Ethics approval from the relevant research ethics committee and informed consent from all patients included in the study was obtained.
The study sample comprised 29 patients (24 male, 5 female), with a mean age of 28.7 years (15.3–46.4 years) who underwent SNU surgery from June 2011 to June 2016 after failed conservative treatment or occult fractures. Patients with previous surgical intervention or without preoperative CT examination were excluded.
In 26 patients, reconstructive surgery was performed, whereas 3 patients underwent 4 corner fusions as a salvage procedure. 10 patients received a non-vascularized bone graft (NVBG) and in 16 patients a vascularized bone graft (VBG) was used in combination with rigid internal fixation.
The histological methodology was presented in detail in the Journal of Anatomy .
Samples were taken during operative intervention. Haematoxylin and Eosin (HE), Azan, Toluidine, von Kossa, and Tartrate-resistant acid phosphatase (TRAP) staining were used to characterize the samples histologically. We determined distribution of Collagen 1 and 2 by immunocytochemistry, while scanning electron microscopy (SEM) was employed to investigate the ultrastructure.
The samples from each patient were subdivided into a proximal, gap, and distal section. Histological parameters were selected according to the existing literature [30,31,32] and to the longstanding experience of our histological experts. Different parameters were defined and evaluated for the presence or absence of each parameter. They were grouped according to their effect on bone healing into parameters with high, partial, and little activity.
Histopathological signs for high bone healing activity included osteoid formation, cell density and cell types, trabecular thickening, the presence of trabecular spikes bordered by osteoblasts (cell lines), neovascularisation, and the presence of collagen 1.
Histopathological signs for bone degeneration consisted of the presence of a sclerotic seal on the fracture site, a dense sclerotic area, few cells, the presence of cysts, the presence of osteoclasts or macrophages, and the presence of a “foam deposit” observed in areas with advanced bone degradation.
Histopathological signs of structure and the consistency of tissue in the gap comprised the presence of blood vessels, status of gap filling, type, and concomitance of collagens.
To quantify the overall biological activity, we calculated the bone healing capacity score (BHC) summarizing the individual parameters. For parameters in the high activity group, every point (presence of parameter) was scored with 2 (as this is evaluated as being better than partial activity). Partial activity was given a score of 1 (as in this case still some healing is taking place), and little activity was awarded a score of − 1 (as this has to be considered negative) .
CT images were acquired with a 64-detector row MDCT scanner (LightSpeed™ VTC, GE Milwaukee) using a standard bone protocol with a slice thickness of 0.625 mm and acquisition parameters of 100 kV and 100 mAs. All CT examinations were performed without a plaster cast.
In general, 2D parameters of bone structure can be evaluated on conventional CT slice images. For our study, we defined the region of interest (ROI) for 2D assessment more precisely to facilitate the measurement of interobserver reliability. In the coronal and sagittal planes, the ROI was centered over the fracture gap and aligned parallel to the fracture line to display the most central part of the proximal fragment using multi-planar reconstructions. For the analysis of SNU location, a 3D volume rendering model of the scaphoid was created.
On 2D-CT reformations, three blinded observers (two senior hand surgeons and one senior radiologist) assessed the following parameters of the proximal fragment:
Trabecular structure was assessed as present if > 50% of the cancellous bone showed signs of typical grid structure compared to the distal fragment (Fig. 1a).
Sclerosis was assessed in comparison to the distal fragment and rated as positive if > 50% of the cancellous part was involved (Fig. 1b).
Fragmentation as a sign of biological and structural breakdown may manifest as partial inner or complete outer fragmentation (Fig. 1c). If fragmentation of cancellous central parts of the proximal fragment exceeded half of the remaining bone and just the outer cortico cartilaginous shelf remained intact, it was defined as inner fragmentation or bone infarction. Outer fragmentation was present if central parts of the articular surface for the radius were separated and the fragmented part measured at least 1/3 of the total size of the proximal fragment. Peripheral wedge fractures and bony ligament avulsions were not rated as fragmentation as these are primary fracture patterns with potential biomechanical or biological impact, but are not signs of necrotic tissue or bone dissolution/infarction.
Based on these parameters of bone structure, we defined four subtypes of SNU (Table 1). This classification system ranges from Type 1 representing normal bone structure to Type 4 reflecting bone infarction with complete disintegration of the proximal fragment. The 2D-CT parameters of trabecular structure, sclerosis, and fragmentation, and the corresponding histologic images are shown in Fig. 2.
3D-CT fracture location
The location of the SNU and the course of the fracture line were related to anatomy and potential blood supply using 3D CT reconstructions. The dorsal interosseous scapholunate ligament attachment was visualized on the 3D model, and depending on the fracture course relative to this anatomical structure the non-unions were divided in three zones (Fig. 3).
SNUs with complete intraarticular cartilaginous sites were defined as white/white zone (WW, proximal), representing anatomical proximal pole non-unions where no separate blood supply can be expected.
Partially, intraarticular fracture lines with some cartilage free areas for ligament attachment and blood supply were allocated to SNUs of the intermediate red/white zone (RW, transition).
In extraarticular fracture lines at the anatomical waist region, a good blood supply of both fragments can be expected. These were defined as SNUs of the red/red zone (RR, waist).
18 cases (62.1%) were found at the completely intraarticular portion of the proximal pole (WW zone) and 11 cases (37.9%) in the other zones (3 RW, 8 RR).
Interrater reliability among the three raters for the 2D-CT parameters of trabecular structure, sclerosis, and fragmentation was assessed using Fleiss’ Kappa (κ) . Values less than 0.5 were rated as poor, between 0.5 and 0.75 as moderate, between 0.75 and 0.9 as good, and greater than 0.90 as excellent reliability .
Descriptive statistics are expressed as mean values and standard deviations or absolute and relative frequencies.
The Kolmogorov–Smirnov test was used for the determination of normal distribution. We assessed differences in BHC, age of SNU, and time to union related to the presence of 2D-CT parameters and to 3D-CT fracture zones with Student’s t tests for independent samples. Pearson correlation coefficients and linear regression analyses were calculated to investigate the relationship between the age of the SNU and the BHC with time to union. We used a significance level of α = 0.05 (two sided). Statistical analyses were performed using SPSS, version 24.0 (IBM Corp, Armonk, NY) and R version 3.3.0.