-score
AUC
✓
×
×
0.5386
0.4938
0.5305
0.5503
0.5265
0.5627
0.5587
0.5294
0.5846
–
–
–
✓
✓
×
0.5383
0.4956
0.5347
0.5479
0.5294
0.5518
0.5592
0.5319
0.5731
–
–
–
✓
✓
✓
0.5652
0.5374
0.5960
0.5978
0.5761
0.6153
0.6373
0.6095
0.6474
0.6404
0.7032
0.6390
Table 3: Ablation study on the Appendix dataset [ 43 ] under 10%, 30%, and 50% cumulative annotation budgets. Each reduced configuration was compared with the complete EvoDiff framework using two-sided paired t -tests, with the five cross-validation folds used as matched pairs. The Upper Bound column reports fully supervised learning on the complete labeled set.
IG Guidance
LOEA
CL
Annotation Ratio=10%
Annotation Ratio=30%
Annotation Ratio=50%
Upper Bound
ACC
F1-score
AUC
ACC
F1-score
AUC
ACC
F1-score
AUC
ACC
F1-score
AUC
✓
×
×
0.6592
0.6051
0.6560
0.7531
0.6859
0.7803
0.8043
0.7275
0.8092
–
–
–
✓
✓
×
0.6631
0.6336
0.6958
0.7894
0.7047
0.7851
0.8383
0.7552
0.8365
–
–
–
✓
✓
✓
0.7048
0.6729
0.7125
0.8126
0.7403
0.8289
0.8656
0.7849
0.8519
0.8937
0.7621
0.8927
Table 4: Ablation study on the TN3K dataset [ 18 ] under 10%, 30%, and 50% cumulative annotation budgets. Each reduced configuration was compared with the complete EvoDiff framework using two-sided paired t -tests, with the five cross-validation folds used as matched pairs. The Upper Bound column reports fully supervised learning on the complete labeled set.
Table S9: Cross-domain generalization of USFM across external validation datasets. USFM was fine-tuned exclusively on source-domain data using training subsets selected by either random sampling or EvoDiff at annotation budgets of 10%, 20%, and 30%, and was subsequently evaluated without target-domain adaptation on seven unseen target-domain datasets, including four public and three private datasets. Performance is reported as balanced accuracy (bAcc). Statistical significance was assessed using two-sided (Random versus EvoDiff) paired t -tests. The Generalization Gain column reports each row’s largest relative gain across the three budgets and is therefore a post-hoc selected maximum, it is not corrected for this selection. For each annotation budget, the higher result between random sampling and EvoDiff is shown in bold.
Budget
Appendix
Breast
Carotid
Liver
Ovary
Thyroid
Senior -/w EvoDiff
Junior -/w EvoDiff
Senior -/w EvoDiff
Junior -/w EvoDiff
Senior -/w EvoDiff
Junior -/w EvoDiff
Senior -/w EvoDiff
Junior -/w EvoDiff
Senior -/w EvoDiff
Junior -/w EvoDiff
Senior -/w EvoDiff
Junior -/w EvoDiff
5%
17.1/2.8
29.6/4.8
19.3/10.3
26.8/14.3
23.3/5.2
28.7/6.5
4.1/1.6
6.5/2.5
31.6/17.1
37.3/20.3
53.0/17.1
61.5/19.8
10%
19.1/5.5
32.9/9.6
49.0/20.6
68.2/28.6
35.6/10.5
43.9/12.9
6.7/3.1
10.5/4.9
78.8/34.4
93.0/40.6
65.8/34.2
76.3/39.6
20%
20.1/11.1
34.7/19.2
73.8/41.2
102.6/57.3
61.0/20.9
75.4/25.8
14.0/6.2
21.9/9.8
99.6/68.9
117.6/81.2
177.2/68.4
205.3/79.2
30%
22.0/16.7
38.0/28.8
86.6/61.8
120.4/85.9
75.4/31.4
93.1/38.7
24.1/9.3
38.0/14.7
138.1/103.3
163.1/121.9
225.0/102.5
260.8/118.8
40%
29.2/22.2
50.4/38.4
131.5/82.4
182.8/114.6
84.9/41.8
104.8/51.6
26.2/12.5
41.2/19.6
228.4/137.6
273.1/164.5
262.4/136.6
304.1/158.4
Table S10: Model-based estimates of Human Annotation Effort (HAE, minutes) required by senior and junior radiologists to achieve matched downstream ResNet-50 performance using random sampling versus EvoDiff-selected annotations. Estimates were recomputed using the expanded six-reader panel. “-/w EvoDiff” denotes random sampling/EvoDiff selection, respectively.
Dataset
Hyperparameter Setting
Warmup Epoch
Warmup LR
Base LR
Minimum LR
Weight Decay
Layer Decay
Appendix [ 43 ]
2
1.25e-3
1.25e-3
1.25e-4
0.001
1.0
BUSI [ 3 ]
2
2.5e-4
1.25e-2
2.5e-7
0.05
0.65
BUS-BRA [ 17 ]
2
2.5e-4
1.25e-2
2.5e-7
0.05
0.65
UDIAT [ 66 ]
2
2.5e-4
1.25e-2
2.5e-7
0.05
0.65
Fatty-Liver [ 7 ]
1
1.25e-3
1.25e-3
1.25e-4
0.001
1.0
Table S11: Hyperparameter settings for different datasets on USFM fine-tuning.
Reader
Experience (years)
Organ
Reading time (min)
Observed count ( n )
Accuracy
Normalized count ( n /30 min)
Mean time (s)
Annotated
Correct
Annotated
Correct
Per image
Per correct
R1 (Senior)
7
Appendix
16.8
380
239
0.629
680
428
2.65
4.21
Breast
21.1
400
326
0.815
569
464
3.16
3.88
Carotid
12.5
400
320
0.800
959
767
1.88
2.35
Liver
20.0
400
369
0.923
600
554
3.00
3.25
Ovary
21.2
500
147
0.294
708
208
2.54
8.65
Table S12: Per-reader annotation performance across the six organ tasks. Six board-certified radiologists (R1–R3 senior, R4–R6 junior) annotated the same image sets under the same protocol. Accuracy is the proportion of correctly annotated images. Normalized counts project each reader’s observed rate to a 30-minute period. “Per image” and “Per correct” denote the mean seconds per annotated image and the mean seconds per correctly annotated image, the latter being the total annotation time in seconds divided by the number of correctly annotated images.
Organ
Real-image reference
s1
s3
s5
s7
s9
Pooled EvoDiff synthetic
Paired t -test vs reference
Appendix
0.624 ± 0.063
0.741
0.786
0.688
0.703
0.744
0.733 ± 0.079
0.176
Breast
0.929 ± 0.014
0.790
0.867
0.901
0.914
0.774
0.847 ± 0.018
<0.001
Carotid
0.657 ± 0.024
0.779
0.731
0.706
0.773
0.763
0.751 ± 0.045
0.018
Liver
0.970 ± 0.053
0.945
0.970
0.963
0.972
0.981
0.966 ± 0.023
0.803
Ovary
0.932 ± 0.031
0.849
0.766
0.852
0.898
0.859
0.844 ± 0.030
0.034
Thyroid
0.853 ± 0.012
0.821
0.762
0.781
0.726
0.847
0.789 ± 0.016
0.007
Table S13: Diagnostic-model recognition of synthetic probes. AUC (mean ± s.d. over five folds) from classifiers trained exclusively on real images, reported across five guidance strengths and for the pooled synthetic images. s1, s3, s5, s7, and s9 denote the guidance strengths used in the EvoDiff generation process. The real-image reference is the AUC of the same models on real images.
Reader
Experience
Appendix
Breast
Carotid
Synthetic Data
Real Data
Acc
Synthetic Data
Real Data
Acc
Synthetic Data
Real Data
Acc
Synthetic (Correct)
Real (Wrong)
Real (Correct)
Synthetic (Wrong)
Synthetic (Correct)
Real (Wrong)
Real (Correct)
Synthetic (Wrong)
Synthetic (Correct)
Real (Wrong)
Real (Correct)
Synthetic (Wrong)
Senior 1
9 years
7
13
12
8
0.475
8
12
12
8
0.500
15
5
10
10
0.625
Senior 2
8 years
2
18
10
10
0.300
2
18
20
0
0.550
4
16
17
3
0.525
Junior 1
3 years
8
12
9
11
0.425
8
12
20
0
0.700
7
13
12
8
0.475
Junior 2
4 years
7
13
11
9
0.450
5
15
13
7
0.450
4
16
12
8
0.400
Table S14: Blinded real-versus-synthetic discrimination. Four radiologists (two senior and two junior) each assessed 20 real and 20 synthetic images per organ. Synthetic images were evenly sampled from five guidance strengths. Accuracy denotes the proportion of correct image-origin judgments.
Organ
Category
n
1
2
3
4
5
Mean ± s.d.
Median
Score ≥ 4
Appendix
Real normal
40
16
11
5
2
6
2.28 ± 1.43
2
20%
Synthetic normal
40
17
8
6
5
4
2.28 ± 1.40
2
22%
Real diseased
40
4
1
4
14
17
3.98 ± 1.25
4
78%
Synthetic diseased
40
9
8
2
10
11
3.15 ± 1.58
4
52%
Breast
Real normal
40
17
13
2
7
1
2.05 ± 1.20
2
20%
Synthetic normal
40
25
5
5
4
1
1.78 ± 1.17
1
12%
Table S15: Blinded reader diagnostic scores by organ and image category. Pooled 1–5 scores from four readers are reported for real and synthetic non-diseased and diseased images, where 1 denotes a confidently non-diseased appearance and 5 a confidently diseased appearance. For each category, n , mean ± s.d., median, and the proportion of scores ≥4 are shown; bold marks the category rated most disease-like within each organ.
Method
Selection model(s)
Params (M)
Cost per round
Peak GPU memory
Random
none
0
0
0
Max-Entropy
ResNet-18 proxy
11.7
19.4 s
66 MB
Core-Set
ResNet-18 proxy
11.7
19.4 s
66 MB
BALD
ResNet-18 proxy, 25 MC passes
11.7
86.5 s
66 MB
VAAL
VAE and discriminator
23.6
104.1 s
1,997 MB
EvoDiff Stage 1 (pretraining)
diffusion UNet and guidance classifier
462
one-time, amortized
N/A
Table S16: Acquisition cost per round, measured on a single NVIDIA A40 for a representative round (BUS-BRA at a 10% budget). EvoDiff is resolved into its pipeline stages; Stage 1 pretraining is a one-time cost amortized across organs, datasets, and rounds, whereas Stages 2 and 3 recur at each round.
Figure S1: Statistical comparison of sample-selection strategies against Random for accuracy across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in accuracy relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S2: Statistical comparison of sample-selection strategies against Random for balanced accuracy across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in balanced accuracy (bAcc) relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S3: Statistical comparison of sample-selection strategies against Random for sensitivity across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in sensitivity relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S4: Statistical comparison of sample-selection strategies against Random for specificity across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in specificity relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S5: Statistical comparison of sample-selection strategies against Random for positive predictive value (PPV) across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in PPV relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S6: Statistical comparison of sample-selection strategies against Random for negative predictive value (NPV) across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in NPV relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S7: Statistical comparison of sample-selection strategies against Random for F1-score across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in F1-score relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S8: Statistical comparison of sample-selection strategies against Random for area under the receiver operating characteristic curve (AUC) across eight datasets and three cumulative annotation budgets. a) Results for the ResNet-50 downstream model. b) Results for the ultrasound foundation model (USFM). In each panel, the three heat maps correspond to cumulative annotation budgets of 10%, 20%, and 30%, rows list the five sample-selection strategies compared against Random (VAAL, Core-Set, Max-Entropy, BALD, and EvoDiff), and columns list the eight internal datasets. In each cell, the upper value is the unadjusted p -value from a two-sided paired t -test on the five matched cross-validation folds, and the lower value is the mean difference in AUC relative to Random, with positive values indicating better performance than Random. The cell colour encodes −log10(p) from light yellow for values near 1 to dark red for values of 3.5 and above, as shown by the colour bar. A black cell border marks an unadjusted p<0.05 , and an asterisk marks significance after Benjamini–Hochberg correction, with q<0.05 within the family of the five strategy-versus-Random comparisons that share a dataset and annotation budget. The testing procedure is identical for both downstream models.
Figure S9: Reader scores for non-diseased images. Four blinded readers rated 10 real and 10 synthetic images per organ on a 1–5 scale (1, confidently non-diseased; 5, confidently diseased). Panels show the appendix, breast, carotid, liver, ovary, and thyroid tasks, respectively. Bars show empirical probabilities, solid curves Gaussian kernel-density estimates, and dashed lines group means; legends report mean ± s.d.
Figure S10: Reader scores for diseased images. The reader study used the same settings as Supplementary Figure S9 .
It is often desirable to generalize medical imaging AI models trained with dense annotations to data acquired from different ultrasound scanners or clinical sites; however, retraining these models with new annotations is often difficult and costly. We examine this challenge in pediatric wrist fracture assessment using point-of-care ultrasound (POCUS), where fractures are common and can be effectively triaged via ultrasound. AI has shown radiologist-level performance for fracture detection, often aided by high-quality bony structure segmentation. However, due to significant domain shifts, models perform poorly on data from other centers or probes, and obtaining segmentation labels across devices is impractical due to manual annotation effort and data privacy concerns. To address this, we propose a target-informed self-supervised pretraining and model-ensemble strategy. Specifically, our approach combines masked image modeling (MIM) and contrastive learning to learn target-domain structural representations without labels, and introduces a confidence-aware infusion head to adaptively integrate predictions. The source dataset, collected with a Philips Lumify probe, contained dense labels, while the target dataset, acquired with a TeleMED portable probe, was unlabeled. The datasets were kept strictly separate throughout the entire process. Our method used labeled source data for supervised training and leveraged target-domain pretraining to improve generalization. On 318 images from 62 pediatric POCUS videos, this approach significantly improved cross-device performance, achieving over 6% Dice improvement on the target domain versus the baseline. These results demonstrate a label-efficient and privacy-preserving approach for cross-device-robust ultrasound AI, offering a framework that can be extended to multi-center studies or federated learning setups.
Yuyue Zhou, Shrimanti Ghosh, Michael +7
Kai Yue · Department of Radiology and Diagnostic Imaging, University of Alberta, 8303 112 St NW, Edmonton, T6G 2T4, AB, Canada
Lung ultrasound (LUS) is a safe and portable imaging modality, but the scarcity of data limits the development of machine learning methods for image interpretation and disease monitoring. Existing generative augmentation methods, such as Generative Adversarial Networks (GANs) and diffusion models, often lose subtle diagnostic cues due to resolution reduction, particularly B-lines and pleural irregularities. We propose A trous Wavelet Diffusion (AWDiff), a diffusion based augmentation framework that integrates the a trous wavelet transform to preserve fine-scale structures while avoiding destructive downsampling. In addition, semantic conditioning with BioMedCLIP, a vision language foundation model trained on large scale biomedical corpora, enforces alignment with clinically meaningful labels. On a LUS dataset, AWDiff achieved lower distortion and higher perceptual quality compared to existing methods, demonstrating both structural fidelity and clinical diversity.
Maryam Heidari, Nantheera Anantrasirichai, Steven Walker +2
University of Bristol, UK · Visual Information Lab, University of Bristol, Bristol, UK · Bristol Medical School, University of Bristol, UK +1
Clinical ultrasound analysis demands models that generalize across heterogeneous organs, views, and devices, while supporting interpretable workflow-level analysis. Existing methods often rely on task-wise adaptation, and joint learning may be unstable due to cross-task interference, making it hard to deliver workflow-level outputs in practice. To address these challenges, we present USTri, a tri-stage ultrasound intelligence pipeline for unified multi-organ, multi-task analysis. Stage I trains a universal generalist USGen on different domains to learn broad, transferable priors that are robust to device and protocol variability. To better handle domain shifts and reach task-aligned performance while preserving ultrasound shared knowledge, Stage II builds USpec by keeping USGen frozen and finetuning dataset-specific heads. Stage III introduces USAgent, which mimics clinician workflows by orchestrating USpec specialists for multi-step inference and deterministic structured reports. On the FMC_UIA validation set, our model achieves the best overall performance across 4 task types and 27 datasets, outperforming state-of-the-art methods. Moreover, qualitative results show that USAgent produces clinically structured reports with high accuracy and interpretability. Our study suggests a scalable path to ultrasound intelligence that generalizes across heterogeneous ultrasound tasks and supports consistent end-to-end clinical workflows. The code is publicly available at: https://github.com/MacDunno/USTri.
Chen Ma, Yunshu Li, Junhu Fu +3
College of Biomedical Engineering, Fudan University, Shanghai, China