CTSpinoPelvic1K: spine, pelvis, ribs and femora in one coordinate frame, annotated for lumbosacral transitional anatomy

Gregory Schwing; Ashley Schehr; Annika Tekumulla; Margret Khoushi; Ryan Christian; Dane Hubers; Faris Mahjoub; Hassan Saad

cs.AISep 19, 2026

CTSpinoPelvic1K: spine, pelvis, ribs and femora in one coordinate frame, annotated for lumbosacral transitional anatomy

Authors: Gregory Schwing, Ashley Schehr, Annika Tekumulla, Margret Khoushi, Ryan Christian, Dane Hubers, Faris Mahjoub, Hassan Saad,

Organizations: Department of Surgery, Detroit Medical Center and Wayne State University, Detroit, Michigan, USA · School of Medicine, Wayne State University, Detroit, Michigan, USA · Department of Radiology, Detroit Medical Center and Wayne State University, Detroit, Michigan, USA

Abstract

Purpose: A vertebra at the lumbosacral junction is named by counting caudally from C2 on whole-spine imaging, but a lumbar case is planned on lumbar-only imaging (T12 to S1), without C2. Abdominopelvic CT holds that span plus the lowest ribs and pelvis. Where a lumbosacral transitional vertebra (LSTV) alters the count, the local anatomy is ambiguous: four rib-free vertebrae may be an L1 with a lumbar rib or an L5 assimilated to the sacrum, and six may be a sixth lumbar vertebra, a T12 with aplastic ribs, or a lumbarized S1. CTSpinoPelvic1K asks whether local morphology resolves it without the count. CTSpine1K's vertebrae and CTPelvic1K's pelvis covered these patients but were never joined; this release joins them on one series and adds the bones neither had. It provides 802 CT records with per-level ribs and femora, levels anchored on the lowest rib-bearing vertebra and S1, and classes for L6, T13, a separate S1 and lumbar ribs, so anomalies are recorded as such. Acquisition and Validation Methods: Records pair CTSpine1K and CTPelvic1K labels on each patient's bone-richest series under a VerSe-native scheme. Validation covered geometric invariants (802/802 pass), rib-vertebra incidence across 5,749 ribs (0.035% offset), and spinopelvic measures matching published values. Data Format and Usage Notes: NIfTI image/label pairs with patient-grouped LSTV-stratified five-fold splits and a loader; archived at https://doi.org/10.5281/zenodo.22139642. Potential Applications: Classifying a vertebra from local features; updating cadaveric morphometry; spinopelvic assessment; opportunistic screening; and, absent a public preoperative lumbar cohort, surgical planning research (377 records prone). Limitations: thoracic ground truth is field-of-view limited; postural angles supine; no held-out test set; ribs are triaged-review pseudolabels; Castellvi grades two-reader consensus on 33 records.

Figures & tables

Collection	Scans	Count L6 Sacrum S1 Pelvis Ribs L. rib Femora Grade
CTSpine1K [ 25 ]	1,005
CTPelvic1K [ 24 ]	1,184
VerSe [ 10 ]	374
RibSeg v2 [ 35 ]	660
TotalSegmentator [ 9 ]	1,204
CTSpinoPelvic1K	802

Table 1: Public CT collections at the lumbosacral junction. Black, class present; outlined, graded only to exclude it [ 34 ] .

Panel of Figure 1 — Figure 1: How the dataset was built. $n$ counts scans, or corrections, at each step.

Check	Pseudolabel	Release
Two numbers on one bone	151 / 584	0 / 0
Rib detached from the spine	37 / 40	0 / 0
Rib in two pieces inside the field	119 / 280	6 / 7
Rib offset from its vertebra	24 / 170	0 / 2
No vertebra within reach	— / 38	— / 11
Rib on a lumbar body	14 / 21	0 / 0

Table 2: Rib checks on the raw pseudolabel and on the release, the same code on both. The first two gated the review; the rest were measured on every record. Pairs are records / ribs; a piece is a second component of at least 50 voxels and 15% of the largest, and a rib that leaves the reconstructed field and re-enters it is not counted as split.

Figure 2: Four phenotypes at the lumbar borders, aligned on the sacrum. Red, lowest rib-bearing vertebra and its rib; blue, sacrum; yellow, the lumbar bodies. Cases 0704, 0428, 0094, 0005, each the most ordinary member of its phenotype by rule. (b) and (c) present the same upper border, a short rib pair under a full pair, and were named differently.

Figure 3: Morphometry by level, T11–L5: median, interquartile range, 5th–95th percentile,

n

at right. (a) End-plate width EPWu. (b) Canal width SCW and depth SCD. (c) Pedicle width PDW, both sides averaged. L5 is withheld in (a) and (c); see text. Reference is Panjabi’s cadaveric series [ 29 , 30 ] , keyed in (a), its SD recovered from the reported SEM; T11–T12 EPWu is digitised from his Figure 4A.

Figure S1: Instrumentation in the release, one exemplar each, drawn through bone. (a) Hip arthroplasty, 66, $n=8$ . (b) Femoral-neck screws, 68, $n=1$ . (c) Sacroiliac screws, 67, $n=1$ . (d) Interbody cages, 63, $n=1$ . Bar, 5 cm.

Figure S2: Spinopelvic measures against standing reference bands (mean

\pm

SD) [ 45 ] .

Figure S3: Records carrying each vertebral level, cranial to caudal. Counts per identifier in Table S1.

Structure	Dice
Sacrum	0.980 $\pm$ 0.003
Left hip	0.970 $\pm$ 0.016
Right hip	0.971 $\pm$ 0.015

Table S1: Held-out fold Dice of the pelvic pseudolabeller against manual CTPelvic1K labels, mean

\pm

SD over the five folds at each fold’s selected checkpoint (nnU-Net validation Dice, ignore regions excluded; results/pseudolabel_dice/ in the repository).

Field	Records	Available
Identifiers and paths, configuration, match type, image/label alignment check	802	100.0%
Patient position, tube potential, section thickness, kernel	802	100.0%
Transitional label and class, annotation origin, instrumentation flags	802	100.0%
Spine series identifier; bone-fraction check	782	97.5%
Scanner manufacturer and model	768	95.8%
Sex	749	93.4%

Table S2: Data types and metadata, and the records populating each, from the released manifest.

id class	records	%
1 C1	2	0.25
6 C6	1	0.12
10 T3	1	0.12
11 T4	9	1.12
12 T5	34	4.24
13 T6	155	19.33

Table S3: Every identifier populated in the release and the number of the 802 records carrying it, counted from the released label volumes. The identifier space runs 0–68 with one hole at the retired 29; 46 and 59 (the thirteenth rib pair) occur in no record.

Medical Image Benchmarks

Medical image benchmarks compare AI methods on curated clinical images, tasks, and evaluation protocols to measure accuracy, robustness, and clinical usefulness.

See all 337 papers

Hybrid Context Forests and Bilateral Viterbi Matching for 3D Rib Centerline Tracing in CT
Oct 12, 2026Quan Wang3D Medical Image Segmentation Computed Tomography
SpineReport: Automated 3D Quantification and Reporting of Lumbar Spine Degeneration on MRI
Jun 8, 2026Nathan Molinier, Adrian A. Marth, Reto Sutter +6Medical Diagnosis 3D Medical Imaging
1NeuroPoly, Institute of Biomedical Engineering, Polytechnique Montreal, Montreal, QC, Canada · 2Mila, Quebec AI Institute, Montreal, QC, Canada · Department of Radiology, Balgrist University Hospital, Zurich, Switzerland +7
MSR:Hybrid Field Modeling for CT-MRI Rigid-Deformable Registration of the Cervical Spine with an Annotated Dataset
Apr 30, 2026Bohai Zhang, Wenjie Chen, Mu Li +8Medical Imaging Medical Image Registration
School of Biomedical Engineering, Southern Medical University, Guangzhou, 510515, China. · Guangdong Provincial Key Laboratory of Medical Image Processing, Guangzhou, 510515, China. · Guangdong Province Engineering Laboratory for Medical Imaging and Diagnostic Technology Guangzhou, 510515, China. +2

Oct 12, 2026J/K move · Enter open · S save

Hybrid Context Forests and Bilateral Viterbi Matching for 3D Rib Centerline Tracing in CT

Quan Wang

3D Medical Image Segmentation Computed Tomography

Open paper AskPDF arXiv

Automatic extraction and anatomical numbering (R01-R12, L01-L12) of all 24 rib centerlines in thoracic CT is essential for unfolded rib visualization, fracture detection, and surgical navigation, but remains challenging due to adjacent-bone confluence, calcified costal cartilage bridges, kyphotic pose variation, and partial visibility of terminal ribs. We present ribline, an open-source three-stage framework combining: (1) an automatically detected lung-apex and vertebral-midline coordinate frame with a kyphosis-aware azimuthal angle for point ordering along each rib; (2) a hybrid voxel detector fusing an anatomically anchored Semantic Context Forest (SCF) with a lightweight 3D U-Net; and (3) a global stage that decomposes the probability skeleton into candidate curves, assigns rib indices via a bilateral Viterbi dynamic program coupled across both hemithoraces, and refines curves via template recovery and normal-plane centering (RibSnake). On the Siemens Corporate Research 2 mm thoracic CT benchmark (47 training and 12 evaluation scans), our pipeline achieves a labeled Chamfer distance of 2.05 mm, labeled clDice of 0.886 (3 mm) and 0.945 (4 mm), 97.2% labeling accuracy (278/286 ribs), and zero false-positive curves on the evaluation split, and 3.64 mm (median 2.39 mm) with 91.8% labeling accuracy across all 59 annotated scans (1,402 ribs) in 5-fold cross-validation. Code: https://github.com/wq2012/ribline Pretrained Models: https://huggingface.co/collections/wq2012/ribline

Abstract

Medical Image Benchmarks

Explore similar work

Hybrid Context Forests and Bilateral Viterbi Matching for 3D Rib Centerline Tracing in CT

SpineReport: Automated 3D Quantification and Reporting of Lumbar Spine Degeneration on MRI

MSR:Hybrid Field Modeling for CT-MRI Rigid-Deformable Registration of the Cervical Spine with an Annotated Dataset