- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper proposes K2S, a label-efficient framework for brain tumor segmentation from undersampled multimodal MRI. The model jointly encodes undersampled k-space and its zero-filled reconstruction. During pretraining, it combines image reconstruction with prediction of an annotation-free spectral prior derived from deviations in the radial k-space energy profile. During fine-tuning, a slice-level lesion-presence branch softly gates TC and ET predictions to suppress false positives. Experiments are conducted on image-derived synthetic k-space from BraTS 2024, using 100 subjects for pretraining, 50 for fine-tuning, 20 for validation, and 20 for testing. The proposed method outperforms several image-domain and pretraining-based baselines and includes component and label-efficiency studies.
Critical strengths
- Task-oriented frequency-domain pretraining. The spectral-prior objective is an interesting attempt to convert an annotation-free k-space statistic into explicit supervision for segmentation-oriented representation learning. The ablation suggests that it improves positive-case Dice and HD95 over reconstruction-only pretraining.
- Effective false-positive suppression. The presence head and soft gating directly address a practical failure mode under limited supervision. They substantially improve TC/ET reliability and contribute to the strong overall results.
Critical weaknesses
-
The central “pathology-specific” and “acquisition-informed” claims are not sufficiently established. The k-space data are synthesized by applying a slice-wise 2D FFT to already reconstructed and preprocessed BraTS images, rather than obtained from raw MRI acquisitions. The experiments therefore exclude coil sensitivities, native phase, acquisition noise, and scanner-dependent effects. Moreover, the zero-filled image is deterministically obtained from the same undersampled k-space, so the results primarily demonstrate the value of dual-domain representations and auxiliary supervision, rather than access to additional acquisition information. The proposed residual also lacks strong null controls. Its GT-free localization precision is only 0.340, and the paper does not compare against random frequency bands, generic high-pass targets, shuffled priors, or conventional edge responses. It is therefore unclear whether the residual captures pathology-specific structure rather than generic contrast or anatomical boundaries.
-
The current ablations and evaluation do not cleanly isolate the source or robustness of the gains. The ablation is cumulative rather than factorial: there is no presence-only variant, no strong baseline equipped with the same presence mechanism, and no parameter-matched dual-domain model without the proposed spectral target. Thus, the contribution of the spectral prior cannot be separated from model capacity, dual-domain processing, or presence supervision. The evaluation further relies on a single 20-subject test split, one acceleration factor, and one mask protocol, without repeated runs, confidence intervals, or statistical tests. These limitations weaken both attribution and robustness claims.
Questions for authors
- Which data split was used for the spectral analysis, GT-guided oracle, modality selection, and hyperparameter choices in Section 3? Were test-set images or annotations used at any stage?
- What exactly is provided to the image encoder: complex-valued data, real and imaginary channels, or magnitude images? Since the zero-filled image is obtained by IFFT from the same k-space, what information is available only to the k-space branch?
- How is “Absent FP” defined, and how many absent TC and ET cases are included in the test set?
- Can the authors provide null controls, including random-band selection, generic Fourier-energy or high-pass targets, shuffled priors, and sensitivity to the number of selected bands and Savitzky–Golay parameters?
- Can the authors report a factorial ablation including reconstruction plus presence supervision, spectral pretraining without presence supervision, and a strong parameter-matched baseline with the same presence head?
9 comments
Statistical Analysis (Continued) and Factorial Ablation
Table R7. Sensitivity analysis of the spectral residual on the independent 40-subject cohort (component-hit rate).
| Parameter Setting | WT / FLAIR | Edema / FLAIR | ET / T1ce | TC / T1ce |
|---|---|---|---|---|
| Number of Selected Bands (Top-K) | ||||
| K = 3 | 0.787 | 0.756 | 0.559 | 0.569 |
| K = 5 (Default) | 0.800 | 0.756 | 0.559 | 0.559 |
| K = 7 | 0.825 | 0.781 | 0.559 | 0.569 |
| K = 10 | 0.825 | 0.762 | 0.608 | 0.608 |
| Savitzky-Golay Window Size | ||||
| Window = 31 | 0.794 | 0.762 | 0.549 | 0.559 |
| Window = 51 (Default) | 0.800 | 0.756 | 0.559 | 0.559 |
| Window = 71 | 0.806 | 0.762 | 0.559 | 0.559 |
| Savitzky-Golay Polynomial Order | ||||
| Order = 1 | 0.806 | 0.762 | 0.559 | 0.559 |
| Order = 2 (Default) | 0.800 | 0.756 | 0.559 | 0.559 |
| Order = 3 | 0.800 | 0.769 | 0.539 | 0.549 |
Response to W2 (a) & Q5. Factorial ablation and parameter matching
Table R1 (also updated as Table 2 in the revised manuscript) includes the previously missing reconstruction + presence + gating control and confirms that our full model outperforms this nearly parameter-matched non-spectral control (48.04M vs. 48.05M parameters). These results demonstrate the independent contribution of spectral supervision beyond model capacity and presence-aware adaptation.
Table R1. Expanded ablation study on the test set (updated as Table 2 in the revised manuscript). All variants share the same reconstruction pretraining. Absent FP denotes the fraction of lesion-absent cases (8 for TC, 8 for ET) with false-positive predictions.
| Variant | Spectral PT | Spectral FT/Inf | Presence | Soft Gating | Avg. Dice ↑ | NE Dice ↑ | NE HD95 (mm) ↓ | Absent FP (TC / ET) ↓ |
|---|---|---|---|---|---|---|---|---|
| Without Spectral Prior (Controls) | ||||||||
| Recon-only baseline | ✗ | ✗ | ✗ | ✗ | 0.544 | 0.525 | 30.24 | 4/8, 4/8 |
| + Presence head | ✗ | ✗ | ✓ | ✗ | 0.540 | 0.519 | 26.65 | 4/8, 4/8 |
| + Presence + Gating (Control) | ✗ | ✗ | ✓ | ✓ | 0.583 | 0.536 | 24.87 | 3/8, 3/8 |
| With Spectral Prior | ||||||||
| + Spectral PT | ✓ | ✗ | ✗ | ✗ | 0.428 | 0.553 | 21.34 | 8/8, 8/8 |
| + Spectral PT + Presence | ✓ | ✗ | ✓ | ✗ | 0.502 | 0.539 | 18.01 | 6/8, 5/8 |
| + Spectral PT + Presence + Gating | ✓ | ✗ | ✓ | ✓ | 0.586 | 0.540 | 17.94 | 3/8, 3/8 |
| + Spectral PT & FT/Inf + Presence | ✓ | ✓ | ✓ | ✗ | 0.562 | 0.553 | 16.62 | 4/8, 4/8 |
| + Spectral PT & FT/Inf + Presence + Gating (Ours) | ✓ | ✓ | ✓ | ✓ | 0.601 | 0.563 | 12.69 | 3/8, 3/8 |
Acquisition robustness and Statistical Reliability
Response to W2(b). Acquisition robustness
Thank you for this suggestion, which helped us enrich the paper with a broader robustness evaluation. Our proposed method consistently ranks first across different acceleration factors and different sampling trajectories. Evaluated across different acceleration factors (R=2, 4, 6, and 8) as well as four sampling patterns without retraining (Table R4), our method achieves the highest average subject Dice in every setting. We have also added a radar plot in Figure 5 (Fig. 5b) of the revised manuscript illustrating average subject Dice across these various undersampled settings.
Table R4. Average subject Dice across seven retrospective acquisition settings on the 20-subject test set.
| Method | Cart. R=2 | Cart. R=4 | Cart. R=6 | Cart. R=8 | 2D Gauss. R=4 | 2D Poisson R=4 | 2D Radial R=4 |
|---|---|---|---|---|---|---|---|
| 3D U-Net | 0.437 | 0.406 | 0.461 | 0.373 | 0.489 | 0.443 | 0.382 |
| nnUNet | 0.491 | 0.487 | 0.497 | 0.521 | 0.439 | 0.454 | 0.470 |
| MedNeXt | 0.508 | 0.519 | 0.516 | 0.401 | 0.458 | 0.440 | 0.428 |
| SwinUNETR | 0.439 | 0.423 | 0.391 | 0.419 | 0.432 | 0.440 | 0.209 |
| AMAES (U-Net-XL) | 0.481 | 0.454 | 0.400 | 0.380 | 0.479 | 0.500 | 0.341 |
| AMAES (MedNeXt-L) | 0.494 | 0.428 | 0.452 | 0.441 | 0.517 | 0.513 | 0.217 |
| KMAE | 0.402 | 0.362 | 0.288 | 0.185 | 0.321 | 0.329 | 0.393 |
| Ours | 0.635 | 0.601 | 0.560 | 0.587 | 0.626 | 0.613 | 0.480 |
Response to W2(c). Statistical Reliability
Following your suggestion, we have added the paired bootstrap 95% confidence intervals (Table R6). Paired bootstrap 95% confidence intervals confirm consistent subject-level gains across all baselines. Using subject-level paired bootstrap resampling (20,000 iterations, n=20), the 95% confidence intervals (CIs) for average subject Dice improvement exclude zero against all seven competing baselines.
Table R6. Paired differences (Ours - baseline) with 95% bootstrap CIs on the 20-subject test set. Positive Dice and negative HD95 favor Ours.
| Baseline Comparison | Δ Avg. Subject Dice ↑ | Δ Avg. NE Dice ↑ | Δ Avg. NE HD95 (mm) ↓ |
|---|---|---|---|
| MedNeXt | +0.082 [0.021, 0.168] | +0.069 [0.015, 0.138] | -9.12 [-21.81, -0.77] |
| nnUNet | +0.115 [0.030, 0.215] | +0.019 [-0.005, 0.046] | +0.37 [-1.42, 2.34] |
| AMAES (U-Net-XL) | +0.148 [0.067, 0.246] | +0.094 [0.052, 0.140] | -7.82 [-18.98, 0.53] |
| AMAES (MedNeXt-L) | +0.173 [0.071, 0.298] | +0.053 [0.035, 0.073] | -9.09 [-19.76, -0.34] |
| SwinUNETR | +0.179 [0.058, 0.314] | +0.016 [-0.018, 0.053] | -13.28 [-27.97, -0.73] |
| 3D U-Net | +0.196 [0.090, 0.318] | +0.089 [0.046, 0.136] | -11.55 [-23.60, -3.15] |
| KMAE (frozen + matched nnUNet) | +0.239 [0.123, 0.366] | +0.101 [0.050, 0.158] | -28.91 [-42.92, -15.65] |
Dual-Domain Clarifications, Null Controls and Statistical Analysis
Thank you for your thorough review and detailed assessment. Below, we clarify our dual-domain formulation, experimental setup, and expanded ablations.
Response to W1(a). Clarification of acquisition-informed learning
We acknowledge that our k-space is synthetically generated from preprocessed BraTS images and does not capture real acquisition effects such as coil sensitivities, native phase, and scanner noise. Accordingly, we do not claim access to additional independent acquisition information through dual-domain inputs; rather, they provide complementary representations with different inductive biases. Importantly, our contribution lies in explicitly modeling lesion-associated spectral deviations rather than simply adding Fourier representations. As shown in Table R5, residual-based spectral supervision outperforms raw Fourier-spectrum supervision in downstream segmentation, supporting the benefit of our proposed spectral prior. We have clarified these limitations and moderated our claims accordingly, while real multi-coil raw k-space validation remains an important direction for future work.
Table R5. Full 2D pretraining target comparison after fine-tuning.
Response to W1(b) & Q4(a). Spectral Null Controls: Localization and Downstream Segmentation
We clarify that our pretraining uses the full spectral residual distribution as the supervision target, while Top-5 band selection is used only for preliminary localization analysis. We evaluate the null controls at two levels:
Localization comparison (Table R2): For frequency-band selection methods, we compare Top-5 selected bands under the same localization protocol. Our residual-based selection achieves higher lesion localization rates than random-band, high-pass, and Sobel-edge controls on an independent 40-subject cohort.
Pretraining comparison (Table R5): Unlike band-selection methods, the raw Fourier spectrum does not explicitly select frequency bands. We therefore compare it as an alternative pretraining target by replacing the spectral residual distribution with the raw Fourier-spectrum distribution, followed by identical pretraining and fine-tuning. Our spectral residual consistently outperforms raw-spectrum supervision across all reported Dice metrics, demonstrating its effectiveness beyond generic frequency-domain supervision.
Table R2. Component-hit rate on an independent cohort.
Response to Q4(b). Parameter sensitivity analysis
As suggested, we evaluated the parameter sensitivity of the spectral residual in Table R7. Localization hit rates remain stable across different numbers of selected bands (K = 3 to 10), Savitzky-Golay window sizes (31 to 71), and polynomial orders (1 to 3).