Bone is a highly vascularized tissue receiving 10 to 15% of the resting cardiac output. Although this aspect of bone used to be overlooked, it has recently come to the front of the stage [1], [2], [3]. Bony blood vessels are necessary for almost all skeletal functions, such as development, homeostasis, and repair [4]. An intimate link between angiogenesis and osteogenesis has been shown in several studies [1], [2], [3], [4], [5]. More precisely, inflammatory responses and neovascularization are critical factors for initiating bone regeneration in case of injury, as blood vessels maintain the high metabolic demand for nutrients and oxygen, and provide pathways for cells to enter the injured areas [6]. Conversely, decreased neovascularization has been linked to avascular necrosis of the femoral head and postmenopausal osteoporosis [2]. Also, imaging of angiogenesis and microvascular morphology is supposed to inform the design of tissue engineering constructs to improve their osteo-integration in an era of “image-based tissue engineering” [[3], [4], [5], [6], [7]]. Bone tumors are also concerned, as angiogenesis might be a prognostic factor in osteosarcomas' natural course and treatment response [8]. Imaging of bone vascularization might therefore bring new scientific data and help to improve patient care in various clinical settings.

The best way to thoroughly observe the microarchitecture of bone is to perform histopathological study of bone, especially using bone clearance, which is time-consuming and destructive [9]. Despite significant advantages such as in vivo and nondestructive evaluation of bone, all imaging techniques suffer from limited spatial resolution. Notably, the detector element size of the current multidetector-computed tomography (MDCT) system is 0.5 mm, whereas an osteon has a diameter of 200–300 µm [2]. High-resolution peripheral quantitative CT (voxel size of 80 µm3; i.e., approximately the size of trabeculae)and micro-CT (voxel size of 19 µm3) are not routinely performed [10,11]. Photon-detector counting CT offers a detector element size of 0.25 mm but is a relatively new technique [12]. Ultra-high-resolution (UHR) CT is a commercially available technique with ultra-small detectors (slice thickness of 0.25 mm), large matrix sizes up to 2048 × 2048 pixels, and reconstruction algorithms with hybrid iterative reconstruction (HIR) and deep learning reconstruction (DLR) techniques that improve spatial resolution (as low as 0.12 mm) while keeping the radiation dose below the dose reference levels [13], [14], [15]. This technology has already been shown superior to conventional CT in depicting trabecular bone microarchitecture and might as well bring new data concerning cortical bone microarchitecture, including its vascularization [11].

Osteoid osteoma (OO) is a benign bone lesion that can be either cortical or intramedullary, containing a nidus with arterial blood supply and surrounding sclerotic bone, also described hypervascular, histologically and using wide area detector CT perfusion [16,17]. However, this bony hypervascularization can also be depicted using unenhanced CT thanks to the “vascular groove sign” (i.e., curvilinear or serpiginous low-density grooves radiating through thickened sclerotic bone from the periosteal surface down to the nidus and corresponding to enlarged feeding arterioles) [16,18].

The purpose of this study was to compare the capability of UHR-CT with HIR, UHR-CT with DLR, and conventional CT in the depiction of cortical vessels, using OO as a model.