BJR
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
British Journal of Radiology (2006) 79, 232-238
© 2006 British Institute of Radiology
doi: 10.1259/bjr/87956832
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Boniatis, I S
Right arrow Articles by Panayiotakis, G S
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Boniatis, I S
Right arrow Articles by Panayiotakis, G S

Full Paper

Quantitative assessment of hip osteoarthritis based on image texture analysis

I S Boniatis, MSc 1 L I Costaridou, PhD 1 D A Cavouras, PhD 2 E C Panagiotopoulos, MD, PhD 3 and G S Panayiotakis, PhD 1

1 University of Patras, School of Medicine, Department of Medical Physics, 265 00 Patras, 2 Technological Educational Institute of Athens, Department of Medical Instrumentation Technology, 122 10 Athens, 3 University of Patras, School of Medicine, Department of Orthopaedics, 265 00 Patras, Greece

Correspondence: George S Panayiotakis.


   Abstract
Top
 Abstract
Introduction
 Methods and materials
 Statistical analysis
 Results and discussion
 Conclusions
 Appendix 1
 References
 
A non-invasive method was developed to investigate the potential capacity of digital image texture analysis in evaluating the severity of hip osteoarthritis (OA) and in monitoring its progression. 19 textural features evaluating patterns of pixel intensity fluctuations were extracted from 64 images of radiographic hip joint spaces (HJS), corresponding to 32 patients with verified unilateral or bilateral OA. Images were enhanced employing custom developed software for the delineation of the articular margins on digitized pelvic radiographs. The severity of OA for each patient was assessed by expert orthopaedists employing the Kellgren and Lawrence (KL) scale. Additionally, an index expressing HJS-narrowing was computed considering patients from the unilateral OA-group. A textural feature that quantified pixel distribution non-uniformity (grey level non-uniformity, GLNU) demonstrated the strongest correlation with the HJS-narrowing index among all extracted features and utilized in further analysis. Classification rules employing GLNU feature were introduced to characterize a hip as normal or osteoarthritic and to assign it to one of three severity categories, formed in accordance with the KL scale. Application of the proposed rules resulted in relatively high classification accuracies in characterizing a hip as normal or osteoarthritic (90.6%) and in assigning it to the correct KL scale category (88.9%). Furthermore, the strong correlation between the HJS-narrowing index and the pathological GLNU (r = –0.9, p<0.001) was utilized to provide percentages quantifying hip OA-severity. Texture analysis may contribute in the quantitative assessment of OA-severity, in the monitoring of OA-progression and in the evaluation of a chondroprotective therapy.


   Introduction
Top
 Abstract
 Introduction
Methods and materials
 Statistical analysis
 Results and discussion
 Conclusions
 Appendix 1
 References
 
Osteoarthritis (OA) is a common joint disease that causes degenerative alterations in the hip as well as other joints [1]. Characteristic radiological manifestation of hip OA includes irregular superolateral, superior or superomedial hip joint space (HJS) narrowing, femoral and acetabular subchondral bone sclerosis, development of marginal osteophytes, as well as femoral and acetabular subchondral cysts formation [2].

Plain film radiography remains the most prevalent imaging modality for diagnosis of hip OA in clinical routine, despite its limited sensitivity compared with innovative imaging techniques, such as CT and MRI [3]. Radiographic assessment of hip OA comprises diagnosis, evaluation of severity, and monitoring of progression of structural alterations related to the disease [4]. A number of qualitative or semi-quantitative grading systems have been proposed for assessing hip OA [5–8], with the Kellgren and Lawrence (KL) grading scale [5] being considered the gold standard despite its deficiencies [9]. A reliable index for monitoring hip OA progression on pelvic radiographs is the progression of HJS-narrowing [6, 10], which may be estimated either manually [11, 12], or by computerized methods [13–15]. The latter are more sensitive, accurate, reproducible, and thus more reliable [16].

Texture analysis refers to algorithms developed to quantify image texture information that may, or may not, be perceived visually [17]. Although texture analysis has been previously employed in examining knee OA by computer processing of radiographic images [18, 19], hip OA has only been investigated in one study by computer analysis (fractal geometry) of digitized histological sections from the femoral head [20]. So far, the quantitative assessment of hip OA has mainly relied on measurements of HJS-width or HJS-area performed on pelvic radiographs [11–15]. To our knowledge, the textural properties of radiographic HJS in OA hips, as well as the capability of computer based radiographic texture analysis in evaluating the severity of hip OA have not been previously investigated.

In the present study, a non-invasive method was developed for analysing the structure of HJS from pelvic radiographs and for evaluating the severity of hip OA, employing computerized texture analysis. In particular, (i) textural features were extracted from the outlined region of each radiographic HJS, (ii) textural-feature thresholds, bearing good correlation to KL scale severity grades, were established for grading OA automatically, and (iii) an index was introduced for evaluating OA-severity.


   Methods and materials
Top
 Abstract
 Introduction
 Methods and materials
Statistical analysis
 Results and discussion
 Conclusions
 Appendix 1
 References
 
Radiographs and patients
32 anteroposterior pelvic radiographs of standing weight-bearing osteoarthritic patients were collected, giving in total 64 hip joint images. All radiographs were retrieved from the medical records of individuals who were candidates for total hip arthroplasty at the Department of Orthopaedics in our Hospital. From the total number of patients, 18 were verified for unilateral and 14 for bilateral hip OA. Patients' ages ranged between 49 years and 83 years with a mean age of 66.7 years. The American College of Rheumatology criteria [21] were used for OA diagnosis.

All pelvic radiographs were obtained using a Siemens X-ray unit (Polydoros 50; Siemens, Erlangen, Germany). Radiographic protocol comprised alignment of the X-ray beam 2 cm above the pubic symphysis, a focus–film distance of 100 cm, tube voltage between 70 kVp and 80 kVp, and use of a fast screen and film cassette (30 cm x 40 cm). Digitization of radiographs was performed at 12 bits (4096 grey levels) and 146 ppi (5.8 pixels mm–1) spatial resolution, using a laser digitizer for medical applications (Lumiscan 75; Lumisys, Sunnyvale, CA) [22]. Digitizer performance was evaluated employing a quality control protocol [23]. All radiographs fulfilled a specific criterion concerning safeguard against variations in hip rotation, introduced by the experienced orthopaedists. According to this criterion, the difference between the widths of projected lesser trochanters on each radiograph should not exceed 8 mm. Measurements on radiographs were performed by custom developed software [24–26].

Three experienced orthopaedists assessed the severity of OA employing the KL grading scale. The KL scale defines five categories of OA-severity (0–4), with KL grades =>2 corresponding to osteoarthritic pathology [5]. Based on the KL scale, patients were grouped into three major OA-severity categories: Normal/Doubtful (KL = 0, 1), Mild/Moderate (KL = 2, 3), and Severe (KL = 4). Accordingly, 18 unilateral-OA patients were assigned to Normal/Doubtful category, 9 to Mild/Moderate and 9 to Severe. The corresponding numbers for the bilateral patients were 0/7/21.

Radiograph enhancement
Pelvic radiographs were first processed by means of custom developed software based on the contrast-limited adaptive histogram equalization (CLAHE) method [27], in order to emphasise the articular margins of the hip joint. The CLAHE method partitioned the images into contextual non-overlapping regions. Within each region the local histogram was obtained, clipped to a specific limit and then histogram equalization was performed within the region. Figure 1Go shows a digitized radiograph enhanced by the implementation of CLAHE algorithm. On each enhanced radiograph two regions of interest (ROIs), one from the osteoarthritic HJS and one from the contralateral normal HJS, were manually outlined by three experienced orthopaedists, in accordance with the method proposed by Conrozier et al [13]. As shown in Figure 2Go, each ROI was defined within an acute angle determined by the patient's standard anatomical landmarks.


Figure 1
View larger version (74K):
[in this window]
[in a new window]
  		Figure 1. Example of(a) an original and (b) the corresponding processed digitized radiograph with the contrast-limited adaptive histogram equalization enhancement algorithm.

 

Figure 2
View larger version (165K):
[in this window]
[in a new window]
  		Figure 2. Hip joint space-region of interest (HJS-ROI) delineation within AOB. A: highest point of the homolateral sacral wing, O: centre of the femoral head, and B: lateral rim of the acetabulum.

 
Texture analysis of radiographic hip joint space
A total of 19 textural features were extracted from each segmented HJS-ROI (see Figure 3Go), utilizing custom developed algorithms. (i) Four textural features were computed from the ROI's grey level histogram [28], (ii) 10 from the ROI's grey level co-occurrence matrix [29] and (iii) five using the ROI's grey level run-length matrix [30].


Figure 3
View larger version (10K):
[in this window]
[in a new window]
  		Figure 3. Grey scale image of hip joint space region of interest(ROI) delineated in Figure 2.

 
Textural feature selection
From the extracted 19 textural features, selection was based on their correlation to an index evaluated for each of the unilateral OA-group patient, employing Equation (1):


Formula 001

where HJSAnormal and HJSApath express the number of pixels corresponding to the manually segmented contralateral normal and osteoarthritic HJS-ROIs, respectively. We have introduced this index, since it quantifies OA-severity by expressing the HJS-narrowing as a percentage of HJS-area difference between the osteoarthritic and contralateral normal HJS. This index is expected to give zero value in case of normal joints, since differences in hip joint spaces have been shown to be negligible in normal individuals [13, 31].

The validity of the proposed HJS-narrowing index was evaluated by examining its correlation with the KL scale, since the latter is considered to be the gold standard for OA-severity assessment. Analysis of HJS-narrowing index performance compared with KL scale is provided in a corresponding paragraph of the Results and Discussion section.


   Statistical analysis
Top
 Abstract
 Introduction
 Methods and materials
 Statistical analysis
Results and discussion
 Conclusions
 Appendix 1
 References
 
The existence of statistically significant differences between osteoarthritic and contralateral normal hips was investigated in the patients of the unilateral OA-group. Differences in HJS-area or in textural features were examined by means of the two-tailed student's paired t-test. To assess the relationship between the HJS-narrowing index and each textural feature extracted from osteoarthritic HJS-ROIs, the Pearson's correlation coefficient was used. To evaluate intraobserver and interobserver reproducibility concerning HJS-area measurements and GLNU calculated values, the coefficient of variation (CV) was used [32]. Accordingly, all radiographs were separately evaluated by each one of the experienced orthopaedists twice, with about a 1 month interval between evaluations. Evaluation scores were used to calculate the CV, which provides (e.g. see Conrozier et al [13]) an assessment of interobserver or intraobserver reproducibility; low coefficient values correspond to high degree of reproducibility. Referring to measured quantities, normality of their distributions was assessed by means of the Lilliefors test [33]. For non-gaussian distributions, a logarithmic transformation (log10) was applied to corresponding data. Matlab Statistics Toolbox and Matlab Curve Fitting toolbox (The MathWorks Inc., Natick, USA) were used for statistical and regression analysis.


   Results and discussion
Top
 Abstract
 Introduction
 Methods and materials
 Statistical analysis
 Results and discussion
Conclusions
 Appendix 1
 References
 
In a digital image, texture is characterized by intensity properties (tone) and spatial inter-relationships (structure) of image pixels, depicting spatial distribution of grey level variations in the image [29, 34]. In a digitized plain radiograph, a two-dimensional spatial distribution of grey-level variation is formed by projecting on a two-dimensional level the three-dimensional spatial distribution of the X-ray attenuation coefficients [17]. In the present paper, textural properties of each analysed radiographic HJS-ROI were attributed to X-ray attenuation, due to superimposed three-dimensional anatomical structures of articular cartilage, posterior acetabular wall and iliac bone. Therefore, the analysed radiographic ROI comprises of either osteoarthritic and/or normal superimposed anatomical components. Consequently, digital image texture analysis attempts to assess the existence and/or severity of structural alterations related to OA.

In patients with unilateral hip OA, statistical analysis revealed the existence of statistically significant differences in 11 (out of 19) textural features values between osteoarthritic and contralateral normal HJS-ROIs. Mean values (± standard deviation (SD)) of significantly differing textural features are presented in Table 1Go. These differences demonstrate textural alterations in radiographic HJS due to OA, which can be attributed to cartilage and subchondral bone tissue alterations associated to the disease. Articular cartilage performs mechanical functions providing transmission and distribution of high loads to underlying bone, maintenance of contact stresses at low levels, reduced frictional resistance to movement and shock absorption with these biomechanical properties being related to cartilage molecular–biochemical composition [35, 36]. Typical OA manifestations concern softening, ulceration, focal disintegration and the final loss of articular cartilage [37]. Alterations in chemical composition of articular cartilage have been associated with remodelling (increased density and stiffness) of subchondral bone in the form of subchondral sclerosis [38, 39]. Taking into account that structural alterations concern only osteoarthritic hips, differentiation of textural properties between normal and osteoarthritic HJS of unilateral OA-patients seems reasonable.


View this table:
[in this window]
[in a new window]
  		Table 1. Mean values(±SD) of statistically significantly differing textural features of contralateral normal and osteoarthritic HJS-ROIs

 
Previous studies on quantitative assessment of hip OA rely on measurements of the width or area of the radiographic HJS [11–15]. In the present study, hip OA-severity was estimated by the introduction of the HJS-narrowing index that evaluates the percentage of HJS-area difference between the osteoarthritic and the contralateral normal hip (Equation (1)). Repeated measurements of the HJS-area concerning the same observer were not found to differ significantly. Intraobserver reproducibility was found on average high for both the HJS area measurements (CV = 3.4%) and the corresponding GLNU values (CV = 3.9%). Similarly, interobserver reproducibility was also found high, 4.2% and 4.4% for HJS-area measurements and GLNU values, respectively. Mean values (±SD) of HJS-area for osteoarthritic and contralateral normal hips were found equal to 33.7 (±20.3) mm2 and 105.0 (±23.8) mm2, respectively. HJS-area values were statistically smaller (p<0.001) in osteoarthritic than in the contralateral normal hips, while previous studies on normal individuals have found no statistical differences between the two hips [13, 31]. HJS-narrowing index was evaluated for each one of the 18 unilateral patients and the mean and standard deviation of the HJS-narrowing index were calculated for the Mild/Moderate and Severe OA categories. Based on these means and standard deviations, classification rules (see Table 2Go) regarding the assessment of OA-severity were introduced (HJS-narrowing index Mean value ± 2SD). Referring to Table 2Go, an osteoarthritic hip was characterized as Severe if its HJS-narrowing index was greater than 77.9, as Mild/Moderate for index values within the interval [11.6, 77.9], and as Normal/Doubtful for OA if index values were lower than 11.6.


View this table:
[in this window]
[in a new window]
  		Table 2. Classification rules for assessment of osteoarthritis severity concerning HJS-narrowing index

 
The introduced classification rules were tested against the KL classification of the unilateral OA patients (Table 3Go). Referring to Table 3Go, all hips corresponding to Mild/Moderate OA-severity category were classified correctly, while classification accuracy of hips with Severe OA was 77.8%, resulting in a relatively high overall classification precision of 88.9%. Taking into consideration that our method relies solely on the assessment of HJS-narrowing, deviations of our results from the KL scale may be attributed to the fact that the KL scale evaluates, besides HJS-narrowing, the presence of osteophytes, subchondral sclerosis, and subchondral cysts.


View this table:
[in this window]
[in a new window]
  		Table 3. Comparison of results obtained by the KL scale and the proposed classification rules concerning HJS-narrowing index

 
Feature selection on the basis of Pearson's correlation coefficients between each of the textural features extracted from osteoarthritic HJS-ROIs and the HJS-narrowing index are summarized in column 4 of Table 1Go. The strongest correlation was found between the HJS-narrowing index and the pathological GLNU textural feature (r = –0.9, p<0.001). This relationship is presented graphically in Figure 4Go. As it can be observed, a regression line, described by Equation (2):


Formula 002

fitted data adequately. The negative slope of the regression line reflects the fact that in advanced stages of the disease, characterized by greater HJS-narrowing and thus higher HJS-narrowing index values, grey level intensities are more uniformly distributed (see Appendix 1) [30] within the region of radiographic HJS.


Figure 4
View larger version (15K):
[in this window]
[in a new window]
  		Figure 4. Hip joint space(HJS)-narrowing index versus pathological grey level non-uniformity (GLNU) textural feature. Solid line is the best line fitted to data points (•). Horizontal solid lines define the regions of osteoarthritis severity. Prediction bounds are denoted by dotted lines.

 
Subsequently, the selected GLNU textural feature was utilized in classification rules concerning the assessment of hip OA-severity. Based on the mean and standard deviation of the GLNU values, which were computed from the normal hips of the unilateral OA-group, a reference threshold value for GLNU, equal to 322.5 was employed (see Table 4Go) for characterizing a hip as either normal (GLNU>322.5) or osteoarthritic (GLNU{els]322.5). Using the contralateral normal hip for establishing thresholds for hip osteoarthritis has been also employed in previous studies. Conrozier et al [13] suggested the establishment of reference values by measuring the HJS-width and HJS-area of the normal hips in patients with unilateral OA, while other studies used the minimum joint space width for classifying a hip as osteoarthritic [7, 40–44]. In the present study, however, a textural-feature based classification rule was employed instead.


View this table:
[in this window]
[in a new window]
  		Table 4. Classification rules for assessment of osteoarthritis severity concerning GLNU textural feature

 
For verification purposes, GLNU values extracted from the HJS-ROIs of the unilateral osteoarthritic patients were subjected to the specific rule. It was found that 17/18 (94.4%) of the contralateral normal hips and a similar number (17/18) of the osteoarthritic hips were characterized correctly. When the same classification rule was applied to the bilateral OA-group, 24/28 (85.7%) hips were correctly characterized as osteoarthritic. In total, an overall classification accuracy of 90.6% (58/64), for assigning normal and osteoarthritic hips to the correct category, was achieved.

Besides characterizing a hip as normal or osteoarthitic, the capacity of GLNU textural feature was tested in establishing rules for differentiating hip OA-severity. Accordingly, rules for distinguishing hip OA-severity were formed, as shown in Table 4Go, that were defined on the basis of the mean and standard deviation GLNU values, obtained for hips assigned by the experienced orthopaedists to the same KL scale severity category. Referring to Table 4Go, a hip was characterized as Severe if its corresponding GLNU value was lower than 137.0, Mild/Moderate if its GLNU value was within the interval [137.0, 322.5], and as Normal/Doubtful for OA if its GLNU value was greater than 322.5. For verification purposes, these classification rules were applied to the unilateral OA-group and results were compared with the KL scale classification (Table 5Go).


View this table:
[in this window]
[in a new window]
  		Table 5. Comparison of results obtained by the KL scale and the proposed classification rules concerning GLNU textural feature

 
Referring to Table 5Go, the highest accuracy (94.4%) was achieved for normal hips. For hips with Severe OA the corresponding value was 77.8%, while for Mild/Moderate hips the accuracy was 88.9%. Finally, an overall classification accuracy of 88.9% (32/36) was achieved. To our knowledge, textural-feature based classification rules have not been proposed in previous hip OA studies.

Finally, the strong correlation of GLNU textural feature to the HJS-narrowing index was utilized to establish means of quantification of hip OA from textural properties of radiographic HJS (via GLNUpath feature) employing Equations (1) and (2). This is important for monitoring the progression of the disease and for assessing the effectiveness of a treatment. Referring to the classification rules of Table 4Go, percentages corresponding to OA-severity categories could be established using Equation (2). Thus, index values for Severe OA were greater than 67.3%, for Mild/Moderate OA within the interval [16.3%, 67.3%] and for Normal/Doubtful OA less than 16.3%.

In this way, a hip may be assigned to an OA-severity scale and its osteoarthritis, if it exists, can be evaluated from its radiographic texture. This is of value because the OA of patients suffering from bilateral-OA, which is often the case, can be now quantified employing Equation (2), whereas OA quantification in a manner similar to Equation (1) applies only to patients with unilateral OA.


   Conclusions
Top
 Abstract
 Introduction
 Methods and materials
 Statistical analysis
 Results and discussion
 Conclusions
Appendix 1
 References
 
Alterations in the radiographic depiction of hip joint space texture, due to osteoarthritis, were evaluated and related to the severity of osteoarthritis, as defined by the KL scale. Specifically, the GLNU textural feature, which was selected considering its strong correlation to the HJS-narrowing index, demonstrated high classification accuracy in distinguishing hip OA-severity categories. In addition, considering the high reproducibility derived for the GLNU, the proposed method may have a contribution in monitoring of OA-progression, as well as in the evaluation of a chondroprotective therapy.


   Appendix 1
Top
 Abstract
 Introduction
 Methods and materials
 Statistical analysis
 Results and discussion
 Conclusions
 Appendix 1
References
 
Description of the grey level non-uniformity (GLNU) textural feature defined by Galloway [30]
The grey-level run is a set of consecutive linearly adjacent pixels having the same grey level value. As length of the run is considered the number of pixels consisting the run. Each element p(i, j) of a grey-level run length matrix represents the number of times an image contains a run of length j for grey level i, in a specific direction.

The mathematical definition of the GLNU textural feature is:


Formula 003

where: p(i, j) is the (i, j)th element of grey level run length matrix, Ng is the number of grey levels in the image and Nr is the number of run lengths in the image.

Equally distributed runs throughout the grey levels, correspond to low values for GLNU and vice versa.


   Acknowledgments
 
The authors thank the staff of the Departments of Orthopaedics and Radiology for their contribution to this work.

Received for publication December 13, 2004. Revision received June 8, 2005. Accepted for publication July 1, 2005.


   References
Top
 Abstract
 Introduction
 Methods and materials
 Statistical analysis
 Results and discussion
 Conclusions
 Appendix 1
 References
 

  1. Stoker DJ. Radiology now. Osteoarthrosis of the hip: one disease or many? Br J Radiol 1977;50:81–3.[Abstract/Free Full Text]
  2. Altman RD, Bloch DA, Dougados M, Hochberg M, Lohmander S, Pavelka K, et al. Measurement of structural progression in osteoarthritis of the hip: the Barcelona consensus group. Osteoarthritis Cartilage 2004;12:515–24.[CrossRef][Medline]
  3. Peterfy CG. Imaging of the disease process. Curr Opin Rheumatol 2002;14:590–6.[CrossRef][Medline]
  4. Ory PA. Radiography in the assessment of musculoskeletal conditions. Best Pract Res Clin Rheumatol 2003;17:495–512.[CrossRef][Medline]
  5. Kellgren JH, Lawrence JS. Radiological assessment of osteoarthritis. Ann Rheum Dis 1957;16:494–501.[Free Full Text]
  6. Altman RD, Fries JF, Bloch DA, Carstens J, Cooke DT, Genant H, et al. Radiographic assessment of progression in osteoarthritis. Arthritis Rheum 1987;30:1214–25.[Medline]
  7. Croft P, Cooper C, Wickham C, Coggon D. Defining osteoarthritis of the hip for epidemiologic studies. Am J Epidemiol 1990;132:514–22.[Abstract/Free Full Text]
  8. Lane NE, Nevitt MC, Genant HK, Hochberg MC. Reliability of new indices of radiographic osteoarthritis of the hand and hip and lumbar disc degeneration. J Rheumatol 1993;20:1911–8.[Medline]
  9. Spector TD, Cooper C. Radiographic assessment of osteoarthritis in population studies: whither Kellgren and Lawrence? Osteoarthritis Cartilage 1993;1:203–6.[CrossRef][Medline]
  10. Altman R, Brandt K, Hochberg M, Moskowitz R. Design and conduct of clinical trials in patients with osteoarthritis: recommendations from a task force of the Osteoarthritis Research Society. Osteoarthritis Cartilage 1996;4:217–43.[CrossRef][Medline]
  11. Lequesne M. Chondrometry: quantitative evaluation of joint space width and rate of joint space loss in osteoarthritis of the hip. Rev Rhum Engl Ed 1995;62:155–8.[Medline]
  12. Hilliquin P, Pessis E, Coste J, Mauget D, Azria A, Chevrot A, et al. Quantitative assessment of joint space width with an electronic caliper. Osteoarthritis Cartilage 2002;10:542–6.[CrossRef][Medline]
  13. Conrozier T, Tron AM, Balblanc JC, Mathieu P, Piperno M, Fitoussi G, et al. Measurement of the hip joint space using computerized image analysis. Rev Rhum Engl Ed 1993;60:105–11.
  14. Conrozier T, Vignon E. Quantitative radiography in osteoarthritis: computerized measurement of radiographic knee and hip joint space. Best Pract Res Clin Rheumatol 1996;10:429–33.[CrossRef]
  15. Gordon CL, Wu C, Peterfy CG, Duryea J, Klifa C, Genant HK. Automated measurement of radiographic hip joint space width. Med Phys 2001;28:267–77.[CrossRef][Medline]
  16. Conrozier T, Favret H, Mathieu P, Piperno M, Provvedini D, Taccoen A, et al. Influence of the quality of tibial plateau alignment on the reproducibility of computer joint space measurement from Lyon schuss radiographic views of the knee in patients with knee osteoarthritis. Osteoarthritis Cartilage 2004;12:765–70.[CrossRef][Medline]
  17. Bocchi L, Coppini G, De Dominicis R, Valli G. Tissue characterization from X-ray images. Med Eng Phys 1997;19:336–42.[CrossRef][Medline]
  18. Lynch JA, Hawkes DJ, Buckland-Wright JC. Analysis of texture in macroradiographs of osteoarthritic knees using the fractal signature. Phys Med Biol 1991;36:709–22.[CrossRef][Medline]
  19. Lynch JA, Hawkes DJ, Buckland-Wright JC. A robust and accurate method for calculating the fractal signature of texture in macroradiographs of osteoarthritic knees. Med Inform 1991;16:241–51.
  20. Fazzalari NL, Parkinson IH. Fractal properties of subchondral cancellous bone in severe osteoarthritis of the hip. J Bone Miner Res 1997;12:632–40.[CrossRef][Medline]
  21. Altman R, Alarcón G, Appelrouth D, Bloch D, Borenstein D, Brandt K, et al. The American college of rheumatology criteria for the classification and reporting of osteoarthritis of the hip. Arthritis Rheum 1991;34:505–14.[Medline]
  22. Lumiscan 75, system specifications. Lumisys Inc. 1998; http://www.lumisys.com/support/manuals.html [Accessed 18 October 2005]
  23. Efstathopoulos EP, Costaridou L, Kocsis O, Panayiotakis G. A protocol-based evaluation of medical image digitizers. Br J Radiol 2001;74:841–6.[Abstract/Free Full Text]
  24. Sakellaropoulos P, Costaridou L, Panayiotakis G. An image visualisation tool in mammography. Med Inform Internet Med 1999;24:53–73.[CrossRef][Medline]
  25. Sakellaropoulos P, Costaridou L, Panayiotakis G. Using component technologies for web-based wavelet enhanced mammographic image visualization. Med Inform Internet Med 2000;25:171–81.[CrossRef][Medline]
  26. Sakellaropoulos P, Costaridou L, Panayiotakis G. A wavelet based spatially adaptive method for mammographic contrast enhancement. Phys Med Biol 2003;43:787–803.
  27. Pizer SM, Amburn EOP, Austin JD, Cromartie R, Geselowitz A, Greer T. Adaptive histogram equalization and its variations. CVGIP (Computer Vision, Graphics and Image Processing) 1987;39:355–68.[CrossRef]
  28. Gonzalez RC, Woods RE, editors. Digital image processing. New Jersey, USA: Prentice-Hall Inc., 2002.
  29. Haralick RM, Shanmugam K, Dinstein I. Textural features for image classification. IEEE Trans Syst Man Cybern 1973;SMC-3:610–21.[CrossRef]
  30. Galloway MM. Texture analysis using gray level run lengths. Comput Graph Image Proc 1975;4:172–9.
  31. Goker B, Sancak A, Arac M, Shott S, Block JA. The radiographic joint space width in clinically normal hips: effects of age, gender, and physical parameters. Osteoarthritis Cartilage 2003;11:328–34.[CrossRef][Medline]
  32. Altman DG, editor. Practical statistics for medical research. London, UK: Chapman & Hall, 1991.
  33. Lilliefors HW. On the Kolmogorov-Smirnov test for normality with mean and variance unknown. J Am Stat Assoc 1967;62:399–402.[CrossRef]
  34. Amadasun M, King R. Textural features corresponding to textural properties. IEEE Trans Syst Man Cybern 1989;19:1264–74.[CrossRef]
  35. Armstrong CG, Mow VC. Variations in the intrinsic mechanical properties of human articular cartilage with age, degeneration and water content. J Bone Joint Surg Am 1982;64:88–94.[Abstract/Free Full Text]
  36. Jeffrey DR, Watt I. Imaging hyaline cartilage. Br J Radiol 2003;76:777–87.[Free Full Text]
  37. Mankin HJ, Mow VC, Buckwalter JA, Ianotti JP, Ratcliffe A. Form and function of articular cartilage. In: Simons S, editor. Orthopaedic basic science. Chicago, USA: American Academy of Orthopaedic Surgeons, 1994;1–44.
  38. Radin EL, Rose RM. Role of subchondral bone in the initiation and progression of cartilage damage. Clin Orthop 1986;213:34–40.
  39. Burr DB, Schaffler MB. The involvement of subchondral mineralized tissues in osteoarthrosis: quantitative microscopic evidence. Microsc Res Tech 1997;37:343–57.[CrossRef][Medline]
  40. Jacobsen S, Sonne-Holm S, Soballe K, Gebuhr P, Lund B. The distribution and inter-relationships of radiologic features of osteoarthrosis of the hip. A survey of 4151 subjects of the Copenhagen City Heart Study: The Osteoarthrosis Substudy. Osteoarthritis Cartilage 2004;12:704–10.[CrossRef][Medline]
  41. Bierma-Zeinstra SMA, Oster JD, Bernsen RMD, Verhaar JAN, Ginal AZ, Bohnen AM. Joint space narrowing and relationship with symptoms and signs in adults consulting for hip pain in primary care. J Rheumatol 2002;29:1713–8.[Abstract/Free Full Text]
  42. Boegard T, Jonsson K. Hip and knee osteoarthritis. Conventional X-ray best and cheapest method. Lakartidningen 2002;44:4358–60.
  43. Ingvarsson T, Hägglund G, Lindberg H, Lohmander SL. Assessment of primary hip osteoarthritis: comparison of radiographic methods using colon radiographs. Ann Rheum Dis 2000;59:650–3.[Abstract/Free Full Text]
  44. Forsberg K, Nilsson BE. Coxarthrosis on the island of Gotland. Increased prevalence in a rural population. Acta Orthop Scand 1992;63:1–3.[Medline]



This article has been cited by other articles:


Home page
 Br. J. Radiol.Home page
I BONIATIS, L COSTARIDOU, D CAVOURAS, I KALATZIS, E PANAGIOTOPOULOS, and G PANAYIOTAKIS
A morphological index for assessing hip osteoarthritis severity from radiographic images
Br. J. Radiol., February 1, 2008; 81(962): 129 - 136.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Boniatis, I S
Right arrow Articles by Panayiotakis, G S
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Boniatis, I S
Right arrow Articles by Panayiotakis, G S

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
BJR DMFR IMAGING  ALL BIR JOURNALS