TRACE-CT learning map

TRACE-CT (../trace-ct, Forgejo BioMedical-IT/trace-ct) is the driving laboratory. This map connects its real, already-done work to the learning chapters so the path is never abstract. All paths below are read-only references — TRACE-CT code is not copied into this repo (D-001).

Already-performed TRACE-CT work → learning chapters

TRACE-CT work	Concept learned	Chapter	Why it matters
RIDER inventory & preflight — `scripts/rider_preflight.py` (9-stage: scan, classify CT/SEG, group acquisitions, select reconstructions, pair repeats, resolve SEG refs, evaluate geometry, assess usability)	DICOM object classification; Study/Series grouping; reconstruction selection; test–retest pairing	1, 2	Real patient-level data handling; the data boundary in practice
DICOM metadata inspection — `rider_preflight.py` (Series/Study/FoR UIDs, kernel, kVp, `SliceThickness`, `ImageType`, `PixelSpacing`, IOP/IPP)	The DICOM hierarchy & identifiers; kernel/thickness as variables	1	The contract (metadata) you must read before the payload (pixels)
CT/SEG linkage — `rider_preflight.py` `build_seg_meta()` / `resolve_seg_references()`; `seg_decode_resample.py` `decode_seg_file()` / `build_source_mask_on_ct_grid()`	Source-image references; per-frame functional groups; segment identity	2	How a SEG names its source CT and why matching is by SOP UID, not assumption
Geometry validation — `seg_decode_resample.py` `compute_slice_normal()`, `build_affine()`, `project_position()`, `orientations_equivalent()`, `spacings_equivalent()`, `positions_equivalent()`; `load_ct_volume_explicit()`	Direction cosines; slice normal; geometric ordering; `EXACT_GRID` vs `RESAMPLED`	2	The whole Part A/B of Ch. 2, implemented and tested
HU conversion — `seg_decode_resample.py` `load_ct_volume_explicit()` (`pixel*slope+intercept`); `rider_radiomics_extraction.py` `clip_intensities()` (clip `[-1000,+400]`)	`RescaleSlope/Intercept`; working in physical HU	1, 3	Never use stored pixels as HU; clip before features
Mask transfer & resampling — `seg_decode_resample.py` `resample_mask_inverse_nn()`, `compute_source_mask_clipping()`, `classify_transfer_pair()`, `verify_determinism()`; `rider_radiomics_extraction.py` `resample_image_isotropic()` (BSpline→1mm), `resample_mask_nn()`, `affine_to_sitk_geometry()`	Linear vs nearest-neighbour; inverse NN; determinism; SimpleITK bridge	2	Masks resample differently from images; same shape ≠ same space
PyRadiomics extraction — `rider_radiomics_extraction.py` `extract_joint_entropy()`, `run_extraction()`; `configs/pyradiomics_params.yaml` (`binWidth 25`, `original`, 3D, label 1)	The radiomics pipeline; pinned preprocessing; GLCM Joint Entropy	3	One feature, done end-to-end reproducibly — the Ch. 3 exercise
Determinism / independent-process comparison — `rider_radiomics_extraction.py` `check_determinism`; `compare_radiomics_runs.py`	Reproducibility of the numbers before any statistics	3, 4	A feature table must be bitwise-reproducible across runs
Visual QC in 3D Slicer — `tools/slicer/open_qc_bundle.py` (2×3 layout, lung window); `scripts/visual_qc/*` (`qc_bundle.py`, `export_rider_qc.py`, `generate_report.py`, `record_review.py`, `aggregate_qc_reviews.py`); `schemas/visual-qc-bundle-v1.schema.json`	Visual QC as real, auditable geometry/overlay verification	2, 3	See the geometry concepts in a viewer; record a defensible verdict
Dependency / environment validation — `scripts/dependency_compute_smoke.py` (numpy/pandas/pydicom/PyRadiomics/SimpleITK; sklearn MICE; CatBoost CPU/GPU; lifelines CoxPH+C-index; CTGAN; TabDDPM upstream; PyTorch diffusion primitives)	A reproducible compute stack; tabular + survival tooling	4	Strong tabular baselines and survival models are first-class, not afterthoughts
Clinical tabular loader & missingness — `scripts/clinical_loader.py` (NSCLC frozen fields, per-field missingness); `results/metrics/nsclc-radiomics-clinical-profile.json`	Survival/tabular data; missingness as a variable	4	The clinical baseline imaging must beat; honest missingness handling

Future learning opportunities (not yet active)

Opportunity	Concept	Chapter	Trigger / constraint
Radiomic robustness (CCC test–retest)	Feature stability under repeat scans	3, 4	RIDER same-day repeats are already paired — ready when Chapter 3 deepens
Reconstruction sensitivity	Feature change across `1.25mm LUNG` vs `5.0mm STANDARD` kernels	3	`classify_source_role()` already tags roles — the Ch. 3 Step-7 experiment
Tabular missingness methods	MICE imputation, leakage-safe within-CV	4	`_run_mice()` smoke-tested; activate with a real modelling task
Survival validation	Cox assumptions, calibration, value beyond baseline	4	`_run_survival()` + NSCLC fields; activate with a radiomics+survival task
CT harmonisation (ComBat)	Scanner/site batch-effect correction & its assumptions	3, 4	Multi-site radiomics; respect ComBat’s assumptions
Synthetic data generation	CTGAN/TabDDPM generative tabular; leakage-safe evaluation	4, 6	Tabular completion direction in `docs/experiment-protocol.md`
Missing-acquisition image synthesis	Conditional image synthesis; beyond-pixel-realism evaluation	6, note	TRACE-CT activates synthesis (D-011); current pre-Spark experiment stays bounded

How to use this map

When a chapter says “TRACE-CT connection”, the concrete path is here.
Open the TRACE-CT file next to the chapter; read it to see the concept implemented, never to copy it here.
Future opportunities are not commitments — they activate only on a project trigger (anti-drift), and never expand the bounded current experiment.