Lumbar spine bone mass as measured by DXA may be elevated by co-existing conditions in or near the spine. Such confounding factors may include osteoarthritis, compression fracture and aortic calcification. In an effort to minimize spurious reporting of spine bone mass resulting from these conditions, the International Society for Clinical Densitometry (ISCD) recommends applying specific criteria to determine whether one or more vertebral bodies is excluded when reporting the final lumbar spine T-score (1). The ISCD recommends excluding vertebrae if DXA or radiographic images indicate a focal structural anomaly affecting some but not all of the visualized vertebrae (1). Vertebrae should also be excluded if an unusual discrepancy in T-scores is noted between adjacent vertebrae (1). Individual T-scores should be within one standard deviation of adjacent vertebrae (1).

In principle, the exclusion criteria improve diagnostic sensitivity of lumbar spine DXA for the detection of osteoporosis. However, interpretation of bone densitometry is subjective. As such, interpreters show only moderate agreement when applying the exclusion criteria to the same bone density studies (2). Focal structural anomalies are the greatest source of interpreter disagreement when excluding vertebrae.

Purpose of the Atlas
The atlas was developed in an initial attempt to improve interobserver agreement and standardization of DXA interpretation when analyzing lumbar spine DXA scans. We are grateful for grant support from the ISCD and assistance from the University of Wisconsin Medical Media Department in the development of this atlas. We anticipate that additional images will be added to the current version to improve its scope and educational quality. Ultimately, we hope this atlas can be used by physicians as a reference tool when interpreting lumbar spine densitometry.
Construction of an atlas required operational definitions of focal structural anomolies (FSA) and T-score discrepancy (TSD). We defined FSA as osteosclerotic or osteolytic processes altering the DXA-measured bone mass or densitometric image of one or more vertebral bodies. FSA may affect all or a portion of one or more vertebrae. Additionally, FSA may be directly adjacent to or overlie one or more vertebrae. We defined a TSD as a T-score that is one or more standard deviates from that of adjacent vertebrae.

Examples of FSA:

  • Vertebral fractures
  • Spine degenerative changes including vertebral osteophytes, degenerative arthritis of the facet joints, spondylolisthesis and scoliosis
  • Hardware in, overlying or underlying vertebrae including wires, cages, screws, staples, navel rings and inferior vena cava filters
  • Opacities adjacent to the lumbar spine including aortic calcification, nephrocalcinosis, barium and calcium tablets
  • Osteolytic processes including lytic lesions, hemangiomas and removal of bone related to laminectomy
  • Osteoblastic processes including Paget’s Disease of bone and bone metastases

Additionally, through consensus group meetings the following teaching points were derived.

  • Do not assume that the L1-L4 T-score is the optimal value to report when analyzing lumbar spine bone mass.
  • When excluding vertebral bodies, always provide the rationale for each vertebra excluded.
  • If spine bone density reveals low bone mass despite focal structural anomalies, such information may be invaluable to the referring clinician, particularly when femur and/or radius imaging is normal or significantly discordant from spine bone density. As such, information regarding low spine bone mass despite FSA should be reported to the clinician, with additional notation that the selected vertebral bone density may be spuriously elevated because of co-existing confounding factors.
  • If two or more vertebrae must be excluded, and femur and/or radius imaging is consistent with a densitometric diagnosis of osteoporosis, consider excluding the entire lumbar spine from the final DXA report.

Navigating the Atlas
For each case in this atlas, experts have applied the ISCD exclusion criteria and determined by consensus those vertebrae affected by FSA. In cases where the location of a particular FSA is noted, we have used “left” and “right” to describe the patient’s left and right side as a means of identifying the anomaly. Each case includes the following:

  • a GE Healthcare or Hologic lumbar spine DXA image with ancillary results
  • Options to exclude individual vertebral bodies
  • Group consensus decisions regarding vertebral body exclusion
  • The rationale for consensus decisions to exclude or include vertebrae
  • Corresponding radiographic or VFA images
  • Additional non-spine bone density regions of interest that assist in determining the validity of the lumbar spine images and bone density results

Sources of Data

  • Images 1-40 were provided by physicians within the University of Wisconsin Osteoporosis Clinic (Madison, WI).
  • Images 101-141 were provided by Dr. Sanford Baim of the Rheumatic Disease Center (Glendale, WI).


  • 1. Hamdy RC, Petak SM, Lenchik L. Which central dual X-ray absorptiometry skeletal sites and regions of interest should be used to determine the diagnosis of osteoporosis? J Clin Densitom 2002;5:S11-S17.
  • 2. Hansen KE, Binkley N, Christian RC, Vallarta-Ast N, Krueger D, Drezner MK, Blank RD. Interobserver reproducibility of criteria for vertebral body exclusion. J Bone Miner Res 2005;20:501-508.
  • Authors of Atlas Version One, released in 2007: Karen E. Hansen,1,2 Neil Binkley,1,2 Robert D. Blank,1,2,3 Diane C. Krueger,1 Irene Golembiewski,2,4 Sanford Baim5
  • 1 University of Wisconsin Osteoporosis Clinical Center and Research Program
  • 2 University of Wisconsin School of Medicine and Public Health
  • 3 William S. Middleton Veterans Affairs Medical Center
  • 4 University of Wisconsin Media Solutions Department
  • 5 Rheumatic Disease Center, Glendale, Wisconsin

Last modified: June 18, 2019