Initial GLOW-FDG release

.gitattributes

@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+GLOW-FDG/fold_0/progress.png filter=lfs diff=lfs merge=lfs -text
+GLOW-FDG/fold_1/progress.png filter=lfs diff=lfs merge=lfs -text
+GLOW-FDG/fold_2/progress.png filter=lfs diff=lfs merge=lfs -text
+GLOW-FDG/fold_3/progress.png filter=lfs diff=lfs merge=lfs -text
+GLOW-FDG/fold_4/progress.png filter=lfs diff=lfs merge=lfs -text

GLOW-FDG/dataset.json

@@ -0,0 +1,14 @@
+{
+    "channel_names": {
+        "0": "CT",
+        "1": "CT"
+    },
+    "labels": {
+        "background": 0,
+        "tumor": 1
+    },
+    "numTraining": 1563,
+    "file_ending": ".nii.gz",
+    "name": "GLOW-FDG-Dataset",
+    "converted_by": "Maximilian Rokuss"
+}

GLOW-FDG/fold_0/checkpoint_final.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:947622aeb87829d9adc5c993b5151fd419f6dd995849987eddb80c2eacb81f3a
+size 819730354

GLOW-FDG/fold_1/checkpoint_final.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aab8a3033565dd31babfe708bef565ea494a837f7bcac09cb3b09c5c5127853b
+size 819730418

GLOW-FDG/fold_2/checkpoint_final.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81e685eb0212811088d87e0892ac0144fa92562fd17131beca1864a41ca1e322
+size 819730354

GLOW-FDG/fold_3/checkpoint_final.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a194b0a45fc0d9c9c1e136939ed8e1b930123a7b741c2f6ef55d1b8ba04fcb5
+size 819730354

GLOW-FDG/fold_4/checkpoint_final.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8be831064af91b597fc72e9071e828541d891954c98958e344e5baf17696aa06
+size 819730418

GLOW-FDG/plans.json

@@ -0,0 +1,211 @@
+{
+    "dataset_name": "GLOW-FDG-Dataset",
+    "plans_name": "nnUNetResEncUNetLPlans",
+    "original_median_spacing_after_transp": [
+        3.0,
+        2.0364201068878174,
+        2.0364201068878174
+    ],
+    "original_median_shape_after_transp": [
+        326,
+        400,
+        400
+    ],
+    "image_reader_writer": "SimpleITKIO",
+    "transpose_forward": [
+        0,
+        1,
+        2
+    ],
+    "transpose_backward": [
+        0,
+        1,
+        2
+    ],
+    "configurations": {
+        "3d_fullres": {
+            "data_identifier": "nnUNetPlans_3d_fullres",
+            "preprocessor_name": "DefaultPreprocessor",
+            "batch_size": 2,
+            "patch_size": [
+                192,
+                192,
+                192
+            ],
+            "median_image_size_in_voxels": [
+                326.0,
+                400.0,
+                400.0
+            ],
+            "spacing": [
+                3.0,
+                2.0364201068878174,
+                2.0364201068878174
+            ],
+            "normalization_schemes": [
+                "CTNormalization",
+                "CTNormalization"
+            ],
+            "use_mask_for_norm": [
+                false,
+                false
+            ],
+            "resampling_fn_data": "resample_data_or_seg_to_shape",
+            "resampling_fn_seg": "resample_data_or_seg_to_shape",
+            "resampling_fn_data_kwargs": {
+                "is_seg": false,
+                "order": 3,
+                "order_z": 0,
+                "force_separate_z": null
+            },
+            "resampling_fn_seg_kwargs": {
+                "is_seg": true,
+                "order": 1,
+                "order_z": 0,
+                "force_separate_z": null
+            },
+            "resampling_fn_probabilities": "resample_data_or_seg_to_shape",
+            "resampling_fn_probabilities_kwargs": {
+                "is_seg": false,
+                "order": 1,
+                "order_z": 0,
+                "force_separate_z": null
+            },
+            "architecture": {
+                "network_class_name": "dynamic_network_architectures.architectures.unet.ResidualEncoderUNet",
+                "arch_kwargs": {
+                    "n_stages": 6,
+                    "features_per_stage": [
+                        32,
+                        64,
+                        128,
+                        256,
+                        320,
+                        320
+                    ],
+                    "conv_op": "torch.nn.modules.conv.Conv3d",
+                    "kernel_sizes": [
+                        [
+                            3,
+                            3,
+                            3
+                        ],
+                        [
+                            3,
+                            3,
+                            3
+                        ],
+                        [
+                            3,
+                            3,
+                            3
+                        ],
+                        [
+                            3,
+                            3,
+                            3
+                        ],
+                        [
+                            3,
+                            3,
+                            3
+                        ],
+                        [
+                            3,
+                            3,
+                            3
+                        ]
+                    ],
+                    "strides": [
+                        [
+                            1,
+                            1,
+                            1
+                        ],
+                        [
+                            2,
+                            2,
+                            2
+                        ],
+                        [
+                            2,
+                            2,
+                            2
+                        ],
+                        [
+                            2,
+                            2,
+                            2
+                        ],
+                        [
+                            2,
+                            2,
+                            2
+                        ],
+                        [
+                            2,
+                            2,
+                            2
+                        ]
+                    ],
+                    "n_blocks_per_stage": [
+                        1,
+                        3,
+                        4,
+                        6,
+                        6,
+                        6
+                    ],
+                    "n_conv_per_stage_decoder": [
+                        1,
+                        1,
+                        1,
+                        1,
+                        1
+                    ],
+                    "conv_bias": true,
+                    "norm_op": "torch.nn.modules.instancenorm.InstanceNorm3d",
+                    "norm_op_kwargs": {
+                        "eps": 1e-05,
+                        "affine": true
+                    },
+                    "dropout_op": null,
+                    "dropout_op_kwargs": null,
+                    "nonlin": "torch.nn.LeakyReLU",
+                    "nonlin_kwargs": {
+                        "inplace": true
+                    }
+                },
+                "_kw_requires_import": [
+                    "conv_op",
+                    "norm_op",
+                    "dropout_op",
+                    "nonlin"
+                ]
+            },
+            "batch_dice": true
+        }
+    },
+    "experiment_planner_used": "nnUNetPlannerResEncL",
+    "label_manager": "LabelManager",
+    "foreground_intensity_properties_per_channel": {
+        "0": {
+            "max": 3337.0,
+            "mean": -26.70451815536975,
+            "median": 35.0,
+            "min": -1441.0,
+            "percentile_00_5": -884.0,
+            "percentile_99_5": 873.0,
+            "std": 253.43406337222342
+        },
+        "1": {
+            "max": 150.53126525878906,
+            "mean": 5.448611647106978,
+            "median": 4.054980278015137,
+            "min": 0.02562030777335167,
+            "percentile_00_5": 0.8588680624961853,
+            "percentile_99_5": 27.83453369140625,
+            "std": 4.639632304482312
+        }
+    }
+}

README.md

@@ -1,3 +1,166 @@
----
-license: cc-by-nc-sa-4.0
----
+---
+license: cc-by-nc-sa-4.0
+tags:
+  - medical-imaging
+  - pet-ct
+  - segmentation
+  - oncology
+  - nnunet
+  - 3d-segmentation
+library_name: nnunet
+pipeline_tag: image-segmentation
+---
+
+<p align="center">
+  <img src="GLOW-FDG/logo.jpg" alt="GLOW-FDG logo" width="240"/>
+</p>
+
+# GLOW-FDG
+
+**G**eneralized cancer **L**esi**O**n **W**hole-body segmentation model for **<sup>18</sup>F-FDG PET/CT**.
+
+GLOW-FDG is an open-source 3D segmentation model for automated whole-body cancer lesion delineation in <sup>18</sup>F-FDG PET/CT. It is built on the [nnU-Net](https://github.com/MIC-DKFZ/nnUNet) framework using the **ResEnc L** architecture and was trained on a curated, multi-institutional corpus of **1,563 FDG-PET/CT scans** spanning lung cancer, head and neck cancer, lymphoma, melanoma, soft tissue sarcoma, prostate cancer, and PET-negative controls. The model was evaluated on **185 external scans** from independent cohorts covering breast cancer, nonmetastatic and oligometastatic lung cancer, head and neck cancer, and metastatic melanoma.
+
+The release contains the **5-fold cross-validation checkpoints** for ensembling.
+
+## Highlights
+
+- Whole-body FDG-PET/CT cancer lesion segmentation across multiple cancer types
+- Dual-head design: a primary lesion head and an auxiliary organ-supervision head (spleen, kidneys, liver, urinary bladder, lung, brain, heart, stomach, prostate, parotid and submandibular glands) to suppress physiologic-uptake false positives
+- Large-scale multi-modal (CT / MR / PET) MultiTalent-style pretraining followed by task-specific finetuning
+- PET/CT misalignment augmentation for robustness to patient motion and registration errors
+- Outperforms publicly available FDG-PET/CT benchmarks on lesion detection and segmentation across five external cohorts; performance approaches inter-observer variability between expert radiation oncologists
+
+## Intended Use
+
+GLOW-FDG is intended for **research use** in automated whole-body FDG-PET/CT cancer lesion segmentation and for the extraction of quantitative PET biomarkers such as total tumor burden (TTB) and total lesion glycolysis (TLG). It is **not** a certified medical device and must not be used as the sole basis for clinical decisions.
+
+### Out-of-scope / Limitations
+
+- Trained on standard-dose FDG-PET/CT; behavior on ultra-low-dose acquisitions has not been validated.
+- Trained only with FDG; not applicable to other tracers (e.g. PSMA, <sup>68</sup>Ga-DOTATATE).
+- Lesions without clear PET visibility may not be reliably detected.
+- Inputs must be a co-registered PET/CT pair with SUV-normalized PET.
+
+## Model Details
+
+| | |
+|---|---|
+| Framework | nnU-Net (3d_fullres, ResEnc L preset) |
+| Inputs | 2 channels: CT (HU) and PET (SUV<sub>BW</sub>) |
+| Output | Binary lesion segmentation mask (auxiliary organ head used during training only) |
+| Target spacing | 3.0 × 2.04 × 2.04 mm |
+| Patch size | 192 × 192 × 192 |
+| Training | 1,500 epochs, batch size 3, SGD with Nesterov momentum 0.99, LR 1e-2 with polynomial decay |
+| Pretraining | MultiTalent-style multi-dataset pretraining on CT/MR/PET, 4,000 epochs, patch 192³, batch 24 |
+| Folds | 5-fold cross-validation checkpoints (intended to be ensembled) |
+
+## Repository Contents
+
+```
+GLOW-FDG/
+  dataset.json           # nnU-Net dataset description
+  plans.json             # nnU-Net plans (architecture, preprocessing, etc.)
+  fold_0/checkpoint_final.pth
+  fold_1/checkpoint_final.pth
+  fold_2/checkpoint_final.pth
+  fold_3/checkpoint_final.pth
+  fold_4/checkpoint_final.pth
+```
+
+## Usage
+
+GLOW-FDG runs through the standard nnU-Net v2 inference API.
+
+### 1. Install nnU-Net
+
+```bash
+pip install nnunetv2
+```
+
+### 2. Download the model
+
+```python
+from huggingface_hub import snapshot_download
+
+model_dir = snapshot_download(repo_id="<org>/GLOW-FDG")
+# model_dir/GLOW-FDG/ now contains dataset.json, plans.json and fold_0..fold_4
+```
+
+### 3. Prepare your data
+
+Each case must contain two co-registered channels following the nnU-Net naming convention:
+
+```
+input_folder/
+  CASE001_0000.nii.gz   # CT (HU)
+  CASE001_0001.nii.gz   # PET (SUV body-weight normalized)
+```
+
+PET intensities **must** be converted to body-weight SUV before inference.
+
+### 4. Run inference
+
+```python
+import torch
+from nnunetv2.inference.predict_from_raw_data import nnUNetPredictor
+
+predictor = nnUNetPredictor(
+    tile_step_size=0.5,
+    use_gaussian=True,
+    use_mirroring=True,
+    perform_everything_on_device=True,
+    device=torch.device("cuda"),
+)
+
+predictor.initialize_from_trained_model_folder(
+    f"{model_dir}/GLOW-FDG",
+    use_folds=(0, 1, 2, 3, 4),
+    checkpoint_name="checkpoint_final.pth",
+)
+
+predictor.predict_from_files(
+    list_of_lists_or_source_folder="input_folder",
+    output_folder_or_list_of_truncated_output_files="output_folder",
+    save_probabilities=False,
+    overwrite=True,
+    num_processes_preprocessing=2,
+    num_processes_segmentation_export=2,
+)
+```
+
+The output is a binary NIfTI mask per case where `1` denotes predicted FDG-avid cancer lesions.
+
+## Training Data
+
+GLOW-FDG was trained on 1,563 FDG-PET/CT scans pooled from public and institutional sources, including AutoPET, HECKTOR, DEEP-PSMA, ACRIN-HNSCC, HN-PET-CT, NSCLC-RadGen, TCIA-STS, SAKK, and the SINERGIA melanoma cohort. All cases were manually reviewed to verify PET–mask correspondence, lesion completeness, and PET visibility. Organ labels for the auxiliary head were generated with [TotalSegmentator](https://github.com/wasserth/TotalSegmentator).
+
+## Citation
+
+If you use GLOW-FDG in your research, please cite:
+
+```bibtex
+@article{fritsak2026glowfdg,
+  title   = {GLOW-FDG: Generalized cancer LesiOn Whole-body segmentation model for 18F-FDG-PET/CT},
+  author  = {Fritsak, Maksym and Rokuss, Maximilian and Maier-Hein, Klaus},
+  year    = {2026}
+}
+
+@misc{rokuss2024fdgpsmahitchhikersguide,
+      title={From FDG to PSMA: A Hitchhiker's Guide to Multitracer, Multicenter Lesion Segmentation in PET/CT Imaging}, 
+      author={Maximilian Rokuss and Balint Kovacs and Yannick Kirchhoff and Shuhan Xiao and Constantin Ulrich and Klaus H. Maier-Hein and Fabian Isensee},
+      year={2024},
+      eprint={2409.09478},
+      archivePrefix={arXiv},
+      primaryClass={eess.IV},
+      url={https://arxiv.org/abs/2409.09478}, 
+}
+```
+
+## License
+
+Released under the [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)](https://creativecommons.org/licenses/by-nc-sa/4.0/) license. The model weights may be used, shared, and adapted for **non-commercial** purposes with attribution; derivative works must be distributed under the same license. Note that some of the underlying training datasets carry their own licenses and data use agreements that may impose additional restrictions.
+
+## Acknowledgments
+
+Developed at the Division of Medical Image Computing, German Cancer Research Center (DKFZ), Heidelberg, in collaboration with University Hospital Zürich (USZ). Built on [nnU-Net](https://github.com/MIC-DKFZ/nnUNet); auxiliary organ labels generated with [TotalSegmentator](https://github.com/wasserth/TotalSegmentator).

config.json

@@ -0,0 +1,3 @@
+{
+    "description": "Dummy config to allow tracking HF downloads."
+}