BAMF - NNUnet Lung and Nodules

models/bamf_nnunet_ct_lung/config/default.yml

@@ -0,0 +1,50 @@
+general:
+  data_base_dir: /app/data
+  version: 1.0
+  description: default configuration for Bamf NNUnet Lung, Nodules and Tumor segmentation (dicom to dicom)
+
+execute:
+- DicomImporter
+- NiftiConverter
+- NNUnetRunner
+- CustomNNUnetRunner
+- LungPostProcessor
+- DsegConverter
+- DataOrganizer
+
+modules:
+  DicomImporter:
+    source_dir: input_data
+    import_dir: sorted_data
+    sort_data: true
+    meta: 
+      mod: '%Modality'
+
+  CustomNNUnetRunner:
+    in_data: nifti:mod=ct
+    nnunet_task: Task777_CT_Nodules
+    nnunet_model: 3d_fullres
+    roi: LEFT_LUNG+RIGHT_LUNG,LEFT_LUNG+RIGHT_LUNG+NODULES
+    export_prob_maps: False
+
+  NNUnetRunner:
+    in_data: nifti:mod=ct
+    nnunet_task: Task775_CT_NSCLC_RG
+    nnunet_model: 3d_fullres    
+    roi: NEOPLASM_MALIGNANT_PRIMARY
+    export_prob_maps: False
+
+  LungPostProcessor:
+    in_rg_data: nifti:mod=seg:nnunet_task=Task775_CT_NSCLC_RG
+    in_nodules_data: nifti:mod=seg:nnunet_task=Task777_CT_Nodules
+
+  DsegConverter:
+    model_name: Bamf NNUnet CT Lung  
+    target_dicom: dicom:mod=ct
+    source_segs: nifti:mod=seg:processor=bamf
+    skip_empty_slices: True
+    json_config_path: models/bamf_nnunet_ct_lung/config/dseg.json
+
+  DataOrganizer:
+    targets:
+    - dicomseg-->[i:sid]/bamf_nnunet_ct_lung.seg.dcm

models/bamf_nnunet_ct_lung/config/dseg.json

@@ -0,0 +1,63 @@
+{
+  "ContentCreatorName": "BAMFHealth^AI",
+  "ClinicalTrialSeriesID": "Session1",
+  "ClinicalTrialTimePointID": "1",
+  "SeriesDescription": "AIMI Lung Nodule Segmentation",
+  "SeriesNumber": "300",
+  "InstanceNumber": "1",
+  "BodyPartExamined": "LUNG",
+  "segmentAttributes": [
+    [
+      {
+        "labelID": 1,
+        "SegmentDescription": "Lung",
+        "SegmentAlgorithmType": "AUTOMATIC",
+        "SegmentAlgorithmName": "BAMF-Lung-CT",
+        "SegmentedPropertyCategoryCodeSequence": {
+          "CodeValue": "123037004",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Anatomical Structure"
+        },
+        "SegmentedPropertyTypeCodeSequence": {
+          "CodeValue": "39607008",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Lung"
+        },
+        "SegmentedPropertyTypeModifierCodeSequence": {
+          "CodeValue": "51440002",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Right and left"
+        }
+      },
+      {
+        "labelID": 2,
+        "SegmentDescription": "Nodule",
+        "SegmentAlgorithmType": "AUTOMATIC",
+        "SegmentAlgorithmName": "BAMF-Lung-CT",
+        "AnatomicRegionSequence": {
+          "CodeValue": "39607008",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Lung"
+        },
+        "AnatomicRegionModifierSequence": {
+          "CodeValue": "51440002",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Right and left"
+        },
+        "SegmentedPropertyCategoryCodeSequence": {
+          "CodeValue": "49755003",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Morphologically Altered Structure"
+        },
+        "SegmentedPropertyTypeCodeSequence": {
+          "CodeValue": "27925004",
+          "CodingSchemeDesignator": "SCT",
+          "CodeMeaning": "Nodule"
+        }
+      }
+    ]
+  ],
+  "ContentLabel": "SEGMENTATION",
+  "ContentDescription": "Image segmentation",
+  "ClinicalTrialCoordinatingCenterName": "dcmqi"
+}

models/bamf_nnunet_ct_lung/dockerfiles/Dockerfile

@@ -0,0 +1,40 @@
+FROM mhubai/base:latest
+
+# FIXME: set this environment variable as a shortcut to avoid nnunet crashing the build
+# by pulling sklearn instead of scikit-learn
+# N.B. this is a known issue:
+# https://github.com/MIC-DKFZ/nnUNet/issues/1281 
+# https://github.com/MIC-DKFZ/nnUNet/pull/1209
+ENV SKLEARN_ALLOW_DEPRECATED_SKLEARN_PACKAGE_INSTALL=True
+
+# Install nnunet and platipy
+RUN pip3 install --no-cache-dir \
+  nnunet 
+
+# Clone the main branch of MHubAI/models
+ARG MHUB_MODELS_REPO
+RUN buildutils/import_mhub_model.sh bamf_nnunet_ct_lung ${MHUB_MODELS_REPO}
+
+# Pull nnUNet model weights into the container for Task777_CT_Nodules
+ENV WEIGHTS_DIR=/root/.nnunet/nnUNet_models/nnUNet/
+RUN mkdir -p $WEIGHTS_DIR
+ENV WEIGHTS_FN=Task777_CT_Nodules.zip
+ENV WEIGHTS_URL=https://zenodo.org/record/8290147/files/$WEIGHTS_FN
+RUN wget --directory-prefix ${WEIGHTS_DIR} ${WEIGHTS_URL}
+RUN unzip ${WEIGHTS_DIR}${WEIGHTS_FN} -d ${WEIGHTS_DIR}
+RUN rm ${WEIGHTS_DIR}${WEIGHTS_FN}
+
+# Pull nnUNet model weights into the container for Task775_CT_NSCLC_RG
+ENV TASK_NAME_NSCLC_RG=Task775_CT_NSCLC_RG
+ENV WEIGHTS_FN=$TASK_NAME_NSCLC_RG.zip
+ENV WEIGHTS_URL=https://zenodo.org/record/8290169/files/$WEIGHTS_FN
+RUN wget --directory-prefix ${WEIGHTS_DIR} ${WEIGHTS_URL}
+RUN unzip ${WEIGHTS_DIR}${WEIGHTS_FN} -d ${WEIGHTS_DIR}
+RUN rm ${WEIGHTS_DIR}${WEIGHTS_FN}
+
+# specify nnunet specific environment variables
+ENV WEIGHTS_FOLDER=$WEIGHTS_DIR
+
+# Default run script
+ENTRYPOINT ["mhub.run"]
+CMD ["--config", "/app/models/bamf_nnunet_ct_lung/config/default.yml"]

models/bamf_nnunet_ct_lung/meta.json

@@ -0,0 +1,139 @@
+{
+    "id": "",
+    "name": "bamf_nnunet_ct_lung",
+    "title": "AIMI CT Lung and Nodules",
+    "summary": {
+        "description": "An nnU-Net based model to segment lung from CT scans",
+        "inputs": [
+            {
+                "label": "Input Image",
+                "description": "The CT scan of a patient.",
+                "format": "DICOM",
+                "modality": "CT",
+                "bodypartexamined": "LUNG",
+                "slicethickness": "2.5mm",
+                "non-contrast": true,
+                "contrast": false
+            }
+        ],
+        "outputs": [
+            {
+                "label": "Segmentation",
+                "type": "Segmentation",
+                "description": "Segmentation lung",
+                "classes": [
+                    "LUNG",
+                    "LUNG+NODULES",
+                    "LUNG+NEOPLASM_MALIGNANT_PRIMARY"
+                ]
+            }
+        ],
+        "model": {
+            "architecture": "U-net",
+            "training": "supervised",
+            "cmpapproach": "3D"
+        },
+        "data": {
+            "training": {
+                "vol_samples": 833
+            },
+            "evaluation": {
+                "vol_samples": 47
+            },
+            "public": true,
+            "external": true
+        }
+    },
+    "details": {
+        "name": "AIMI CT Lung",
+        "version": "1.0.0",
+        "devteam": "BAMF Health",
+        "authors": [
+            "Soni, Rahul",
+            "McCrumb, Diana",
+            "Murugesan, Gowtham Krishnan",
+            "Van Oss, Jeff"
+        ],
+        "type": "nnU-Net (U-Net structure, optimized by data-driven heuristics)",
+        "date": {
+            "code": "17.10.2023",
+            "weights": "28.08.2023",
+            "pub": "23.10.2023"
+        },
+        "cite": "Murugesan, Gowtham Krishnan, Diana McCrumb, Mariam Aboian, Tej Verma, Rahul Soni, Fatima Memon, and Jeff Van Oss. The AIMI Initiative: AI-Generated Annotations for Imaging Data Commons Collections. arXiv preprint arXiv:2310.14897 (2023).",
+        "license": {
+            "code": "MIT",
+            "weights": "CC BY-NC 4.0"
+        },
+        "publications": [
+            {
+                "title": "The AIMI Initiative: AI-Generated Annotations in IDC Collections",
+                "uri": "https://arxiv.org/abs/2310.14897"
+            }
+        ],
+        "github": "https://github.com/bamf-health/aimi-ct-lung"
+    },
+    "info": {
+        "use": {
+            "title": "Intended Use",
+            "text": "This model is intended to perform lung, nodules and tumor segmentation in CT scans. The model has been trained and tested on scans aquired during clinical care of patients, so it might not be suited for a healthy population. The generalization capabilities of the model on a range of ages, genders, and ethnicities are unknown."
+        },
+        "analyses": {
+            "title": "Quantitative Analyses",
+            "text": "The model's performance was assessed using the Dice Coefficient and Normalized Surface Distance (NSD) with tolerance 7mm, as specified in the CT Liver segmentation task in the Medical Segmentation Decathlon challenge. The model was used to segment cases from the IDC collection TCGA-LIHC [1]. Nine of those cases were reviewed and corrected by a board-certified radiologist and a non-expert. The analysis is published here [2]",
+            "tables": [
+                {
+                    "label": "Label-wise metrics (mean (standard deviation)) between AI derived and manually corrected CT lung annotations",
+                    "entries": {
+                        "Dice: Radiologist": "0.97 (0.07)",
+                        "NSD: Radiologist": "0.96 (0.09)",
+                        "Dice: Non-expert": "0.97 (0.08)",
+                        "NSD: Non-expert": "0.95 (0.11)"
+                    }
+                }
+            ],
+            "references": [
+                {
+                    "label": "DICOM-LIDC-IDRI-Nodules",
+                    "uri": "https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=6885436"
+                },
+                {
+                    "label": "The AIMI Initiative: AI-Generated Annotations for Imaging Data Commons Collections",
+                    "uri": "https://arxiv.org/abs/2310.14897"
+                }
+            ]
+        },
+        "evaluation": {
+            "title": "Evaluation Data",
+            "text": "The model was used to segment cases 89 from the IDC [1] collection TCGA-LIHC [1]. Nine of those cases were randomly selected to be reviewed and corrected by a board-certified radiologist. The model predictions, and radiologist corrections are published on zenodo [3]",
+            "references": [
+                {
+                    "label": "Imaging Data Collections (IDC)",
+                    "uri": "https://datacommons.cancer.gov/repository/imaging-data-commons"
+                },
+                {
+                    "label": "TCGA-LIHC",
+                    "uri": "https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=6885436"
+                },
+                {
+                    "label": "Image segmentations produced by the AIMI Annotations initiative",
+                    "uri": "https://zenodo.org/records/10009368"
+                }
+            ]
+        },
+        "training": {
+            "title": "Training Data",
+            "text": "The training dataset consists of  1565 CT liver annotations taken from TotalSegmentator [1] (N=1204) and Flare21 Liver Dataset [2] (N=361).",
+            "references": [
+                {
+                    "label": "TotalSegmentator Dataset",
+                    "uri": "https://zenodo.org/records/10047292"
+                },
+                {
+                    "label": "Flare21 Dataset",
+                    "uri": "https://zenodo.org/records/5903672"
+                }
+            ]
+        }
+    }
+}