Predicting the mechanical hip–knee–ankle angle accurately from standard knee radiographs: a cross-validation experiment in 100 patients

Figures & data

Figure 1. Measurement of the Hip–knee–ankle angle on full limb radiograph. The hip–knee–ankle angle (HKAA, in green) is measured between 2 axes (in red). One axis runs from the middle of the femoral head to the middle of the femoral notch, and a second axis from the middle of the tibial notch to the middle of the talar head.

Definitions of the femoral axis from left to right (top):

Fem1—mid-shaft at approximately 10 cm proximal of the femoral notch + mid-shaft in the area where the meta- and epiphysis meet (van Raaij et al. 2009, Iranpour-Boroujeni et al. 2014, Zampogna et al. 2015).

Fem2—mid-shaft at approximately 10 cm proximal of the femoral notch + center of the femoral notch (Felson et al. 2009, van Raaij et al. 2009, McDaniel et al. 2010, Sheehy et al. 2011).

Fem3—mid-shaft at approximately 10 cm proximal of the femoral notch + base of the tibial spines (Kraus et al. 2005, Hinman et al. 2006, Issa et al. 2007, McDaniel et al. 2010, Navali et al. 2012, Iranpour-Boroujeni et al. 2014, Zampogna et al. 2015).

Fem4—mid-shaft at approximately 10 cm proximal of the femoral notch + middle of tibial plateau (McDaniel et al. 2010).

Definitions of the tibial axis from left to right (bottom):

Tib1—mid-shaft at approximately 10 cm distal of the base of the tibial spines + mid-shaft in the area where the meta- and epiphysis meet (van Raaij et al. 2009, Iranpour-Boroujeni et al. 2014, Zampogna et al. 2015).

Tib2 Mid-shaft at approximately 10 cm distal of the base of the tibial spines + center of the femoral notch (McDaniel et al. 2010).

Tib3—mid-shaft at approximately 10 cm distal of the base of the tibial spines + base of the tibial spines (Kraus et al. 2005, Hinman et al. 2006, Issa et al. 2007, Colebatch et al. 2009, van Raaij et al. 2009, McDaniel et al. 2010, Sheehy et al. 2011, Navali et al. 2012, Iranpour-Boroujeni et al. 2014, Zampogna et al. 2015).

Tib4—mid-shaft at approximately 10 cm distal of the base of the tibial spines + middle of tibial plateau (McDaniel et al. 2010).

The 2 pictures on the left show the measurement of the FTA using method 2 for the femoral axis and method 1 for the tibial axis on a standard AP knee radiograph from the present data set.

Figure 2. Measurement of the femoro-tibial angle (FTA) on standard knee radiographs.

Tib4

Fem4

Tib3

Fem3

Tib2

Fem2

Tib1

Fem1

Table 1. Correlations between femoro-tibial angle and hip–knee–ankle angle as reported in the literature

Landmarks	Radiograph	Pearson correlation	Reference
Fem3 + Tib3	AP extended	0.26	Zampogna et al. 2015
Fem1 + Tib1	AP extended	0.71
Fem3 + Tib3	AP extended	0.81	Colebatch et al. 2009
Fem2 + Tib2	PA semi-flexed ^a	0.50	McDaniel et al. 2010
Fem3 + Tib3	PA semi-flexed ^a	0.65
Fem4 + Tib4	PA semi-flexed ^a	0.55
Fem3 + Tib3 ^b	PA semi-flexed ^a	0.64
Fem2 + Tib3	PA semi-flexed ^a	0.59
Fem3 + Tib3	PA semi-flexed ^a	0.86	Issa et al. 2007
Fem2 + Tib3	PA semi-flexed ^a	0.66	Felson et al. 2009
Fem1 + Tib1 ^c	PA semi-flexed ^a	0.76	Iranpour-Boroujeni et al. 2014
Fem3 + Tib3	PA semi-flexed ^a	0.68
Fem3 + Tib3	PA semi-flexed ^a	0.75	Kraus et al. 2005
Fem3 + Tib3	Full-limb	0.65
Fem3 + Tib3	Full-limb	0.88	Hinman et al. 2006
Fem2 + Tib3	Full-limb	0.34	van Raaij et al. 2009
Fem1 + Tib1	Full-limb	0.65
Fem2 + Tib3	Full-limb	0.88	Sheehy et al. 2012
Fem3 + Tib3	Full-limb	0.93	Navali et al. 2012

^aPositioning aided with Synaflexer frame.

^bSlight variation where the tips of the tibial spines are used instead of the base.

^cSlight variation where the bottom point at the femur is determined using the middle femoral condyles instead of the shaft.

Table 2. Pearson correlation coefficients and intra-class correlations (ICC) between FTA and HKAA measurements (across all pairs of images)

Method	Pearson correlation	ICC (95% CI)
Fem1 + Tib1	0.88	0.87 (0.82–0.91)
Fem1 + Tib3	0.86	0.86 (0.80–0.90)
Fem1 + Tib4	0.86	0.86 (0.80–0.90)
Fem2 + Tib1	0.90	0.90 (0.85–0.93)
Fem2 + Tib2	0.87	0.86 (0.80–0.90)
Fem2 + Tib3	0.89	0.89 (0.84–0.92)
Fem2 + Tib4	0.89	0.89 (0.84–0.92)
Fem3 + Tib3	0.84	0.83 (0.76–0.88)
Fem4 + Tib4	0.83	0.82 (0.74–0.87)

Figure 3. Bland–Altman plot depicting the error between the observed HKAA (gold standard) and the predicted HKAA in the cross-validation setting. Negative numbers represent the degree of varus alignment and positive numbers represent the degree of valgus alignment. The solid line depicts the mean error and the dotted lines the 95% confidence interval.

Table 3. Pearson correlation coefficients and intra-class correlations (ICC) between FTA and HKAA measurements (cross-validation experiments) 

Method	Pearson correlation	ICC (95% CI)
Fem1 + Tib1	0.88	0.87 (0.82–0.91)
Fem2 + Tib1	0.90	0.90 (0.85–0.93)
Fem2 + Tib3	0.89	0.89 (0.84–0.92)
Fem2 + Tib4	0.89	0.87 (0.80–0.91)

van Raaij T M, Brouwer R W, Reijman M, Bierma-Zeinstra S M A, Verhaar J A N. Conventional knee films hamper accurate knee alignment determination in patients with varus osteoarthritis of the knee. Knee 2009; 16(2): 109–11.

 PubMed Web of Science ®Google Scholar

Iranpour-Boroujeni T, Li J, Lynch J A, Nevitt M, Duryea J. A new method to measure anatomic knee alignment for large studies of OA: data from the Osteoarthritis Initiative. Osteoarthr Cartil 2014; 22(10): 1668–74.

 PubMed Web of Science ®Google Scholar

Zampogna B, Vasta S, Amendola A, Uribe-Echevarria Marbach B, Gao Y, Papalia R, Denaro V. Assessing lower limb alignment: comparison of standard knee X-ray vs long leg view. Iowa Orthop J 2015; 35: 49–54.

 PubMedGoogle Scholar

Felson D T, Cooke T D V, Niu J, Goggins J, Choi J, Yu J, Nevitt M C. Can anatomic alignment measured from a knee radiograph substitute for mechanical alignment from full limb films? Osteoarthr Cartil 2009; 17(11): 1448–52.

 PubMed Web of Science ®Google Scholar

McDaniel G, Mitchell K L, Charles C, Kraus V B. A comparison of five approaches to measurement of anatomic knee alignment from radiographs. Osteoarthr Cartil NIH Public Access 2010; 18(2): 273–7.

 PubMed Web of Science ®Google Scholar

Sheehy L, Felson D, Zhang Y, Niu J, Lam Y-M, Segal N, Lynch J, Cooke T D V. Does measurement of the anatomic axis consistently predict hip–knee–ankle angle (HKA) for knee alignment studies in osteoarthritis? Analysis of long limb radiographs from the multicenter osteoarthritis (MOST) study. Osteoarthr Cartil NIH Public Access 2011; 19(1): 58–64.

 PubMed Web of Science ®Google Scholar

Kraus V B, Vail T P, Worrell T, McDaniel G. A comparative assessment of alignment angle of the knee by radiographic and physical examination methods. Arthritis Rheum 2005; 52(6): 1730–5.

 PubMed Web of Science ®Google Scholar

Hinman R S, May R L, Crossley K M. Is there an alternative to the full-leg radiograph for determining knee joint alignment in osteoarthritis? Arthritis Rheum 2006; 55(2): 306–13.

 PubMedGoogle Scholar

Issa S N, Dunlop D, Chang A, Song J, Prasad P V, Guermazi A, Peterfy C, Cahue S, Marshall M, Kapoor D, Hayes K, Sharma L. Full-limb and knee radiography assessments of varus–valgus alignment and their relationship to osteoarthritis disease features by magnetic resonance imaging. Arthritis Care Res 2007; 57(3): 398–406.

 Web of Science ®Google Scholar

Navali A M, Bahari L A S, Nazari B. A comparative assessment of alternatives to the full-leg radiograph for determining knee joint alignment. Sports Med Arthrosc Rehabil Ther Technol 2012; 4(1): 40.

 PubMedGoogle Scholar

Colebatch A N, Hart D J, Zhai G, Williams F M, Spector T D, Arden N K. Effective measurement of knee alignment using AP knee radiographs. Knee 2009; 16(1): 42–5.

 PubMed Web of Science ®Google Scholar