Current Osteoporosis Reports

, Volume 10, Issue 1, pp 42–47

Can PET-CT Imaging and Radiokinetic Analyses Provide Useful Clinical Information on Atypical Femoral Shaft Fracture in Osteoporotic Patients?


    • Department of Radiology, Division of Neuro-Interventional RadiologyUniversity of Washington: Harborview Medical Center (HMC)
  • Charles H. ChesnutIII
    • Department of RadiologyUniversity of Washington Medical Center
Current Therapeutics (SL Silverman, Section Editor)

DOI: 10.1007/s11914-011-0088-6

Cite this article as:
Chesnut, C.H. & Chesnut, C.H. Curr Osteoporos Rep (2012) 10: 42. doi:10.1007/s11914-011-0088-6


Atypical femoral shaft fractures are associated with the extended usage of nitrogen-containing bisphosphonates as therapy for osteoporosis. For such fractures, the positron emission tomography (PET) procedure, coupled with computerized tomography (CT), provides a potential imaging modality for defining aspects of the pathogenesis, site specificity, and possible prodromal abnormalities prior to fracture. PET-CT may assess the radiokinetic variables K1 (a putative marker for skeletal blood flow) and Ki (a putative marker for skeletal bone formation), and when combined with PET imaging modalities and CT skeletal site localization, may define the site of such radiokinetic findings. Further studies into the clinical usage of PET-CT in patients with atypical femoral shaft fractures are warranted.


Atypical femoral shaft fracturePositron emission tomography-CTMicrocracksOsteoporosisNitrogen-containing bisphosphonatesRadiokinetic analyses


The past 5 to 6 years have seen increasing scientific attention to a rare but disabling medical condition, atypical femoral shaft fracture (AFSF), potentially linked to long-term nitrogen-containing bisphosphonate (N-BP) usage for osteoporosis. Although the clinical morbidity of such AFSFs is well defined, the precise etiology and pathogenesis of these femoral shaft fractures remains unclear. This article presents an argument for the usage, in osteoporotic patients with AFSF, of positron emission tomography coupled with computerized tomography (PET-CT) as modalities providing imaging and radiokinetic analyses. Such analyses will contribute to an understanding of the pathogenesis of these fractures, and potentially to an improved clinical management of AFSF.

Atypical Femoral Shaft Fracture

These subtrochanteric/diaphyseal femoral shaft fractures occur most frequently in the proximal one third of the femur, but may exist anywhere in the subtrochanteric/supracondylar (diaphyseal) area of the femoral shaft, and are usually associated with minimal trauma [1•]. Such fractures may be complete, extending transversely or obliquely across the femoral shaft, or incomplete as a transverse radiolucent line across the lateral, and occasionally medial, cortex [1•]. A rather characteristic radiologic appearance is noted for AFSF including thickening of the periosteum and cortices (compared with the site and appearance of the typical femoral neck or trochanteric fractures usually associated with osteoporosis, or to the appearance of the typical subtrochanteric/diaphyseal fractures associated with trauma); the incomplete fractures are radiographically quite similar to a stress fracture [1•]. Complete, and incomplete, AFSFs may be either unilateral or bilateral, will frequently require orthopedic surgical management, and are associated with obvious clinical morbidity [1•].

The recent attention to such AFSFs has occurred in the setting of extended N-BP usage as treatment for osteoporosis, usually continuously for greater than 3 years. Although an association between extended BP usage and AFSF has been considered, a causal relationship between BP usage and such fracture has not been proven. In this regard, conflicting findings have been reported with a number of studies [2, 3] suggesting no association between usage of N-BP and AFSF, whereas others have indicated a significantly increased risk of AFSF with N-BP usage [4, 5]. However, even in studies describing a significant association between N-BP usage and AFSF [4, 5], it is noted that the benefit of N-BPs in reducing osteoporotic fractures at spine and hip exceeds the risk of an AFSF possibly associated with N-BP usage [1•, 4, 5].

Questions Related to the Occurrence of AFSF

A number of questions may be asked related to the occurrence of such an AFSF, particularly in regard to the possible association of AFSF with N-BP therapy:
  1. 1.

    Is the occurrence and pathogenesis of AFSF related to such pathophysiologic factors as impaired microdamage repair of cortical microcracks [1•, 4]? Such microcracks may be associated with stress fractures. As stress fractures heal by normal bone remodeling [1•], a suppression of bone resorption (and secondarily, with coupling, of bone formation) may impede healing; N-BP therapy will reduce bone resorption, and would then be expected to reduce healing of stress fractures, with subsequent coalescence of such stress fractures to completed fractures. If, however, the occurrence of AFSF is related primarily to factors (eg, biomechanical forces) usually associated with typical, usually traumatic, femoral shaft fractures, healing would be expected to be by endochondral ossification, with a lesser effect of N-BP therapy on such healing. At this stage of our understanding, the concept of impaired microdamage repair of impending or completed stress fractures due to N-BP usage would seem convincing as a pathogenetic explanation for the occurrence of AFSF.

  2. 2.

    Why do AFSFs, as stress fractures or completed fractures, occur, apparently solely, at the diaphyseal femoral site, with then a second fracture, stress, or completed occurring at a corresponding anatomical site of the contralateral femur? Although biomechanical forces exerted on the femoral shaft, both for an initial AFSF and a matching contralateral AFSF, may partially explain the site specificity of AFSF, it is also quite possible that disruption of other biologic processes (eg, skeletal blood flow or bone remodeling in osteoporotic patients receiving prolonged N-BP therapy) may also contribute to their unique regional occurrence at the femoral shaft.

  3. 3.

    If disruption of such biologic processes as skeletal blood flow/bone remodeling contributes to the pathogenesis of AFSF, is it possible that there is a prodromal, incipient state of such disruption prior to the occurrence of AFSF, either as the initial fracture, or as a contralateral fracture after the initial fracture. In other words, is there an “early warning” of changes in skeletal blood flow and/or bone remodeling that takes place at the AFSF site prior to the occurrence of either the stress, or completed, fracture? Such an “early warning” could obviously serve as a clinical alert for risk of a subsequent fracture.


In Regard to These First and Second Questions, Can PET-CT Imaging and Radiokinetic Analyses Provide Information Related to the Pathogenesis, Pathophysiology, and Site Specificity of AFSF? A Discussion of the Evidence

Assumptions Involved in Answering This Question

To answer this question one must assume that disruption of such biologic processes as skeletal blood flow and/or bone remodeling contributes to the postulated pathogenesis and pathophysiology of AFSF (a disruption of healing of cortical microdamage in N-BP–treated patients), to the apparent unique site specificity of AFSF, and possibly to prefracture changes at the femoral shaft. If such an assumption is then true, can PET-CT imaging and radiokinetic analyses provide site-specific assessment of skeletal blood flow and bone remodeling? In that regard, prior to addressing the primary question raised here of the utility of PET-CT in evaluating AFSF, a brief review of the evidence for assessment of these parameters with PET-CT is in order.

Assessment of Bone Remodeling and Skeletal Blood Flow with 18F-Fluoride PET

Bone remodeling (turnover), and its components of bone resorption and bone formation, has classically been assessed by the invasive and site-specific iliac crest bone biopsy with histology, or with serum/urine biomarkers of bone turnover, which provide a global, rather than regional, measurement of skeletal metabolism. Skeletal blood flow may be indirectly measured, along with skeletal metabolic activity (from which it has been difficult to distinguish in routine radionuclide bone scans), utilizing the skeletal uptake of bone-seeking radiopharmaceuticals such as 99mTc-labeled diphosphonates and 18F-fluoride.

The imaging modality of 18F-fluoride PET with kinetic analyses can noninvasively, and regionally, assess the delivery of the radionuclide 18F-fluoride to bone, and its incorporation into bone. Its basic principle is the deposition of 18F-fluoride in mineralizing osteoid, as fluoride replaces the hydroxyl ion in the hydroxyapatite crystal (forming fluorapatite rather than hydroxyapatite). The parameters of interest (regional bone flow, and bone remodeling, specifically bone formation) may be measured by assessment of the radiokinetic-derived rate constant K1 (putative skeletal blood flow) and the macroparameter Ki (putative skeletal bone formation): such parameters are derived from a three-compartment model originally proposed by Hawkins et al. [6], and are further confirmed as indicative of such biological functions by subsequent animal and clinical metabolic bone disease studies [711, 12••].

Therefore, the 18F-fluoride radionuclide as a skeletal imaging agent can potentially be utilized with the PET technology, in combination with kinetic analysis, to image and quantitate areas of increased osteoblastic bone formation and areas of increased radionuclide delivery (and by implication, increased blood flow). When these imaging and radiokinetic analyses parameters are combined with CT, the specific anatomical, and regional, skeletal site of the changes in skeletal vascularity and metabolism can be identified.

18F-Fluoride PET Results at Different Skeletal Sites

Further studies [13, 14] have used 18F-fluoride PET to examine skeletal kinetics (K1 and Ki) at differing trabecular (lumbar vertebrae) or cortical (distal humerus) skeletal sites; significantly greater K1 (putative skeletal blood flow) and Ki (putative skeletal bone formation) were noted at the trabecular vertebrae than at the cortical humerus. Skeletal heterogeneity can then be expected in K1 and Ki between different skeletal sites, with predominantly trabecular or cortical bone.

18F-Fluoride PET Findings in Osteoporotic Patients Treated with N-BP, or Anabolic Osteoporosis Therapies

If 18F-fluoride PET may be used in the evaluation of AFSF occurring in patients receiving N-BP, is there evidence that osteoporosis therapies, either N-BP, or anabolic therapies, affect the radiokinetic parameters quantitated with 18F-fluoride PET? In this regard two studies have examined the effects of N-BP therapy on 18F-fluoride PET parameters: Frost et al. [15] reported a significant decrease in vertebral Ki but no change in vertebral K1 in postmenopausal osteopenic or osteoporotic women after 6 months of treatment with the N-BP risedronate, and Uchida et al. [16] reported a significant reduction in standardized uptake value (SUV; an alternative method of quantifying skeletal 18F-fluoride uptake with 18F-fluoride PET) at the spine and hip in postmenopausal women with glucocorticoid-induced osteoporosis treated for 12 months with the N-BP alendronate. Frost et al. [17••] also reported the effects over 6 months of the anabolic 1–34 fragment of human parathyroid hormone teriparatide on 18F-fluoride PET skeletal kinetics, and SUV, in postmenopausal osteoporotic women. Spine Ki values significantly increased but with no significant change in K1; total hip and femoral shaft SUV values significantly increased. Therefore, these results suggest that the radiokinetic and SUV variables of 18F-fluoride PET may define a skeletal effect of N-BP (reduction in Ki and SUV values) and of teriparatide (increased Ki and SUV) at different skeletal sites.

If the Pathogenesis of AFSF Is Related to a Disruption of Healing of Cortical Microdamage (Cortical Microcracks), Is There Evidence That Such Microdamage/Microcracks Can Be Assessed with 18F-Fluoride PET?

Li et al. [18] studied the imaging of bone microdamage/microcracks with the 18F-fluoride PET technology in a rat ulnar loading model. It was concluded that 18F-fluoride PET imaging detected the bone microdamage produced in this in vivo model, that the location of increased 18F-fluoride radiotracer accumulation at sites of microdamage correlated with the location of microdamage as detected by histologic and autoradiographic parameters, and that 18F-fluoride PET “shows promise” as a noninvasive technology for imaging bone microdamage. Although this study was directed primarily at imaging the localization of the 18F-fluoride radiotracer by PET at sites of microdamage, rather than in determining K1 and Ki as parameters of 18F-fluoride PET-derived radiokinetics, Li et al. [18] provide evidence of the 18F-fluoride PET technology’s abilities to identify and localize microdamage/microcracks in vivo.

Can PET-CT Imaging and Radiokinetic Analyses Provide Information Related to the Pathogenesis, Pathophysiology, and Site Specificity of AFSF? The Summary of the Evidence

  • The attributes of 18F-fluoride PET encompass imaging of 18F-fluoride skeletal uptake, putatively defining the radiokinetic parameters of skeletal blood flow (K1) and bone formation (Ki), and, utilizing the accompanying CT, identifying the location of such imaging and radiokinetic findings.

  • As well, 18F-fluoride PET defines an apparent heterogeneity in terms of skeletal blood flow and bone formation at trabecular, or cortical, bone sites (axial spine, or appendicular humerus and femur).

  • Also, 18F-fluoride PET radiokinetic parameters such as Ki respond appropriately to various osteoporosis therapies that either suppress bone resorption (N-BPs such as risedronate or alendronate), or stimulate bone formation (teriparatide).

  • Lastly, cortical bone microdamage/microcracks may be imaged with 18F-fluoride PET, at least as defined in animal studies.

If Such Findings Are Confirmed, How Might 18F-Fluoride PET Imaging and Radiokinetic Parameters Be Utilized in the Assessment and Management of AFSF? Given the Above Findings Regarding Imaging and Radiokinetic Analyses of 18F-Fluoride PET, One Would Hypothesize That, in Osteoporotic Individuals with AFSF and Treated with N-BP, the 18F-Fluoride PET-CT Procedure Would Demonstrate

  • An increased uptake of 18F-fluoride at the AFSF fracture site, with a subsequent imaging “hotspot” (in nuclear medicine parlance) on imaging, and increased K1 (increased blood flow) and increased Ki (increased bone formation), at the diaphyseal femoral shaft fracture site.

  • One would also hypothesize increased 18F-fluoride uptake (and a “hotspot”), and increased K1 and Ki at the contralateral, but unfractured, femoral shaft site matching (as defined by CT) the AFSF site, although the increased 18F-fluoride uptake/increased K1 and Ki would be of lesser magnitude than that seen at the AFSF site. Also, increased 18F-fluoride uptake and increased K1 and Ki would not be seen at other skeletal trabecular or cortical sites (spine, extremities, etc.) that might be included in a total body PET scan.

  • As well, such an increased 18F-fluoride uptake/increased K1 and Ki would not be expected at diaphyseal femoral shaft sites in osteoporotic individuals treated with non-N-BP osteoporosis pharmacotherapies, or in osteoporotic individuals receiving no osteoporosis therapies.

If Such Findings Are Confirmed, How Are Such Findings Explained in Terms of the Postulated Pathogenesis of AFSF in Osteoporotic Patients Receiving Prolonged Treatment with N-BPs?

The above findings, particularly at the contralateral unfractured site, would be explained by the hypothesis that in osteoporotic patients receiving prolonged N-BP therapy the expected healing of cortical microcracks normally occurring at the femoral shafts (presumably due to substantial biomechanical loading at the diaphyseal femoral shaft with day-to-day normal activities) is impeded. Such unhealed microcracks would then progress to stress fracture and/or completed AFSF. The PET findings of an increased 18F-fluoride uptake/imaging “hotspot,” and an increased K1 and Ki, would then indicate an attempt of the bone to either repair itself if an AFSF has occurred, or to prevent progression of microcracks to stress fracture to completed fracture if an AFSF has not yet occurred. Such an attempt is most frequently unsuccessful due to the presence of N-BP in the bone impeding the normal bone remodeling healing process of microcracks/stress fractures, as noted on page 3 of this discussion. The bone increases blood flow (K1) and bone formation (Ki) at the site of the microcracks/stress fracture, even though N-BP therapy has suppressed bone resorption and to a lesser extent bone formation.

The postulated findings of increased 18F-fluoride uptake/imaging, and increased K1 and Ki, at both the AFSF site and the contralateral site, then reinforces the concept of impaired microdamage repair with N-BP usage as a significant etiologic factor in the pathogenesis of AFSF, particularly if such 18F-fluoride PET findings are not seen in the femoral shafts of osteoporotic patients not treated with N-BP. Although such PET-CT findings, in the absence of bone biopsy with histology and/or autoradiography, do not definitively confirm the presence of unhealed microcracks at the contralateral femoral shaft of patients with AFSF, such a possibility, and the subsequent pathophysiologic events, seem most reasonable.

Do the Above-Postulated Findings, If Confirmed, Provide Any Information on the Apparent Site Specificity of AFSF?

The findings on 18F-fluoride PET-CT of increased 18F-fluoride uptake/imaging “hotspot,” and increased K1 and Ki, only at the AFSF fracture site and the contralateral femoral shaft site, but at no other trabecular or cortical skeletal sites, reinforces the apparent site specificity of AFSF. However, the PET-CT procedure alone provides no mechanistic explanation for the apparent site specificity (ie, no information on the biomechanical forces responsible for the apparent preponderant cortical microcracks at the diaphyseal femoral shaft). In this regard, it should be noted that the biomechanical properties of the proximal femur including the proximal femoral shaft are determined to a great extent by bone geometry. Further quantitative CT analysis may provide an assessment of hip geometry, with uniquely an assessment of cortical bone separate from trabecular bone. Such biomechanical properties as average cortical density, cross-sectional moment of inertia, etc., may be determined [19, 20]; as well, finite element modeling [21] of the CT data may also be utilized to provide additional analysis of the mechanical behavior of the femoral bone segment being examined. Therefore, with additional analyses, the CT portion of the PET-CT scan may conceivably provide substantial information on the mechanical competence of the proximal femur, and of the likelihood of fracture.

In Response to the Third Question Originally Posed (Page 5 of This Discussion), If the Above-Postulated Findings are Confirmed, Can PET-CT Serve As a Clinical Predictor of Risk of AFSF in Patients Treated with N-BP?

If an increased 18F-fluoride uptake/“hotspot,” and increased K1 and Ki, are noted on PET at a contralateral femoral shaft site corresponding on CT to the site of the AFSF on the opposite femoral shaft, without evidence on CT of either a stress fracture or completed fracture, this implies a prodromal abnormality of blood flow and bone remodeling occurring prior to the occurrence of such a stress or completed fracture. Such findings would suggest the potential usage of 18F-fluoride PET-CT to predict a future stress or completed fracture prior to its occurrence (ie, as an “early warning” of abnormalities at the contralateral site that may indicate a risk for subsequent fracture). Such an “early warning” signal would then indicate such appropriate clinical measures as reduced loading of the femoral shaft by reduced physical activity, etc.

As well, if such prodromal attributes of 18F-fluoride PET-CT are confirmed, a logical next step would then be to consider its usage to define a potential AFSF risk in osteoporotic patients receiving prolonged N-BP therapy. Quite obviously financial costs, and availability, would affect such usage.


As noted in this discussion, there are numerous previous studies describing the clinical utility of the 18F-fluoride PET technology in osteoporotic patients. The assessment of apparent skeletal blood flow and bone turnover, specifically bone formation, utilizes the 18F-fluoride PET imaging and K1/ Ki parameters, and is coupled with the CT technology to identify the site of the assessed parameters. Based upon such previous studies, we present here a hypothesis that the 18F-fluoride PET-CT technology, potentially including CT assessment of bone geometry, biomechanical properties, and finite element modeling, may provide pertinent information regarding the pathogenesis and apparent site specificity of the AFSFs associated with prolonged N-BP usage. As well, there is a potential clinical application for the 18F-fluoride PET-CT technology to identify a risk for AFSF prior to the occurrence of such a fracture. Further research trials exploring the clinical utility of 18F-fluoride PET-CT, and its cost-effectiveness, are indicated.


No potential conflicts of interest relevant to this article were reported.

Copyright information

© Springer Science+Business Media, LLC 2012